Wednesday, July 23, 2014

Unemployment and Labor Force Participation: Revisiting the Puzzle

The US unemployment rate has been painfully high in the Great Depression and its aftermath, but the high unemployment rates of the early 1980s look even worse--at least at first glance. However, the 1970s and 1980s were a time when a rising share of the adult population, and especially women, were entering the (paid) labor force, while the last few years are a time when the share of the adult population is in the labor force is declining .  Indeed, it's been a standard concern in the last few years that the official unemployment rate is not capturing the true pain in the labor market, because the official unemployment rate only includes those who are looking for work--not those who have become discouraged and given up looking. What's is the interaction between the unemployment rate and the labor force participation rate telling us?

Here'a figure showing the post-World War II unemployment rate. The monthly peak of the unemployment rate at 10% in October 2009 was second-highest of any post World War II recession, behind the peak of 10.8% in November and December 1982. In addition, from December 1979 to August 1987, a period of almost eight years, the monthly unemployment rate exceeded 6%. More recently, the unemployment rate first rose above 6% in August 2008, and as of June 2014 was at 6.1%. Thus, it's plausible that that the current stretch of monthly unemployment above 6% will last a little more than six years--which is dismal, but still with some way to go before matching the high unemployment rates that prevailed during most of the 1980s.

However, this comparison doesn't feel quite fair. After all, during the 1970s and 1980s, the labor force participation rate (that is, the share of adults who either had jobs or were unemployed and looking for jobs) was rising from 60% to about 67%. Since the start of the Great Recession, the labor force participation rate has fallen from 66% down to about 63%. An unemployment rate falling back below 6% was more comforting in the 1980s, with a rising share of adults working, than it would be in 2014, with a falling share of adults working. Here's a figure showing the labor force participation rate in the post World War II period.

To investigate these issues, the Council of Economic Advisers has just published "The Labor Force Participation Rate Since 2007: Causes and Policy Implications." The rise in the labor force participation rate from about 1960 up through the mid-1990s was driven both by the baby boom generation reaching working age, and the dramatic entry of women into the (paid) labor force. The decline since about 2000 is largely because the rising proportion of women in the labor force levelled off, and the aging of the baby boom generation is raising the number of retirees. Here's a figure from the report showing the labor force participation rate for men and women separately.

The CEA also notes that during every recession, the labor force participation rate tends to fall a little, as some people give up looking for jobs and thus aren't counted as officially "unemployed." After presenting its own analysis, and showing that it fits fairly well with previous studies of the subject, the CEA notes: "Up until the beginning of 2012 the [labor force] participation rate was generally slightly higher than would have been predicted based on the aging trend and the standard business cycle effects. But in the last two years, the participation rate has continued to fall at about the same rate even though the unemployment rate has been declining rapidly."

In other words, the drop in the labor force participation rate up through about 2012 was explainable based on the aging of the population and the common patterns during a recession. (Here's a post from April 2012 that refers to a study making this point.) In this telling, the real puzzle is not why the labor force participation rate fell up to 2012, but why it has continued to fall so quickly since 2012. To explain what is different about the period since 2012, compared with the period after previous recessions, the CEA report focuses most heavily on how long-term unemployment has been different in the aftermath of the Great Recession. For example, here's one figure showing the share of the total unemployed who have been out of work more than 27 weeks. In past recessions, this share peaked at about 20% of the total unemployed. In the Great Recession, the share of long-term unemployed peaked at above 40% of all unemployed, and even now remains at historically high levels. (Here's a post from July 2013 on the legacy of long-term unemployment.)

Another measure of long-term unemployment is the average duration of unemployment. Again, remember that only those who are actually looking for a job are counted as officially unemployed, not those who have become discouraged and stopped looking. In the last few recessions, the average length of unemployment peaked at around 20 weeks. In the Great Recession, it peaked at about 40 weeks, and is still at a discomfitingly high 35 weeks.

The CEA report looks at some other potential reasons for why labor force participation has dropped off, like a rising share of the 16-24 year-old age bracket being in school, and other breakdowns by age, gender and disability status. At least to me, the most important bottom line is not to focus on the somewhat sterile argument of who should "really" be counted as unemployed. After all, the Bureau of Labor Statistics does also collect statistics on "discouraged" workers and "marginally attached" workers. Instead, the key message is that the lower unemployment rate includes a much larger than usual share of long-term unemployed. Waiting for the U.S. economy to re-absorb these workers has not been showing good results.