Four years after the onset of the Great Recession, labor market outcomes in the U.S. remain depressed. The fraction of 16- to 64-year-old individuals who are employed fell from above 72 percent in 2007 to less than 67 percent in 2009 and remains stuck there. The unemployment rate rose from 4.5 percent to 10 percent and still hovers above 8 percent. And the fraction of unemployed workers who have been looking for a job for more than six months has increased to a share not seen in the United States in at least 60 years. The Atlanta Fed’s Center for Human Capital Studies hosted a conference last weekend, organized by Richard Rogerson (Princeton University), Robert Shimer (University of Chicago) and the Atlanta Fed’s Melinda Pitts that explored why the employment losses were so large and why the labor market recovery has been so weak. Examining these questions is important because different hypotheses about the nature of the recession suggest that different policy interventions may help to accelerate the recovery.
The paper “On the Importance of the Participation Margin for Labor Market Fluctuations” by Michael Elsby, Bart Hobijn, and Ayşegül Şahin offered some suggestions on how to think about the disparate behavior of the unemployment rate and labor force participation rate during the last couple of years. While the unemployment rate has steadily fallen back towards its historic levels, labor force participation has fallen, keeping the employment-population ratio constant. At some level, this movement suggests that the decline in labor force participation has acted as a relief valve for the unemployment rate. Using evidence on the gross flows of workers between employment, unemployment, and out-of-the-labor-force, Elsby and his coauthors question that interpretation. Instead, relatively few unemployed workers have dropped out of the labor force during the recovery, reflecting the high desire to work among the current stock of unemployed individuals.
A number of papers offered specific hypotheses about the reason for the large and persistent deterioration in labor market outcomes and tested those hypotheses using a variety of methodologies and datasets. For example, the paper “What Explains High Unemployment? The Aggregate Demand Channel” by Atif Mian and Amir Sufi explored the implications of the negative shock to household balance sheets that followed the collapse in house prices. They document that employment in the nonconstruction, nontraded sector declined most in U.S. counties that experienced the largest adverse shock to house prices, while the decline in the traded goods sector occurred equally nationwide. If wages and prices were flexible, we would expect the balance sheet shock to reduce the demand for nontraded goods and raise the supply of labor and hence employment in the traded good sector. The fact that this did not happen is evidence that wages and prices have not adjusted. They infer that roughly two-thirds of the total employment losses can be attributed to the balance sheet shock, in combination with wage and price rigidities.
A second hypothesis is that the recovery has been so weak because of underlying adverse trends in the U.S. labor market. “Manufacturing Busts, Housing Booms, and Declining Employment: A Structural Explanation” by Erik Hurst, Matt Notowidigdo, and Kerwin Charles shows how the ongoing decline in the demand for less educated men in manufacturing has generated a negative trend in labor market outcomes for these workers for three decades. This trend continued unabated during the years after the 2001 recession but was masked by the housing boom, which lifted employment for less-skilled workers for another five years. This observation is relevant for how one interprets the time series changes in labor market outcomes. If we view the housing boom as an aberration that is unlikely to resume, it is inappropriate to compare current labor market outcomes with those just preceding the onset of the Great Recession.
The paper “The Trend is the Cycle: Job Polarization and Jobless Recoveries” by Nir Jaimovich and Henry Siu focuses on a related but distinct long-term phenomenon in the U.S. labor market: job polarization. This refers to the fact that the U.S. labor market increasingly consists of low- and high-paying jobs with relatively few middle-income jobs. While this ongoing change has been noted by other researchers, Jaimovich and Siu show that this long-term evolution has not been occurring at a slow and steady rate but rather has been concentrated during aggregate downturns. They argue that the recent phenomenon of jobless recoveries is simply a reflection of the fact that these are the periods in which middle income jobs are disappearing, never to be brought back.
On the other hand, “The Labor Market Four Years Into the Crisis: Assessing Structural Explanations” by Jesse Rothstein explores and finds little direct evidence for a number of specific structural channels that might explain the weak recovery. For example, there are no identifiable sectors of the U.S. economy with strong wage growth, which suggests that the shortage of suitable workers is probably not a large constraint on employment growth.
A third hypothesis is that the weak recovery reflects an increase in economic uncertainty, which induces firms to wait rather than hire and invest. “Measuring Economic Policy Uncertainty” by Scott Baker, Nicholas Bloom, and Steve Davis proposes a novel methodology for quantifying the overall level of economic uncertainty and the portion of uncertainty that is induced by economic policy. They show that both measures of uncertainty have been elevated since the onset of the Great Recession and have scarcely recovered during recent years. “Uncertainty, Productivity and Unemployment in the Great Recession” by Edouard Schaal examines how an increase in uncertainty affects labor market outcomes in the context of a job search model. He focuses on one measure of uncertainty, the cross-sectional variability of sales growth rates across business establishments, which increased sharply in 2008 but has since subsided. Because of this finding, Schaal finds that the model can account for a large deterioration in labor market outcomes at the time of the shock but that it cannot explain why the deterioration has been so persistent.
A final hypothesis is that the weak recovery reflects disincentive effects of new tax and transfer programs that have been introduced since the onset of the recession. One aspect of this that has attracted particular attention is the extension of unemployment benefits. “The Effect of Unemployment Insurance Extensions on Reemployment Wages” by Johannes Schmieder, Till von Wachter, and Stefan Bender uses evidence from Germany to explore this hypothesis. They show that extending unemployment benefits by six months causes approximately a one-month increase in the amount of time it takes an individual to return to work. This extension has two effects on the wage of workers when they return to work. On the one hand, the additional time to look for a job allows workers to find better jobs. On the other hand, workers’ skills tend to decline during an unemployment spell. On net, these effects roughly cancel so extended benefit programs do not have a large impact on average wages.
The framework that most economists use to study the behavior of unemployed workers is search theory. Robert Hall’s paper “Viewing the Observed Acceptance Decisions of Job-Seekers through the Lens of Search Theory” analyzes detailed data on the job finding process for a sample of unemployed workers in New Jersey from 2009 in the context of this theory to assess how well the theory can provide a consistent explanation for observed behavior. Previous work had suggested that this framework has problems in accounting for observed job acceptance decisions, but Hall shows that with a few simple modifications, the framework offers a consistent explanation of how workers behave given labor market conditions.
The discussions at the conference questioned the usefulness of labels like deficient demand, structural unemployment, and cyclical unemployment. These terms mean different things in different contexts and do not clarify the key causal factors. Explanations such as “employment is slow because uncertainty is high” could easily fit under any of these banners. Instead, isolating the key changes that have taken place in the U.S. economy, and then scrutinizing the factors that have influenced how those changes have affected the labor market, would be more conducive to arriving at answers.