September’s cost-of-living report put the annual inflation rate at 3.7 percent, a larger increase than the month before, with the spike in energy prices reported as a big driver. The question before the Federal Reserve’s December meeting is different: what a rate increase would do to that price path. The reporting supplies the observation; the decision requires an interventional fact. The gap between the two is the analytical problem, and it is visible throughout the material — in the committee’s internal division, in the market’s December bet, and in the declining purchasing power that Mark Zandi of Moody’s Analytics describes among lower- and middle-income Americans.

Causal structure

The reporting supports a specific causal structure, and the shape of that structure determines what the data can prove. The energy component traces to an external shock: Dan North of Allianz Trade characterizes the energy-price spike as “getting worse” since the war with Iran started eight months ago and describes a “wrong way” sign on his charts — “That inflation line is going up, not down.” The demand side runs through heavy investment in artificial intelligence, which NPR’s Scott Horsley reported is contributing to higher inflation and that “no amount of tapping on the brakes by the Fed seems likely to slow,” and through consumer spending that has so far been resilient. Between them sits the channel the central bank can operate: the federal funds rate, which policymakers raised at their September meeting and which reaches prices through demand and through the housing market — a sector Horsley described as in a “really deep, long-lasting slump,” with mortgage rates north of 7 percent and with housing costs that reach the price index with a lag.

In this structure, the war is the exogenous shock, energy prices mediate its effect on the index, the rate is the candidate intervention, demand and housing are the transmission channels, inflation is the outcome — and, through the policy reaction function, a cause of the rate itself. That last arrow is the decisive feature. The committee does not observe inflation as a neutral experimenter; it responds to it. Most members have signaled they expect interest rates to go higher this year, a prior hold drew dissents from members who wanted an increase, and the September increase was unanimous — the record of a policy rule in action. Because policy responds to inflation, the rate and the price index are jointly determined, and a reading of their co-movement is an association, not an effect.

In the terms of the causal literature, the September report answers an observation-level question: what does observing a 3.7 percent annual increase tell us about the state of prices? The December decision poses an intervention-level question: what happens to the price path if the committee sets the rate higher? Observational data carry an interventional answer only under identifying assumptions, and the relevant ones fail here. The reaction function and the unmeasured shocks that move both prices and policy open backdoor paths from the rate to inflation, and the backdoor criterion is not met with the observed variables. The standard empirical strategy for isolating monetary-policy effects — the surprise component of policy announcements as priced in fed funds futures immediately around their release — is not supplied. The one source of quasi-exogenous variation the article does describe, Horsley’s report that investors are betting the committee will “keep a low profile” at Wednesday’s meeting with Election Day near and raise rates in December, fails the exclusion restriction for an instrument if the calendar also shapes expectations and prices directly, which is the claim any political-calendar reading of the decision would make.

Four absent-arrow assumptions carry that verdict: that the war reaches policy only through prices, that energy prices do not respond to the rate within the decision window, that market expectations of policy do not feed back into prices within the decision window, and that the calendar affects the price path only through the rate. The first two are consistent with the reporting. The third fails on the reporting’s own terms: investors are betting on a December hike, and a priced-in policy path moves financial conditions before the committee acts, so the expectations channel the market’s bet makes operative is itself a backdoor path — an additional reason the effect is not identifiable, rather than an assumption that can be met. The fourth is load-bearing and the least defensible: the calendar is common knowledge, and if it shapes expectations it shapes prices directly, in which case the post-election data cannot be read as a clean test of the committee’s reaction function. The verdict from the material in hand is an identification failure: the causal effect of a December hike on inflation is not identifiable from this record. What the two intervening months of data can do is update the observation-level question of whether “the upward trend in inflation continues” — the frame the reporting attributes to the committee itself.

Counterfactual claims are further still from the evidence. What the September print would have been had the committee raised rates earlier is a counterfactual question, and it demands a structural model of how rate changes propagate through demand, housing, and expectations. The reporting supplies no such model, and the price index alone cannot generate one. Whether any interventional estimate would license such a counterfactual is itself contested in the causal literature; the practical point here is that neither the interventional nor the counterfactual claim can be substantiated from the article’s material. What the September report can show is persistence — and persistence is exactly what the committee can measure, and exactly what it cannot act upon with tested effect. The energy shock driving the index is not, on North’s account, abating; it is “getting worse,” and a rate increase does not reach a war.

The decision as a negotiation

Read as a negotiation, the December decision has a structure that explains both its timing and its division. The relationship track is strained on two fronts. Internally, the committee is, as the article reports, “not necessarily on the same page about how” to bring prices under control: a prior hold drew dissents from members who wanted a rate hike, the September increase was unanimous, and Brian Mulberry of Zacks Investment Management anticipates “another 9-3 type of a vote,” possibly more divided, “based on people’s strong opinions of where the direction of inflation is going.” Externally, the relationship with markets is priced: investors are betting on a December hike, so a hold would be the surprise. And the relationship with the public runs through lived experience rather than aggregates. Zandi describes lower- and middle-income Americans as having “a real struggle,” with purchasing power actually declining; he says this “likely goes to explain to a large degree why people just don’t feel very good about the economy,” where they are “pretty upset.” Horsley expects “a pretty solid GDP report” for the third quarter while adding that it is little comfort “if you’re struggling to put gas in the car and groceries in the cupboard”: “you can’t eat GDP.” The FRED real gross domestic income series — the wage-income side of output, a measure the expected output figures do not themselves capture — stood at 32,752.528 on a raw basis as of Oct. 28, 2026, and supplies the documented gap between aggregate and lived measures.

The positions are narrow and known: hold at Wednesday’s meeting, hike or hold in December. The interests behind them are inferred from institutional mandates and the reporting rather than confirmed in the principals’ own words. The Wednesday hold serves two interests at once: institutional legitimacy — Horsley reported investors betting that the committee will try to keep a low profile with Election Day near — and information, since two more months of data will arrive before the December meeting. The December hike serves the credibility interest the members have voiced; Federal Reserve Chairman Kevin Warsh and his colleagues, the article reports, “have all said they are determined to bring prices under control.” Dissenting members’ interest in earlier action is documented in their recorded dissents. Households’ interest, on Zandi’s analysis, is purchasing power at a moment when, as NPR host Steve Inskeep noted in the interview, prices are growing more rapidly than the average worker’s wages. These are structural inferences, flagged as such.

The option space is wider than the binary suggests, because the communicative envelope is the real negotiation surface — the credibility channel runs through expectation formation, which is what makes the market’s pricing of a December hike not merely a prediction but part of the transmission mechanism. A December hike paired with an explicit decomposition of the inflation report — separating the war-driven energy component from the demand-driven component — would let the committee demonstrate resolve on the part of the index that interest rates can reach while acknowledging the limit of the instrument on the part they cannot, serving the credibility interest and the public-legitimacy interest in the same move. The risk is that a public decomposition reads as the committee carving the inflation target into rate-relevant and rate-irrelevant components — a split that, if the energy component persists, becomes a contested public object and erodes the credibility anchor the option is meant to protect. An announced decision rule linking December action to the two months of data would convert a binary surprise into a conditional commitment: markets receive the predictability they are pricing, and the committee receives visible resolve without premature commitment. On the objective-criteria side, the article supplies standards a December decision can be tested against without a contest of will: the published cost-of-living series itself; the GDI series capturing the wage side; mortgage rates north of 7 percent as evidence of transmission already at work; the precedent of the unanimous September vote; and market pricing as an aggregation of investor expectations. What the criteria cannot supply is the causal warrant — they are observation-level standards, and their limit is the same as the data’s.

The alternatives, assessed concretely, set the walk-away points. For the committee, the alternative to a December hike is a continued hold with hawkish language: its value is optionality and the avoidance of tightening into a housing slump; its cost is the credibility interest, since a hold against a continuing trend — and against a priced market expectation — must be read as consistency rather than contradiction. The substrate records the precedent that makes that reading available: the prior hold drew dissents and was followed by a unanimous hike — the data-dependent skip pattern in which a hold paired with forward guidance carries the direction without acting on it. The pattern, however, is retrospective: it is only as good as the action that follows it, and a hawkish hold that is not the prelude to December action defers rather than protects the credibility interest. For the dissenting minority, the documented alternative to consent is renewed public dissent, a tool already used at the prior meeting. For markets, the counterparty position is the pricing itself: the December hike is being bet on now, so a quiet hold without a data-based justification is the outcome the market is not positioned for. The counterparty, in other words, is not a negotiator making demands but a market aggregating expectations, and the committee’s credibility is the currency of settlement.

The probability of a December hike

The December decision can also be forecast transparently from a reference class. The class: FOMC pauses whose stated rationale is timing and data — an election and the need for two more observations — rather than a changed view of inflation; cases in which most members have signaled higher rates for the year, and in which markets price a hike for the meeting after the pause. The class is constructed from the reporting itself rather than drawn from a published database of comparable episodes, so its “base rate” is an analyst judgment, not a documented frequency. The class’s defining features set that prior: the pause’s rationale is satisfied by action at the next meeting, the members have already signaled the direction, and December is the last meeting within the window the members themselves referenced in saying they expect rates “higher this year.” The modal resolution, in other words, is the hike.

The inside-view signals in the article adjust the estimate in both directions. Toward the hike: North’s characterization of the energy trend as worsening points to the intervening reports reinforcing rather than reversing the trend; the investors’ December bet marks the market’s expectation; and the passage of the election removes the timing constraint the reporting identifies with the hold. Away from the hike: the war-driven component is not rate-sensitive in the short run, so a hike reaches only part of the reported increase; the housing slump indicates transmission already in place; and Zandi’s purchasing-power analysis documents a distributive cost concentrated on lower- and middle-income households. The pulls are offsetting but not symmetric — the timing rationale resolves in December by construction, while the composition concern does not. Starting from a majority base rate — the class definition and the market’s bet put the prior above 50 percent — an upward adjustment of roughly five to ten points for the trend-persistence and signaling evidence and a downward adjustment of roughly five to ten points for the composition and transmission constraints yield a range of roughly 60 to 70 percent for a December hike. The width is deliberate: the article supplies no market-implied figure, and the decisive data do not yet exist.

The update triggers can be stated in advance. If the two intervening cost-of-living reports show the annual rate at or above 3.7 percent, with the energy component persisting, the estimate moves toward the top of the range. If energy prices reverse — for example, on a change in the war’s trajectory — or the committee’s public communications soften, it moves down. The reference class also names the outcome most likely to be misread: a hawkish hold can initially be read as the standard data-dependent skip — the substrate’s own precedent of a hold, dissents, and a unanimous hike to follow — but that reading is retrospective; if December does not deliver the action it signaled, the hold converts into a contradiction of both the priced expectation and the members’ signaled direction, and that is the combination the market would have to reinterpret rather than confirm.

The December meeting is forecastable; its consequence is not, from this record. The data tell the committee what is happening to prices. The reference class tells it what it will probably decide. The causal structure tells it what it cannot know a hike will achieve. A December increase, if it comes, will be an act of resolve priced in advance and tested against two months of data that can establish persistence — but not the effect of the rate on the energy shock that is driving the index.

Analytical techniques used in this piece

This analysis applies the methods below. Each links to a short, plain-English explainer you can read and reuse.

Causal DAG
Maps cause and effect as an explicit directed graph, exposing confounders and mediators (Pearl).
Principled Negotiation
Works a negotiation from interests, options, and objective criteria rather than positions.
Probabilistic Forecasting
Puts calibrated probabilities on what happens next.