The Atlanta Fed's macroblog provides commentary and analysis on economic topics including monetary policy, macroeconomic developments, inflation, labor economics, and financial issues.

Authors for macroblog are Dave Altig, John Robertson, and other Atlanta Fed economists and researchers.

« Weighing In on the Recent Discrepancy in the Inflation Statistics | Main | Labor Force Participation and the Unemployment Threshold »

May 10, 2013

Behavior’s Place in the Labor Force Participation Rate Debate

It's not often that the mainstream media is interested in the nuances of labor market statistics, so last week’s debate over the meaning of labor force participation rates (LFPR) in the pages of the Washington Post and the Wall Street Journal was music to this labor economist's ears.

Sparked by an article by Ben Casselman in his April 29 Wall Street Journal Outlook column, the ensuing back and forth (here, here, here, and here) between Casselman and the Post’s Jim Tankersley focused on what has become a central preoccupation in assessing the likely course of the labor market: Is the recent decline in the labor force participation rate the result of structural factors (e.g., an aging population) or cyclical ones (such as weak economic conditions)? Almost contemporaneously, Bill McBride declared in his recent Calculated Risk blog, "…most of the [recent] decline in the participation rate was due to changing demographics...as opposed to economic weakness."

The changing pattern of labor force participation has been a topic of discussion among economists for some time—for example, see my Federal Reserve Bank of Atlanta Economic Review article—and both Tankersley and Casselman agree that the long-run secular decline in participation is a matter worthy of independent concern. But the Federal Open Market Committee has substantially raised the stakes on disentangling longer-run trends from short-run cyclical (and presumably temporary) movements in labor force participation. It’s done this by introducing into its policy deliberations concepts like unemployment thresholds and qualitative assessments on “substantial” labor market improvement.

Casselman, in an October 2012 WSJ article, cites work by my colleagues at the Chicago Fed, who find that while more than two-thirds of the decline in LFPR between 1999 and 2011 is accounted for by changes in the age distribution of the population, "…over the 2008-2011 period...only one-quarter of the...decline of actual LFPR...can be attributed to demographic factors."

This conclusion—that three-quarters of the decline in the LFPR since the beginning of the Great Recession can be attributed to cyclical factors—is supported by other research. Colleagues at the Kansas City Fed and at the Board of Governors concur that the vast majority of the decline in the LFPR since 2008 is the result of cyclical factors. Even economists outside the Federal Reserve System acknowledge the significant role of cyclical factors in the LFPR decline (for example, see the analysis by economists at the Deutsche Bank).

But there is a critical third piece to the LFPR puzzle that most of these studies ignore. In addition to changing demographics (which have, for example, been associated with a rising share of retirement-age individuals in the total population) and cyclical effects (for example, the tendency for participation to fall when wage growth is tepid or job opportunities scarce), there are also behavioral changes afoot—a point Casselman makes in his final installment of the Post/WSJ debate. For example, individuals of near-retirement age may extend their participation as a result of significant, unexpected declines in wealth. Or women with young children—a demographic group typically less likely to participate in the labor market—may increase participation if a partner loses a job during an economic downturn. In both cases, participation rates for these demographic groups would not fall by as much as expected in response to high unemployment rates alone.

Work that I've done with Fernando Rios-Avila, a colleague at Georgia State University, finds that more than 100 percent of the fall in the LFPR since 2008 is accounted for by the condition of the labor market (cyclical factors), but these particularly strong cyclical forces were countered by increased tendencies to participate (behavioral changes). In other words, if individuals hadn't stepped up to the plate and exhibited even stronger labor force participation behavior than before the recession, the LFPR would be even lower than it is.

To illustrate the role that changing behavior played in the LFPR decline during the Great Recession, the chart below illustrates how this decline can be separated into a trend component (demographics), a cyclical component (strength of the labor market), and a behavioral component. The solid black line reflects the actual LFPR in March of each year calculated using the Current Population Survey, which is the survey data used by the U.S. Bureau of Labor Statistics to calculate the monthly labor force statistics. The orange line reflects the trend estimate of the LFPR using only demographic data (such as the age distribution of the population) through 2007, projecting out to 2012. As many others have pointed out, changing demographics—the aging of the baby-boom generations, if you will—explains only about 30 percent (in this example) of the actual post-2007 decline in LFPR.

But the chart also reveals something that may be underappreciated. Including a measure of labor market conditions in the projection of the LFPR, as well as a depiction of prerecession behavior (the green line), indicates that the LFPR should be much lower than it actually is. The message from this exercise is that the actual LFPR in 2012 was above what would have been projected had each demographic group exhibited the same labor force participation behavior after the recession as before the recession.

As it turns out, women, ethnic minorities, older people, and individuals with small children were much more likely to participate in the labor market after the recession than before it. These workers are often referred to as "added workers," or workers who join the labor force to make up for lost income elsewhere in the household. As I noted above, if these demographic groups had not increased their participation in the labor force, the aggregate LFPR would be much lower than it is.

What the chart tells us is that the cyclical factors affecting labor force participation are even more important than generally imagined. However, it is also true that the inevitable march of time will continue to put powerful downward pressure on labor force participation. Indeed, our research predicts only a modest rise in the LFPR if labor markets rebound to prerecession conditions. Our results suggest that relative to the the average LFPR over the years 2010–12, the average LFPR over the years 2015–17 will rise by about a third of a percentage point—again, if the labor market returns to prerecession conditions. Though higher than today, this level would still leave the LFPR considerably lower than it was before the recession, primarily reflecting the continued downward pressures of aging baby boomers.

Photo of Julie HotchkissBy Julie Hotchkiss, a research economist and policy adviser in the Atlanta Fed’s research department

May 10, 2013 in Economics , Labor Markets | Permalink


TrackBack URL for this entry:

Listed below are links to blogs that reference Behavior’s Place in the Labor Force Participation Rate Debate :


Your analsysis is right. In 2008, we published in the Journal of Applied Economic Sciences a paper that predicted the LFPR fall in 2010 (http://ideas.repec.org/a/ush/jaessh/v3y2008i3(5)_fall2008p203-222.html). We based our prediction on demography. Currently, the same model shows that the LFPR should return to 64% in 2013-2014 (http://mechonomic.blogspot.co.at/2013/05/the-rate-of-participation-in-labor.html) .

Posted by: kio | May 14, 2013 at 04:41 AM

That there are wildly differing takes on the underlying cause for the decline in the LFPR is no big deal. But when Fed economists differ so drastically something seems amiss. See, for example, the linked Philly Fed article from Nov. 2013, which attributes the entire drop in the LFPR since the beginning of 2012 to demographics.


Posted by: Peter Thom | May 10, 2014 at 07:06 AM

LFPR is 62.8 in April 2014, a 36 year low mark, not since 1978. If the LFPR were that of April 2000, then U3 unemployment rate would be around 12.5%. over 20 million workers would be unemployed instead of 9.753 million. That said, Japan, Italy, France and Germany have much lower rates. I wonder if the median household living costs, basic needs, has pushed the LFPR up in this country. The Economic Policy Institute publishes a Basic Family Budget Calculator showing Topeka, Kansas, at the median, a four person household needs $63,364 to get by. The median income for a working age family is $63,967 in 2010, down from $69,233 in 2000, a drop of 7.6%. (This from State of Working America, Income) Half of working age families, therefore, do not have the basic income to meet their basic family expenses. That would drive up LFPR. Though the U.S. has the highest disposable income per capita, $40,045, it does not distribute well the abundance, driving more than would be normal into the labor force. What would be normal? Preferences change as do needs.

Posted by: Ben Leet | May 13, 2014 at 02:51 PM

Post a comment

Comments are moderated and will not appear until the moderator has approved them.

If you have a TypeKey or TypePad account, please Sign in

Google Search

Recent Posts



Powered by TypePad