2011 Annual Report
The Natural Rate of Unemployment Remains Around 6 Percent
Our research on workers losing and gaining jobs strongly suggests that the long-term unemployment rate has not shifted permanently higher. Rather, labor markets are just adjusting more slowly because of lackluster leading economic growth and low labor market turnover.
Some people think that the unemployment rate will permanently remain as high as it has been recently, arguing that individuals are unemployed because their skills no longer meet the needs of employers looking to hire—a “mismatch” narrative. Other people think unemployment will eventually move lower as the economy continues to recover. The narrative here is that high unemployment mainly reflects the overall weak recovery. These two stories for today’s elevated unemployment rate have very different implications for monetary policy.
Which story to believe—is the labor market just structurally different than it used to be, with different skills in demand? Or is it mainly a matter of weak labor demand?
To answer these questions, economists estimate the natural rate of unemployment, compare it to the current level, and infer slack from the difference in the two rates. However, measuring the long-term trend of the unemployment rate that would occur without further shocks to the economy is not a simple task.
At the Federal Reserve Bank of Cleveland, we use the fact that the unemployment rate can be expressed as a combination of the flows of workers into and out of unemployment. Using this relationship, the natural rate of unemployment is the product of long-term trends in worker turnover.
The Great Recession did significantly affect the short-term flows—job-separation rates increased, while job-finding ones did the opposite. But this has not changed our estimates of long-term trends in these flows or, thus, our view of the long-term trend in the unemployment rate. The sharp rise in flows out of unemployment and the decline in worker flows into employment are fully consistent with the depth of the recession and the gradual recovery.
In our model, the long-term job-finding rate has declined greatly over the past decade, but it has been offset by the long-term decline in the separation rate. This implies a relatively stable long-term trend in the unemployment rate. Indeed, the long-term trend in unemployment rates estimated from our model has remained between 5.5 and 6 percent over the past decade or so, even during the depths of the Great Recession.
The model interprets today’s elevated rate as temporary: It reflects the prolonged aftereffects of the deep recession that have stunted economic growth. Moreover, the model suggests that the unemployment rate should move down over time and converge to its natural rate in the long run as the effects of the shocks that led to the recession diminish.
If the natural rate of unemployment has not risen, why has the labor market recovery been so sluggish?
First, the pace of job turnover has been slowing. A more dynamic labor market speeds up the adjustment process, moving the economy forward more quickly. However, worker flows now stand at historic lows—and have been trending down for several decades. Thus, the reshuffling process is not as dynamic as it has been in the past, slowing the convergence in the unemployment rate toward the long-term trend. This low level of labor turnover explains, in part, the relatively muted rebounds we have seen in labor markets over the past three cycles.
Second, the overall strength of the recovery, as measured by GDP growth, has been the weakest in the post–World War II era. A weak economy leads to weak job creation and to slow improvement in cyclical unemployment.
At this point, there is little evidence that the trends in labor market flows will change markedly over the near term. It will still take unemployed workers longer to find new jobs, and the rate of people leaving their jobs will remain low. But the slack that we see today in the labor market should be reduced if the economy continues to make substantive progress.