Is this note, we assess whether the current inflationary episode in the US and in the EA is “resolved” through the lens of an IMF paper (One Hundred Inflation Shocks: Seven Stylized Facts” by Anil Ari, Carlos Mulas-Granados, Victor Mylonas, Lev Ratnovski, and Wei Zhao). Overall, both in the US and the EA a re-acceleration of inflation is unlikely but the historical evidence cautions against considering inflation as fully done. Recent communication from the Fed and the ECB is in line with these findings.
What the paper does
The goal of the paper is to study a broad set of historical inflation episodes spread across time, country, and cause. The authors identify 100 inflation episodes, roughly defined as periods in which inflation rises from a moderate level to over 3%. Over half of the inflation events occurred following the 1973 and 1979 oil crises. Other episodes are mostly associated with demand surges or exchange rate depreciations. The empirical strategy is to compare the change in macroeconomic conditions around an inflation event.
Main results
Inflation is persistent. 40% of inflation episodes in the sample lasted more than 5 years, which the authors label “unresolved”. Among those episodes that did resolve within five years, the average period lasted 3 years. For 90% of unresolved episodes, inflation declined materially within the first three years after the shock but then reaccelerated.
Inflation can be addressed by effective policy measures without large real costs. Resolved inflation episodes were associated with significantly tighter monetary policy in terms of higher real rates, lower broad money growth, and stronger positive Taylor rule deviations (see Figure 1). The blue bars indicate the mean difference in pre-shock vs post-shock policy stance, averaged across inflation episodes. The red lines indicate the median difference. Unsurprisingly, it appears rather the case that countries with unresolved inflation episodes had particularly expansionary policy stances.
Figure 1. Average monetary policy stance during resolved (<= 5 years) and unresolved inflation episodes.
Real wage growth is negative during inflation episodes, stronger disinflation does not come with weaker output growth. Real wage growth looks very similar across resolved vs unresolved inflation episodes, though nominal wage growth is significantly higher in unresolved episodes. This hints at a potential role of wage-price spirals in persistent inflation regimes. Countries with shorter inflation episodes did not experience weaker real output growth or higher unemployment, although it is unclear if this indicates that successful disinflation policies have little negative side effects or simply reflects that countries with shorter inflation episodes faced more favorable shocks to begin with.
A critique of the paper
The paper provides a comprehensive historical record of inflation episodes and documents several useful stylized facts. In particular, many inflation episodes are not short-lived and there is a significant risk of reacceleration. The US experience of the Great Inflation is often presented as an isolated example of double-peak inflation. In reality, it is the rule rather than the exception.
Beyond the stylized facts, there is no identification of causality. As the authors acknowledge, the focus of the paper is to document correlations. Due to data limitations, they are unable to control for many confounding factors such as differences in policy regimes, shock sizes, or other idiosyncratic factors. This limits the conclusions that can be drawn from the analysis.
Historical evidence is useful, but changes in economic structure limit the lessons to be learned. While inflation in the 1970s arguably originated from similar types of shocks as (part of) the post-pandemic inflation, the sectoral composition of the economy and the degree of policy credibility have changed. These crucial factors mean that the importance of oil/gas price shocks and the risk of monetary policy mistakes may be different today. Further, economic integration has increased over the past 50 years and today’s policy makers need to pay more attention to spillover effects from foreign inflation shocks and corresponding policy responses.
Is inflation “resolved” in the US and in the EA?
“Almost, be careful” is the answer. Using Ari et al. (2023) to assess whether the US and the EA can be classified as “resolved” brings mixed results. On the one hand, it is pretty clear that both regions have made progress and that monetary policy is sufficiently restrictive (see Figure 1). However, in other dimensions, neither the US nor the EA can be classified as “resolved”. For instance, Figure 2 shows that past episodes were associated to a positive structural primary fiscal balance, while in fact fiscal policy remains way too loosen both in the US and in the EA (we have been vocal in this dimension). Therefore, overall the answer is “the US and the EA are “almost” resolved but not quite yet”.
Figure 2. Average structural primary fiscal balance during resolved (<= 5 years) and unresolved inflation episodes.
Policy implications
The historical record supports a carefully calibrated, data-dependent approach to policy normalization in response to recent signs of disinflation among advanced economies. The policy implications of the paper are as follows. Meeting-by-meeting policy decisions appear preferable over forward guidance policies, especially during a disinflation period with upside risks to the inflation forecast. Policy communication remains key and needs to convincingly convey the determination to return inflation to target, which should keep inflation expectations anchored. Any calls for premature policy loosening for fear of negative consequences for real output or employment should be weighed against the evidence of insignificant medium-term real costs in periods of successful disinflation.