A plain-English guide to monetary policy frameworks.

A monetary-policy framework is the Federal Reserve's articulated approach to the dual mandate of maximum employment and stable prices. It is not a model of the economy, not a forecasting tool, and not a rule the Committee has to follow. It is the Committee's public statement of how it intends to translate incoming data into a policy stance. Frameworks tend to be durable across cycles and Chairs; Powell's 2020 framework, the Flexible Average Inflation Targeting (FAIT) regime, was the first formal framework change since 2012. Warsh's comment at yesterday's press conference that forward guidance is "not well suited" to current conditions, combined with his task-force announcement, is the strongest public signal in 14 years that a new framework is coming. This is the explainer on what each framework is, what changed and why, and what Warsh has hinted at.

What a framework actually is

The Federal Reserve articulates its framework in a formal document called the "Statement on Longer-Run Goals and Monetary Policy Strategy". The document was first published in 2012, then re-issued in 2020 with the FAIT change, and reviewed every five years. The 2025 review window opened in early summer; Warsh's task forces are part of that work.

A framework specifies three things:

• The objective. Inflation target, employment target (or "broad-based, inclusive" goal), and any explicit cross-mandate priority.
• The reaction function. How the Committee will respond to deviations from the objectives. Whether misses on inflation are "made up" through subsequent under-shoots (FAIT), or whether the Committee responds symmetrically to over- and under-shoots (traditional IT).
• The communication strategy. Statement language conventions, the SEP, and forward-guidance norms.

Five US frameworks since 1979

1. Monetarist (1979-1987). Paul Volcker took over with double-digit inflation and announced a shift to targeting monetary aggregates (M1, M2) rather than the policy rate. The framework was technically about controlling the money supply; in practice it produced extreme rate volatility and the 1981-1982 recession. Inflation was broken; the framework was abandoned within four years.
2. Implicit inflation targeting (1987-2012). Under Greenspan and the early Bernanke, the Committee operated with no explicit numerical target but used the policy rate to keep inflation in a 1.5 to 2.5 percent zone. The framework was widely understood by markets but never formalised in writing. The 2008 crisis exposed the limits of unspecified reaction-function communication.
3. Explicit 2 percent target (2012-2020). Bernanke's last major institutional reform: a published 2 percent inflation target paired with a symmetric reaction function. The framework worked well in the 2014-2019 sample but produced a persistent under-shoot relative to target, which the 2020 review identified as a credibility problem.
4. FAIT (2020-2025). Powell's response to the under-shoot problem. The Flexible Average Inflation Targeting regime accepted modest over-shoots after periods of under-shoot, on the principle that the Committee could "make up" for missed inflation by allowing inflation above target temporarily. The framework was designed for a low-inflation, low-growth world; the 2021-2022 inflation surge revealed its weaknesses. The 2025 review formally ended FAIT as the operative regime.
5. Pending (2025-2027). The current review window. Warsh's stated framework views point in specific directions that we walk through below.

What FAIT actually did, and why it ended

The FAIT framework had three core mechanics: a 2 percent average inflation target rather than a point target, an expanded employment objective ("broad-based and inclusive"), and an asymmetric reaction function that tolerated over-shoots after under-shoots. The framework was designed around the 2014-2019 problem: persistently below-target inflation despite a tight labour market.

Three problems with FAIT in the 2021-2024 sample:

1. The "make-up" was opaque. The Committee never specified how much above-target inflation it would tolerate or for how long. When inflation peaked at over 7 percent in 2022, the "average" framing did not anchor expectations because no one knew what window the average was being computed over.
2. Asymmetric tolerance compromised credibility. The framework explicitly told markets the Committee was willing to be late on inflation. When inflation arrived, the implicit promise of patient tolerance became a credibility cost rather than a flexibility benefit.
3. The employment objective was too broad. "Broad-based and inclusive" was politically attractive but operationally unmeasurable. The Committee had to fall back on the unemployment rate as the actual gauge, which did not differ materially from the pre-FAIT framework.

The alternatives the review is examining

• Return to traditional 2 percent IT. A symmetric point target with no make-up provision. The simplest path forward; the 2012 framework with the FAIT modifications stripped out. Most of the published academic research argues for this.
• Taylor-rule benchmarking. An explicit reference to a Taylor-rule-style published rate path, with the Committee's actual policy described as deviating from the rule in specific quantified ways. This would anchor the reaction function in published math; the cost is the Committee loses discretion to respond to circumstances the rule cannot capture.
• Price-level targeting (PLT). Targets the price level rather than the inflation rate, which effectively commits the Committee to make-up periods in both directions. PLT has strong academic support but has been implemented nowhere in G10.
• Nominal GDP targeting (NGDPT). Targets the level of nominal output rather than inflation; the Committee adjusts policy to keep nominal GDP on a pre-announced trend. NGDPT has been advocated by a small group of economists since the 2010s; it is theoretically elegant but operationally hard to communicate.
• Multiple-mandate explicit weighting. An explicit numerical weighting on inflation vs employment (e.g., 70 percent inflation, 30 percent employment in the reaction function). Most G10 frameworks do something like this implicitly; making it explicit would be a meaningful institutional change.

Warsh's published views

Kevin Warsh has written extensively on framework design, going back to his time on the Board in the 2009-2011 period. His published positions, summarised:

1. Forward guidance is a credibility tax. Warsh has argued that explicit commitments to future policy paths (calendar- or threshold-based) reduce Committee flexibility without proportionate benefits to expectations anchoring. Yesterday's "not well suited" comment was the most direct public version of this view.
2. Higher r-star is structural. Warsh has cited Lubik-Matthes and supply-side productivity gains as evidence that the neutral rate is structurally higher than the 2010s estimates. The longer-run dot drift in this week's SEP is consistent with this view.
3. The dot plot is "false precision". Warsh's decision yesterday not to submit a dot was the first public expression of this view in an institutional capacity. He has previously argued for either eliminating the dot plot or restructuring it to publish a single central-tendency range rather than individual dots.
4. Statement language should be shorter. Yesterday's 130-word statement (down from 341) is the implementation of this view. Warsh has written that statement language proliferation increased credibility costs without information benefits.

What the 2026 framework is likely to look like

A reasonable forecast based on Warsh's published views and the task-force structure he has announced:

• Symmetric 2 percent point target, with no make-up provision.
• Explicit weighting in favour of price stability when the mandates conflict, consistent with Warsh's statement yesterday that "price stability is the Fed's priority".
• Reduced forward-guidance discipline. The Committee preserves more discretion meeting-to-meeting; the SEP becomes the primary quantitative communication channel, and the statement becomes shorter.
• Restructured SEP with central-tendency ranges replacing individual dots, or a hybrid where individual dots remain but the Chair does not participate.

Why the framework matters for cross-asset pricing

The framework is the most durable input to long-end yields, structural FX rates, and equity-risk premia. A framework that prioritises price stability over employment translates mechanically into a higher longer-run policy rate (the longer-run dot moves up), a higher terminal Fed funds rate embedded in the OIS curve, and a stronger structural dollar in the smile framework. None of those re-prices on a single SEP; all of them re-price gradually as the framework is announced, debated, and implemented.

The market will read framework signals over the next 12 to 18 months. The first major public hook will be Warsh's Jackson Hole speech in August, traditionally the venue where Fed Chairs articulate framework-level views. Watch that speech as the highest-stakes communication event between now and the formal 2027 framework publication.

What frameworks do not tell you

1. The next meeting's decision. Frameworks describe the reaction function in general terms; meeting-specific actions depend on data. A new framework can coexist with any specific policy action.
2. Whether the Committee follows it. Frameworks are aspirational documents. The Committee retains discretion to deviate, and historically has deviated when circumstances required.
3. The political support. Frameworks change with administrations only with great difficulty. The 2020 framework survived two administrations; the 2026 framework will be designed under a single administration and is likely to last longer.

Related reading

• Yesterday's analysis applies this directly: USD/JPY broke 161. Japan escalated to step-five language.
• The forward-guidance piece Warsh just called "not well suited": A plain-English guide to forward guidance
• The r-star companion: A plain-English guide to r-star
• The SEP and dot plot: A plain-English guide to the SEP and the dot plot