A Central-Bank Forecast That Is Willing to Be Proven Wrong

By Scott Shields – Contributing Writer – Capitol Times Media - From Conversations and Material of Zhu Weisha. Learn more about Zhu Weisha here at Capitol Times Media's July Magazine Issue

From Verifiable Finance to Verification-Based Credit: Why Everything Changes When the Mode of Credit Changes.

Can We Anticipate Warsh? appears at first to be an essay about Kevin Warsh and the Federal Reserve. Its more important subject, however, is the methodology of prediction itself.

Most market commentary predicts outcomes. Inflation rises, so analysts price a hike. Employment weakens, so they price a cut. A central-bank chair changes a phrase, and markets immediately recalculate the probability of the next meeting.

That approach is useful, but it can reduce analysis to a continuous exercise in guessing the next move.

This essay chooses a different object of prediction.

Rather than asking primarily whether Warsh will raise rates at the next meeting, it attempts to identify his reaction function: what facts are likely to cause him to act, what changes in the information set could cause him to revise his judgment, and which principles are likely to remain stable through short-term market fluctuations.

That is the paper’s central analytical move.

A forecast is no longer a single point estimate. It becomes a conditional structure that can be tested repeatedly over time.

This matters because a wrong call on one rate decision need not imply that the underlying model of the policymaker was wrong. Conversely, correctly guessing one meeting proves very little about whether the analyst actually understood the decision process.

The essay therefore separates several questions that financial commentary often collapses into one.

Were the underlying facts correct?

Was the transmission mechanism properly understood?

Was the judgment supported by evidence available at the time?

Were the conditions under which the forecast would fail stated in advance?

This is what the paper means by “verifiable.” The concept is deliberately modest: facts should be verifiable, processes should be reconstructable, and judgments and forecasts should state their grounds so that later evidence can test them.

The most important feature of this framework is that it is applied first to the authors themselves.

The essay repeatedly distinguishes between what Warsh has actually said, what he has actually done, what the authors infer from those facts, and what he has not addressed. That discipline is especially important when the subject is a policymaker who has spent decades operating inside financial markets and public institutions.

The aim is not to claim superior insight into Warsh. It is to register an interpretation before the evidence is complete and then allow subsequent events to confirm, weaken, or falsify it.

The paper’s strongest original hypothesis is its treatment of transparency.

Warsh’s preference for a “quieter Fed” does not necessarily imply a less transparent central bank. The essay proposes instead what might be called a re-timing of transparency: less communication before a decision, shorter communication at the moment of decision, but better documentation of the reasoning afterward and greater scope for later reconstruction.

This leads to an especially useful distinction between predictive transparency and ex-post verifiability.

Predictive transparency attempts to tell markets in advance what the central bank is likely to do.

Ex-post verifiability asks a different question: after the decision, can outsiders reconstruct what facts were available, what competing explanations were considered, what risks were accepted, and why one course of action was chosen?

The strength of this hypothesis is not its terminology. It is its falsifiability.

If the Fed simply speaks less, publishes less, and leaves a thinner record, the hypothesis fails. It is supported only if reduced forward signaling is accompanied by stronger reasoning, more traceable disagreement, and better ex-post review.

The essay’s treatment of artificial intelligence and productivity is similarly disciplined.

Rather than assuming that AI-driven growth is automatically inflationary or automatically disinflationary, the paper distinguishes between a demand shock and an improvement in supply capacity.

If AI investment first raises demand for energy, semiconductors, skilled labor, and financing while productivity gains arrive slowly, monetary pressure could be tighter.

If productivity and potential output rise materially, however, stronger growth need not imply overheating.

That distinction illustrates the paper’s broader method: do not infer policy from an aggregate number before understanding the mechanism that produced it.

The September rate discussion provides the clearest real-world test of that method.

As of September 1, 2026, the author leans toward no rate increase in September. Importantly,

however, the paper labels this explicitly as a scenario call, not as one of its institutional forecasts, and states the conditions under which that judgment must be reconsidered.

That separation is crucial.

Whether the September call proves right or wrong is one question.

Whether Warsh actually responds to changing facts in a manner consistent with the reaction function inferred from his public framework is another.

The two should not be graded as though they were the same prediction.

This is also where the essay makes its sharpest criticism of conventional central-bank forecasting.

A high PCE reading does not mechanically imply a rate hike.

The first questions should be: Why is PCE high? Energy? Housing? Wages? Services? Tariffs? Aggregate demand? Is the increase persistent? Will it generate second-round effects? Is the policy rate the appropriate instrument for the problem?

Only after the causal structure is understood does the forecast have a defensible basis.

In that sense, the paper is not anti-market and not anti-forecasting.

It is opposed to a particular form of forecasting in which markets guess the Fed from its rhetoric, the Fed observes the market reaction to that rhetoric, and both sides become increasingly dependent on signals partly generated by the other.

The alternative proposed here is simple but demanding: look at the facts, reconstruct the process, state the grounds for judgment, and allow later evidence to revise the conclusion.

For that reason, Can We Anticipate Warsh? is best understood not as an ordinary central-bank commentary, but as a pre-registered analytical baseline.

Read today, it is an interpretation of Warsh’s monetary-policy framework.

Read several months from now, it becomes a scorecard.

Read several years from now, it may become something more unusual: a record of what was knowable on September 1, 2026, before the later outcomes were known.

That is the essay’s deeper contribution.

Its most important claim is ultimately not that it can anticipate Kevin Warsh.

It is that a serious forecast should make itself vulnerable to future facts.

One correct call proves little. Only a repeated sequence of future actions consistent with the reaction function written down in advance would justify confidence in the model.

The first object being tested, therefore, is not Warsh.

It is the analysis itself.