Discipline, Not a Decision

A Bayesian/Game-Theoretic Reading of Kevin Warsh's Jackson Hole Address, the Corrected Empirical Record, and a 2026–2030 Reaction-Function Framework

Revised and expanded analysis — August 28, 2026

Preface to the Revision

The version of this paper prepared before Chairman Warsh delivered his remarks anticipated, correctly, that he would resist conventional forward guidance while still communicating a great deal about his priorities. What the earlier draft could not know is precisely how he chose to resolve that tension, because he chose to resolve it more starkly than anticipated: rather than replacing calendar guidance with an implicit reaction function, Warsh explicitly declined to offer one, telling the audience at the Kansas City Fed's Jackson Hole Economic Policy Symposium that a published rule for policy "works better in theory than in practice," and that the Federal Reserve's models are not yet reliable enough to bind the Committee to one. His own formulation was that he stands "committed to a discipline, not to a decision."

This revision corrects that record, incorporates the market's and the data's response through the close of trading on August 28, 2026, and reframes the whole argument in explicitly Bayesian and game-theoretic terms, including a formal reaction-function framework spanning Warsh's term through 2030 and probability-weighted scenarios for inflation, the September 15–16 Federal Open Market Committee decision, Treasury yields, and the dollar. All data below are drawn from the Federal Reserve, the Bureau of Economic Analysis, the Bureau of Labor Statistics, the U.S. Treasury, and contemporaneous wire and financial reporting, cited by source and date throughout, with full references at the end.

I. What Warsh Actually Said: The Corrected Record

Warsh's address, titled "In Our Time" and delivered at 10:00 a.m. Eastern time on Friday, August 28, at the Kansas City Fed's symposium in Moran, Wyoming, was his first appearance at Jackson Hole as chair, and only his second major public statement on monetary policy since the Federal Open Market Committee's July 28–29 meeting, at which the Committee held its target range at 3.50 to 3.75 percent over three dissents in favor of a quarter-point increase — the first instance of three same-direction dissents on the Committee in roughly a decade (CNBC, Kiplinger).

Two features of the speech matter most for the analysis that follows. First, Warsh explicitly rejected both forward guidance and a published reaction function, arguing that a rule-based commitment to a specific interest-rate path "works better in theory than in practice" because the Fed's understanding of the economy's underlying structure is still incomplete. He framed this not as evasion but as candor about the limits of the Committee's models, while promising that he and his colleagues will work over his term to "construct more reliable models and more robust rules to guide policy decisions" (CNBC, Bond Buyer). Second, and in some tension with that restraint, Warsh was unambiguous about his diagnosis of the current conjuncture: inflation is too high, the labor market is at full employment, and financial conditions are not restrictive. On inflation specifically, he stated that "while this summer's PCE and CPI readings were better than expected, they do not tell me that underlying trends have meaningfully improved," and that "we must be confident that underlying inflation is moving to our objective, clearly and at sufficient speed. Otherwise, we have work to do" (CNBC). He separately said he "would be hard pressed to describe broad financial conditions as restrictive" (Bond Buyer).

Warsh grounded that judgment in figures he cited directly: personal consumption expenditures (PCE) inflation running at 3.7 percent over the past twelve months and at a 4.1 percent annualized pace over the past six months — a marked acceleration in the underlying trend relative to the trailing-year figure — against a 2 percent objective he repeatedly called "a firm, fixed target" (Benzinga, Bloomberg). He also devoted a substantial section of the speech, under the heading "Preparing for Future Policy Conjunctures," to artificial intelligence, describing business capital expenditure in equipment and intangibles as growing at roughly 9 percent over the past four quarters — its fastest pace since 2021 — with more than half of that growth attributable to the AI buildout; S&P 500 profits up more than 20 percent over the past year; corporate credit spreads near the low end of their historical range with strong issuance; and annualized token sales at the two leading AI labs exceeding $100 billion, up more than 500 percent in a year (Federal Reserve, Euronews, Decrypt).

The upshot is that the earlier draft's characterization of Warsh as having replaced calendar guidance with a "highly recognizable policy reaction function" overstated what he did. He did not commit to a rule. What he offered instead was closer to what game theorists would call a stated preference ordering without a committed strategy: a clear statement of what outcome he wants (inflation convincingly returning to 2 percent), combined with a refusal to specify the policy function that maps future data realizations into future rate decisions. That combination — clear objective, opaque function — is the single most important structural fact about the Warsh Fed, and it is the organizing problem for the rest of this analysis.

II. The Market's Answer: Warsh Cannot Escape the Reaction-Function Problem by Declining to State One

The market treated the speech as decisively hawkish despite Warsh's explicit refusal to guide. According to CME Group's FedWatch tool, the probability traders assigned to a 25-basis-point hike at the September 15–16 meeting jumped from roughly 34–35 percent on Thursday, August 27, to approximately 57.5 percent by Friday's close — essentially a coin flip that tilts toward tightening (CNBC). Prediction markets moved even further: Polymarket odds of a 2026 rate hike rose to as high as 69 percent, up from the mid-50s the day before, while Kalshi traders assigned roughly 48 percent odds to a September hike specifically (Benzinga, CNBC). The policy-sensitive two-year Treasury yield rose approximately 8 to 9 basis points to about 4.31–4.32 percent, its highest level since late July, while yields further out the curve were comparatively little changed — evidence, discussed further below, that the market repriced near-term policy rather than the long-run inflation regime (CNBC, Benzinga, Investing.com). The U.S. Dollar Index gained roughly 0.4 to 0.5 percent to around 99.50, putting the dollar on pace for a weekly gain of nearly 1 percent that erased the prior week's "debasement trade" losses (FXStreet, Investing.com). Equities were more ambivalent: the S&P 500 slipped modestly on the day even as it closed out a positive week overall, a sign that investors absorbed the hawkish signal without a disorderly repricing of risk assets (CNBC).

This is the empirical core of the paradox this paper is built around, and the corrected record sharpens rather than dissolves it. Warsh said, in substance, that he will not tell markets his reaction function because the Fed's models are not good enough to bind him to one. Markets responded by constructing a reaction function for him anyway, inferring from his inflation diagnosis and his "work to do" language that a September hike is now more likely than not priced at even money. In Bayesian language, Warsh supplied a strong signal about his prior beliefs — that the current data represent adverse news relative to the Fed's inflation objective — without supplying the likelihood function that would map future data into a specific policy response. Markets filled that gap themselves, using their own model of Warsh's reaction function, which is precisely the "hall-of-mirrors" dynamic Warsh has warned against: the Fed's words shape market prices, and those prices will now be read by the Fed, and by outside observers, as if they were an independent read on the economy, when a meaningful share of the move is mechanically downstream of the chairman's own rhetoric.

The strategic logic here is closer to a signaling game than to conventional forward guidance. Warsh's speech functions as a costly signal of type — a hawkish policymaker revealing information about his own preferences and tolerance for inflation risk — without committing to an action rule contingent on future states of the world. Markets, as rational Bayesian receivers, updated their posterior over Warsh's type (more hawkish than previously priced) and, absent an explicit reaction function, substituted the modal type-consistent action (a hike) as their best guess of what a hawkish Warsh does when data are ambiguous. That substitution is exactly what a game-theoretic treatment would predict when a principal reveals type but withholds strategy: the agents she is signaling to will impute the strategy most consistent with the revealed type. Warsh's stated goal of a "quieter" Fed is, on this reading, unlikely to be achieved by withholding reaction functions alone, because market participants will construct and price an inferred reaction function regardless of whether the Fed publishes one. The chairman's own words on this point—describing his approach as a "discipline, not a decision"—capture the intended posture, but the market's job is to price decisions, and it will use whatever discipline it can observe as a proxy for the decision rule.

III. The Corrected Data Record Through August 28, 2026

On inflation, the July Personal Consumption Expenditures report, released by the Bureau of Economic Analysis on August 26, showed headline PCE inflation at 3.7 percent year over year and core PCE (excluding food and energy) at 3.3 percent, both firmer than consensus expected; the monthly core reading has now printed 3.3, 3.4, 3.3, and 3.3 percent in April through July, which is the four-month stall in underlying disinflation that Warsh's six-month annualized 4.1 percent figure was designed to highlight (Bureau of Economic Analysis; Yahoo Finance/BEA data). The July Consumer Price Index, released by the Bureau of Labor Statistics on August 12, showed headline CPI up 3.4 percent year over year and core CPI up 2.5 percent, both a tenth of a point softer than June, with shelter accounting for roughly two-thirds of the monthly increase and energy prices volatile amid the ongoing Middle East conflict (Bureau of Labor Statistics; CNBC; NBC News).

On employment, the July jobs report showed nonfarm payrolls declining by 23,000 — a sharp reversal from the roughly 34,000 average monthly gain over the prior year — while the unemployment rate held at 4.1 percent (Bureau of Labor Statistics; Wichita Liberty/BEA-BLS synthesis). A separate annual benchmark revision to the payroll count, timed to coincide with the run-up to Jackson Hole, registered a further modest downward adjustment to the level of employment, reinforcing the sense among several FOMC participants that labor-market momentum has cooled even as the unemployment rate itself remains low (FXStreet, "Forex Today"). Warsh's own reading, delivered in the speech, was that the labor market remains "consistent with full employment," even as he called the inflation side of the mandate "more concerning" — an explicit statement of asymmetric concern across the dual mandate that matters for the reaction-function reconstruction in Section VI (Investing.com).

On financial conditions, Warsh's own remarks — not the earlier draft's secondhand characterization — are the best evidence: he told the audience that corporate credit spreads are "near the low end of their historical averages" with strong issuance volumes, that business capital expenditure in equipment and intangibles is growing near a five-year high, and that S&P 500 profit margins are "quite elevated, relative to history" (Federal Reserve transcript). Those are Warsh's own words, not an outside inference, and they support his conclusion that policy is not obviously restrictive by the normal channels through which monetary tightening is supposed to bite.

On the fiscal side, the Treasury's regular August refunding, announced by the Department earlier in the month, will raise $125 billion through new three-year, ten-year, and thirty-year issuance — $58 billion, $42 billion, and $25 billion respectively — to refund roughly $96.3 billion of maturing debt, raising about $28.7 billion in new cash (U.S. Department of the Treasury). More consequential for the fiscal-monetary interaction this paper emphasizes was a surprise announcement on August 19: Treasury Secretary Scott Bessent said the Department would at least double the size of its buyback operations in longer-dated nominal coupon securities — the 10-to-20-year and 20-to-30-year sectors — from a $2 billion to a $4 billion maximum per operation, beginning September 9 and running through the November 4 refunding (U.S. Department of the Treasury; CNBC; Council on Foreign Relations). Treasury officials have also indicated Bessent could draw on a Treasury General Account that has been built up to roughly $950 billion to fund a larger buyback program if he judges conditions warrant it (CNBC). Reuters and CNBC reporting both describe this as an unusual departure from Treasury's long-standing "regular and predictable" issuance doctrine, prompted by a multi-week "buyers' strike" in the long end of the curve; the initial announcement briefly pushed the ten-year yield down 5.7 basis points and the thirty-year down 9 basis points before skepticism about Bessent's actual firepower reasserted itself (CNBC).

On energy and geopolitics, the conflict between the United States, Israel, and Iran remains the through-line affecting the inflation data Warsh is reading. Brent crude has risen more than 38 percent in 2026 amid recurring disruption to transit through the Strait of Hormuz, with national average U.S. gasoline prices oscillating in the $4.00-to-$4.10-per-gallon range over the summer (CNBC/AAA data; TradingEconomics). The July FOMC minutes record that "developments over the intermeeting period were influenced by the conflict in the Middle East," that oil prices rose as a result, but that "inflation compensation moved little in response to higher oil prices" — evidence the Committee itself has already begun using to distinguish a supply-side energy shock from a more troubling de-anchoring of inflation expectations (Federal Reserve, FOMC Minutes, July 28–29, 2026).

IV. The Forward-Guidance Paradox, Restated

The original critique of Warsh — that his rejection of forward guidance was itself a form of forward guidance — required amendment once his actual remarks were available, but the underlying insight survives in a sharper form. Warsh did not, as the earlier draft assumed, construct an implicit rule. He constructed something game theorists would recognize as a partially separating signal: enough information to reveal that his type is hawkish relative to market expectations, but not enough to specify the mapping from future data to future action. The paradox is therefore not that his silence was itself guidance — it demonstrably was not, in the narrow sense of specifying a rule — but that his silence was informationally indistinguishable from guidance in its market effect, because markets do not require an explicit rule to trade a hawkish signal; they only require a credible update to their estimate of the chairman's reaction function, which his tone, his selective statistics, and his "work to do" language supplied in abundance.

This distinction matters for how the framework should be built going forward. A model of "Warsh forward guidance" that treats his speech as announcing a rule would be a category error; what should instead be modeled is Warsh as a Bayesian type whose preferences are becoming better known to the market with each communication, while his policy function — the mapping from data to decisions — remains deliberately unspecified and is instead inferred, meeting by meeting, from incoming data and from the chairman's evolving rhetoric. That is a fundamentally different — and, in an important sense, more honest — way to run a central bank than either Powell-era "data dependence" rhetoric or classical Taylor-rule commitment. But it is not obviously more stabilizing, because it pushes the burden of constructing the reaction function onto the market itself, which will do so noisily and with its own behavioral biases, including a tendency to overweight the most recent, most quotable line from the chairman.

V. The Lucas Critique, Applied to Central-Bank Communication Itself

Warsh's own instinct — that publishing a rule invites a form of the Lucas critique, in which private agents adapt their behavior to the announced rule and thereby undermine the historical relationships the rule was built on — is defensible in principle. Historical estimates of the relationship between the federal funds rate and inflation outcomes were generated under a regime of increasingly explicit forward guidance; there is no guarantee those relationships are stable if the Fed reverts to a discretionary regime, or vice versa. But Warsh's proposed solution — discretion without a stated rule — does not escape the deeper problem, because market expectations do not require an official rule to form; they only require data and communication, both of which Warsh continues to supply liberally. The practical question this paper's framework must answer is therefore not guidance versus no guidance, but how the Fed can supply enough structure to keep market-inferred reaction functions anchored to fundamentals, rather than to the volatility of the chairman's most recent public appearance.

VI. A Formal Reaction-Function Reconstruction, 2026–2030

Because Warsh has declined to publish a rule, any reaction function attributed to him is necessarily a reconstruction from revealed preference rather than a stated commitment, and it should be treated as a probabilistic model of his behavior rather than a description of Fed policy. Four structural features of the reconstructed function can be read directly out of the evidence in Sections I through III, described here in words rather than as an explicit formula, consistent with this paper's standing convention of avoiding mathematical notation.

First, Warsh weights the inflation leg of the dual mandate more heavily than the employment leg at the current juncture, and has said so explicitly by calling inflation "more concerning" while describing the labor market as consistent with full employment. A reconstructed Warsh function should therefore treat further softening in payrolls as a much weaker trigger for easing than persistent core PCE readings above roughly 3 percent would be as a trigger for tightening — an asymmetric, inflation-dominant reaction function, at least for as long as unemployment stays near its current low level.

Second, Warsh privileges the level and trend of underlying, rather than headline, inflation, and has revealed a preference for six-month annualized core measures — hence his choice to cite the 4.1 percent six-month PCE figure rather than the more favorable twelve-month reading — as the metric most likely to move him toward tightening. A reconstructed function should weight the most recent two to three months of core data more heavily than the trailing-twelve-month comparison the earlier draft of this paper implicitly used.

Third, Warsh treats financial-conditions looseness — tight credit spreads, strong equity performance, robust capital expenditure — as evidence that the current policy rate is not doing much restraining work, independent of what the federal funds rate itself is doing. This is consistent with a policymaker who would be willing to raise rates even without further inflation deterioration, on the theory that the existing rate has simply not been transmitted into tighter conditions. A reconstructed function should therefore include a financial-conditions term that can push toward tightening even when the inflation and employment data are merely stable rather than worsening.

Fourth, and this is the most novel feature relative to the pre-Warsh Fed, artificial intelligence enters the reconstructed function as a source of genuine two-sided uncertainty about the economy's noninflationary growth capacity rather than as a straightforward hawkish or dovish input. Warsh's own language — describing AI as a potential "new factor of production" whose productivity effects the Fed "cannot answer yet" — suggests he is deliberately keeping this variable out of the near-term reaction function while a new Fed task force on productivity and jobs examines it, with findings not expected before year-end and explicitly stated to have no bearing on near-term decisions (Euronews).

Extending this reconstruction across Warsh's likely term — a chairmanship that began in May 2026 and would run, on the standard four-year term, into 2030 — requires treating the four features above as reasonably stable priors about Warsh's type, updated meeting by meeting as new data arrive, rather than as a fixed rule. The most defensible forward-looking statement is therefore not a prediction of specific rate levels through 2030, but a statement about the character of the regime: a Fed that will remain inflation-dominant in its revealed weighting for as long as unemployment stays in the low fours, that will resist publishing a rule even as its five internal task forces (on communications, balance-sheet policy, data, productivity and jobs, and inflation frameworks) report out through late 2026 and into 2027, and that will treat each Jackson Hole address, each FOMC statement, and each significant Warsh public appearance as a fresh, freshly-informative signal about his evolving reaction function rather than as confirmation of a previously announced path — precisely because he has foreclosed the latter.

VII. Probability-Weighted Scenarios: The September 2026 FOMC Decision

Treating the September 15–16 meeting as the first live test of the reconstructed reaction function, three scenarios can be assigned rough, evidence-consistent probability weights as of August 28, 2026, understanding that these weights will move with each subsequent data release, most importantly the August employment report and the August CPI report, due September 11, just days before the meeting.

The modal scenario, carrying perhaps 45 to 55 percent weight given the roughly 57.5 percent CME-implied probability of some hike but the absence of any Warsh commitment, is a 25-basis-point increase to a 3.75-to-4.00 percent target range, delivered alongside an unusually candid Summary of Economic Projections given that September is one of the Fed's four dot-plot meetings. This outcome is most likely if the August employment report shows stabilization rather than further deterioration and if August CPI or the incoming core PCE trend fails to show the deceleration Warsh said he has not yet seen. A second scenario, carrying perhaps 35 to 40 percent weight, is a hold at 3.50 to 3.75 percent accompanied by an explicitly hawkish statement and dot plot signaling a high probability of tightening at the October or December meeting — in effect, the Committee using the statement language rather than the rate itself to manage expectations, consistent with Warsh's general preference for judgment over pre-commitment. This scenario becomes more likely if the labor-market data deteriorate further, giving the more dovish members of the Committee, several of whom dissented toward easing earlier in the cycle before the Iran-related energy shock complicated that view, more room to argue for patience. A third, low-probability scenario, weighted at perhaps 10 to 15 percent, is a hold with no clear hawkish signal, which would most likely follow either a sharp downside surprise in employment or a ceasefire-driven collapse in oil prices materially undercutting the inflation case Warsh built his speech around; given the fragile, on-and-off character of the Iran conflict, an abrupt de-escalation cannot be ruled out and would materially soften the near-term inflation read Warsh emphasized (Federal Reserve, FOMC Minutes; CNBC).

VIII. Probability-Weighted Scenarios: Inflation Through Year-End

On core PCE inflation, the most probable path over the next two prints, carrying a weight of roughly 50 percent, is continued stagnation in the 3.2-to-3.4 percent range, consistent with the flat four-month run already observed and with an energy shock that is real but, per the July FOMC minutes, not yet visibly bleeding into inflation compensation. A second scenario, weighted at roughly 25 to 30 percent, is a further modest acceleration toward or above 3.5 percent, most likely to occur if the six-month annualized trend Warsh emphasized proves to be the leading indicator rather than a statistical artifact, or if a renewed escalation in the Iran conflict pushes oil and gasoline prices materially higher than the already-elevated levels observed through the summer. A third, more optimistic scenario, weighted at roughly 20 to 25 percent, is a resumption of disinflation toward 3.0 percent by year-end, most plausible if shelter costs — which have accounted for roughly two-thirds of recent monthly CPI gains — continue to ease as they showed some signs of doing in the July CPI report, and if energy prices stabilize or retreat on a durable, rather than temporary, reduction in Middle East tensions.

IX. Probability-Weighted Scenarios: Treasury Yields and the Dollar

For the two-year Treasury yield, now trading near 4.31 to 4.32 percent, the most probable near-term path, weighted at roughly 50 percent, is a range-bound drift between roughly 4.20 and 4.45 percent through the September meeting, with the outcome of that meeting itself the dominant catalyst for a breakout in either direction. A hawkish September outcome — either an actual hike or an explicitly hawkish hold — would be consistent with a push toward or above 4.50 percent, a scenario this paper weights at roughly 30 percent given the market's already-substantial repricing since Friday. A dovish surprise, weighted at roughly 20 percent, would likely pull the two-year back toward 4.00 to 4.10 percent. Longer-dated yields are governed by a more genuinely two-sided process, because the ten-year and thirty-year sectors are now simultaneously exposed to Warsh's inflation signaling and to Bessent's buyback program, whose first operations begin September 9. Because the two forces point in different directions — hawkish Fed signaling pushes long yields up through higher expected short rates and inflation compensation, while larger Treasury buybacks push long yields down through reduced net supply and improved liquidity — the most probable outcome for the ten-year, weighted at perhaps 45 percent, is a comparatively contained trading range relative to the more volatile two-year sector, with the two forces substantially offsetting one another through the fall. A scenario in which Bessent's buybacks prove insufficient relative to financing needs, allowing the fiscal-supply effect to dominate and push the ten-year higher despite the offsetting purchases, carries perhaps 30 percent weight, consistent with Evercore ISI's public skepticism that the buyback program, at $4 billion per operation against roughly $5.5 trillion in outstanding 20- and 30-year debt, has the scale to durably move the market (CNBC). A scenario in which the buyback program, potentially reinforced by draws on the nearly $950 billion Treasury General Account, meaningfully compresses term premium carries the remaining roughly 25 percent weight.

For the dollar, now at a two-week high near 99.50 on the Dollar Index after Friday's roughly half-percentage-point gain, the modal scenario, weighted at roughly 45 percent, is continued firmness into September as the market keeps a hawkish Fed repriced, particularly if the European Central Bank proceeds with a widely-expected September rate increase of its own — which would partially offset dollar strength through relative-rate channels — and if the Bank of England remains constrained by weaker growth expectations (FXStreet, Yahoo Finance). A scenario in which the dollar gives back a substantial share of Friday's gains, weighted at roughly 30 percent, would most likely follow either a September hold or data between now and the meeting that undercut the inflation narrative. A tail scenario, weighted at roughly 25 percent, involves a return of the "debasement trade" — the broad skepticism about U.S. fiscal sustainability and Fed independence that pressured the dollar and lifted gold in the week before Jackson Hole — reasserting itself if Bessent's fiscal interventions are read by markets as evidence of Treasury pressure on the Fed rather than as ordinary debt management, a risk this paper's fiscal-dominance discussion in Section X treats as understated in the original draft.

X. Fiscal Dominance and the Treasury-Fed Interface, Updated

The original draft's distinction between Treasury debt issuance and central-bank money creation remains analytically correct and worth preserving: Bessent's operations do not mechanically expand the money supply in the way the earlier, cruder critique implied. But the empirical record through August 28 makes the fiscal-monetary interaction considerably more concrete and more urgent than the original draft's abstract treatment suggested. Bessent's decision to announce a surprise doubling of long-bond buybacks on August 19 — bypassing the normal quarterly refunding channel for such changes, according to Treasury officials themselves — and his acknowledgment on CNBC that operations could run even larger than the new stated minimum, both signal an activist Treasury willing to intervene directly in the long end of the curve at the same time the Fed chair is signaling tighter near-term policy (CNBC). Citi's own commentary ahead of the Jackson Hole speech explicitly flagged that investors would be watching "whether his views would align with those of Treasury Secretary Scott Bessent," and that greater apparent coordination could itself compress long-term yield volatility — precisely the kind of signal a Bayesian market participant should treat as information about the joint fiscal-monetary regime, not simply about the Fed in isolation (yournews.com/Citi commentary).

This is the clearest evidence yet that the fiscal-dominance risk this paper's framework must price is not a distant, hypothetical concern but an active, live dynamic: a Treasury running a near-$1 trillion cash buffer and an enlarged buyback program, operating alongside a Fed chair who has explicitly declined to coordinate his communications with the Treasury's own signals — Seeking Alpha's pre-speech commentary noted analysts did not expect Warsh to comment on Bessent's buyback plans at all, for fear of appearing "political" (Seeking Alpha) — is a textbook setting for the market to misread a debt-management signal as a monetary signal, or vice versa, exactly the confusion Section IX's yield-scenario analysis above is built to price around rather than assume away.

XI. AI as a Genuine Two-Sided Bayesian Variable

Warsh's own numbers substantially strengthen, rather than merely restate, the AI section of this paper's original analysis. A four-quarter capital-expenditure growth rate of roughly 9 percent, the fastest since 2021 and more than half attributable to AI infrastructure; S&P 500 profit growth above 20 percent; and annualized AI token sales at the two largest labs exceeding $100 billion, up over 500 percent in a year, together describe an investment boom of a scale that plausibly does shift the economy's noninflationary growth capacity, exactly as the productivity-shock hypothesis in the original framework proposed (Federal Reserve; Decrypt; Euronews). But the same data are equally consistent with the asset-price-boom hypothesis: concentrated gains, a single chipmaker's ability to both supply and finance a large share of its own customer base, and reporting that a large majority of generative-AI companies remain commercially unsuccessful even as capital continues to pour into the sector, all point toward the possibility that today's elevated profit margins and tight credit spreads partly reflect speculative enthusiasm rather than realized productivity gains (Decrypt). Warsh's own posture — creating a dedicated task force on productivity and jobs, explicitly stating its findings will not affect near-term policy, and continuing to "watch the change in growth rates, the second derivative" of both capital expenditure and earnings — is itself consistent with a policymaker treating this as a genuinely unresolved Bayesian hypothesis rather than a settled input, which is the correct posture and one this paper's framework should adopt as well, resisting the temptation to resolve the AI question in either direction before the evidence supports doing so.

XII. Energy, Iran, and the Second-Round Inflation Question

The Iran-Israel-U.S. conflict remains the paper's most important recurring geopolitical variable and continues to complicate the inflation read at the center of Warsh's speech. Brent crude's more than 38 percent rise in 2026, driven by recurring disruption to Strait of Hormuz transit, and national gasoline prices that have oscillated near $4.00 to $4.10 per gallon through the summer, are the proximate cause of the energy volatility embedded in both the CPI and PCE releases discussed above (CNBC/AAA; TradingEconomics; NBC News). The July FOMC minutes' explicit finding that inflation compensation "moved little" despite the oil-price increase is the single most important piece of evidence available to the Committee that this remains, for now, a first-round supply shock rather than a second-round de-anchoring of expectations — precisely the distinction this paper's Bayesian framework treats as the central open question for energy-driven inflation (Federal Reserve, FOMC Minutes). The appropriate policy variable, as in the original draft, remains the probability that the energy shock becomes embedded in wage- and price-setting behavior rather than the oil price itself, and the July minutes' finding is evidence — not yet conclusive, but genuinely informative — that this probability remains low as of late July, a reading the Committee will need to revisit before September 16 given the conflict's continued, unresolved character.

XIII. Final Assessment

The corrected record sharpens rather than resolves the central tension this paper set out to examine. Kevin Warsh did not, as initially reported and analyzed, construct an implicit reaction function to replace explicit forward guidance. He explicitly declined to construct one at all, on the stated grounds that the Fed's models remain too unreliable to bind the Committee to a rule, while nonetheless communicating an unmistakable diagnosis — inflation too high, labor market fully employed, financial conditions too loose — that markets have priced with nearly as much conviction as if he had announced a rule outright. The gap between what Warsh intended to communicate and what markets extracted from his remarks is itself the most important data point in this analysis, because it demonstrates that a central bank cannot escape the burden of managing expectations simply by declining to publish a formal reaction function. Markets will construct one regardless, from whatever signal is available, and the quality of that market-constructed function — whether it tracks the Fed's true, still-undisclosed priorities or diverges from them — is now substantially outside the Fed's direct control.

Four structural features of the Warsh Fed, reconstructed here from revealed preference rather than stated commitment, should anchor expectations through the remainder of his term: an inflation-dominant weighting of the dual mandate for as long as unemployment stays low; a preference for near-term, six-month annualized core inflation measures over trailing-twelve-month comparisons; a willingness to treat loose financial conditions as an independent argument for tightening; and a deliberately unresolved, two-sided treatment of artificial intelligence as a productivity variable that could shift the neutral rate in either direction. Layered onto this reconstructed monetary reaction function is an increasingly assertive fiscal actor — a Treasury Secretary now conducting surprise interventions in the long end of the Treasury curve with a nearly $1 trillion cash buffer behind him — whose signals markets must learn to separate from the Fed's own, and an unresolved geopolitical energy shock whose second-round inflation consequences remain, on the best evidence available through July, contained but not resolved.

The September 15–16 FOMC meeting is therefore best understood not as a single up-or-down test of Warsh's credibility, but as the first of what will be a long sequence of Bayesian updates, in which each employment report, each inflation print, each Bessent buyback operation, and each subsequent Warsh public appearance further refines — without ever fully revealing — the reaction function this paper has attempted to reconstruct. The discipline Warsh has committed to is real. Whether it produces better decisions than the forward guidance it replaces will not be knowable from this speech alone. It will be knowable only from the pattern of decisions the Committee actually makes as 2026 gives way to 2027, and as the five internal task forces Warsh has commissioned report their findings into a policy framework that, as of August 28, 2026, remains explicitly and deliberately unfinished.

References
Federal Reserve Board, "Keynote remarks by Chairman Warsh at the 2026 Jackson Hole Economic Policy Symposium," August 28, 2026, https://www.federalreserve.gov/newsevents/speech/warsh20260828a.htm
Federal Reserve, Minutes of the Federal Open Market Committee, July 28–29, 2026, https://www.federalreserve.gov/monetarypolicy/fomcminutes20260729.htm
Federal Reserve, Minutes of the Federal Open Market Committee, June 16–17, 2026, https://www.federalreserve.gov/monetarypolicy/fomcminutes20260617.htm
U.S. Bureau of Economic Analysis, "Personal Income and Outlays, July 2026," released August 26, 2026, https://www.bea.gov
U.S. Bureau of Labor Statistics, "Consumer Price Index News Release," July 2026 results, released August 12, 2026, https://www.bls.gov/news.release/cpi.htm
U.S. Bureau of Labor Statistics, Employment Situation, July 2026
U.S. Department of the Treasury, "Quarterly Refunding Statement of Deputy Assistant Secretary for Federal Finance Brian Smith," August 2026, https://home.treasury.gov/news/press-releases/sb0590
U.S. Department of the Treasury, "Treasury Announces Increased Sizes of Nominal Long-End Liquidity Support Buybacks Beginning September 9," August 19, 2026, https://home.treasury.gov/news/press-releases/sb0607
CNBC, "Fed Chairman Kevin Warsh delivers his key Jackson Hole speech Friday. Here's what to expect," August 27, 2026
CNBC, "Kevin Warsh Jackson Hole federal reserve inflation," August 28, 2026
CNBC, "September Fed decision is now a coin flip as rate hike odds increase post Warsh," August 28, 2026
CNBC, "Stock market today live updates," August 27–28, 2026
CNBC, "Bessent could tap near $1 trillion Treasury General Account to fund bond buybacks," August 24, 2026
CNBC, "Treasury announces upscaled buyback operation for longer-term debt, sending yields lower," August 19, 2026
CNBC, "CPI inflation report July 2026," August 12, 2026
CNBC, "Fed meeting today live updates," July 29, 2026
Bond Buyer, "Markets see rate hike coming following Warsh's speech," August 28, 2026
Benzinga, "Fed September Hike Odds Spike to 59% After Warsh's Jackson Hole Speech," August 28, 2026
Euronews, "Warsh flags inflation concerns as he rejects Fed forward guidance," August 28, 2026
Decrypt, "Fed Chair Warsh Calls AI a 'Hinge Point in History'—4 Key Things He Said," August 28, 2026
FXStreet, "US Dollar Index rallies as Warsh puts September rate hike back on the table," August 28, 2026
FXStreet, "Forex Today: Fed Chair Warsh speech at Jackson Hole, NFP benchmark revision to lift volatility," August 28, 2026
Investing.com, "Dollar firms on hawkish Warsh at Jackson Hole, heads for weekly gain of nearly 1%," August 28, 2026
Council on Foreign Relations, "What the Treasury's Buyback Surprise Says About the Bond Market," August 20, 2026
Seeking Alpha, "Warsh's First Jackson Hole Speech," August 27, 2026
NBC News, "July 2026 CPI report: Inflation remained stubborn," August 12, 2026
Yahoo Finance / Wichita Liberty, syntheses of BEA and BLS July 2026 releases, August 26–28, 2026
yournews.com, "Dollar Holds Near One-Week High as Markets Await Warsh's Jackson Hole Remarks," August 28, 20

Translate

Friday, 28 August 2026