Where Are We in the Labor Market Recovery?
The effects of the recent recession have been especially bad for the labor market. With the current estimate of real GDP only 0.6 percent lower than its prerecession peak, most of the loss in real GDP over the course of the recession has been recovered, but payroll employment is still about 5.4 percent less than its pre-recession peak. The U.S. economy has been generating only 86,000 new nonfarm payroll jobs a month on average since the beginning of 2010. Though blurred somewhat by the hiring of temporary Census workers, total private nonfarm payrolls gives a similar picture; firms on average created 106,000 jobs in the first eleven months of 2010.
The recovery in payroll employment so far is relatively weak by historical standards. In previous recessionary episodes, it took almost 23 months for payroll employment to return to its pre-recession peak. The current recovery presents a stark contrast; even after 35 months, we are still 5.4 percent below the previous peak. The slow recovery might be due to the unusually long duration of the last recession. On average, recessions last about 10 months, but the last one lasted 18 months.
However, the unusually long duration doesn’t seem to explain the sluggish recovery in payroll employment entirely. One problem is the timing of the recovery. In all previous recessionary episodes, the end of the decline in payrolls coincided with the official end of the recession on average, about 10 months. After the last recession, however, payroll employment reached its trough in 24 months, half a year after the official end of the recession.
Another aggregate measure that we can look at to judge the strength of the recovery is the unemployment rate. The unemployment rate increased from about 5 percent in December 2007 to 10.1 percent in October 2009. Since then, it has been relatively steady above 9.5 percent, without showing much sign of improvement. The recovery so far measured by unemployment also presents a stark outlier relative to the behavior of the unemployment rate throughout previous cycles. Not only was the extent of the increase in the unemployment rate large, it has also been persistent. This level of persistence is not entirely surprising: Unemployment usually lags behind the recovery in output. On average, the unemployment rate peaks about three months after the end of the recession. In the current cycle, it took the unemployment rate four months after the end of the recession end to reach its peak. However, the picture since then is really troubling; we have not seen the unemployment rate drop below 9.5 percent since July 2009. Thus, two main measures of the aggregate labor market suggest that the labor market recovery is exceptionally weak so far.
Another disturbing feature of the weak labor market recovery so far, is the presence of the unusually high ratio of long-term unemployed. The ratio of unemployed workers who are unemployed for more than six months among the total unemployed has sharply increased over the last recession to unprecedented levels. As of May 2010, 46 percent of the unemployed reported to be out of work for more than six months, almost double the any previous recession peak in this measure in the data. Over the past six months, this fraction declined somewhat, but it still points to a pervasive long-term unemployment problem. Another measure, we can look at to see the sign of persistent unemployment is the relatively small fraction of the unemployed who are on temporary layoffs. Especially in manufacturing, some employers used to lay off their employees when the demand for their products was low, with the implicit (sometimes even explicit) understanding that they will be recalled when business conditions improve. This type of temporary adjustment in firms’ employment had a common cyclical pattern in the past. However, during the last recession, if anything, temporary layoffs declined, as a fraction of the total unemployed. Once again, this might imply that employers did not anticipate economic conditions to improve as quickly or else they were willing to go through a costly rehiring without committing to their previous employees. Thus, both of these measures indicate that long-term unemployment was a significant part of the unemployment picture in this cycle and it has not improved drastically yet.
One can think of various reasons why employment has not increased significantly in the post-recession period. Individuals working part-time for economic reasons tell part of the story. The share of workers working part-time due to slack work is at an all-time high. The share of workers who can only find part-time work although they would prefer full-time employment rose steadily throughout the recent recession as well and has just recently shown signs of leveling off. Firms can utilize those workers along the intensive margin before rehiring again. A possible evidence of that strategy could be seen in average hours data. After hitting a low-point in the 2nd quarter of 2009, average weekly hours have been steadily increasing while employment is little changed as we showed above over the same period. Average weekly hours of production employees have risen by 1.5 percent since the second quarter of 2009 as have average weekly hours of all employees. Over the same period, output per worker, or productivity, increased by almost 6 percent. Putting that gain in context, over the previous post-recessionary period (2003:Q3 to 2007:Q4), productivity increased by about 7 percent. Thus in a little over a year, productivity gains already almost equal the productivity gains of the previous post-recessionary period, despite the fact that employment growth has been sluggish.
Taken together, these data suggest weak labor demand will persist until employers exhaust the current ranks of workers who are willing to put in more hours. These conditions are indicative of a loose, rather than tight, labor market. Labor market tightness is measured by comparing job vacancies to the unemployment rate. A tight labor market is one in which the ratio of vacancies to unemployment is high. The historical pattern is clear: in tight labor markets, the probability of finding a job increases as there are a large number of vacancies relative to the pool of people looking for work. Conversely, in a looser labor market, the ratio of vacancies to unemployment is lower which translates to a lower probability of finding employment. Indeed, the current job finding rate is at an historical low while the labor market is also experiencing one of its weakest periods in several decades.