Inflation and Its Troubled History

Summary

• A new NBER working paper documents the dynamics of inflation, inflation expectations and other key macroeconomic variables after inflation surges.
• It finds 112 previous episodes like today’s, and, on average, inflation takes two years from its peak to halve and a further four years to normalise fully.
• Typically, inflation surges precede long-term changes in inflation expectations, rises in interest rates, lower growth, and higher unemployment.
• For the US, the results imply headline CPI will remain above 4% until Q2 2024 and that the Fed is behind the curve from a historical perspective.

Introduction

Inflation is here – and back with a vengeance. Sadly, it also appears to have a begrudging character. How long it can resist the Whac-A-Mole approach of central banks and governments, though, is yet to be determined. A look back at past inflation surges may give us the answer.

A new National Bureau of Economic Research (NBER) working paper does exactly this and documents three facts. First, inflation after the initial surge is persistent, taking 3-4 years to normalise. Second, both short- and long-term inflation expectations increase, with the former taking two years to converge on realised inflation and the latter persisting throughout disinflation. And lastly, policy responses typically include nominal interest rate hikes, with no tightening of real rates or fiscal balances.

Data and Methodology

The authors look at official inflation and expectations data in 55 countries – 23 developed markets (DM) and 32 emerging markets (EM) – over the last 30 years. To measure short-term expectations, they use one-year-ahead forecasts, and for long-term expectations 5y5y forwards.

They also collect other macroeconomic data, such as:

• Real gross domestic product (GDP) and the unemployment rate from the World Bank’s World Development Indicators (WDI).
• Fiscal balances, revenue and expenditure over GDP from the IMF’s World Economic Outlook (WEO).
• Short-term interest rates from national sources.

Armed with this data, they ask the following: what typically happens to inflation, expectations and other key macroeconomic variables following inflation surges?

Inflation surges are defined as price rises in the right tail (90^th percentile) of the distribution. Typically, this means a 3.8pp rise in inflation in the first year, a doubling from 3.7% pre-surge. There have been 112 such episodes since 1990, most in EM countries (Chart 1).

Inflation Dynamics

Focussing first on the official consumer price index (CPI), the authors document the dynamics of inflation during those 112 episodes. On average, a surge is characterised by a large spike in inflation during the first year, peaking at around 5pp. This is followed by a prolonged period of disinflation, front-loaded in the first two years, but taking a further four years to normalise (Chart 2).

Applying the authors’ calculations to the US, Year Zero was April 2021. Loosely speaking, then, inflation should have peaked around April 2022, when core and headline CPI reached 6.1 and 8.2%, respectively. Using the last 112 inflation surge episodes as a template, it means headline inflation should roughly halve by 2024 to 4% (Chart 3) and only return to target by 2028.

Inflation Expectations

How do inflation expectations evolve when inflation surges? This has been a hot topic of discussion over the last couple of years. Expectations are a key determinant in central bank inflation models. If individuals expect higher inflation, prices follow suit. For those who buy into the expectations channel, higher inflation expectations are bad news – they keep current inflation higher for longer.

Starting with short-term inflation expectations, which in this paper refers to changes in the one-year-ahead forecast, inflation surges increase expectations, but only by a fifth of the magnitude of actual inflation (Chart 4). The forecast errors, or differences between the predictions and realised inflation, then subside after three years (e.g., by April 2024 in the US).

As for longer-term inflation expectations, which are average 5y5y forward forecasts, these typically increase by only 0.3pp in Year One and Two. In contrast to short-term expectation, however, they stay high for a further four years. In the US, 5y5y forwards have risen 0.25pp since 1 April 2021 and, according to the results, should stay there until 2027.

Monetary and Fiscal Policy Dynamics

Naturally, inflation surges will be followed by higher nominal interest rates. Indeed, the authors document an average policy rate increase of 2.6pp following an inflation surge — around half the increase in inflation (Chart 5). Loosely speaking, the US policy rate should then have been closer to 4–4.25% in April rather than 0.75–1%.

However, notably, inflation surges since 1990 were mainly observed in EM countries, where inflation and policy rates are typically higher. And issues with credibility and stability may make it harder to control inflation surges. So, extrapolating the results to the US is imperfect.

For fiscal policy dynamics, fiscal balances typically deteriorate by 0.5pp of GDP the year following the inflation surge. This, according to the authors, is because expenditure mildly increases and revenues contract following the surge. Two years after the surge, revenues recover and expenditure starts contracting, which strengthens the fiscal balance (Chart 5).

For the fiscal policy results, it is best to focus on the relative movements rather than the absolute values. Governments were a key driver of the current inflation surge which, looking at Chart 5, is not usually the case. However, it seems plausible that expenditures may rise to support individuals through rising prices, and revenues fall as the labour market deteriorates.

Growth and Unemployment

Lastly, the authors track economic growth and unemployment during inflation surges. Their results show a recessing influence on growth during the first couple of years, alongside an increase in unemployment that peaks after three years (Chart 6).

Clearly, the dynamics of the economy will be a function of policy. In the US, real GDP growth has declined since April 2021, but the pandemic has distorted the status quo. In the last decade pre-Covid, average real growth was 2.3%. In Q2 2022, this figure was 1.8%. Activity is dropping but usually bottoms out the year after interest rates peak (Charts 5 and 6). It may, therefore, be that the dynamics in Chart 6 are delayed relative to previous episodes.

Bottom Line

The results from the paper are unsurprising, but they offer a reliable forecast for many countries. The impact of inflation is unlikely to lessen soon, and the negative effects on the US economy from taming it are probably yet to be felt.

These results are an average across many different inflation episodes across many different countries. They do not factor in the origins of inflation, the institutions, the types of policies and other contexts. They do, however, paint an interesting average picture.