Paper prepared for presentation at the Fall 2014 Brookings Panel on Economic Activity, to be presented on September 12, 2014. thank Robert Hall, David Romer, Jim Stock, and Justin Wolfers for helpful comments on an earlier draft, and Erik Larsson, Dennis Mawhirter, and Tyler Petersen for excellent research assistance. authors also benefited from conversations and collaboration with Andrew Figura, Charles Fleischman, Chris Nekarda, and David Ratner. views expressed in this paper do not necessarily reflect the views of the Board of Governors of the Federal Reserve System, the Federal Reserve Bank of Cleveland, or other members of their staffs.
I. Introduction
More than five years after the Great Recession ended, the labor market has, by many metrics, finally shown substantial improvement. unemployment rate is now nearly 4 percentage points below the peak reached in late 2009, and the number of nonfarm payroll jobs has returned to pre-recession levels. , one lingering concern is the ongoing decline in the labor force participation rate and the concomitant absence of a significant rise in the percentage of the working-age population who are employed. particular, the labor force participation rate has fallen from about 66 percent of the population in 2007 to about 63 percent over the first half of 2014, while the employment-to-population ratio currently stands at 59 percent, only about 12 percentage point above its low point in the wake of the recession (figure 1).
To an important extent, this decline in the labor force participation rate likely reflects the ongoing influence of the aging population that was the focus of a Brookings Paper written nearly a decade ago, whose authors included a subset of us (Aaronson et al., 2006). , in that paper, we predicted further declines in the participation rate over the subsequent decade as the population continued to age and the baby-boom generation continued to enter their retirement years. , population aging cannot account for the entire decline in the aggregate participation rate, and the deep recession that was precipitated by the financial crisis, along with the slow economic recovery that has followed, have led some observers to ask whether cyclical factors may have played an important role as well - and, if so, whether many individuals who dropped out of the labor force because they became discouraged about their job prospects may eventually re-enter the workforce as the labor market continues to strengthen.
The answers to these questions have important implications for government policies. much of the decline in the participation rate can be reversed (or a further decline prevented) by a sufficiently tight labor market, then policymakers should arguably take the low level of the participation rate into account in designing countercyclical policy actions. , some of the decline in the participation rate may not be amenable to countercyclical policies. We will refer to this portion of the decline as "structural" in nature, and these structural factors present a different set of challenges for policymakers. the extent that they are caused by obstacles faced by individuals who would like to work or disincentives to work, policymakers would be well-advised to look for other ways to mitigate them. , some of these structural factors may be unpreventable (such as aging of the population) or undesirable to reverse (such as higher school enrollment rates among the young), leading to a slower growth rate of potential output. , despite the significant policy implications, the range of views on these questions is surprisingly wide.
In this paper, we provide an in-depth analysis of the sources of the decline in the participation rate over the past decade or so. primary aim is to assess explanations for this recent decline in participation. , since participation rates have been falling for some demographic groups since well before the recession began, at times our analysis necessarily extends to earlier periods in order to properly frame more recent developments.
We take two separate approaches to assessing the recent decline. , we examine a number of specific explanations for the decline in participation using a variety of analyses. begin by assessing the importance the of aging of the population, which a priori seems likely to have been a significant contributor to the decline in the aggregate participation rate. then examine measures of degree of labor market attachment and use cross-state panel regressions to assess how much of the decline in participation may be the product of the recession and subsequent slow recovery that could be reversed as the economy strengthens further. We then turn to assessing the importance of several other developments that have been noted elsewhere, such as the decline in participation for teenagers and young adults and less-educated prime-age adults, and changes in retirement and disability rates, with an eye towards understanding how these developments have affected the decline in aggregate participation over the short and medium run, and whether they might be reversed going forward.
In order to account for these (and other) factors in a unified framework, our second approach to assessing the decline in labor force participation builds off of a model of the participation rate that had its genesis in the earlier Brookings paper mentioned above. model attempts to simultaneously capture the contributions of aging, the business cycle, other measurable factors - such as changes in life expectancy, educational attainment, Social Security generosity, and marriage and fertility rates - and birth-cohort-specific factors that we have not so far identified.
Combining the results from these different approaches, our overall assessment is that much - but not all - of the decline in the labor force participation rate since 2007 is structural in nature. a result, while policymakers can view some of the current low level of the participation rate as indicative of labor market slack beyond that indicated by the unemployment rate alone, they should not expect the participation rate to show a substantial increase from current levels as labor market conditions continue to improve. , as we show in the final section of the paper, projections from our model point to further declines in the trend participation rate over the next decade or so.
II. Alternative views of the recent decline in labor force participation
The prominence of the decline in the labor force participation rate since 2007, along with its importance for policymakers, has fueled a substantial flow of recent research, as well as more general commentary, on this topic. an extensive review of this literature is beyond the scope of this paper, as a prelude to our own analysis we first provide a relatively brief summary of some of the recent research that has attempted to shed light on the causes of the recent decline in participation, highlighting the wide range of often contradictory conclusions reached in these studies.1
In our view, observers should not have been particularly surprised by the fact that the labor force participation rate has declined noticeably over the past seven years. noted above, our earlier Brookings paper, which was written prior to the financial crisis, had highlighted a number of factors that were likely to put downward pressure on labor force participation over the subsequent decade, and indeed, as shown by the dashed line in figure 1, the predictions we made in that paper turned out to track the decline in the actual participation rate quite closely. said, we readily admit that the severe recession complicates the interpretation of the participation rate decline, and, more generally, we would advise against taking an overly strong signal about the sources of the decline in the aggregate participation rate from our previous forecasts. particular, although the traditional view on movements in labor force participation over the business cycle generally emphasized the absence of a substantial cyclical response, the breathtaking drop in labor demand in 2008 and 2009 may mean that this time really is different. that regard, the severity of the Great Recession and the subsequent slow pace of the economic recovery have led some researchers to interpret the decline in participation as having a large cyclical component.
A recent paper by Erceg and Levin (2013) provides a prominent example of this line of thought. first point out that labor force projections made by the BLS in November 2007 went badly off track over the next several years, both for the aggregate participation rate and for several key demographic groups.2 supplement that observation with a cross-state regression showing a statistically significant negative correlation between changes in state-level participation rates for prime-age adults from 2007 - 2012 and changes in state-level unemployment rates for this demographic group between 2007 and 2010. conclusion from this analysis is that "the aggregate decline in prime-age LFPR can be fully explained by the persistent shortfall in labor demand" (p. 15), suggesting that the current level of the unemployment rate significantly understates the extent of labor market slack. suggestive, a number of caveats pertain to that analysis, as we discuss in section IV, including that their analysis covers a short time period and does not make use of information on the relationship between the unemployment and participation rates in previous episodes.
Another paper in this vein is Hotchkiss and Rios-Avila (2013), who argue "that the dramatic decline in labor force participation during the Great Recession is more than explained by deteriorating labor market conditions (cyclical factors)" (p. 1). acknowledging the downward pressure on participation from the aging of the population, they claim that the participation rate would be even lower had individuals not altered their behavior in structural ways following the recession. particular, constructing a counterfactual participation rate path using a reweighting approach suggested by DiNardo et al. (1996), they find that, all else equal, greater educational attainment and a reduction in fertility in the wake of the recession caused labor force participation to be higher in recent years than it would have been in the absence of these behavioral responses, masking some of the effects of weak labor demand on the participation rate.
Although Hotchkiss and Rios-Avila attempt to quantify the importance of behavioral responses in participation to some extent, they do not allow for differential participation rates across different cohorts, which we find to be an important factor in the model we present later in the paper. addition, like Erceg and Levin, these authors base their estimates on a short sample period and thus do not make use of the behavior of the participation rate in other episodes. , they use average weeks worked as an indicator of labor market conditions, which, because of its mechanical relationship with the participation rate, likely biases upward their estimate of the effects of the recession on the change in labor force participation.
At the other end of the spectrum, Kudlyak (2013) uses a barebones version of the cohort-based model we present later in the paper and shows that the actual participation rate in 2012 was quite close to an estimate of the trend participation rate constructed from a model that includes only age-gender fixed effects and birth-year/gender fixed effects, and above a model that takes into account the cyclical deviation of employment from its trend. she cautions that the estimated cohort effects may be influenced by both structural and cyclical factors, she interprets her results as suggesting that most of the decline in the participation rate is accounted for by the trend. , it is difficult to assess that interpretation because she does not include other factors that might cause changes over time in the propensity of different demographic groups to participate in the labor force, as we do below.
Other authors come out somewhere in the middle. example, Aaronson et al. (2012) estimate a model that allows cohort effects and the coefficients on other controls to differ by age, sex, and educational attainment and find that nearly half of the decline in the participation rate between 2000 and 2011 reflected demographic factors. , the Council of Economic Advisors recently analyzed potential sources of the decline in participation since 2007 and attributed half of the decline to aging, one sixth to "typical" cyclical weakness, and the remainder to other pre-existing trends or other factors associated with the severity of the recession. SPAN ID="hue59">A separate analysis by Hall (2014) comes to a similar conclusion, but traces much of the decline beyond that of aging to a combination of an increase in disability recipients and the expansion of the food stamp program, both of which discourage participation by implicitly taxing earnings. Finally, pure time-series methods, such as those employed by Van Zandweghe (2012), Barnes et al. (2013), and Reifschneider et al. (2013), attribute between one half and two-thirds of the decline in participation since 2007 to trend movements, although of course such analyses say little about the underlying sources of a declining trend participation rate.
All of these research papers provide a useful perspective on recent changes in the labor force participation rate. , as Kudlyak concludes in her paper, "more research is needed that would explicitly model and account for the factors that influence the labor force participation decision of different demographic groups" (pp. 40-41). is what we attempt to accomplish in this paper. particular, we think the most promising approach to analyzing participation rate movements would ideally incorporate insights from the voluminous literature on the factors that affect the labor force participation rate. example, researchers have shown that, in addition to changes in the age distribution of the population and cohort-specific effects, the labor force participation rate is importantly driven by factors specific to labor supply among particular groups, declining market opportunities (for instance labor market polarization) and wage growth, import competition, disability, and retirement decisions. empirical model that accounts for at least some of these factors is more likely to accurately decompose participation rate movements into cyclical and the structural components than a model that is agnostic about these factors. , such a model can also enable a more precise identification of driving forces behind structural participation rate trends than embodied in the literature to date.
III. The contribution of changes in the age distribution
Perhaps the determinant of the aggregate participation rate that is easiest to analyze is the changing age distribution of the population. short, as is well known, the population as a whole has been aging, putting downward pressure on the participation rate as the large baby-boom generation moves into age groups that traditionally have low participation rates.
A "shift-share" calculation of the contribution of aging to the recent declines in the labor force participation rate (LFPR) is straightforward, but, as is often the case, the numbers can differ depending on what one chooses to use as the baseline. first column of numbers in table 1 shows three such choices.
Note that the first row of that table shows the participation rate data that have been adjusted for periodic changes to population controls and the redesign of the CPS in 1994, in order to provide a more consistent measure of participation, and, to allow for calculations by single-year ages, built up from the CPS microdata and seasonally adjusted. See section VI for more details.) differs slightly from the published rate; the latter declined 3.1 percentage points between 2007:Q4 and 2014:Q2, as compared to 2.8 percentage points in our data. will refer to the adjusted-basis aggregate rate throughout the paper, although several of the analyses will use unadjusted detailed data.
Type | 2007Q4 to 2014Q2 | 2000Q4 to 2007Q4 | 1990Q4 to 2000Q4 | 1976Q4 to 1990Q4 |
1. Actual LFPR | -2.8 | -1.3 | +0.2 | +4.5 |
---|---|---|---|---|
2. Constant LFPR | -1.5 | -0.8 | -0.3 | +0.4 |
3. Constant Share | -1.3 | -0.7 | -0.3 | +0.8 |
4. Chain-type | -1.3 | -0.7 | -0.1 | +0.7 |
Row 2 of the table holds within-age participation rates constant at their 2007:Q4 levels and lets the population share of each age evolve as in the data. ,
(1) Contribution of aging = age,sexLFPRage,sex,2007Q4* (shareage,sex,2014Q2 - shareage,sex,2007Q4),
"age" refers to single years of age from 16 to 79 and those older than 79 are treated as a single group. this calculation, aging contributed 1.5 percentage points to the total decline of 2.8 percentage points in the aggregate participation rate between the fourth quarter of 2007 and the second quarter of 2014.3
Row 3 holds the population share of each age constant and lets the age-specific participation rates evolve as in the data. portion of the actual change in the participation rate that is not accounted for by this counterfactual is the contribution of aging. is,
(2) Contribution of aging = (LFPR2014Q2 - LFPR2007Q4) - age,sexshareage,sex,2007Q4 * (LFPRage,sex,2014Q2 -
LFPRage,sex,2007Q4).
= 2014Q2 - age,sexshareage,sex,2007Q4 *LFPRage,sex,2014Q2.
By this calculation, aging contributed 1.3 percentage points to the total decline of 2.8 percentage points in the aggregate participation rate.
These two calculations differ, essentially, in their treatment of the cross terms
( LFPRage *
shareage) and, of course, are
sensitive to the particular dates chosen for fixing the weights. 4 shows a third option, a chain-type calculation, in which the age-specific participation rates are held constant only month by month.4,
(3) Contribution of aging =
,sexm=Oct 2007Jun 2014LFPRage,sex, m+1+LFPRage,sex,m2*shareage,sex,m+1-shareage,sex,m
By this calculation, aging contributed 1.3 percentage points to the total decline of 2.8 percentage points in the aggregate participation rate.
Note that none of these formulas depend directly on changes in within-age participation rates as such over the period in question. , equations (1) and (2) implicitly assume that the levels of the participation rates in the base year chosen for the calculation are appropriate. the right-most column of the table illustrates, in some contexts the results can be sensitive to that choice. (3) does not require the choice of a base year. this sense, that calculation is more robust, and is our preferred formulation. it happens, however, the three calculations in table 1 do not differ greatly when it comes to the post-2007 period: all indicate that the changing age distribution of the population has been a substantial component of the decline in aggregate labor force participation over this period, with our preferred calculation attributing nearly half of the observed decline to this source.5
The role of the age distribution is of particular interest both because it has been the focus of much of the recent commentary and because its evolution going forward is relatively easy to predict. , one could perform a similar exercise along several other dimensions, such as educational
attainment or marital status, both of which would contribute positively to the change in the participation rate in recent years and thus offset some of the effects of population aging.6 attempt to quantify the contributions of these and other factors in section VI.
IV. Assessing the recent cyclicality of labor force participation
As described in the previous section, a simple shift-share analysis suggests that nearly half of the decline in participation since 2007 can be attributed to aging. natural explanation for declining participation over the past seven years is the persistently weak labor market. , it would not be unreasonable to hypothesize that the deep recession and slow pace of recovery led an unusually large number of persons to temporarily drop out of the labor force in recent years because they were discouraged about their job prospects. so, these persons could return to the labor force when economic conditions improve sufficiently. section assesses how much the participation rate appears to be cyclically depressed using a number of different approaches.
Reasons for non-participation
We begin with an examination of the reasons labor market nonparticipants report for being out of the labor force. CPS asks nonparticipants several questions aimed at identifying discouragement, including whether an individual wants a job, is currently available to work and has looked for a job in the past 12 months (classified by the BLS as "marginally attached"), and whether he or she has a job-related reason for not currently looking for work.7 these individuals are not officially counted in the labor force, the data on gross labor market flows indicate that they have a relatively high probability of moving into the labor force (we discuss the evidence on gross labor market flows data in section VII).8
As indicated in figure 2, three successively stringent measures of labor market discouragement, defined by the questions noted above, did increase during the Great Recession and the early stage of recovery, consistent with the explanation of temporary labor market withdrawal by individuals who faced poor job prospects during cyclical downturns. , the share of the population in each of these categories is relatively small, ranging from 0.3 percent for discouraged workers to slightly below 1 percent for the marginally attached and to 212 percent for those who say that they want a job. , judging by the behavior of the broadest "want job" category, the rise in discouragement could at its peak explain at most 34 percentage point of the decline in the labor force participation rate since the end of 2007.
The "want job" category appears to have lagged somewhat the unemployment rate during the Great Recession and peaked at the end of 2012, three years later than the unemployment rate.
This observation suggests that the cyclical component of labor force participation might lag changes in the unemployment rate, an issue that we investigate more formally below. recently, all three measures have declined somewhat, but nevertheless remain elevated, suggesting that the labor force
participation rate was cyclically depressed in 2014:Q2 by perhaps 12 percentage point due to labor market discouragement as broadly defined.9
Evidence from cross-state comparisons of the reasons for nonparticipation suggests a smaller effect of discouragement on the aggregate participation rate. shown in the top panel of figure 3, the ratio of the state-level shares of nonparticipants who want a job in 2013 relative to the five-year period prior to the financial crisis varies positively with the ratio of state unemployment rates over those periods, consistent with a positive relationship between unemployment and discouragement. , as shown in the bottom panel, there appears to be no such cross-state relationship between similarly-defined ratios of the labor force participation rate and the unemployment rate. possible interpretation is that the share of nonparticipants who say that they want a job, and the cyclical variation in that share, is too slight to contribute very much to the overall cyclicality in the aggregate participation rate.10
Of course, some amount of discouragement may manifest itself in ways that are unlikely to show up in these measures. example, business-cycle conditions almost certainly affect individuals' decisions to enroll in school, apply for disability insurance, retire, or to stay home and take care of house or family. many of these individuals may not report themselves as wanting a job. , determining how much of these movements reflect cyclical and how much structural factors is difficult. will return to some of these issues later in the paper.
Using cross-state variation in the recent changes in unemployment and participation rates
An alternative way to assess the cyclicality in labor force participation in recent years is by exploiting more formally cross-state variation in labor market conditions and the participation rate. particular, because the severity of the recession varied considerably across states, the correlation between state-level movements in unemployment rates and participation rates may help to identify the effects of the business cycle on participation. , our 2006 Brookings Paper included such an analysis, and Erceg and Levin (2013) lean heavily on this estimation framework to support their claim that cyclical factors account for a large portion of the decline in participation since 2007.
Dependent variable
![]() |
(1) 16 to 24 years | (2) 25 to 54 years, men | (3) 25 to 54 years, women | (4) 25 to 54 years, all | (5) 55+ years | (6) all |
![]() |
-0.95*** | -0.36* | -0.19 | -0.29** | -0.44 | -0.50*** |
![]() |
(0.22) | (0.20) | (0.14) | (0.14) | (0.33) | (0.16) |
Constant | 0.08 | -0.59 | -0.46 | -0.48 | 3.78** | -0.30 |
Constant, standard error | (1.24) | (0.88) | (0.78) | (0.65) | (1.44) | (0.85) |
Observations | 51 | 51 | 51 | 51 | 51 | 51 |
R-squared | 0.23 | 0.13 | 0.03 | 0.12 | 0.08 | 0.25 |
Dependent variable
![]() |
(1) 16 to 24 years | (2) 25 to 54 years, men | (3) 25 to 54 years, women | (4) 25 to 54 years, all | (5) 55+ years | (6) all |
![]() |
-0.27 | -0.35 | 0.29 | -0.06 | -0.45 | -0.28 |
![]() |
(0.34) | (0.28) | (0.20) | (0.20) | (0.42) | (0.22) |
![]() |
1.25** | 0.01 | 0.88*** | 0.43 | -0.02 | 0.41* |
![]() |
(0.51) | (0.35) | (0.32) | (0.30) | (0.42) | (0.23) |
Constant | -0.50 | -0.60 | -0.87 | -0.67 | 3.79** | -0.49 |
Constant, standard error | (1.18) | (0.94) | (0.69) | (0.64) | (1.47) | (0.88) |
Observations | 51 | 51 | 51 | 51 | 51 | 51 |
R-squared | 0.32 | 0.13 | 0.17 | 0.18 | 0.08 | 0.28 |
We begin by examining cross-state regressions along the lines of those used by Erceg and Levin - that is, by regressing the state-specific change in the participation rate for various demographic groups from 2007 to 2013 on the change in the aggregate unemployment rate from 2007 to 2010, where the latter proxies for the severity of cyclical conditions during the Great Recession.11 shown in column (6) in panel A of table 2, the estimated cyclical parameter for the aggregate participation rate is -0.50. prime-age workers (those 25 to 54 years old, column 4), the estimated cyclical parameter is -0.29, and the cyclicality in participation for prime-age men is greater than for prime-age women. estimated cyclical effect is especially large for youths, and sizable, but imprecisely estimated for older individuals (55+ years) - groups which Erceg and Levin do not analyze. constant terms, which can be thought of as an estimate of the trend decline over the 2007-2013 period, are negative but estimated imprecisely; the one exception is for older individuals for whom the estimated trend is positive and statistically significant. that the average state unemployment rate rose by 5.0 percentage points between 2007 and 2010, this analysis could be interpreted as suggesting that cyclical factors contributed 212 percentage points (or 90 percent) to the decline in the aggregate participation rate between 2007 and 2013.
Erceg and Levin (2013) augment their cross-state regressions with a variable measuring the decline in the state unemployment rate from 2010 to 2012. follow this idea but add in one more year of data and include the decline in the state unemployment rate from 2010 to 2013 in our regressions shown in panel B of table 2. results, which are for prime-age adults only (not shown), continue to show a negative and statistically significant effect of the change in the unemployment rate from 2007 to 2010, but essentially no effect of the change in the unemployment rate from 2010 to 2012, which they interpret as evidence that the cyclical decline in participation is highly persistent and thus will respond with a considerable lag to improvements in labor market conditions. contrast to their findings, in our results for all ages the coefficient on the change in the unemployment rate between 2007 and 2010 remains negative but is much smaller in absolute value (and less precisely measured). addition, the coefficient on the change in the unemployment rate between 2010 and 2013 is positive and statistically significant. result could still be consistent with the lagged effects posited by Erceg and Levin (with an even greater lag between the rise in unemployment and the fall in participation than they find), but it also could reflect structural influences that reduced both state unemployment rates and state participation rates over that period. generally, however, the results from these cross-section regressions do not seem particularly robust to the choice of age group or included years.
Using state-level panel data over a longer period
The regression results in Table 2 are suggestive of important dynamic linkages between unemployment and participation, but further analysis is called for. , the apparent dynamic relationship points to the need for the explicit inclusion of lagged values of the unemployment rate in the regression specification and the use of panel data. contrast with the regressions in Table 2, the panel regressions do not assume that the change in unemployment from 2007 to 2010 is the most relevant quantity for all states. latter point may be important because not all states experienced peak unemployment in 2010, and the pace of unemployment increases and declines differed across states on a year-by-year basis. , identifying linkages between participation and unemployment from changes over a single cyclical episode risks confusing them with other idiosyncratic events, such as more generous unemployment insurance benefit durations, or other distortions to labor supply such as the increase in foodstamp take-up noted by Hall (2014). data provide additional variation that can lead to more precise estimates of cyclical elasticity of participation and its dynamics.
In light of these concerns, we next estimate panel state-level regressions using the following specification:
![]() |
where s represents state fixed effects,
t are time fixed effects,
s are state-specific time trends, and
X is a vector of covariates related to demographics.12
Dependent variable: LFPR | (1) 1978-2013 | (2) 1978-2013 | (3) 1978-2013 | (4) 1978-2013 | (5) 1990-2013 | (6) 1978-2007 |
LFPRt-1 | 0.44*** | 0.35*** | 0.39*** | |||
LFPRt-1, standard error | (0.03) | (0.03) | (0.03) | |||
Unemp. ratet | -0.17*** | -0.01 | -0.07 | -0.08** | -0.02 | -0.10** |
Unemp. ratet, standard error | (0.05) | (0.05) | (0.04) | (0.04) | (0.04) | (0.04) |
Unemp. ratet-1 | -0.10*** | -0.06* | -0.07* | -0.06 | -0.08** | |
Unemp. ratet-1, standard error | (0.03) | (0.03) | (0.04) | (0.05) | (0.04) | |
Unemp. ratet-2 | -0.09*** | -0.07** | -0.04 | -0.03 | -0.03 | |
Unemp. ratet-2, standard error | (0.03) | (0.03) | (0.04) | (0.05) | (0.04) | |
Unemp. ratet-3 | -0.25*** | -0.17*** | -0.07** | -0.10** | -0.08** | |
Unemp. ratet-3, standard error | (0.04) | (0.03) | (0.03) | (0.04) | (0.04) | |
Contrib. of trend, 2007 to 2013 | -2.34 | -1.10 | -1.45 | -1.35 | -1.53 | |
Implied cyclical shortfall, 2014Q2 | -0.12 | -1.40 | -1.02 | -1.08 | -1.02 | -1.18 |
Observations | 1836 | 1683 | 1683 | 1683 | 1224 | 1377 |
Number of states | 51 | 51 | 51 | 51 | 51 | 51 |
Demographic controls | NO | NO | YES | YES | YES | YES |
As shown in column (1) of table 3, the cyclical parameter 0 is -0.17 when only the contemporaneous unemployment
rate is included in the specification.13 the contribution of trend to be the contribution of the year fixed effects and state time trends (that is,
everything but the unemployment rate), this specification implies that trend participation fell by 2.3 percentage points from 2007 to 2013. the cyclical shortfall in the participation rate in 2014Q2 to be
0 multiplied by the cyclical shortfall in the unemployment rate as estimated by the CBO (the difference between the actual unemployment rate and the CBO's estimate of the long-run natural rate, a gap that was
about 0.7 percentage point in 2014Q2), the estimated cyclical shortfall in the participation rate from this specification is only -0.1 percentage point.
The next column adds three lags of the unemployment rate. , these lags tend to be statistically significant at conventional levels, are negative in sign, and in some cases are quite large.14 fact, the magnitudes of the later lags (the third lag in particular) are so large as to raise questions about mis-specification, which we will return to in a moment. , taking this specification at face value, the implied cyclical shortfall in participation in 2014Q2 is nearly 112 percentage points and the contribution of the trend decline in participation falls to about 1 percentage point.15 demographic controls (column 3) reduces the estimate of the cyclical shortfall somewhat and boosts the contribution of trend (for these calculations, the contribution of the demographic controls are included in the trend). , at 1 percentage point the current cyclical shortfall is still sizable.
Motivated by the implausibly large effect of longer lags of the unemployment rate, we next include one lag of the participation rate. , lags of the participation rate could matter if labor force participation is a persistent state, for example, if hysteresis effects or transition costs are important.16 this case, our finding that lags of the unemployment rate are sizable could reflect the effect of the cycle on the participation rate in previous years, rather than direct effects of lagged unemployment rates. , when we include one lag of the participation rate (column 4), the coefficients on lags of the unemployment rate are smaller and the pattern appears more plausible (earlier lags tend to be larger, and the magnitude of the third lag is more than halved).17 , in this specification, the implied current cyclical shortfall in participation is about the same as the previous specification (1 percentage point), although the contribution of trend is a bit larger.18
Finally, we have explored the robustness of these findings to different sample periods. column 5 we limit the sample to 1990 and later. contribution of the trend to the post-2007 decline and the estimate of the current cyclical shortfall in participation are fairly similar to the estimates from the full sample. we limit estimation to the pre-2007 period, the estimate of the current cyclical shortfall is again similar to previous estimates.
To summarize, our investigation of the relationship between state-level unemployment and participation rates over a multi-decade period suggests that the cyclical shortfall in the participation rate in 2014Q2 is between -0.1 percentage point (no lags of the participation or unemployment rates) and -1.0 percentage point (one lag of the participation rate and 3 lags of the unemployment rate). estimated decline in trend participation between 2007 and 2013 is between 114 and 112 percentage points when lags of the participation and unemployment rates are included. , the estimate of the current cyclical shortfall is remarkably similar regardless of whether we limit our sample to earlier or later periods. contrast to the cross-state regressions presented in the earlier section, regressions that use a longer panel of state-level data suggest that trend factors have played a much larger role in explaining recent declines in participation. , they also suggest that the current cyclical shortfall is sizable.
We conclude this section by cautioning that the estimated link between state-level labor force participation rates and state-level unemployment rate can be interpreted in several possible ways. , the participation rate changes could be related to discouragement and thus likely to reverse as the economy strengthens further. , the recession might have merely accelerated trend declines that would have happened anyway (e.g., retirements) and hence are unlikely to reverse. , the recession may have caused cyclical declines in the participation rate that will eventually become permanent (e.g., retirement or disability). , it is conceivable that the correlation might be spurious, if, for example, states with a larger share of older population (and thus more subject to participation declines due to aging) are more prone to housing booms and busts, and thus more severe downturns. this sense, empirical models that explicitly control for some of the above mentioned factors are better equipped to distinguish between trend, as we have defined it, and cycle.
V. explanations for the decline in participation
Taking the results of the previous sections at face value, of the 234 percentage point decline in participation since 2007, aging appears to explain roughly 114 percentage points of the decline, while 14 to 1 percentage point may be attributable to a weak labor market (and the lagged effect of the weak labor market over the last few years). these results are suggestive, they do not constitute a comprehensive decomposition, and they leave many questions unanswered. instance, they cannot inform us on the extent to which the remaining decline represents a continuation of secular declines in participation in some age groups that pre-dated the recession and would have occurred even in the absence of a recession. , the top panel of figure 4 plots age and education-specific participation rates, and shows that participation rates were falling for younger persons (16-24 years old) and prime-age individuals (24-54 years old) without a college degree for many years prior to the recession. In contrast, there has been an ongoing upward trend in participation rates for individuals 55years and older). can the previous results speak to whether the severity of the recession and subsequently slow labor market recovery may have induced atypically large responses on some margins of participation, such as retirement and disability insurance receipt, which are less likely to be reversed going forward. this section, we speculate on the likelihood of these possibilities.
The decline in participation for teenagers and young adults
As highlighted by the bottom panel of figure 4, the labor force participation rate for 16-24 year olds has declined from about 66 percent in 2000 to roughly 55 percent in mid-2014. fact, the decline in participation for this group has been so large that it accounts for roughly 40 percent of the downtrend in aggregate participation over this 15-year period. this trend has been documented by a number of researchers, a cohesive explanation has remained elusive.19
We begin with an examination of whether an increase in schooling or schooling intensity can explain the bulk of the decline in participation for this age group. with education as an important explanation, the drop in labor force participation among youths coincides with a general rise in their school enrollment rates (figure 5).20 younger adults (19-24), the rise in school enrollment reflects rising college enrollment, likely in response to the historically high college earnings premium (e.g. Autor, 2014). high school age students (16-18), rising annual average school enrollment rates mostly reflect higher schooling rates in the summer.21
Figure 5 also indicates that youths who report being enrolled in school in the CPS have lower labor force participation rates on average, suggesting that the rise in enrollment helps to explain some of the decline in participation for this age group. , participation rates have been falling for both enrolled and non-enrolled individuals, indicating that rising enrollment cannot explain the entirety of the overall decline. quantify the contribution of rising enrollment to declining participation, we construct a counterfactual participation rate by fixing enrollment rates at their 1985 levels and letting participation rates for enrolled and non-enrolled 16-24 year olds evolve as they did (plotted as the counterfactual participation rate in figure 6). counterfactual suggests that rising enrollment accounted for about one-half of the decline in participation for this group over the past three decades and about one-quarter of the decline since 2000.22 similar exercise for 2007-2012 implies that the rise in enrollment can explain 114 percentage points of the 434 percentage point decline in participation for 16-24 year olds over this period, or about 0.2 percentage point of the decline in the aggregate participation rate.
Of the remaining 312 percentage point decline in the youth participation rate since 2007, a simple decomposition suggests that about 212 percentage points is attributable to the decline in participation of enrolled persons, and 1 percentage point is due to the decline in participation of
non-enrolled persons.23 enrollees, a portion of the decline may be attributable to increased educational intensity, such as greater time spent on homework
or other extracurricular activities, which itself may have been induced by the elevated college earnings premium.24
Indeed, according to estimates from the American Time Use Survey (which only begin in 2003), for 16-24 year olds who reported being enrolled in high school or college, the average time spent daily on education-related activities increased by nearly 8 percent from 2003-2007 to 2008-2013 for high
school students and by 15 percent for college students.
While education-related reasons (such as the rising returns to a college degree) seem likely to be an important explanation for the decline in participation of high-school and college-age individuals over the past two decades, they cannot explain the decline in participation for non-enrolled persons.25 said, the two phenomena may be related. example, the flip side of the increased demand for college-educated workers has been a decrease in demand for some adult workers lower on the educational scale, which may have displaced them into lower-skilled sectors, thus increasing competition for jobs in the low-skilled labor market (e.g. retail sales and food service) and crowding out younger job seekers. explore this idea more carefully in the next section. , another on-going source of crowd-out may be the increasing population share of less-educated adult immigrants, as some evidence suggests that the displacement effect of immigration on the employment of younger persons is much larger than on the employment of prime-age adults (Smith, 2012).
The role of polarization in the secular decline in participation for less-educated adults
Returning to the top panel of figure 4, prime age males without a college degree have experienced a long secular decline in their participation rates, joined by prime age women without a college degree beginning in the early 2000s. declines have been the subject of a considerable literature reaching back to the 1980s. early literature, which focused on prime age men, identified declining labor market opportunities for low-skilled workers, manifested in stagnant real wage growth, as the likely explanation (e.g. Juhn, 1992). , since the 1990s, changes in labor demand have not been characterized by a monotonic increase in the demand for skilled workers, but rather by a decline in labor demand for occupations that have tended to be "middle-paying" or middle-skill jobs, and a concurrent increase in the both the share employed in higher-paying jobs (for better educated persons) and the share employed in lower-paying jobs (for less educated persons), e.g. Autor 2010.26 this polarization explain the decline in labor force participation among these workers over the past decade or two?
Polarization in labor demand, driven by exogenous technological changes and globalization, seems at least a plausible candidate explanation for some of the secular decline in participation among less-educated individuals. idea is that polarization, while increasing demand for better-educated
workers, displaces some less-educated (non-college) workers who were employed in middle-type jobs. these, some are able to transition to high-type jobs, some transition into the lower-paying service sector (perhaps displacing lower-skilled workers), and some may temporarily or permanently drop out
of the labor force, as the decline in demand for their labor pushes their offer wages below their reservation level.27 force withdrawal is likely to be most
acute for less-educated adults, since they are most likely to have been employed in middle-type or lower-type jobs.28
Panel A of figure 7 shows the share of all employed persons in the CPS who report being employed in "high-type" occupations which tend to be better paid and require more education (managerial, professional, and technical occupations), "middle-type" occupations which tend to require less education but pay reasonably well among jobs with similar educational requirements (non-managerial, non-supervisory machine operators, production workers, or office administration workers), "lower-type" occupations which tend to pay lower wages (for instance, non-managerial - non-supervisory service sector jobs in hospitality or retail), and other non-managerial, non-supervisory jobs which also tend to require less education but pay better than service sector jobs (construction, extraction and transportation occupations).29 has been documented in much other work, the share of persons employed in middle-type jobs has been falling, while the shares employed in low-skilled service sector jobs and jobs that tend to require a college degree have been increasing.30 particular, over this period, the share of prime age men without a college degree employed in middle-type jobs has fallen while the share employed in lower-type jobs has increased significantly (panel B of figure 7), and these trends have been starker than for more-educated persons (not shown).
While it is difficult to prove that polarization of labor demand caused a substantial portion of the observed decline in labor force participation among less-educated individuals, exploratory econometric evidence is supportive of the hypothesis. 4 shows results from cross-state regressions in which the dependent variable is the change in the participation rate for various demographic groups, and the right-hand side variable of interest is the change in the share of all adults in middle-type jobs, which is a proxy for the degree of polarization in the state.31 suggested by the theory outlined above, participation rates for less educated adults (columns 2 and 3) fell more in states with a greater decline in middle-type employment shares, and the elasticity is greater for these groups than for the participation rates for more educated adults (columns 4 and 5).32 , participation rates for higher-educated persons also fell more in states that experienced greater polarization, potentially consistent with the Beaudry, Green, and Sand observation that the effects of polarization have recently begun to creep up the education distribution. , column 5 of Table 4 shows that participation of 16-24 year olds fell more in states that experienced greater polarization, consistent with the earlier discussion of the decline in participation among younger persons (i.e. rising college wage premium leading to substitution towards more schooling and crowd-out by adults in the low-wage service sector).
Admittedly, this evidence is indirect. , polarization in labor demand is one of the most striking developments in the labor market over the last few decades, and it would be surprising if such a pervasive change has not left a noticeable imprint on aspects of labor supply, including participation rate trends.
25-54, less ed., Male | 25-54, less ed., Female | 25-54, more ed., Male | 25-54, more ed., Female | 16-24 | |
Change in share employed in middle-type jobs, pp. | 0.34 | 0.47 | 0.14 | 0.33 | 0.71 |
Standard error | (0.16) | (0.17) | (0.07) | (0.15) | (0.16) |
A final thought on the topic concerns the relationship between polarization and the cyclical response of participation in the wake of the financial crisis. , polarization is a long-term phenomenon. said, some research suggests that it accelerates during recessions (Jaimovich and Siu, 2012). with this phenomenon, the participation rate for less-educated prime-age men has also exhibited a stair-step pattern, discretely falling during the recession and failing to recover thereafter (figure 4, panel B). this pattern represents an actual acceleration of the process or just the reemergence of the trend during a cyclical downturn, the result is the same - polarization does not unwind as the economy expands. to the extent that polarization explains the decline in labor force participation, we would not expect to see an increase in participation among these workers as the labor market improves.
This discussion is not meant to preclude other possible explanations for the decline in participation among prime age workers, prime age men in particular. alternative is the increase in labor market opportunities for females, which could have resulted in a shift in the relative gender balance for providing household income. , this theory has to contend with the declining participation of less-educated women since the early 2000s.
The role of retirements and disability
As shown in Section III, the changing age distribution of the population can account for nearly half of the decline in the labor force participation rate since 2007. this section we examine two of the age-related factors contributing to declining participation rate in recent years: retirement and disability.
Retirement
Assessing the role of retirements for labor force participation since the start of the Great Recession is particularly challenging for several reasons. , the leading-edge of the baby-boom generation reached age 62 - the minimum age to receive Social Security retirement benefits - in 2008, concomitantly with the onset of the recession. , we would have expected an upswing in retirements to begin even absent the recession, which complicates the attempt to distinguish structural and cyclical factors. , while the recession likely affected many retirement decisions, the direction of the effect is uncertain. the one hand, staggering job losses during the recession and a subsequent lack of employment opportunities undoubtedly led some individuals to enter retirement sooner than planned. the other hand, the financial crisis, with its associated losses in housing and stock market wealth, wreaked havoc on household balance sheets, likely causing some individuals to delay their retirement.
Evidence for the expected increase in retirements associated with the aging of the baby-boom cohort is clear. can be seen in panel A of figure 8, and as has been reported elsewhere, the share of the working-age population reporting as retired in the CPS has risen by over a percentage point since 2007.33 , the dashed line, which holds the fraction retired at each age constant at their 2007 levels, illustrates that actual retirements during the recession and early recovery have lagged what would have been predicted by the changing age composition alone. , within-age retirement rates declined modestly between 2007 and 2011, offsetting much of the effect of aging during this period.34
However, as can be seen in panel B of figure 8, within-age retirement rates actually started declining around the late 1990s, likely due to a combination of institutional changes in social security and pension plans, increasing levels of education among older individuals, and longer lifespans.35 is no obvious impact of the Great Recession and its aftermath on within-group retirement rates, and thus no clear evidence in favor of a dominant effect from either increased early retirements due to discouragement or deferred retirements due to lower-than-expected asset accumulation.36 again at panel A, since 2012 - which is also the year when the first baby boomers reached full retirement age - actual retirements as measured by the CPS have risen broadly in line with predicted retirements on the basis of the changing age composition, with the share of retirees in the US population rising about 0.3 percentage point per year.
While the CPS data are a particularly useful source of data on retirement for our purposes, since they come from the same survey as the published labor force participation rate, they are not the only source of data on the issue. C of figure 8 plots Social Security retirement recipients as a percent of the working-age population; this measure also shows a sharp upturn in retirements in recent years. contrast to the CPS series, however, Social Security retirements started to turn up during the recession. possible interpretation of these patterns is that some individuals who lost their jobs began to collect Social Security retirement benefits during the recession (perhaps before their full retirement age) but nevertheless remained in the labor force to search for a new job.37
Disability
Turning to disability, the thin line in panel A of figure 9 depicts CPS self-reported nonparticipation in the labor force due to disability. due to disability as a percent of the US working-age population has been edging up by about 0.1 percentage point annually over the past decade or so, which appears to be mostly due to structural factors, with only a little evidence of cyclicality.38
While the CPS definition of disability does not depend on the receipt of SSDI and there is a difference in the levels of the CPS data and the administrative data on the receipt of SSDI benefits (the thick black line), their trajectories track each other reasonably well. 39 on the administrative data, once we account for the increase in SSDI receipts resulting from the rise in the social security full retirement age from 65 to 66 by concentrating on persons 64 and under, there is only scant indication of a cyclical increase in benefit receipt.
That said, applications for Social Security disability insurance benefits (the dotted line) did step up during the recession and only began to ease in late 2012. increase in applications alone could have resulted in a decrease in labor force participation, as applicants often view themselves as effectively disqualified from working (Autor, 2011). , we would expect any such effect to be reflected in the CPS disability measure, which as mentioned, does not seem to have risen more quickly after 2008 than before. the contrary, disability in the CPS seems to have continued to rise in recent years even as applications and actual benefit receipts have stabilized, a discrepancy for which we have no ready explanation.
Taken together, the increase in retirements and disability can account for a large portion of the rise since the recession in the share of individuals who are out of the labor market and don't want a job. , the available evidence suggests that these increases are primarily continuations of longer-term trends and so have largely been driven by structural influences. will discuss the future prospects for a cyclical movement of people from retirement and disability back into the labor force in section VII. here we finish by noting that once we account for retirement and disability, there would seem to be only a little room for a cyclical recovery among other nonparticipants who say they do not want a job. is illustrated in panel B of figure 9, which plots the cumulative change since 1995 in the share of the 16-and-over population reporting themselves out of the labor force and not wanting a job. can be seen, this series turns up sharply after the end of the Great Recession, and then rises fairly steeply in the recovery. , once we eliminate retirees, the series rises more slowly, and once we also eliminate those identifying themselves as out of the labor force due to disability, the series appears nearly flat, accounting for perhaps a 12 percentage point decrease in the participation rate since the recession - a not unusual movement, for instance, compared to the increase in nonparticipation in the years after the much smaller recession in the early 2000's.
VI. Declining participation through the lens of a model: Updating the model from the 2006 Brookings paper
While the analyses above go a long way towards quantifying the extent that aging, the business cycle, and an assortment of other factors can explain recent declines in participation, this factor-by-factor approach is limited in its ability to decompose the aggregate decline into cyclical and
structural components in a clean, integrated and consistent fashion. , outside of the contributions of aging and the cycle, it is difficult to use these analyses to project the path of aggregate labor force participation over the next decade. this section, we turn to an alternative, but
complementary, approach based on an updated version of the model introduced in Aaronson and others (2006).
Description of the model
In this cohort-based, demographically disaggregated model, we combine the changing age distribution with various factors explaining within-age changes in participation. refer the reader to the earlier Brookings paper for a general description of the model and its motivation.40 , we lay out the model briefly here, as the specification has changed significantly between 2006 and now.41
Form and estimation
The model has the form
![]() |
a = age (in single years), 16 to 7942
t = calendar time (in quarters)
s = sex
A = an age-and-sex-specific constant, i.e., an "age effect"
K = a birth-year-and-sex-specific constant, i.e., a "cohort effect" 43
X = a vector of variables that may vary by age, time and/or sex
= a vector of coefficients, which generally vary by both age and sex. coefficients are constrained to be zero for some age-sex groups.
= an i.i.d. error term.
We draw the data on participation rates by age and sex from the micro CPS files, adjust the raw series to account for changes in the survey and changes in the population controls, and seasonally adjust them.44 these reasons, the aggregate participation rates implied by these age/sex-specific participation rates differ slightly from the published aggregate rates.
We include ten variables in the vector X:
A. The aggregate unemployment rate gap, divided into positive and negative components in order to allow for asymmetric responses to tight and loose labor markets. use the long-term natural rate estimated by the CBO to define the unemployment rate gap. baseline specification includes the contemporaneous gaps and lags at 4-, 8-, and 12- quarters. have also estimated the model using only the contemporaneous unemployment rate gap and with other lag lengths; all yield similar results.45
B. The aggregate personal bankruptcy rate, as a percent of the population. principal, household wealth should influence participation decisions, most notably for retirement. , these effects are difficult to identify in aggregate data, possibly because holdings of wealth are so skewed. personal bankruptcy rate is intended to represent changes in household wealth at the most relevant parts of its distribution.
C. The percent of each age-sex group with a college degree. rates differ significantly by education, which is typically attributed to a combination of higher returns to market work and, especially at older ages, the lower physical demands usually associated with occupations requiring greater education. we include only college attainment explicitly, this variable is intended to represent the patterns in educational attainment in general, which are highly correlated over time with college attainment. tabulate this variable from the micro CPS data and include these demographically disaggregated college degree rates for ages 27 and over for both sexes.
D. Life expectancy conditional on survival to each age, 55 to 79. /SPAN>Because mortality and morbidity at older ages tend to improve together, this variable is intended to represent changes in both.46 life expectancy should increase participation by raising the level of assets necessary to finance retirement at any given age. morbidity - better health at higher ages - should be associated with a lower disutility of participation.
E. The Social Security "pay-out rate." is the average fraction of the Primary Insurance Amount (PIA) a person would receive if he or she were to retire at a particular age.47 persons below the statutory normal retirement age, a higher value implies a smaller penalty for retiring early. older persons, a higher value of the variable implies a greater reward for delaying retirement.
F. Marriage and young children. associations between labor force participation and the presence of young children appear to vary by marital status, and vice-versa. therefore include three variables to capture this interaction, at least crudely: percentage of women who are married with a child less than 6 years old, the percentage of women who are not married with a child less than 6 years old, and the percentage of women married without a child less than 6 years old.48
G. The ratio of the effective minimum wage, adjusted to account for state-level minimum wages that are above the federal level, relative to average hourly earnings.49
H. The ratio of the median hourly wage rate for ages 16-19 to the median hourly wage rate for ages 25 plus.50 variable is intended to reflect movements in the relative demand for teenagers, perhaps due to the factors discussed above in section V.
I. The ratio of summer to non-summer school enrollment rates, by age and sex, among teenagers. variable represents the expansion of schooling more generally since the mid-1990s, which, as noted above, has mainly occurred in the summer months. 51
J. The number of Social Security Disability Insurance recipients, by age and sex.52
In general, this model attempts to capture, in a parsimonious manner, many of the factors that we touched upon earlier in the paper. age effects should capture the contribution of the aging of the population. asymmetric unemployment rate gap should capture cyclical effects. additional right-hand side variables are included to account for some of the additional considerations discussed previously, such as retirement (proxied by our variables for life expectancy, Social Security generosity, and age effects in general), disability (proxied by the number of SSDI recipients), trends in school enrollment (proxied by the ratio of summer to non-summer enrollment), and possibly polarization.54
The model is estimated by least squares as a panel of 128 equations with cross-equation restrictions: one equation for each age/sex combination with the cohort effect for a given birth cohort and sex assumed to be constant across equations.55 coefficients on the right-hand-side variables may vary freely across age/sex combinations except for the many cases where they are constrained to be zero as noted above.56 57
As is well known, age, cohort, and time effects are not separately identified, and a tractable model requires choices to be made. cohort and age effects are identified (with the cohort effects for 1975 normalized to zero) in our model because we assume that there are no time effects that are not captured by the RHS variables. earlier Brookings paper and Fallick and Pingle (2007) argued that the cohort dimension is historically both important and meaningful, and that, therefore, restricting the time dimension to elucidate the cohort dimension was worthwhile.
This is especially true for women. Durand (1948) observed, "As they grew older, each successive generation of women seems to have retained the greater propensity to be in the labor force which it developed in early adulthood, and so the higher percentages of labor force members have gradually been transmitted throughout the age groups from the late 20's to the early 60's." can be seen in figure 10, which shows the participation rates for three age groups of women (35-44, 45-54, and 55-64).58 horizontal axis shows the birth year for the middle age of the group. this way, each birth cohort is vertically aligned with itself at different ages. importance of the cohort dimension can be seen in the inflection points, which occur within each age group at approximately the same birth cohorts. apparent importance of differences in the participation rates of successive cohorts means that the evolution of the aggregate participation rate ought to be at least partly predictable. , of course, one would like to model all of the economic, technological, and social factors that led to these "greater propensity[ies] ... developed in early adulthood", but that has so far not proved possible.59
, when it comes to the projections we show in section VII, restricting the time dimension to observables allows us to make reasonable, or at least clearly defined, assumptions about how the various factors represented by the RHS variables will evolve. recognize that any time-varying influences not captured by the included RHS variables may contaminate the cohort effects, causing the projections from the model, which propagate the cohort effects through the age distribution over time, to suffer accordingly.
Recent cohorts
Estimates of the cohort effects for cohorts that appear for only a few years at the beginning or end of our sample period are likely to suffer from endpoint bias if estimated as part of the system. importantly, the cohort effects for those born between 1992 and 1998 would currently be estimated
using only data since the onset of the Great Recession. only might this give a misleading impression of how much of the recent movements in participation are trend as opposed to cycle, but because the cohort effects are assumed to persist for a lifetime, this bias could prejudice the model's
predictions for future years. mitigate this problem, we estimate the model excluding the first and last 10 cohorts from the data, for estimation of both their cohort effects and the coefficients on the right-hand side variables. cohort effects for these cohorts are then linearly extrapolated from
the adjoining 10 cohort effects.60 is one of the changes from the 2006 model, in which we included these cohort effects in the estimation but constrained
them to evolve slowly.
Model estimates
We estimated the model on quarterly data over the period 1976:Q1 to 2014:Q2. 11 shows the estimates from the baseline model.61 solid line in the figure shows our calculated value for the actual labor force participation rate.62 dotted line shows the fitted values from the model. the most part, the model tracks the actual rate reasonably closely. dashed line shows the model's estimate of the trend participation rate, which, as noted above, we define as the fitted values with the unemployment gap and the detrended personal bankruptcy rate set to zero. difference between the fitted and trend values is the model's estimate of the aggregate cyclical response of the participation rate, which is shown separately in figure 12.
According to the model's estimates, the trend participation rate fell 2.1 percentage points between 2007:Q4 (at which point the participation rate was a little above trend) and 2014:Q2, accounting for three-quarters of the total decline of 2.8 percentage points in the aggregate participation
rate over this period. of 2014:Q2, the model estimates that continued weakness in the labor market was holding the participation rate down by 0.2 percentage point, with the currently low bankruptcy rate contributing a bit more, for a total cyclical response of 0.3 percentage point; in
addition, the participation rate was 0.3 percentage point lower than the model could explain.63 Of course, by construction the model cannot capture any
change in the cyclical sensitivity of participation that may have been present in the recent episode, a point to which we will return below.)
Compared to our analysis in section IV, the model estimate of the current cyclical element is about 14 percentage point smaller than a strict read of the increase in the number of discouraged workers would suggest, and 34 percentage point smaller than implied by the comparable lag structure in our
panel data analysis.
The decline in the aggregate trend combines the changing age distribution of the population with disparate movements in age/sex-specific trends (which include the influence of the cohort effects). 13 summarizes these trends for several broad age/sex groups, the general tenor of which have been well-recognized elsewhere. rates for teenagers and young adults have been trending downward for some time, with the downtrend for teenagers (not shown separately) being particularly steep. for prime-aged men have continued their long and gradual downward trend, joined relatively recently by prime-aged women, whose marked increase in participation ended around 1990.64 contrast, participation rates for older persons have been rising since the mid-1990s. age-specific trends combine with the changes in the age distribution to produce the overall downward aggregate trend.65
We are comfortable attributing some of these age-specific trends to particular explanatory variables. , some of the variables in the model are highly correlated with each other, most notably fairly monotonic variables like education levels and longevity, so the model's attributions should be interpreted with caution. , as noted above, we intended some variables to represent a broader set of related factors. example, we included conditional life expectancy in the model but assume that increases in life expectancy represent declines in age-specific morbidity as well.
That said, looking over the longer sweep of the past four decades, the model attributes a large part of the rapid increase in prime-aged women's participation rates through around 1990 to an increase in the share of women who are unmarried or without young children and to rising educational attainment, counteracted to some extent by increases in disability. , it leaves some of the increase to be accounted for by changes in cohort-specific proclivities to participate, which may reflect changes in such diverse elements as societal attitudes, workplace technologies, and reproductive technologies, to the extent that these are not already reflected in the right-hand side variables. , the model attributes some of the long-standing decline in male participation in this age group to an increase in disability rolls, but leaves the bulk of the downtrend to be "explained" by the cohort effects. cohort effects are not much of an explanation, of course; they are more of a description. suspect that the downtrend in cohort effects reflects diverse factors such as increased availability of other sources of income (such as transfer programs and labor income from other members of their households), and changes in the structure of labor demand (e.g., deindustrialization and polarization) that were unfavorable to many prime-aged men. future work we will attempt to quantify some of these factors.
Among persons of both sexes aged 55 and up, the range for which the option of retirement is most salient, the model attributes much of the marked increase in participation rates since the mid-1990s to a number of factors, including changes to Social Security rules and increased levels of education, but the largest contributor is the increasing life expectancy of men. entrance into this age group of female cohorts with greater attachment to the labor force also contributed, on net, over this period. rising contribution of cohort effects is estimated to have stopped around 2010.
Participation rates for teenagers and young adults have been falling since about 1990. model attributes the trend decline primarily to falling cohort effects, reinforced to a small extent by increasing school enrollment among teens and counteracted to some extent by changes in fertility and marriage patterns.66 model clearly has the greatest difficulty explaining the behavior of this group.
What of the model's interpretation of the changes since 2007? the most part, the model attributes this decline to the same factors that have been at work over the previous two decades: changing age distribution, falling cohort effects, rising disability rolls, and a net cyclical weakening have
pushed participation down more than changing patterns of marriage and fertility and increases in education and life expectancy have pushed them up (table 5).67 steepening of the downtrend over this period is mainly a function of the changing age distribution, as more of the large baby-boom cohort has moved farther into the age range in which the largest drop-offs in participation rates are observed, although there is also some
steepening of the decline in cohort effects, especially among prime-aged women.
Source | Contribution |
---|---|
Age Distribution | -1.3 |
Cohort Effects | -1.7 |
Unemployment Rate Gap | -0.3 |
Bankruptcy Rate | -0.2 |
Percent with College Degree | +0.2 |
Life Expectancy | +0.3 |
Social Security Pay-Out Rate | +0.1 |
Marriage x young children | +0.8 |
Minimum Wage | 0.0 |
Teenage Wage Ratio | 0.0 |
Summer Enrollment Ratio | 0.0 |
Disability Insurance | -0.5 |
Model Residual | -0.2 |
Econometric studies have traditionally found that the aggregate participation rate varies little (about its trend) over the business cycle. a glance at the movements around recessions and recoveries in figure 11, it should come as no surprise that our model finds a similarly small amount of cyclicality in the aggregate. the model's estimates, a sustained 1 percentage point increase in the unemployment rate, all else equal, currently would be expected to reduce the participation rate by something on the order of 0.2 percentage point.
As noted above, some of the RHS variables in the model may themselves vary with the state of labor demand, and it is possible that the model does an inadequate job of attributing this indirect influence to the unemployment rate. order to gain a sense of the potential size of any misattribution, we
attempted to isolate the components of the suspect variables - bankruptcy, college attainment, marriage x young children, teenage/adult wage ratio, summer/nonsummer enrollment, and SSDI recipiency - orthogonal to the cyclical state of the labor market by regressing them on the unemployment rate
variables. then replaced these variables in the model with the residuals from these regressions. resulting version of the model would attribute an additional 0.3 percentage point of the decline in aggregate participation since 2007 to labor market weakness, by increasing the estimated
contribution of labor market strength in 2007.
Robustness tests
Given the severity of the Great Recession and the drawn-out nature of the recovery, one may suspect that the cyclical sensitivity of participation during this cycle to have been greater than has been typical historically. course, no time-series model could reliably identify a change in the cyclical
coefficients in a single episode, and there is no indication in the model's residuals of nonlinearities with respect to the severity of a downturn in previous recessions. , of particular concern in our model is that the presence of "excess cyclicality" in the past 7 years may influence the
estimated cohort effects, especially for those cohorts with relatively few years in the data so far.
As described above, we attempted to mitigate any possible endpoint biases by estimating cohort effects only for cohorts with at least 10 years of available data and extrapolating cohort effects for the remaining cohorts. the length of the current period of labor market weakness, perhaps this is not
adequate. is a trade-off, of course, between limiting endpoint bias by reducing the number of cohorts included in the estimation and limiting the currency of the information used to estimate the model, and the proper balance is difficult to know. explored the robustness of the model's estimates to
issues of endpoint bias in the cohort effects in two ways.
First, we varied the number of cohorts whose effects are extrapolated by varying the minimum number of years of data a cohort has to have available in order to be included in the estimation. turns out that the estimated degree of cyclicality is not sensitive to varying the number of cohorts
estimated. The variation in the estimated cyclical departure from trend in 2014:Q2, for example, amounts to well less than 0.1 percentage point.) contrast, the estimated trends were sensitive to this choice, mostly, although not exclusively, from variation in the estimated cohort effects. 14 shows
the (model-data-consistent) actual participation rate and the trend participation rates from versions of the model that extrapolate differing numbers of cohorts. the number of cohorts extrapolated from the baseline model's 10 makes little difference. , as we increase the number of cohorts
extrapolated, the estimated downward trend becomes less steep, and as a result, more of the decline in participation over the past ten years is left unexplained by the model. can see this more easily in figure 15, which shows the residuals from the model over the current cycle.
That said, the elimination of so many cohorts from the estimation risks missing important information in the behavior of those born more than 25 years before the recession began--in particular much of the steep decline in the participation of young people that well-preceded the recession.
Second, we kept the number of extrapolated cohorts constant at 10, but varied the estimation period. this way, we can remove the Great Recession and its aftermath from the estimation completely. course, endpoint bias in this dimension can work in either direction.
For example, if the housing boom of the mid-2000s pushed the participation rates of some groups above their trend, as might be suggested by Charles, Hurst, and Notowidigdo (2013), then ending the estimation with 2007 may bias the model's trend upward. tried ending estimation in each of 2005:Q2,
2007:Q2, 2009:Q2, and 2010:Q2, in addition to the baseline of 2014:Q2.
As shown in figure 16, the model's estimated trend falls less steeply as we move the end date of estimation earlier, regardless of whether doing so ends the estimation in a period with a stronger or weaker labor market.68 a result, as when we reduced the number of cohorts included in the estimation (i.e., increased the number of cohorts extrapolated), reducing the number of years used in the estimation causes the fit of the model over the post-crisis period to deteriorate and leaves more of the net decline in participation unexplained.
As with the first exercise, the model's estimate of the sensitivity of the aggregate participation rate to the unemployment rate varies little as the end date of estimation changes. , the same is not true for the bankruptcy rate. the Great Recession in the estimation period, the coefficients on the bankruptcy rate are small and, when aggregated, negative. , and not surprisingly given the relatively small and transient previous movements in the (detrended) bankruptcy rate, the large losses of the recent housing bust and financial crisis provide the main identification of these coefficients in the baseline estimation.
VII. Future Prospects
To summarize the results so far, the model attributes three-quarters of the decline in the labor force participation rate since 2007 to demographics and other factors mostly captured by the cohort effects. have argued that the latter probably represent factors such as labor market polarization and rising disability rolls, which are likely to be structural. raises the question of how much scope there is for participation to recover as the economy improves. the unemployment rate continues to move toward its natural rate, we would expect the participation rate to move back toward its trend (which could simply entail the participation rate moving sideways or declining less quickly relative to its downward trend). , the line between being out of the labor force and unemployed can be fluid, and it is possible that some people who we might consider to have dropped out of the labor force for structural, rather than cyclical, reasons could re-enter if appropriate opportunities arose. , the participation rate was above the model's estimate of the trend just prior to the financial crisis, when the unemployment rate was low. in this section we discuss some ways to think about likely future outcomes.
As noted above, the participation rate on aggregate has not exhibited sizable cyclical fluctuations in the past. to our model, at current levels, a 1 percentage point decrease in the unemployment rate, taken alone, would result in about a 0.2 percentage point increase in the participation rate (although our panel regressions suggest an effect up to twice as large, depending on our specification). labor force participation among certain groups, notably teenagers, does move more closely with the business cycle, these groups constitute a relatively small share of the population, and therefore impart only modest cyclicality to the aggregate participation rate. the unemployment rate were to continue to decline at its pace of recent years of about 14 percentage point per quarter, it would reach the CBO's estimate of the long-term natural rate of unemployment in 2015:Q1. this assumption, and given the lags in the baseline specification, the model would predict the participation rate to fall to its trend level about a year later (see below).
Might we see a greater cyclical rebound than usual?
However, the severity of the recent recession raises the question of whether the historical degree of cyclical sensitivity of participation is a good guide to the current situation. answering this question, it helps to examine the underlying dynamics of labor force participation as illustrated by the gross monthly labor force flows in the CPS.69 can be seen in figure 17, it turns out that the nearly acyclical nature of the participation rate is the product of two opposing factors. monthly rate at which individuals enter the labor force for employment (NE flow) declines as the unemployment rate increases (panel A), no doubt as a result of reduced employment opportunities.70 at the same time, monthly transitions from outside of the labor force into unemployment (NU flow) increase (panel B). may be that the increase in NU flows is also related to poor labor market demand, as individuals who ordinarily would have moved from out of the labor force into a job (such as recent graduates) endure a spell of unemployment when opportunities are poor.71 the explanation, combining the two flows results in a pattern of overall flows from not in the labor force into the labor force that is indeed acyclical (panel C), which would be consistent with our seeing at most a modest inflow of workers into the labor force as the unemployment rate falls.
That said, the most recent data points (the squares) suggest, if anything, a somewhat lower rate of transition into the labor force than is typical at this point in the cycle, although one that is well within the range of past experience. is possible that, as the recovery continues, transitions
into the labor market will be on the correspondingly high side of historic norms, so that on balance the participation rate will move back to a more typical cyclical position.
However, another possibility is that the relatively modest flows into the labor force at this point in the cycle are indicative of a greater amount of structural damage, with the result that the labor force rebounds less.
A disaggregated view of the potential cyclical rebound
We can also think about whether some of the withdrawal from the labor market that appears structural might actually be reversed. instance, as discussed earlier, the share of nonparticipants who report that they want a job is unusually high. many of them were to return to the labor force, this could result in a greater cyclical rebound than the model currently expects.
A variety of evidence suggests that individuals reporting themselves as out of the labor force but wanting a job do have greater labor force attachment than those who are out of the labor force and do not want a job. can be seen in panel A of figure 18, the 12-month transition rate into the labor force for a person who is out of the labor force and wants a job is over 40 percent, whereas the transition rate for a non-retired, non-disabled person who is out of the labor force and does not want a job is just above 20 percent.72
Correspondingly, individuals who are out of the labor force but want a job also have spent more time working in the recent past than those who are out of the labor force and don't want a job. the March Annual Social and Economic Supplement to the CPS we can identify those who report themselves as marginally attached in the basic March survey and then examine their work history during the prior calendar year using the Supplement. analysis shows that in recent years about 30 percent of those reporting themselves as wanting a job in March worked at least one week in the prior year, compared to less than 10 percent for those who do not want a job. reporting themselves as wanting a job were also more likely to have looked for work in the prior year (over 25 percent in recent years, compared to about 2 percent for those not wanting a job). said, these individuals were still less likely to have looked for work or worked in the prior year than those who reported themselves unemployed at the time of the March survey (about 57 percent of whom engaged in work and or search).
Another way to evaluate the attachment of those not in the labor force but who want a job is to look at their impact on wages. these individuals are moving in and out of the labor force, even if they search and work less than those who are in the labor force, then they would be expected to put downward pressure on wages. cross-state evidence suggests as much: regression of the change in the state median hourly wage on the share of the state population that is out of the labor force and wanting a job (not shown), the share out of the labor force not wanting a job and the unemployment rate indicates that those who want a job do put statistically significant downward pressure on wages. 1 percentage point increase in the share of the population that wants a job holds down wage growth by 0.7 percentage point.73
Taken together, this evidence suggests that those who are marginally attached to the labor force in the broad sense do form a potential pool of workers that we might expect to see return. said, because they constitute a relatively small share of the population, their movement into labor force would
likely boost participation by only about 12 percentage point.
Expectations for the future participation rates of younger workers, recent retirees, and the recently disabled
Beyond those who identify themselves as marginally attached, there are other groups of workers whose exit from the labor force may have been prompted by the recession and weak labor market recovery, although in our discussion of the model we attributed their exit to structural factors.
One such group is teenagers and young adults, whose participation rates have dropped dramatically. noted previously, increased enrollment can explain part of the behavior of recent cohorts. the extent that lower participation rates for youths are attributable to a rising college premium and greater educational attainment, there are potentially positive effects on the labor force attachment of these cohorts later in life, since participation rates are higher on average for more educated persons. , the positive effects of additional education could offset the lost benefits of early work experience. , we have also witnessed a sharp drop in participation among nonenrollees, which seems more likely to be explained by other factors. this case, the lack of early labor market experience may have deleterious effects on later labor market experiences for these individuals. the decline in participation among teens and young adults since 2007 does not primarily appear to be related to a cyclically-induced increase in school enrollment, we probably should not expect the participation rates of these cohorts to rebound substantially beyond the usual effects of aging into their prime working years. said, it is difficult to pin down the extent to which the low participation of recent cohorts was affected by the deep recession, and so in the projections of the participation rate trend in the next section, we present alternative projections that partly speak to this question
Other groups that might seem to form a potential pool of participants include recent retirees and the disabled. , the evidence suggests that individuals who retire or become disabled return to the labor market in fairly small numbers, so that even if some of those who transitioned into these
state in response to poor job opportunities re-enter as the labor market strengthens further, the effect on the labor force participation rate is likely to be small.
As can be seen in panel A of figure 19, over 90 percent of individuals in the survey who report themselves as not in the labor force due to retirement in a given month are in that same state the following year (among those who can be matched); fewer than 4 percent of them are in the labor force a
year later (panel B).74 flows do suggest transitions from retirement into the labor market are cyclical, apparently rising when expansions are under way and
declining as they peak. , even if 12-month flows from retirement into the labor force increased to their pre-recession high of about 4 percent, this would boost the aggregate participation rate by only 0.1 percentage point, all else equal.
Given this evidence, we would only expect greater-than-usual re-entry into the labor force if there was something unique about the current episode. instance, if many more workers below age 62 had retired than is typical in less severe recessions, we might expect larger flows from retirement into the labor force, since these workers appear to have greater labor force attachment. , as shown earlier, retirement rates for those under the normal retirement age have not increased since 2007.
The picture is similar for those who report themselves as not in the labor force due to disability: recent years, nearly 75 percent of individuals who have reported themselves as disabled (figure 19, panel C) and out of the labor force in a given month remained in that state the following year. , the second most common exit from disability is into retirement (not shown), accounting for 12 percent of transitions, while another 6 percent remain out of the labor force and say they do not want a job, but no longer report themselves as disabled. told, only 6 to 7 percent of one-year transitions from disability have been into the labor force (panel D), a rate that appears to be trending down, perhaps due to population aging. striking, data from the Social Security Administration suggest that on average over the 2000's fewer than 12 percent of recipients had their benefits terminated each year due to recipients' earned income, amounting to just a few basis points on the labor force participation rate).75
One final note with respect to retirement in particular, and to a lesser extent disability: the movement of individuals into these states has only a temporary effect on the trend. age 66, nearly 60 percent of individuals are retired and by age 70 about 70 percent are. , as people age, what might have started as a premature retirement turns into an expected retirement, and the participation rate converges to its trend. is not to say that there is no cost if people were retiring or entering disability early. are welfare costs to the individuals involved as well to society as a whole. the implications for the aggregate labor force participation rate over time are muted.
Model projections
The previous analysis suggests that the cyclical increase in the participation rate as the labor market continues to improve will be at most a little larger than usual, with the most likely entrants to be from the ranks of those who are out of the labor force but want a job and from youths who obtained additional education in response to the weak labor market. any cyclical effect, the participation rate in coming years will be determined by its trend. therefore now turn to the model's forecast of the aggregate participation rate.
First, however, we calculate the implications of projected changes in the age distribution alone using equation 1 above, holding the age-specific participation rates constant at their 2014:Q2 values. use the "middle" projections from the Census Bureau to project the evolution of the age distribution from 2014:Q2 on. 1 of table 6 shows the result of this exercise. changing age distribution alone would be expected to lower the aggregate participation rate a further 2.6 percentage points over the next 10 years.
Projections | 1 Equation (1) | 2 Model: Hold incoming cohort effects and most variables constant | 3 Model: Hold incoming cohort efects contstant and extrapolate all other variables |
2014:Q2 | 62.8 | 63.1 | 63.1 |
2015:Q2 | 62.6 | 62.9 | 63.0 |
2016:Q2 | 62.3 | 62.5 | 62.7 |
2017:Q2 | 62.0 | 62.0 | 62.4 |
2018:Q2 | 61.8 | 61.5 | 62.1 |
2019:Q2 | 61.6 | 61.1 | 61.8 |
2020:Q2 | 61.3 | 60.6 | 61.5 |
2021:Q2 | 61.0 | 60.2 | 61.3 |
2022:Q2 | 60.7 | 59.8 | 61.0 |
2023:Q2 | 60.4 | 59.3 | 60.7 |
2024:Q2 | 60.2 | 58.9 | 60.4 |
Next, we use the baseline model to project the aggregate participation rate, allowing the age-specific trend participation rates to continue to evolve. terms of the model, the evolution of those age-specific trends depends on the paths of the cohort effects for future cohorts and of the various explanatory variables. course, any assumptions for these paths are highly speculative. , they may provide a sense of the range of reasonable projections.
We examine two scenarios. both, we assume that the unemployment rate falls at its pace of recent years of about 14 percentage point per quarter until it reaches the CBO's estimate of the long-term natural rate of unemployment in 2015:Q1; we hold the unemployment rate gap at zero thereafter. project the bankruptcy rate from a linear regression on the unemployment rate gap. both scenarios, we set the cohort effects for incoming cohorts to the value estimated for the most recent (as of 2014:Q2) cohort of 16-year-olds (i.e., no further declines), extrapolate a continued increase in life expectancies, and allow the Social Security payout rate to evolve according to current law. , the treatment of the other variables differs between the two scenarios.76
The first scenario, shown in column 2, to some extent isolates the implications of the aging of the population by holding the right-hand side variables not mentioned above constant at their last observed levels. scenario differs from the calculation using equation 1 along several dimensions, but the most important is that, although it holds the incoming cohort effects constant, it allows the existing cohort effects to continue to move through the age distribution as cohorts continue to age. is this assumption that is mainly responsible for the steeper downward trajectory of the participation rate: the young cohorts have lower estimated cohort effects than older ones, they pull down the within-age group participation rates as they age. the second scenario we (linearly) extrapolate forward each of the variables held constant in the first scenario. result of this exercise is shown in column 3. this case, the decline in the aggregate participation rate is similar to that in column1, as factors mentioned earlier - notably increasing longevity and educational attainment and changes in marriage and fertility patterns - continue, offsetting the propagation of the cohort effects.
Beginning from the scenario in column 3, we also imitate the earlier robustness exercises by varying either the number of cohorts included in the estimation or the estimation period. For simplicity, figure 20 shows the baseline projection (omitting the most recent 10 cohorts from the estimation and estimating the model through 2014:Q2) extrapolating the RHS variables as in column 3 of table 6, along with two alternatives: one in which the estimation omits the most recent 15 cohorts and one in which the estimation ends in 2007:Q2. By 2024, the projected participation rates from the two alternatives lie 212 and 2 percentage points, respectively, above the baseline projection. The differences evident in figures 14 and 16 are amplified as the higher estimated cohort effects propagate to older ages.77 so, both alternative scenarios eventually predict further declines in the participation rate from recent values.
Year | Authors' Model | Congressional Budget Office | Bureau of Labor Statistics | Social Security Administration | International Monetary Fund |
---|---|---|---|---|---|
2012 | 63.7 | 63.7 | 63.7 | 63.7 | 63.7 |
2013 | 63.4 | 63.3 | 63.5 | 63.3 | 63.3 |
2014 | 63.1 | 62.9 | 63.3 | 63.1 | 63.0 |
2015 | 63.0 | 62.7 | 63.1 | 63.2 | 63.0 |
2016 | 62.7 | 62.5 | 63.0 | 63.2 | 62.9 |
2017 | 62.3 | 62.4 | 62.7 | 63.2 | 62.8 |
2018 | 62.1 | 62.2 | 62.5 | 63.3 | 62.6 |
2019 | 61.8 | 62.0 | 62.3 | 63.3 | 62.3 |
2020 | 61.5 | 61.8 | 62.0 | 63.3 | - |
2021 | 61.2 | 61.5 | 61.8 | 63.1 | - |
2022 | 61.0 | 61.3 | 61.6 | 62.9 | - |
Note: Author's projections are for the annual average participation rate, as are those from the CBO, the SSA, and the International Monetary Fund. BLS projections are for the annual average trend participation rate.
Table 7 compares the projection for the annual average participation rate, based on our second scenario (column 3 above), to projections developed by several government or international agencies. 2014, our projection of the labor force participation rate, at 63.1 percent is similar to that of the other projections. , over the next eight years, we project the participation rate to decline 214 percentage points to 61 percent, a steeper decline than that projected by the BLS, CBO or SSA. said, the differences in the projections are not as stark as in the 2006 Brookings Paper, in which our 10-year ahead projection for 2015 was 212 percentage points below that of the next lowest forecast.
VIII. Conclusions
The evidence we present in this paper suggests that much of the steep decline in the labor force participation rate since 2007 owes to ongoing structural influences that are pushing down the participation rate rather than a pronounced cyclical weakness related to potential jobseekers' discouragement about the weak state of the labor market - in many ways a similar message as was conveyed in the 2006 Brookings Paper. prominently, the ongoing aging of the baby-boom generation into ages with traditionally lower attachment to the labor force can, by itself, account for nearly half of the decline. addition, estimates from our model, as well as the supplementary evidence on which we report, show persistent declines in participation rates for some specific age/sex categories that appear to have their roots in longer-run changes in the labor market that pre-date the financial crisis by a decade or more.
In particular, participation rates among youths have been declining since the mid-1990s, in part reflecting the higher returns to education documented extensively by other researchers, but also, we believe, some crowding out of job opportunities for young workers associated with the decline in middle-skill jobs and thus greater competition for the low-skilled jobs traditionally held by teenagers and young adults. "polarization" effects also appear to have weighed on the participation of less-educated prime-age men and, more recently, prime-age women. contrast, increasing longevity and better health status, coupled with changes in social security rules and increased educational attainment, have contributed to an ongoing rise in the participation rates of older individuals, but these increases have not been large enough to provide much offset to the various downward influences on the aggregate participation rate.
That is not to say that all of the decline in labor force participation reflects structural influences. cohort-based model suggests that cyclical weakness was depressing the participation rate by about 14 percentage point in 2014:Q2, while evidence from cross-state regressions suggests that the contribution of cyclical weakness could be as much as 1 percentage point. greater cyclicality evidenced in the cross-state regressions could be capturing some of the features of the current labor market we discussed outside the context of the model, such as the unusually high level of those out of the labor force who want a job, or any unusual cyclicality in youth participation or retirement.
Looking ahead, demographics will likely continue to play a prominent role in determining the future path of the aggregate labor force participation rate. youngest members of the baby-boom generation are still in their early fifties, and thus the effects of population aging will continue to put downward pressure on the participation rate for some time. , on our estimates, the continued aging of the population alone will subtract 212 percentage points from the aggregate participation rate over the next ten years. the overall downtrend could be even larger if some of the negative trends evident for particular age-sex groups persist.
Of course, considerable uncertainty attends these projections. we can be reasonably sure that the domestic population will age according to Census projections, the future pace of immigration will undoubtedly influence the age distribution of the population, as immigrants are more likely to be in their prime working years than in their 50s or 60s. , future trends in participation for specific demographic groups are difficult to predict. model has had little success in accurately capturing changes in teenage participation rates, and given the opposing effects of increased school enrollment and polarization, future changes in participation for currently younger cohorts seem especially uncertain. , while a further uptrend in the participation rates for older individuals seems likely, the pace of that uptrend is difficult to predict.
Moreover, our analysis does not account for general equilibrium feedbacks that may mitigate future declines in participation. instance, as the downward trend in participation restrains the growth in the labor force, firms may react by increasing real wage rates or otherwise making work more
attractive - say, by making work arrangements more flexible - offsetting the influence of aging and other factors.
In the end, however, we see further declines in the aggregate labor force participation rate as the most likely outcome. improvements in labor market conditions may stem that decline temporarily as discouraged workers are pulled back into the job market, and, indeed, it would not be surprising if the participation rate moved above its trend for a time. the longer-term, however, the downward influences on the aggregate labor force participation rate will likely dominate, restraining trend growth in the aggregate labor force and in the growth rate of GDP.
Finally, accepting our conclusion that the aggregate participation will likely decline further over the next decade, it is worth considering the potential implications of this development. first-order effect, mentioned above, is that--holding trends in population growth (such as migration), average hours worked, and productivity fixed--the nearly 214 percentage point decline in the aggregate participation rate we project over the next decade will continue to hold down trend output growth by a little less than 12 percentage point per year through the end of the decade. implication is that, as the growth in the labor force slows, the "break-even" level of monthly job gains required to hold the unemployment rate unchanged month-to-month will be lower than decades past. our calculations, over the next decade somewhere between 50,000 and 75,000 jobs per month will be needed to maintain an unchanged unemployment rate, well less than the amount needed in the 1990s). course, these calculations are greatly dependent on the general equilibrium concerns discussed above. optimistically, for instance, as baby boomers continue to retire, job vacancies may rise in sufficient numbers to mitigate some of the secular downtrend in participation for younger adults and less-educated workers.
Aaronson, Daniel, Jonathan Davis, and Luojia Hu. 2012. "Explaining the Decline in the U.S. Labor Force Participation Rate." Chicago Fed Letter, Federal Reserve Bank of Chicago, March.
Aaronson, Daniel, Kyung-Hong Park, and Daniel Sullivan. 2006. "The Decline in Teen Labor Force Participation." Federal Reserve Bank of Chicago Economic Perspectives 2006Q1: 2-18.
Aaronson, Stephanie, Bruce Fallick, Andrew Figura, Jonathan Pingle, and William L. Wascher. 2006. "The Recent Decline in Labor Force Participation and its Implications for Potential Labor Supply." Brookings Papers on Economic Activity, 1:2006.
Acemoglu, Daron and David H. Autor. 2011. "Skills, Tasks, and Technologies: Implications for Employment and Earnings." Handbook of Labor Economics, Vol. 4.
Arellano, Manuel and Stephen Bond. 1991. "Some Tests of Specification for Panel Data: Monte Carlo Evidence and an Application to Employment Equations." Review of Economic Studies, Vol. 58, No. 2, pp. 227-297.
Autor, David H. 2010. "The Polarization of Job Opportunities in the U.S. Labor Market: Implications for Employment and Earnings." Policy Brief by the Center for American Progress and The Hamilton Project, April 2010.
Autor, David H. 2011. "The Unsustainable Rise of the Disability Rolls in the United States: Causes, Consequences, and Policy Options." NBER Working Paper No. 17697.
Autor, David H. 2014. "Skills, Education, and the Rise of Earnings Inequality Among the `Other 99 Percent."' Science 34(6186): 843-851.
Autor, David H. and Mark G. Duggan. 2003. "The Rise in the Disability Rolls and the Decline in Unemployment." Quarterly Journal of Economics, 118(1), February.
Autor, David H. and Mark G. Duggan. 2006. "The Growth in the Social Security Disability Rolls: A Fiscal Crisis Unfolding." Journal of Economic Perspectives 20(3): 71-96.
Autor, David H. and David Dorn. 2013. "The Growth of Low-Skill Service Jobs and the Polarization of the US Labor Market." American Economic Review 103(5): 1553-1597.
Autor, David H., David Dorn, and Gordon H. Hanson. 2013. "Untangling Trade and Technology: Evidence from Local Labor Markets." NBER Working Paper No. 18938.
Balleer, Almut, Ramon Gomez-Salvador, and Jarkko Turunen. 2009. "Labour force participation in the euro area: a cohort based analysis." ECB Working Paper no. 1049, May 2009.
Balleer, Almut, Ramon Gomez-Salvador, and Jarkko Turunen. 2014. "Labour force participation across Europe: a cohort based analysis." Empirical Economics. Vol. 46(4): 1385-1415.
Barnes, Michelle L., Fabià Gumbau-Brisa, and Giovanni P. Olivei. 2013. "Cyclical versus Secular: Decomposing the Recent Decline in U.S. Labor Force Participation." Federal Reserve Bank of Boston, Public Policy Briefs, No. 13-2.
Barnichon, Regis, and Andrew Figura. 2013. "Declining Labor Force Attachment and Downward Trends in Unemployment and Participation." Finance and Economics Discussion Series 2013-88. Board of Governors of the Federal Reserve System (U.S.).
Beaudry, Paul, David A. Green, and Benjamin M. Sand. 2013. "The Great Reversal in the Demand for Skill and Cognitive Tasks." NBER Working Paper No. 18901.
Benito, Andrew, and Philip Bunn, "Understanding labour force participation in the United Kingdom." Bank of England Quarterly Bulletin, 2011 Q1, pp. 36-42.
Blau, David M. and Ryan M. Goodstein. 2010. "Can Social Security Explain Trends in Labor Force Participation of Older Men in the United States?" Journal of Human Resources, 45(2), March.
Board of Trustees of the Federal Old-Age and Survivors Insurance and Disability Insurance [OASDI] Trust Funds. 2014. "2014 Annual Report of the Board of Trustees of the Federal Old-Age and Survivors Insurance and Disability Insurance Trust Funds." Washington, DC.
Bosworth, Barry P. and Gary Burtless. 2010. "Recessions, Wealth Destruction, and the Timing of Retirement." Center for Retirement Research at Boston College, CRR WP 2010-22.
Charles, Kerwin Kofi, Erik Hurst, and Matthew J. Notowidigdo. 2013. "Manufacturing Decline, Housing Booms, and Non-Employment", NBER Working Paper No. 18949.
Clark, Kim B., and Lawrence H. Summers. 1982. "Labour Force Participation: Timing and Persistence." Review of Economic Studies, Vol. 49(2): 825-844.
Congressional Budget Office. 2014. "An Update to the Budget and Economic Outlook: 2014 to 2024." August 2014. Washington, DC.
Council of Economic Advisers. 2014. "The Labor Force Participation Rate Since 2007: Causes and Policy Implications." July.
DiNardo, John, Nicole M. Fortin, and Thomas Lemieux. 1996. "Labor Market Institutions and the Distribution of Wages, 1973-1992: A Semiparametric Approach." Econometrica, Vol 64, p. 1001-1044.
Duggan, Mark G. and Scott A. Imberman. 2009. "Why Are the Disability Rolls Skyrocketing? The Contribution of Population Characteristics, Economic Conditions, and Program Generosity." In Health at Older Ages: The Causes and Consequences of Declining Disability among the Elderly, NBER conference volume, David M. Cutler and David A. Wise, editors; pp 337-379.
Durand, John D., The Labor Force in the United States 1890-1960. New York: Social Science Research Council, 1948.
Duval-Hernández , Robert, and Pedro Orraca Romano. 2009. "A Cohort Analysis of Labor Participation in Mexico, 1987-2009." IZA Discussion Paper No.4371, August 2009.
Elsby, Michael W.L., Bart Hobijn, and Aysegul Sahin. 2013. "On the Importance of the Participation Margin for Market Fluctuations." Federal Reserve Bank of San Francisco Working Paper Series, 2013-05.
Erceg, Christopher J., and Andrew T. Levin. 2013. "Labor Force Participation and Monetary Policy in the Wake of the Great Recession," IMF Working Paper 13/245, July 2013.
Fallick, Bruce and Jonathan Pingle. 2007. "A Cohort-Based Model of Labor Force Participation." Finance and Economics Discussion Series 2007-9. Board of Governors of the Federal Reserve System (U.S.).
Fujita, Shigeru. 2014. "On the Causes of Declines in the Labor Force Participation Rate." Federal Reserve Bank of Philadelphia, Research Rap.
Gorodnichenko, Yuriy, Jae Song, and Dmitriy Stolyarov. 2013. "Macroeconomic Determinants of Retirement Timing." NBER Working Paper No. 19638.
Hall, Robert. 2014. "Quantifying the Lasting Harm to the U.S. Economy from the Financial Crisis." NBER Macroeconomic Annual 2014.
Helman, Ruth, Nevin Adams, Craig Copeland, Jack VanDerhei. 2014. "The 2014 Retirement Confidence Survey: Confidence Rebounds - for Those With Retirement Plans." Employment Benefit Research Institute, Issue Brief No. 397.
Hotchkiss, Julie L., and Fernando Rios-Avila. 2013. "Identifying Factors behind the Decline in the U.S. Labor Force Participation Rate." Business and Economic Research, Macrothink Institute, Vol. 3, No. 1, pp. 257-275.
International Monetary Fund, "United States: Selected Issues", IMF Country Report No. 14/222, July 2014,
Jaimovich, Nir and Henry E. Siu. 2012. "The Trend is the Cycle: Job Polarization and Jobless Recoveries." NBER Working Paper No. 18334.
Judson, Ruth A. and Ann L. Owen. 1999. "Estimating Dynamic Panel Data Models: A Guide for Macroeconomists." Economic Letters, Vol. 65, No. 1, pp. 9-15.
Juhn, Chinhui. 1992. "Decline of Male Labor Market Participation: The Role of Declining Market Opportunities." Quarterly Journal of Economics 107(1): 79-121.
Kawata, Hiroshi and Saori Naganuma. 2010. "Labor Force Participation Rate in Japan." Bank of Japan Review 2010-E-7, December 2010, pp. 1-9.
Kudlyak, Marianna. 2013. "A Cohort Model of Labor Force Participation." Economic Quarterly, Federal Reserve Bank of Richmond, Vol. 99, No. 1, pp. 25-43.
Maestas, Nicole. 2010. "Back to Work: Expectations and Realizations of Work after Retirement." Journal of Human Resources 45(3): 718-748.
Mastrobuoni, Giovanni. 2009. "Labor supply effects of the recent social security benefit cuts: Empirical estimates using cohort discontinuities." Journal of Public Economics, 93(11-12), December.
Morisi, Teresa K. 2008. "Youth Enrollment and Employment During the School Year." Monthly Labor Review February 2008: 51-63.
Morisi, Teresa K. 2010. "The Early 2000s: A Period of Declining Teen Summer Employment Rates." Monthly Labor Review May 2010: 23-35.
Mueller, Andreas I., Jesse Rothstein, and Till M. von Wachter. 2013. "Unemployment Insurance and Disability Insurance in the Great Recession." NBER Working Paper No. 19672.
Nekarda, Christopher J. 2009. "A Longitudinal Analysis of the Current Population Survey: Assessing the Cyclical Bias of Geographic Mobility." Working Paper. http://chrisnekarda.com/2009/05/a-longitudinal-analysis-of-the-current-population-survey/.
O'Leary, Paul, Gina Livermore, and David C. Stapleton. 2011. "Employment of Individuals in the Social Security Disability Programs." Social Security Bulletin, Vol. 71, No. 3, pp. 1-10.
Ramey, Gary and Valerie A. Ramey. 2010. "The Rug Rat Race." Brookings Papers on Economic Activity Spring 2010: 129-199.
Reifschneider, David, William L. Wascher, and David W. Wilcox. 2013. "Aggregate Supply in the United States: Recent Developments and Implications for the Conduct of Monetary Policy." Finance and Economics Discussion Series 2013-77. Board of Governors of the Federal Reserve System.
Smith, Christopher L. 2011. "Polarization, Immigration, Education: What's Behind the Dramatic Decline in Youth Employment?" Federal Reserve Board of Governors Finance and Economics Discussion Series 2011-41.
Smith, Christopher L. 2012. "The Impact of Low-Skilled Immigration on the Youth Labor Market." Journal of Labor Economics 30(1): 55-89.
Toossi, Mitra. 2011. "A Behavioral Model for Projecting the Labor Force Participation Rate." Monthly Labor Review, May, pp. 25-42.
Toossi, Mitra. 2013. "Labor Force Projections to 2022: The Labor Force Participation Rate Continues to Fall." Monthly Labor Review. Bureau of Labor Statistics. December 2013. Washington, DC.
Van Zandweghe, Willem. 2012. "Interpreting the Recent Decline in Labor Force Participation." Economic Review, Federal Reserve Bank of Kansas City, First Quarter 2012, pp. 5-34.