June Unemployment and Jobless Recoveries

Yesterday, the government released unemployment data that indicates continued weakness in labor markets.  The unemployment rate fell to 9.5%, but only because 652,000 workers dropped out of the labor-force (Timothy R. Homan and Steve Matthews, Bloomberg, "Employment Shortfall Sets the Stage for Slowdown in U.S. Economic Recovery," see also my posts, "The Little Engine Who Could?" and "The Unemployment and Labor-Force Participation Rates").  The only good news was that private sector employers added 83,000 jobs.  The previous day, the government released data on initial unemployment claims that showed that claims increased by 13,000 to 476,000, and that the four week moving average of initial unemployment claims increased 3,250 to 466,500.  Many economists believe that initial claims are a leading indicator in the movement of unemployment rates.  Robert Gordon's research finds that recessions soon hit bottom after the four week moving average peaks.  True to form, shortly after the average peaked in April 2009, the economy resumed growth.

The above graphs compare the four week moving average of initial unemployment claims and the unemployment rate of the last four recessions.  The current recession is clearly deeper than the previous three recessions.  The graphs also show that employment recovered more slowly in the last three recessions and many have dubbed the phenomenon as jobless recoveries.

James Hamilton of Econbrowser proposes an explanation in which he finds that changes in recoveries may be more cosmetic than real ("Jobless recoveries"). The Social Security Amendments of 1958 exempted unemployment insurance from income taxation.  To benefit employees, corporations who had the choice of reducing a worker's weekly hours or temporarily subjecting them to layoff, chose layoffs which increase the unemployment rate and their workers' wages net of taxes.  The situation reversed with the Revenue Act of 1978, which subjected unemployment benefits to partial taxation under the income tax law, and the Tax Reform Act of 1986, which made unemployment benefits taxable as ordinary income.  Now workers do no derive a tax benefit from being unemployed compared to having their work week reduced.  Since the mid 1980s, temporary layoffs have become a less important component of the unemployment rate.  Hamilton includes a graph which adjusts the unemployment rate for temporary layoffs which he subtracts from the unemployed.  With this adjustment, improvements in the unemployment rate lag in all recoveries.  Brad DeLong, the first to comment on Hamilton's post, and Arnold Kling, who responds in his own blog post both note that this change in definitions also increases the severity of the current recession compared to others (Econlog, "Temporary Layoffs in Postwar Recessions").

Kling adds an alternative hypothesis: fewer industries are subject to large inventory fluctuations reducing the number of workers subject to temporary layoff.  Let me add a third: regime uncertainty as described by Robert Higgs has slowed investment and recovery (The Beacon, "Regime Uncertainty—Now Maybe People Will Take the Idea Seriously".  Are these dueling hypotheses or all partial explanations?  That question can only be resolved empirically.