“Folks would say, well, if we go to 4 percent, why not go to 6 percent. It’d be very difficult to tie down expectations to 4 percent.” Former Federal Reserve Chair Ben Bernanke, cited in David Wessel, “Alternatives to the Fed’s 2 percent inflation target,” 2018.
“I doubt that a higher inflation target would be viewed as consistent with the Federal Reserve’s Congressional mandate to pursue price stability.” Former Federal Reserve Bank of New York President William C. Dudley, Speech, April 18, 2018.
With inflation significantly above target in most advanced economies, there are renewed calls for central banks to raise their targets from 2% to 3% or 4%, in order to limit the prospective costs of disinflation. Proposals for a higher inflation target initially surfaced more than a decade ago: the goal was to provide central banks with more room to lower interest rates to counter recessions. The subsequent long period during which monetary policy rates were stuck at (or even below) zero markedly strengthened the case.
In this post, we review the benefits and costs of a higher inflation target. Reflecting the reduced equilibrium real rate of interest (r*), a target viewed as optimal in the 1990s (when inflation targeting became widespread) now seems too low. At the same time, the costs associated with modestly higher inflation remain a subject of intense debate. Both theory and experience suggest that higher inflation adds to uncertainty. In addition, the fact that people exhibit “rational inattention” when inflation is low suggests a risk that heightened inflation can become embedded in inflation expectations.
Regardless of the balance between the costs and benefits of raising the inflation target, our view is that central banks ought not be able to choose their inflation targets. As former Chair Bernanke’s opening citation suggests, the key problem with such discretion is the slippery slope. If households and firms come to expect that a central bank will opportunistically raise its inflation target to avoid the economic sacrifice associated with disinflation, inflation expectations will no longer be anchored at the target (whatever it is).
Against today’s background of widening calls for a higher inflation target, the perception of such discretion probably raises the ongoing costs of disinflation. Put simply, many observers seem to think that the Federal Reserve will be satisfied if inflation falls to 3% and will not be willing to pay the further costs of reaching its 2% target.
Limiting the “inflation expectations ratchet”—avoiding perceptions of opportunistic central bank discretion—requires a governance framework that allows for a shift in the target only when conditions truly warrant. Our preferred solution to this is clear: give elected officials a key role in setting inflation targets. Examples of such constrained central bank governance include the mechanisms in the United Kingdom and New Zealand , where the governments set the specific numerical target for inflation.
In the United States, the co-equal status of the executive and legislative branches complicates rules for constraining central bank discretion. As a minimum, we would counsel the Federal Reserve to follow an approach that it now employs regarding the possible introduction of a central bank digital currency: namely, the Fed should announce that it will not alter its inflation target without the explicit support of both the legislative and executive branches, ideally in the form of legislation.
Such a constraint would have clear benefits. First, it would unambiguously raise the hurdle for target changes. Second, it would underpin central bank independence by making it clear that the central bank will be held accountable for meeting the government-endorsed target. Combined, these benefits raise credibility, which helps to anchor inflation expectations, and lowers the costs of disinflation.
Key benefit of a higher inflation target
The case for a higher target is based on the desire to reduce the frequency and duration of zero-policy-rate episodes. Implicit in this goal is unease both over the effectiveness of unconventional monetary policies and, in the case of asset purchases, difficulties in returning balance sheets to their original size.
Low r* adds markedly to the risk that policy rates will fall to zero. For example, in a modified Taylor rule in which the policy rate responds one-for-one to the unemployment gap, when r* falls from 2% to 1%, the equilibrium nominal policy rate falls from 4% to 3%. Because of the zero bound, this reduces the room for monetary policymakers to stimulate aggregate demand by cutting interest rates.
A low r* has clear implications for the frequency and duration of spells at the zero-bound. To see, we can look at the frequency with which the unemployment gap—the difference between the unemployment rate and the noncyclical rate of unemployment—exceeds a given threshold. When that threshold is surpassed, the modified Taylor rule calls for a policy rate at or below zero. Looking at the following chart, and focusing on the period since 2000, we see that this gap exceeds 3% more than 25% of the time, compared to the 17% frequency that it exceeded 4%.
U.S. unemployment gap (percent), 1960-2022
Source: FRED.
Had we assumed the Fed’s balanced approach rule, in which the policy rate moves two-for-one with the unemployment gap, r* of 1% implies a threshold for the unemployment gap of a mere 1½%. In practice, the unemployment gap exceeded this level a whopping 37% of the time—roughly equal to the zero-bound frequency implied by simulations using the Fed’s FRB/US model (see Table 3 here).
Possible costs of a higher inflation target
Raising the inflation target entails costs. One concern is the potential for increased uncertainty that can make household and business decisions less efficient. Indeed, economists have long observed a correlation between the level and the volatility of inflation, with the latter serving as a proxy for uncertainty. For a sample of more than 120 countries with average inflation up to 10%, the chart below shows that a one-percentage point addition to average inflation is associated with a 0.64-percentage-point rise in the annual standard deviation of inflation. Subsequent theoretical analysis has shown that increases in trend inflation naturally lead to greater volatility in a standard macro model (see here).
Average Annual CPI Inflation versus Standard Deviation of Annual Inflation, 1970-2021
Note: The sample of 121 countries includes those with average inflation up to 10% and with at least 10 annual observations. Three outliers are not depicted. Source: World Bank (FP.CPI.TOTL.ZG).
Recent research highlights behavioral risks associated with raising the inflation target. At low levels, there is little evidence that people pay much attention to inflation. This “rational inattention” is consistent with former Chair Greenspan’s famous definition of price stability as “the state in which expected changes in the general price level do not effectively alter business or household decisions” (see page 53 here). Absent a notable rise of inflation, it also is consistent with stable inflation expectations that are anchored at the target.
However, as inflation rises, both theory and evidence suggest people will respond with increasing sensitivity (see, for example, Bracha and Tang). This, in turn, will drive up inflation expectations, making it more difficult and more costly for the central bank to achieve its target. That is, when inflation rises sufficiently, people start to pay attention.
Where is this attention threshold? Using the frequency of online searches for the word “inflation” as a proxy, Korenok, Munro and Chen (KMC) estimate country-by-country attention thresholds beyond which behavior changes. The chart below—which reproduces their Figure 8—plots the estimated thresholds against each country’s average level of inflation. On average, a one-percentage-point increase in average inflation is associated with a 0.67-percentage point rise in the attention threshold.
Estimated Inflation Attention Thresholds and Average Inflation Rates, 2004-May 2022
Source: Figure 8 in Korenok, Munro and Chen, Inflation and attention thresholds, September 2022. The figure includes those 28 out of 37 countries in the Google Trends sample where the estimates showed either no or little inflation sensitivity of online search activity below the estimated threshold and marked sensitivity of search activity above that level. We thank Oleg Korenok, David Munro, and Jiayi Chen for graciously providing their data.
KMC and others emphasize the importance of keeping inflation targets below the level that triggers attention as this poses a risk that inflation expectations will rise. While their U.S. point estimate of 3.55% (red circle in the chart) would seem to offer space to raise the inflation target above 2%, we suspect that this is illusory. Instead, looking at the cross-country evidence, it appears that as households and businesses become accustomed to higher average inflation, their attention threshold rises. Knowing this, however, people may well come to expect that a central bank with discretion will raise its inflation target repeatedly to exploit the observed inattention gaps as attention thresholds climb.
We view this story as a version of the unstable slippery slope logic highlighted by former Chair Bernanke in the opening citation. Students of the 1980s discussion about central bank independence will recognize it as a form of the time consistency problem. Based on that analysis, overcoming the problem of the inflation expectations ratchet requires limiting a central bank’s discretion.
Finally, we note that there is little empirical evidence to support one of the most widely cited costs of higher inflation: namely, that higher inflation increases cross-sectional price dispersion, making the allocation of resources less efficient. Indeed, recent research finds no evidence that inefficient price dispersion was greater in the U.S. high-inflation period of the late 1970s. Instead, an increased frequency of price changes in periods of higher inflation limited price dispersion even as the magnitude of price adjustments was stable over decades. (See Nakamura, Steinsson, Sun, and Villar.)
What to conclude?
Had we known in the 1990s about long bouts at the zero bound for policy rates, many economists (including ourselves) probably would have argued for a higher inflation target of (at least) 3% rather than 2%.
However, unless elected officials express clear and steady support for a higher inflation target, it would lack credibility. This is of particular importance wherever the central bank’s legal mandate is for price stability. As the opening citation from former FRBNY President Dudley notes, it is far from clear that Congress would find a higher U.S. inflation target consistent with the price stability mandate that it set in 1977 for the Federal Reserve.
Moreover, there is an enormous difference between the initial choice of an inflation target decades ago when inflation was reasonably low and stable, and a decision to raise it now when inflation is both far above the target and quite volatile. Central banks have invested heavily in establishing the credibility of the 2% inflation target that is now nearly universal among advanced economies. The attention-related slippery slope considerations that we associate with central bank target-setting discretion could easily undermine this reputational capital and make it more costly for central banks to secure economic and price stability.
For these reasons, we also disagree with recent suggestions that the Federal Reserve use its next strategic review—which could begin in a year or so—to initiate a discussion about the appropriate level of its inflation target (see, for example, Furman). Making the quantitative target a part of the strategic review would only strengthen perceptions that the target itself is a proper subject of central bank discretion. For those (like us) who expect inflation to remain above 3% a year from now, such discretion poses a threat to confidence in the Fed’s 2% target that adds to the costs of disinflation.
Going forward, countries should aim for a framework in which the central bank takes its quantitative target as a legislative mandate that it carries out independently with transparency and accountability using the tools that the government provides. As experience with sustainably independent central banks indicates, such constrained discretion is the best way to maximize sustainable growth and employment.