US Recession: A Matter of When, Not If

Summary

• Prominent economist Larry Summers looks at the labour market to argue the Fed is unlikely to reduce inflation without causing a recession.
• Wage inflation is reaching its highest point in over 40 years, and recessions typically follow such peaks.
• His views concur with another economic heavyweight, former Bank of England MPC member David Blanchflower, who stresses the importance of consumer confidence surveys in predicting recessions.

Introduction

The literature on recession risk models is dense. You have higher-frequency financial market indicators (e.g., yield curve and equity volatility) alongside lower-frequency macroeconomic data (e.g., labour market tightness and inflation). And then you have sentiment measures that lie somewhere in between (e.g., credit market sentiment and business and consumer confidence surveys). One of the most reliable predictors is the yield curve, from which we have built a recession model.

Across the spectrum of indicators, dials show a rising probability of recession. A new NBER working paper co-authored by Larry Summers focuses on the signals coming from the labour market. Specifically, information contained within vacancy and quit rates implies wage inflation could reach 7% by December 2022. Alongside a low unemployment rate, that puts the probability of recession within the next two years near 100%.

How Tight Is the US Labour Market?

The paper builds on earlier work in which the authors collect monthly data from 1990 to 2021 on the following measures of labour market slack:

• Unemployment rate: U-3, from BLS.
• Prime-age employment ratio: Employment-to-population ratio for adults between the ages of 25 and 54, from BLS.
• Vacancy rate: Number of openings divided by the labour force, from BLS JOLTS.
• Quits rate: Nonfarm quits rate, from BLS JOLTS.

They then show how vacancy and quit rates are the best predictors of wage inflation, i.e., labour market slack. And based on these two firm-side labour market indicators, they create a demand-side measure of the unemployment rate. This measure has historically been a very good estimate of the actual unemployment rate. And it suggests the ‘true’ unemployment rate was 1.2-1.5% in January 2022 – far lower than the 4% reported (Chart 1).

Wage Inflation

According to the best available wage data from the Federal Reserve Bank of Atlanta, wage inflation in February 2022 reached a series high of 6.5% (Chart 2). Even if the vacancy-to-unemployment ratio moves lower to 1.5 (currently 1.7) and CPI drops to 6%, the authors expect wage inflation to push above 7% by yearend (Chart 3). Even under the most conservative estimate, wage inflation will not fall below 5.7%.

Recession Predictions

Summers and his co-authors use the above forecasts for wage inflation alongside different scenarios of the unemployment rate to provide a probability of a recession occurring within the next two years. These probabilities are based on information from previous recessions going back to the 1950s. Their discoveries point towards a high chance of recession in the next two years.

Assuming wage inflation remains above 5%, then irrespective of the unemployment rate, the probability of recession over the next two years is 66%. However, if the unemployment rate stays at 4%, or even increases to 5%, there is a 100% chance of recession (based on historical occurrences). That is, the current wage inflation and the unemployment rate have always preceded a recession within two years. If, however, wage inflation falls below 5%, the probability of recession then ranges between 60% and 70%.

The authors note three previous occasions where the Fed has achieved a ‘soft landing’. That is, bringing inflation down from current levels without inducing a recession. These were 1965, 1984 and 1994. They note, though, that on each occasion, the unemployment rate was above the current level, wage inflation was lower, and the interest rate was above the inflation rate (Chart 4). In their opinion, this only highlights the likelihood of a much harder landing than ever.

How Soon Could We Expect a Recession?

The low-frequency nature of key macroeconomic data leads to a curious realisation: we rarely know we are in a recession until we are exiting it. For example, it took the NBER a year, until December 2008, to call the start of the Great Recession as December 2007. So what economic data can we look towards that may give us a leading insight?

Former Bank of England MPC member David Blanchflower recently published a series of papers explaining why consumer and business survey data provide more economic insights than any other quantitative data. One feature of consumer surveys, such as the University of Michigan Consumer Sentiment Index, is that significant drops can predict downturns 18 months ahead of NBER. The index has fallen nearly 30 points since its peak in April 2021 (Chart 5).

To identify a recession ahead of the NBER, though, Blanchflower also requires employment to fall (CPS or NFP) and the unemployment rate (U3) to rise. Currently, nonfarm payrolls have risen consistently almost every month since March 2020, and the unemployment rate continues to fall (Chart 6).

The key, then, is whether and when either of these measures start to turn. For Summers, the probability of a recession over the next year rises to 64% when the unemployment rate reaches 5% (Chart 7). Moreover, in the words of Blanchflower, a recession is imminent if consumer surveys have fallen more than 10 points since their peak, and the unemployment rate has increased 0.3pp in consecutive months.

Bottom Line

Inflation will be a sticky problem for the Fed. With wage inflation still coming through, and housing inflation yet to peak, the problem is unlikely to see resolution without a significant economic slowdown. For context, there has never been a quarter where average wage inflation was higher than 5% and average unemployment was below 5% that escaped a recession in the next two years. Both thresholds are now in play.

So then, the question is not if, but when? Our latest models put the probability of a recession at 62% within the next 12 months. Summers has it at roughly 50%. When will we know? Well, if history repeats, as soon as we start to see NFP and the unemployment rate turn, you can be confident we are entering a recession. For preppers, we have created a recession portfolio…