Recovery has started, but without the jobs | Econ Matters

I saw an opinion piece by Robert Reich last week in which he argued that the US recovery is not real until we see jobs recovering. Lots of people have turned to the game of computing what job gains would have to be realized to get back to the employment peak before the recession started in December 2007.
The point is this: We thought this recession was remarkable because of the near-financial collapse we experienced but now are learning that the incidence of the recession was borne disproportionately in employment reductions. In addition, those employment declines are barely showing any signs of rebound. Although the recovery is underway, it is being led by productivity gains, while the labor market seems to be undergoing some reallocation that is taking a while to show up.

Nationally, nonfarm employment recorded an increase of 909,000 workers in 2010. Employment gains were recorded in six of the 12 months. This is not the accelerating increases that people had hoped for, especially because employment is 7.7 million jobs below the January 2009 peak of 138 million jobs. Columbia’s job picture is still not clear because the last data point is from June 2010. One thing we can say is that from June 2009 to June 2010, Columbia employment increased 0.6 percent. Compared with the 2001 recession, employment gains are much slower. One year after the business cycle trough, we saw Columbia’s employment increase 1.8 percent.
To compare apples to apples, consider the changes in employment and output at the national level between June 2009 and June 2010. Relatively speaking, Columbia fared better than the US as national nonfarm payroll employment declined 0.4 percent between June 2009 and 2010. Meanwhile, real GDP increased 3 percent from the second quarter of 2009 and the second quarter of 2010.
So what can account for the output growth? The answer to the question is productivity gains. The amount of output per worker rose 3.4 percent during the first year of the national recovery.
Going back farther, it seems that employment has been bearing the brunt of changes in the US economy since 2007. Professor Lee Ohanian of UCLA decomposed the change in output that occurred during the 2007-2009 recession and compared it to the averages of postwar US recessions. He found that employment typically fell by 3.8 percent between business cycle peaks and business cycle troughs. For the recent downturn, employment declined 6.7 percent. If one looks at hours worked, peak-to-trough declines averaged 3.2 percent in post-war recessions but fell 8.7 percent in the 2007-2009 recession. Thus, the employment picture has been worse than average for more than three years.
With these facts in hand, it is natural to want an explanation — moreover, if the explanation offered indicates that a policy treatment is in order so that people can get back to work.
In my view, the best explanation recognizes that employment reflects a match between businesses and workers. These matches depend on the workers’ skill sets and the needs of companies. No skills were lost during the recession, which left changes in the needs of businesses.
This broad characterization covers a host of issues, generally captured by some reallocation. The reallocation story could be sectoral (like a shift from manufacturing to services) or spatial (from the Midwest to the South). I am not aware of research offering a convincing explanation for the role that financial factors played in affecting the sectoral or geographical dispersion of jobs.
What this framework does provide is a way to assess what existing policies or potential policies can address the mismatch. For example, extending unemployment insurance benefits is akin to throwing sands in the gears of the matching process. It is difficult for me to see how the Fed’s QE II can increase the rate at which matches occur.
Lastly, raising the minimum wage probably reduces the likelihood that matches will be realized. The bottom line is that it is important to understand the employment matching problem. Without such an understanding, the fixes are not easy.