Federal Reserve · SEP Median Dots · 2015–2024

What the Fed Projected for Its Own Rate

Two FOMC median-dot vintages against the realized federal funds rate · percent

Dec 2015 dot for 2018
3.3% → 2.4%
Mar 2021 dot for 2022
0.1% → 4.3%
Realized rateDec 2015 projection (too high)Mar 2021 projection (too low)
Source: FOMC Summary of Economic Projections, median federal funds rate (dots); FRED (realized rate). Author’s calculations.nikkhosravipour.com
Figure 1 — Two FOMC median-dot projections against the realized federal funds rate. December 2015 (red) projected the rate climbing toward 3.3% by 2018, above what the committee delivered; March 2021 (blue) projected 0.1% through the 2021 inflation, far below the rate that arrived.

Kevin Warsh took office as the seventeenth chair of the Federal Reserve on May 22, 2026.1 At the close of his first meeting of the Federal Open Market Committee, held on June 16 and 17, he disclosed that he had submitted no dot to the Summary of Economic Projections, the quarterly exercise in which each participant marks the policy rate he judges appropriate in the years ahead. No chair had abstained since the projections began in 2012. Warsh paired the abstention with a review, commissioning task forces to examine both the dot plot and the post-meeting press conference, the two channels through which the contemporary Fed states its intentions.2

The case the review will weigh is visible before any of it is argued. In March 2021 the median participant projected a federal funds rate of 0.1% for the end of 2022. The rate arrived at 4.33%.3 Across the four meetings of 2021 the median dot for year-end 2022 moved from 0.1% to 0.1% to 0.3% to 0.9%, while the realized path, once inflation forced the committee's hand, climbed past four percent. The dots did not lead the rate. They followed it. The same shape repeats one year out: the median projection for year-end 2023, first set at 1.0% in September 2021, was overtaken at meeting after meeting by the rate the committee would actually choose. The failure is not a matter of direction. Through the prior decade the dots ran the opposite way, projecting a rate that climbed toward 3% by the end of the decade while the committee stopped short and then cut (Figure 1). What misfires in both directions is the premise shared by every vintage, that the committee can fix today the rate it will choose years from now.

The dot plot entered the Fed's communications in January 2012, as the centerpiece of a turn toward transparency meant to make policy more predictable and therefore more effective.4 Each participant supplies a path, and the median is read, inside the markets and out, as the committee's collective forecast of its own conduct. That reading is the instrument's purpose. By publishing where the policy rate is expected to travel, the Fed reaches the long rates and financial conditions through which a change in the overnight rate actually transmits, shaping them in advance of any move. The dot plot is a forecast of the committee's own future, offered to anchor the expectations that govern everything downstream of it.

A point forecast of the policy rate presumes that the committee can know its own future reaction. That reaction depends on shocks that have not yet landed and on judgment the committee has not yet exercised. Hayek drew the relevant distinction in his Nobel lecture of 1974: in any sufficiently complex order, prediction is confined to the general pattern of what will form, and does not reach particular events.5 A policy rate two years out is such an event. The dot plot states it to a tenth of a percentage point, presenting as a precise point prediction what the institution can hold only as a pattern. This is the error the 2021 vintage exhibits, a committee forecasting its own discretion and missing, because that discretion had not yet been exercised and the shocks that would govern it had not yet arrived.

Guidance of this kind did real work once. With the policy rate pinned at its floor after 2008, the committee's words were the only instrument it had left, and it used them deliberately: the calendar commitment of August 2011 to hold the rate low at least through mid-2013, and the threshold commitment of December 2012 to hold until unemployment fell below 6.5% or projected inflation rose above 2.5%.6 Each lowered longer rates by binding the Fed to a path the market could price. Retiring the dot plot discards a channel that demonstrably functioned when the rate could not move. The value was bounded, though, and the bound is the point. Guidance at the floor stood in for a rule the Fed did not otherwise possess, managing expectations in place of a commitment that would have constrained the committee in both directions. Once inflation arrived, the substitute failed in exactly the direction the 2021 dots record. A genuine rule would have forced the rate up on a schedule the committee could not revise at will. The dots, carrying no such force, trailed the data instead.

Retiring an instrument that projects false precision is therefore defensible on its own terms. A point forecast of a pattern-predictable variable misleads whether or not the committee intends the deception, and the record of 2021 is what that misleading looks like in practice. Defensibility is not sufficiency. For all its pretension the dot plot supplied one thing of value, a public and quantified record of the committee's own expectation, against which its later conduct could be measured by anyone. The 2021 vintage indicts the Fed only because the dots were published. Remove the instrument and put nothing binding in its place, and the committee is left holding the discretion while the yardstick is gone. That is the same exchange the 2025 framework review already made once, when it abandoned average inflation targeting and replaced it with nothing the Fed can be held to.7 Whether the task forces break the pattern or extend it turns on one question, and it is not whether the dot plot survives. It is whether anything binding takes its place. Absent that, the chairman will have removed the pretence of a path and kept the latitude that the pretence, at least, left on the record.

Footnotes

  1. Kevin Warsh was confirmed as chair of the Board of Governors on May 13, 2026, and sworn in on May 22, 2026, succeeding Jerome Powell. "Kevin Warsh confirmed as Fed chair, succeeding Jerome Powell," CNN Business, May 13, 2026. ↩
  2. The Summary of Economic Projections has published participants' federal funds rate projections, the "dot plot," since January 2012. At his June 17, 2026 press conference Warsh confirmed that he had submitted no projection ("I did not submit a dot for me"), the first chair to abstain, and announced task forces to review the dot plot and the post-meeting press conference. CNBC, "Fed Chair Warsh expected to withhold 'dot' from central bank's interest rate outlook," June 16, 2026, and "Fed meeting recap: Warsh announces task forces to overhaul major Federal Reserve operations," June 17, 2026. ↩
  3. Median dots are the federal funds rate projections, median value at the midpoint of the target range, from Table 1 of each Summary of Economic Projections, meetings of March 2021 through December 2023 (Board of Governors of the Federal Reserve System, federalreserve.gov/monetarypolicy). Year-end 2022 across the 2021 meetings: 0.1%, 0.1%, 0.3%, 0.9%. Realized rate is the effective federal funds rate (FRED series DFF): year-end 2022 near 4.33%, within the 4.25–4.50% target range set December 14, 2022; year-end 2023 near 5.33%, within the 5.25–5.50% range held from July 2023. Source for Figure 1. ↩
  4. The federal funds rate projections were added to the SEP at the January 24–25, 2012 meeting, alongside the Committee's first explicit 2% longer-run inflation objective. ↩
  5. F. A. Hayek, "The Pretence of Knowledge," Nobel Memorial Prize lecture, December 11, 1974, in Les Prix Nobel en 1974 (Stockholm: Nobel Foundation, 1975); reprinted in American Economic Review 79(6) (December 1989), 3–7. The distinction between "mere pattern predictions" and the prediction of particular individual events is drawn there, and earlier in "The Theory of Complex Phenomena" (1964). ↩
  6. FOMC statements of August 9, 2011 (calendar guidance, the rate to stay exceptionally low "at least through mid-2013") and December 12, 2012 (threshold guidance, the so-called Evans rule, conditioning the rate on unemployment above 6.5% and projected inflation no more than 2.5%). ↩
  7. The Fed's 2025 framework review formally abandoned Average Inflation Targeting, the asymmetric framework adopted on August 27, 2020, without replacing it with a binding benchmark. See the companion analysis, Discretion Over Rules: Federal Reserve Policy Departures and the Inflation Surge of 2021–2022. ↩

Get new essays and research by email — subscribe on Substack (opens in new tab)