The 8-4 Print Markets Are Reading Wrong
The consensus take on the April 29-30 FOMC vote, the one circulating across desks and the financial press since the statement dropped, treats the 8-4 dissent count as a forecast disagreement. Bowman and Waller wanted 25bp of cuts because they read the labor cooling as ahead of schedule. Daly and Goolsbee dissented in the opposite direction because the March CPI print at 3.3% YoY and core services ex-housing running 4.1% annualized over three months told them the disinflation has stalled. Four dissents, two on each side, the highest dissent count in decades. The chart in every morning note shows a Fed split between two coherent camps, each with a defensible model, each waiting for data to vindicate them. Powell holds the median. The dot plot still shows 75bp of cuts by year-end. Fed funds futures at the CME are pricing roughly 60bp through December. Everyone goes home.
That reading is wrong, and it is wrong in a specific and expensive way. The 8-4 vote is not a forecast fight. The committee, if you read the March SEP carefully and listen to what the dissenters have actually said in speeches over the past six weeks, agrees on the destination within about 50bp. The terminal rate guesses cluster between 2.75% and 3.25% on a two-year horizon. Bowman's January Cato speech and Daly's April 2 SF Fed remarks describe almost identical end-states. What they disagree on is sequencing: who moves first, how fast, and which side of the dual mandate eats the political cost when the move is wrong. That is not a forecasting disagreement. That is a coordination problem, and coordination problems have a specific failure mode that the consensus narrative is missing entirely.
Coordination Problems Have a Failure Mode
A forecasting disagreement is resolvable through data. One side gets confirmation, the other capitulates, the median moves, the curve reprices in an orderly way. The April 2024 episode, when the doves quietly folded after three consecutive hot CPI prints, is the canonical example. Spread compression, breakeven repricing, MOVE index drifting back to the low 90s, life continues. A coordination problem does not resolve that way. In a coordination problem, every player's optimal move depends on what they expect the others to do, and the equilibrium can flip discontinuously when expectations shift. The textbook failure mode is well known to anyone who has read Cooper and John on coordination games or watched a bank run from the inside: the system can sit in a stable but suboptimal equilibrium for months, then jump to a new equilibrium in a single meeting once someone defects loudly enough to convince the others they will be the last to move.
The Fed in May 2026 sits in that stable-but-suboptimal pocket. Powell, Jefferson, Cook, and the four other median voters want to cut, but only if at least two of the regional dissenters cosign so the move looks like a committee decision rather than a Powell pivot. The hawks want to hold until core PCE prints below 2.6% for two consecutive months, which on current trajectory means September at the earliest, but they will defect to a cut sooner if they believe Powell is going to cut anyway and they want to shape the dovish language. The doves will tolerate a hold in May if they believe a June cut is locked in; they will dissent more aggressively if they suspect the June meeting will produce another hold. Four dissents in April is the tell. It is not noise around a consensus. It is the first visible crack in a coordination equilibrium that has been quietly fragile since the December 2025 SEP downgrade.
The specific failure mode here is what game theorists call equilibrium jumping. Markets price the current equilibrium, which is roughly: 25bp cut at the July meeting, two more by year-end, terminal at 2.875%. That path is consistent with fed funds futures, with the implied path from SOFR futures through Q2 2027, and with the TIPS breakeven curve which has 5y5y inflation pinned at 2.34%. If the coordination breaks, the path does not move gradually. It snaps to a new equilibrium, either a hold-and-hike path that the hawks have been quietly modeling or an emergency-50bp path that Waller hinted at in his April 18 Brookings remarks. The MOVE index at 88 is pricing neither. That is the trade.
Historical Pattern: 1991, 1995, 2022 - and What Followed
The historical record on Fed coordination breaks is short but consistent, and the consensus framing tends to forget that the August 1991 episode, the closest analog to what is happening now, was misread in real time in exactly the way April 2026 is being misread. Greenspan's FOMC voted 7-3 at the August 20, 1991 meeting, with Mullins, LaWare, and Forrestal dissenting in different directions, which at the time was treated by the WSJ and the Fed-watcher community as a sign of forecast disagreement about whether the post-Gulf War recovery was real. It was not. The Greenspan papers, declassified in 2018, show that the August dissents were a fight about who would absorb the political cost of being wrong on the credit crunch, and the equilibrium broke within ten weeks. The FOMC delivered 75bp of cuts between September 1991 and December 1991, the 10Y fell from 7.90% to 6.71%, and the curve steepened by 130bp. Anyone positioned for an orderly resolution got run over.
August 1995 is the cleaner case. The July 6 FOMC had cut 25bp on a 9-3 vote, with Lindsey, Yellen, and Blinder dissenting against further easing because they read the soft landing as already complete. The August 22 meeting held, but the minutes when they released showed a committee that had effectively agreed to disagree about sequencing rather than direction. The market read the hold as hawkish. It was not. By December 1995 the Fed had cut another 25bp, by January 1996 another 25bp, and the 30Y bond rallied from 6.86% to 5.96% over five months. The 1995 episode is the polite version of a coordination resolution: slow, telegraphed, but still a 90bp move in the long bond that nobody had priced when the August dissent surfaced.
September 2022 is the impolite version, and it is the one that should worry anyone long duration here. The September 21, 2022 FOMC delivered 75bp on a unanimous vote, but Bullard had spent the prior two weeks pushing publicly for 100bp, and the dissent never showed up in the vote because Powell privately conceded the dot plot. The published SEP that day raised the terminal by 100bp. TLT lost 11.4% in the next six weeks. The 10Y went from 3.49% to 4.24%. The coordination failure that meeting was not visible in the vote count; it was visible in the SEP revision and the speed of the repricing afterward. The lesson from 2022 is not that hawks win. The lesson is that when the coordination breaks, the move comes in the dots and the language, and the curve adjusts before the next meeting rather than at it.
Three episodes, three different directions, one shared structure. The dissent vote in each case was read by markets as a forecast disagreement that would resolve over multiple meetings. In each case it resolved in one. The repricing was 90 to 130bp in the relevant tenor, took less than ten weeks, and disproportionately punished whoever was positioned for the current equilibrium to persist. April 2026 has more dissents than any of those three episodes.
Where the Repricing Will Hit First
The first place to look is not the front end. Fed funds futures at the CME are already pricing roughly 60bp of cuts by December, which is approximately what the dot plot suggests, and the gap is narrow enough that even a clean coordination resolution would only reprice the December 2026 SOFR contract by maybe 20bp. That is not where the trade is. The trade is in the belly and the long end, where the curve is pricing a glide path that requires the coordination to hold. The 5Y at 4.18% and the 10Y at 4.31% imply a terminal rate around 3.25% with a stable 100bp term premium, which is internally consistent only if the Fed delivers the dot plot path within plus or minus 25bp. If the coordination breaks in either direction, both tenors reprice by 50 to 80bp within a meeting cycle. TLT, currently around $86, has historically moved 7-9% on repricings of that magnitude. The 2s10s at plus 13bp is the second tell. Coordination breaks resolve through the long end, not the front, because the front is anchored to the next meeting and the long end is anchored to the credibility of the entire path. The curve steepening trade is the cleanest beta to the dove resolution, and the curve flattening trade is the cleanest beta to the hawk resolution. The fact that both can be put on for almost identical option premium tells you the market is not picking sides. It is pricing the equilibrium to hold.
Breakevens are the second pressure point and the more interesting one. 5y5y breakeven at 2.34% is pricing the Fed credibly returning inflation to target on the dot plot path. That is a heroic assumption with CPI at 3.3% and gold at $4,613, which is itself a tell about how the marginal global allocator views fiat credibility. If the doves win the coordination fight and the Fed cuts faster than the dot plot, 5y5y breakevens jump 30 to 50bp and TIPS reals stay roughly flat, meaning nominals get hit hardest in the 7-10Y sector. If the hawks win and the Fed holds, breakevens come in 20-30bp but reals back up, meaning the long end gets hit through a different channel. There is no clean equilibrium in which 5y5y breakevens stay at 2.34%. They are the cleanest expression of the trade, and they are the cheapest hedge against either resolution.
The MOVE index at 17.99 on the VIX equivalent for rates, which printed 88 last week, is the third anomaly and the one that makes the position attractive on a risk-reward basis. Realized vol on the 10Y over the past 90 days has run around 95 annualized on the MOVE-equivalent measure. Implied is pricing a continued grind. That gap, roughly 7 points, is the market pricing the current equilibrium as stable. In each of the three historical coordination breaks, implied vol on the comparable measure was within 5 points of realized vol the month before the break and 30 to 60 points above realized one month after. Gamma is cheap here for a reason that does not survive contact with the historical base rate.
What May 12 CPI Actually Tells Us
Yesterday's print, the April CPI release at 3.3% headline and 3.4% core, was treated by the wires as confirming the stagflation-stable narrative. Core goods deflation continues at minus 0.3% annualized, services ex-shelter ran 4.1% three-month annualized, shelter is still grinding lower but at a slower pace than the OER models suggested in late 2025. The headline number was in line with the Cleveland Fed nowcast. The two-year yield moved 3bp on the release. The S&P closed up 18bp. Nothing happened, which is exactly the problem.
What the April CPI actually does is remove the most plausible exogenous resolver of the coordination problem. If April had printed at 3.0% or 3.6%, one side of the FOMC would have gotten enough cover to fold, and the May 13-14 meeting next month would resolve into a clean 7-2 or 8-1 vote. A 3.3% print, the third in a row within 15bp of trend, gives neither side what they need. The hawks cannot point to acceleration. The doves cannot point to confirmed disinflation. Both sides go into the June meeting with the same information set they had in April, which means the coordination problem persists for at least another meeting cycle and likely two. That is bullish for the persistence of the current equilibrium in the very short term and bearish for an orderly resolution in the medium term. The longer the equilibrium holds without an exogenous resolver, the larger the jump when it eventually breaks. Powell knows this. Read his April 16 Economic Club of Chicago remarks carefully, particularly the section on "two-sided risks to credibility," which is Fed-speak for I am aware my committee is stuck.
The specific data point to watch is not next month's CPI. It is the May 28 release of the April PCE, which the hawks have been telegraphing as their threshold. If core PCE prints at or above 2.9%, Daly and Goolsbee will dissent again in June and the four-vote split holds. If it prints below 2.7%, Bowman and Waller likely fold and the committee resolves dovish without a coordination break. The dangerous outcome is the one most likely on current trajectory: a print between 2.75% and 2.85%, ambiguous enough that both sides claim vindication, four dissents in June, and the equilibrium stretches to July. Each additional meeting the coordination holds without resolution raises the implied magnitude of the eventual repricing by, on the 1991-1995-2022 base rate, roughly 15 to 25bp per meeting. That math is why the gamma is cheap and why the trade is in the belly and the breakevens rather than the front end.
There is a subtler signal in the Powell speech worth flagging. He used the phrase "committee deliberations" four times in a 28-minute address. The standard count in his speeches over the prior eighteen months is one or two. Powell does not pad. When he repeats a phrase, he is conditioning the audience for what the next set of minutes will contain, and the next set of minutes, due May 21, are going to read more fractious than the statement suggested. Kashkari's interview with Bloomberg on May 6 confirmed this from the other direction: he described the committee as having "healthy disagreement on tactics," which is precisely the language a coordination-aware chair would workshop to soften the public read of a vote count that broke 8-4. Jefferson has stayed quiet, which on past pattern means he is the swing.
Watch the May 28 PCE print, and watch how the regional Fed presidents speak in the ten days after it.
Explore these indicators together: Chart Household Financial Obligations Ratio, CPI (All Urban), and 3 more on the Indicators Dashboard