Unemployment Is a Labor Market Mismatch
After an unexpected upswing at the beginning of this year, the labor market appears to be back in the doldrums again. The month-on-month employment increases of 200,000-240,000 workers in December, January and February created a sense of optimism in the minds of many economists, including this one, that the economic recovery had finally seeped through to the labor market. The subsequent drop in the employment numbers to slightly more than 100,000 has left us scrambling for an explanation.
In the absence of policy changes and initiatives, the most obvious explanation for the observed volatility in the jobs numbers is that the labor market is responding primarily to short-term stimulus, in the form of seasonal changes in demand and better weather conditions. These are hardly likely to spur the economy towards a sustained long-term recovery. So what are the best policy prescriptions for these times?
Identifying the problem is the first step. As traditional economic theory suggests, hiring is ultimately a matching game. Every month when the Bureau of Labor Statistics reports the unemployment rate, the underlying assumption in the minds of most consumers of the report, is that firms created fewer jobs and therefore hiring was low. Less well understood is the idea that while the jobs exist, firms may be unable to find workers to fill those positions.
What is needed for the employment numbers to rise is not only job or vacancy creation, but also an adequate supply of workers that the firm views as good, productive matches for the jobs created. This idea is neatly summarized in a theoretical construct titled the Beveridge Curve, after the economist William Henry Beveridge. The Beveridge Curve shows the relationship between the jobs vacancy rate and the unemployment rate. The curve typically slopes downward since higher unemployment would be associated with fewer vacancies. However, if the curve moves outwards, away from the origin, this would suggest that a given level of vacancies is associated with a higher level of unemployment. This implies a less efficient labor market caused by mismatches between available jobs and the unemployed.
The graph below uses data from the BLS for the job openings rate and the U-6 rate, a broader measure of unemployment or underemployment that typically includes discouraged workers, marginally attached workers, and part-time workers. The graph breaks down the available data into the recession years, as well as the intervening years. It also shows trends since the trough of the current recession in June 2009.
In the recent 2001 recession that lasted from March to November, the average job openings rate was 3.2 percent and the U-6 rate was 8 percent. In the intervening years between that recession and the current one, the job openings rate averaged 2.8 percent while the unemployment rate increased to 9.1 percent. Hence there was a slight flattening of the curve during this period as we had fewer vacancies and higher unemployment. This trend became more pronounced in the current recession in which the job openings rate declined significantly to 2.3 percent, while the U-6 rate soared to 14.8 percent. Hence there has been a clear decline in the job creation rate in the current recession and a significant increase in the U-6 measure of unemployment.
The most interesting part of the chart is the outward movement of the curve since June 2009 to the current period. As mentioned earlier, this clearly suggests that the labor market is becoming less efficient at matching workers to available jobs. In October 2008, a vacancy rate of approximately 2.5 percent was associated with 9 percent underemployment. In October 2011, a vacancy rate of 2.5 was associated with 16 percent underemployment.
There are a number of reasons why we might see such an outward shift in the curve. First, if people have been unemployed for a long period, then firms are more reluctant to hire them as they have a negative perception about their skills and human capital. The current recession has seen an unprecedented increase in the rate of long-term unemployment. More than 42 percent of the unemployed have been out of work for 27 weeks or longer. Prolonged unemployment may be a self-reinforcing problem, as longer unemployment spells - driven in part by more generous unemployment insurance - make workers less likely to find new jobs.
Second, labor immobility could be a factor. Workers who are underwater in their mortgages may not be able to sell their homes and move to places where jobs are available. There is also some evidence to suggest that mortgage modification policies that have enabled workers to make lower mortgage payments based on their current income have changed the incentives for households to relocate from poor to better labor markets. With 28.6 percent of U.S. mortgage holders underwater, the drain of the mortgage market on labor markets should not be underestimated.
Whatever the reason behind the outward shift of the Beveridge Curve, we currently have six unemployed or underemployed workers for every job opening, nearly double that at the start of the recession. The data suggests that if we could match each worker to a job vacancy, we could fill more than 3.7 million jobs. Of course, some level of frictional unemployment is inevitable, so this is an exaggeration. But clearly, addressing the issues that hinder the matching of workers to firms could go a long way towards addressing the problem.
