Warsh’s Fed: from guiding expectations to managing them without commitment

Beyond a somewhat more restrictive tone in response to the uptick in inflation, the most significant change appears to lie in a redefinition of the Fed’s communication framework. The market reaction suggests that investors are anticipating a Fed that is less inclined to commit to future interest rate moves and that will provide less forward guidance, although the anti-inflationary bias of its first meeting helped dispel potential doubts about its political independence.

In the last two decades, the Fed has believed that providing hints about the future path of interest rates – forward guidance – was the most effective and efficient communication framework for monetary policy, although it was not without risks or costs. The logic is that the effects of interest rates depend both on current decisions and on expectations about their trajectory, so influencing these amplifies the transmission mechanism. As Bernanke explained, this required reducing communicative ambiguity: the better agents could anticipate the Fed’s reaction function, the more directly they would incorporate this behaviour into prices and decisions, thus promoting macro-financial stability, especially in a low inflation and rate environment.1

The dot plot, introduced in 2012, serves this purpose by translating an implicit reaction function into an observable signal.2 Although it is neither a collective decision nor a formal commitment, it reveals how FOMC members incorporate their macroeconomic forecasts into interest rates. However, the dot plot can become overly influential and hinder a proper adjustment of expectations, especially in contexts with greater uncertainty and dispersion. Recent evidence suggests that while it contains useful information and can improve the average accuracy of forecasts, it may also cause anchoring effects if agents continue to place too much emphasis on projections which may have been rendered obsolete by the economic reality.3 Moreover, the market tends to interpret the median as if it were the Fed’s intention, even if this is not formally the case. This risk increases when there is significant dispersion in the dot plot, as is currently the case, because the median can conceal deep disagreements and convey a false sense of consensus.4

Warsh has expressed scepticism about the usefulness of forward guidance, as he believes that both this and the forecasts of macro-financial variables can overly constrain central bankers. His critique does not reject transparency, but it does question whether more communication always enhances monetary policy transmission. As he explained in his review of the transparency of the Bank of England’s Monetary Policy Committee, transparency is desirable, but it does not necessarily mean that everything should be made more explicit and visible.5 In his view, an «overly talkative» central bank can give the market a false sense of precision about the future, excessively influence financial prices, and limit its own flexibility. The lesson that seems to be drawn from mistakes such as characterising inflation as «transitory» in 2021 is that the Fed should not always draw more attention to its deliberations, but rather should acknowledge its limitations, for instance, when predicting economic variables.6

Warsh put that vision into practice at his first meeting, where the shift in communication was more significant than the decision on rates. The statement was shorter and more direct, and it removed the explicit reference to future guidance. At the press conference, he insisted that the statement should be limited to presenting the facts «as best we can judge them» and refused to provide anything akin to future guidance, arguing that it did not suit the current situation. Additionally, he chose not to publish his own projection in the dot plot.

This shift in the communication paradigm poses a challenge, as economic agents perceive communication as being reasonably effective. According to a Brookings survey of academics and economists (from both think tanks and private companies), the Fed’s communication is rated highly, and tools such as the press statement and conference are widely appreciated.7 The dot plot, despite its limitations, is still considered useful by the majority. The risk, therefore, is that altering a framework that works reasonably well without clearly explaining what will replace it may end up reducing clarity without gaining sufficient flexibility.

The cost of shifting to a new, more opaque communication regime and greater flexibility in setting interest rates may be reduced predictability and increased volatility. The market reaction following the latest FOMC meeting provides a good illustration of what can happen under this new regime, although it was also influenced by a hawkish bias which investors did not anticipate. The 2-year treasury yield – the segment most sensitive to monetary policy – rose by just over 13 bps on the day of the meeting. Although the FOMC kept rates unchanged and the movement was not particularly sharp from a historical perspective, it was nevertheless significant: it exceeded 1.6 standard deviations from the average movements observed on FOMC meeting days since 1976. If only meetings without rate changes since Bernanke’s arrival in 2006 are considered, the movement was almost 2 standard deviations above the average and ranked among the most intense, specifically in the top 2.5% out of some 140 meetings.

Warsh’s interpretation of the variables which define the Fed’s dual mandate (price stability and full employment) is also relevant. On inflation, his approach hints at a Fed that is guided less by direction and more by the assessment of the data at each meeting. In his public statements, he has argued that the data used to judge inflation are imperfect and has expressed a preference for measures that better distinguish noise from the underlying trends, such as trimmed means. These metrics exclude extreme price swings and can provide a more stable reading of underlying inflation than traditional measures that only remove food and energy, as well as better estimating future inflation based on empirical evidence.8

However, trimmed measures also have their limitations: they are useful for estimating trends over long time horizons but may not perform so well for short periods.9 This is relevant today: if shocks – such as the tariff and energy shocks – are isolated, trimmed inflation measures can adequately filter out the noise; however, if they become widespread across components and affect expectations, they tend to underestimate inflationary pressure. In the current context, these measures do not present a particularly worrying diagnosis, although this did not prevent the FOMC from showing greater concern about inflation than in previous months.

Regarding the labour market, Warsh has been attentive to the impact of artificial intelligence (AI) on productivity and, consequently, on the natural or equilibrium interest rate. The statement even included a reference to productivity growth and capital investment. Various studies suggest that technological advances can have ambiguous effects on inflation and the natural rate of interest: an increase in productivity may exert disinflationary pressures, but it can also raise the equilibrium rate if it boosts investment and aggregate demand.10 For now, the latest total factor productivity data do not show any clear acceleration in recent quarters or compared to previous decades (although the inflation data do indicate that the high investment in AI is increasing demand and, with it, consumer prices for related items). This uncertainty could support Warsh’s argument for reducing the publication of forecasts on unobservable indicators, which are particularly difficult to estimate during times of change, such as potential growth, labour market potential, or the natural rate of interest.

Finally, balance sheet policy could represent another major shift at Warsh’s Fed. If he translates his historical scepticism towards asset purchases into a more rapid balance sheet reduction process, he may encounter the operational limits that the Fed faced previously when it withdrew liquidity too quickly. As governor, Warsh was sceptical about the ability of bond purchases to boost the economy and has even mentioned the need for a new «agreement» between the Fed and the Treasury. As demonstrated by episodes such as the taper tantrum of 2013 (which highlighted the costs of unclear communication about stimulus withdrawal) or the tensions in the repo market in 2019 (which showed the central role of reserves in the current financial system), it will be necessary to tread very carefully in order to avoid disrupting financial stability.

Ultimately, Warsh’s Fed is moving towards a monetary policy that is less guided and more data dependent. The exact extent of the change in communication policy will depend on his ability to transform his ideas on managing expectations and implementing monetary policy into a framework that is intelligible and accepted by the market. Making the reaction function less explicit could increase the Fed’s flexibility in a more uncertain environment, but it also requires the institution’s credibility and independence to be strengthened. The challenge lies in communicating the new framework without replacing what is considered false precision with equally unnecessary opacity – the risk being that this reduced guidance could lead to more volatility than clarity.