In their paper, “Why has the U.S. economy recovered so consistently from every recession in the past 70 years?” (National Bureau of Economic Research, Working Paper 27234, May 2020), Robert E. Hall and Marianna Kudlyak use Current Population Survey data to study the recovery path of the unemployment rate after each negative economic shock that has occurred over the past 70 years (1949–2019). Normally, an economic recession results in a spike in unemployment. Hall and Kudlyak discover that the unemployment rate has declined in a predictable way, at an average of 0.55 percentage point annually after every recession since 1949. This decline has occurred notwithstanding the causes of the recession or the monetary and fiscal policy tools implemented to combat the recession.
The causes of recessions are many and heterogeneous. However, the authors of this study find that the recovery of the unemployment rate is actually homogenous—it declines by about 0.55 percentage point each year until reaching between 3 and 5 percent. The authors discuss the “widespread” belief that the recovery from the financial crisis of 2008 took longer than other recoveries, such as the 2001 recession or the recessions of the early 1980s. In fact, the growth rate of gross domestic product was slower to recover following the 2007–09 recession, but the unemployment rate recovered in a way similar to that of the other recessions during the 70-year period. When the unemployment rate is 10 percent, though, the recovery takes longer. Hall and Kudlyak try to explain why the decline in the unemployment rate was both slow and linear after each recession during the 1949–2019 period.
The authors use Job Openings and Labor Turnover Survey data and other sources to study different types of job loss: layoffs, job destruction, worker displacement, and unemployment insurance claims. All of these measures of job loss tend to spike after a negative shock to the economy; however, they all recover rather quickly, while the aggregate unemployment rate takes much longer. In addition, the authors argue that individual jobseekers seem to find employment faster than the overall unemployment rate declines. Although these findings suggest that individual people find jobs more quickly than the unemployment rate indicates, this can be explained by the fact that people often fluctuate from being employed to unemployed until they find the right job, which adds to the persistence of a higher unemployment rate.
The aggregate unemployment rate also takes time to recover because there are “frictions in rebuilding unemployment” that cause the rate to decline slowly. For example, many people seeking employment creates an adverse selection problem for firms that want to hire new employees, because they have to vet all of the job applicants. This raises the costs of the hiring process, which results in fewer jobs being offered. The credit market provides another example. During a recession, banks are more reluctant to lend money, which raises the cost of capital, making it more difficult for firms to expand and hire.
Although Hall and Kudlyak use various models to try to explain the linearity of the unemployment rate’s recovery, none could explain it fully. The models instead predicted a convex recovery path, meaning that the unemployment rate would recover in larger increments at the beginning and smaller increments as the unemployment rate approached the range of 3 to 5 percent.
This paper is timely, as the COVID-19 pandemic has pushed the United States into a recession—the Business Cycle Dating Committee of the National Bureau of Economic Research determined that the U.S. economy reached a peak in February 2020, marking the beginning of an economic downturn. The unemployment rate rose from a 2019 annual average of 3.7 percent to 14.7 percent in April 2020 (the rate was 11.1 percent in June 2020). Although this recession is highly unusual because of the pandemic, studying past recessions and recoveries can nevertheless help us understand the present one.Download PDF »