How Much Do Fed Officials Actually Disagree?
Data current through the April 29, 2026 FOMC meeting. Updated when new readings are published by Tsang and Yang at digitecon.org following each FOMC meeting minutes release.
Latest reading: 0.2311 (April 29, 2026 FOMC meeting).
Down 0.020 from the prior meeting reading of 0.2507 (March 18, 2026).
What Is the FOMC Dissent Index?
The FOMC Dissent Index measures the level of hidden dissent among Federal Open Market Committee members at each meeting, based on the words used in the meeting record rather than the recorded vote. It is the “Hidden Dissent” measure proposed by Kwok Ping Tsang and Zichao Yang in Agree to Disagree: Measuring Hidden Dissent in FOMC Meetings.
The index ranges from 0 (strong support across the committee) to 1 (substantial opposition). It is constructed from FOMC meeting minutes, allowing a measurement of internal disagreement to be made within about three weeks of each meeting — well before the five-year embargo on full transcripts is lifted.
The authors’ premise is that formal dissent votes are rare — only about 7% of all FOMC votes since 1976 have been “No” votes — and therefore understate the actual extent of disagreement inside the committee. By analyzing the language of the meeting record, the index aims to capture disagreement that does not appear in the recorded vote tally.
The April 2026 FOMC Minutes
The April 28–29, 2026 meeting maintained the target range for the federal funds rate at 3.5 to 3.75 percent. The vote was 8 in favor, 4 against.
- Stephen Miran dissented in favor of a 25-basis-point cut, citing concerns that “the current policy stance was overly restrictive in a situation of downside risks to the labor market.”
- Beth Hammack, Neel Kashkari, and Lorie Logan supported maintaining the target range but did not support inclusion of an easing bias in the statement, preferring “a more two-sided characterization of the Committee’s future interest rate decisions.”
On inflation, “the vast majority of participants” noted “an increased risk that inflation would take longer to return to the Committee’s 2 percent objective than they had previously expected.” A “majority of participants” highlighted that “some policy firming would likely become appropriate if inflation were to continue to run persistently above 2 percent.”
On the labor market, “most participants” judged that recent data suggested “stabilization in the labor market.” “Several participants” commented that the low rates of job growth were “not necessarily indicative of labor market fragility, as they could be roughly commensurate with the recent slow growth in the labor force,” while “a few participants” pointed to “the possibility that the low rates of job growth were a sign of labor market fragility.”
On the policy path, “Many participants” indicated they “would have preferred removing the language from the postmeeting statement that suggested an easing bias.” Meanwhile, “several participants” said it would likely be appropriate to lower the target range “once there are clear indications that disinflation is firmly back on track or if solid signs emerge of greater weakness in the labor market.”
Read the full minutes: federalreserve.gov
How the Hidden Dissent Index Is Measured
Tsang and Yang first train a deep-learning model on FOMC transcripts from 1976 to 2018. For each meeting, the model compares each member’s statements against those of the FOMC chair — used as a proxy for the meeting’s adopted policy position — and learns to predict, from the language differences, whether the member voted YES or NO on the meeting’s policy action. The output for each member is a score between 0 and 1, representing the model’s estimated probability that the member would have voted to dissent based on the language of their statement alone. These individual scores are averaged at the meeting level to produce a transcript-based hidden dissent reading.
Because FOMC transcripts are released with a five-year delay, a second model is trained to estimate the same measure from FOMC minutes, which are released about three weeks after each meeting. The minutes-based model takes the participants’ views and policy discussion sections of the minutes as input, and is trained to predict the meeting-level transcript-derived score. The authors report that the two series correlate at 0.848, allowing the minutes-based index to serve as a timely proxy for the underlying transcript measure.
This page focuses on the minutes-based readings.
Dissent Votes vs. Hidden Dissent
Formal dissents are rare and binary — a member either casts a “No” vote or does not. The hidden dissent measure is continuous, derived from the language of the meeting record. It reflects the model’s estimated probability that the underlying transcripts would show dissent-flavored language patterns, calibrated against the historical relationship between recorded NO votes and member-level transcript divergence.
The authors document that the two signals can diverge substantially. A meeting can produce a unanimous vote while the language of the discussion shows extensive disagreement, and a meeting with multiple recorded dissents can show comparatively narrow disagreement across the room as a whole.
Tsang and Yang interpret this divergence as evidence that the official voting record is too coarse to detect subtle but economically important differences in policy preferences across the committee.
What Drives Hidden Dissent
Hidden dissent is strongly related to macroeconomic conditions. It rises when inflation is high or rising, consistent with the idea that high inflation creates a policy dilemma between curbing prices and avoiding recession. The connection to unemployment is weaker in aggregate, but becomes significant once members’ formative-year experiences (Great Depression, Great Inflation, World War II) are taken into account.
Hidden dissent reflects differences in policy preferences more than differences in economic outlook. Using data from the Summary of Economic Projections, the authors show that disagreement over the appropriate path of the federal funds rate is a strong predictor of hidden dissent, while disagreement over projected unemployment, GDP, and inflation is not. This is reinforced by a strong correlation between hidden dissent and the optimal policy perturbation measure of Barnichon and Mesters (2023).
How Markets Respond to Hidden Dissent
Hidden dissent has measurable effects on financial markets. Using the minutes-based index, the authors find that higher hidden dissent is followed by lower S&P 500 returns, higher VIX readings, higher 10-year Treasury yields, and higher perceived interest rate risk over the 15 days following the release of FOMC minutes. These effects hold after controlling for the hawkish or dovish sentiment of the minutes, indicating that hidden dissent is a distinct information channel for financial markets, separate from policy tone.
Limitations
The authors note two limitations of the current measure. First, the index is direction-agnostic: it captures the level of disagreement but does not distinguish whether dissenting members are arguing for tighter or looser policy.
Second, after the Federal Reserve’s 2008 framework shift to include tools beyond the federal funds rate (quantitative easing, forward guidance, balance sheet operations), members may disagree not only about the direction of policy but about the appropriate tool. The model, trained primarily on rate-direction dissents, may not fully capture disagreements about the approach to policy.
When Are the Next FOMC Minutes Released?
The next FOMC meeting minutes is expected to release on July 8, 2026 at 2:00 p.m. Eastern Time, covering the June 16–17, 2026 meeting. The Federal Open Market Committee publishes meeting minutes on the third Wednesday following each scheduled meeting. The full schedule for the current cycle is shown below.
| FOMC Meeting | Minutes Release Date | Status |
|---|---|---|
| January 27–28, 2026 | February 18, 2026 | Released |
| March 17–18, 2026 | April 8, 2026 | Released |
| April 28–29, 2026 | May 20, 2026 | Released |
| June 16–17, 2026 | July 8, 2026 (projected) | Upcoming |
| July 28–29, 2026 | August 19, 2026 (projected) | Upcoming |
| September 15–16, 2026 | October 7, 2026 (projected) | Upcoming |
| October 27–28, 2026 | November 18, 2026 (projected) | Upcoming |
| December 8–9, 2026 | December 30, 2026 (projected) | Upcoming |
Meeting dates are set by the Federal Open Market Committee. Minutes are released approximately three weeks after each policy decision; projected release dates above are calculated from that rule and may be adjusted by the Committee, particularly around holidays. The official calendar — with confirmed release dates added after each release — is published at the Federal Reserve’s FOMC calendar.
Citation and Data
Tsang, Kwok Ping and Yang, Zichao, Agree to Disagree: Measuring Hidden Dissent in FOMC Meetings (May 09, 2025). Journal of Economic Dynamics and Control, volume 180, 2025[10.1016/j.jedc.2025.105197], Available at SSRN: https://ssrn.com/abstract=4546049 or http://dx.doi.org/10.1016/j.jedc.2025.105197
The data shown on this page is sourced from the authors’ regularly updated dataset at digitecon.org/hidden-dissent-index. Readings on this page are republished from that dataset following each update.
