The Atlanta Fed's macroblog provides commentary and analysis on economic topics including monetary policy, macroeconomic developments, inflation, labor economics, and financial issues.
- BLS Handbook of Methods
- Bureau of Economic Analysis
- Bureau of Labor Statistics
- Congressional Budget Office
- Economic Data - FRED® II, St. Louis Fed
- Office of Management and Budget
- Statistics: Releases and Historical Data, Board of Governors
- U.S. Census Bureau Economic Programs
- White House Economic Statistics Briefing Room
June 09, 2016
It’s Not Just Millennials Who Aren't Buying Homes
In recent years, much attention has been focused on the growing tendency of millennials to rent. Theories for the decrease in homeownership among young adults abound. They include rising student debt levels that crowd out additional borrowing, a tendency to live in more urban areas where the cost to buy is relatively high, a generally tougher credit environment, and even shifts in the perception of homeownership in the wake of the housing bust. The ideas have been widely debated, and yet no single factor seems to neatly explain the declining share of the millennial population opting to buy a house. (See this webcast by the Atlanta Fed's Center for Real Estate Analytics for a discussion of these issues.)
To the extent that these factors are true, they may be affecting the decisions of other generations as well. Chart 1 below shows the overall average homeownership rate and homeownership rates by age group from 1982 to 2015. It's clear that homeownership rates have declined for everyone during the past 10 years, not just for millennials.
In fact, homeownership among young Generation Xers has fallen by a bit more than the millennial generation since the housing peak—declining 11 percentage points since 2005 compared with a decline of 9 percentage points for those under 35 years old.
Another interesting point of comparison is the mid-1980s to mid-1990s, a period in which the United States had a relatively stable share of owner-occupied housing of around 64.0 percent. During the subsequent housing boom, the homeownership rate climbed to a peak of 69 percent in 2004, only to fall back down to 63.7 percent in 2015, a level similar to that prevailing before 1995. However, each age group under age 65 has a somewhat lower homeownership rate than their same-aged peers had during the 1986–94 period.
The fact that the average U.S. homeownership rate is close to rates seen in the mid-1980s and mid-1990s while homeownership rates within age groups (under 65) are currently lower than their respective averages in the mid-1980s to mid-1990s suggests that factors other than age may be affecting the average person's decision to buy or rent.
To investigate what else may be going on, charts 2 and 3 show homeownership rates by family type and race. Between 2005 and 2015, the trend mirrors what's happening by age group. The tendency to own a home has been falling for all family types and races over the past decade. In general, economic incentives (or cultural attitudes) appear to have shifted the population toward renting and away from buying.
However, the picture is quite different when you compare homeownership rates by family type and race to the pre-1995 period. While homeownership rates within age groups are generally lower today, married couples, one-person households, and nonmarried, multiperson households were all more likely to own their home in 2015. Homeownership rates across race (except for blacks) were also higher in 2015 than in 1994.
So how do we interpret the fact that the overall homeownership rate is close to its average in the 1986 to 1994 period? Are millennials to blame? Yes. But so is everyone else under the age of 65. The data suggest that whatever is affecting millennials' homeownership decisions is applicable to older individuals as well. Further, it seems there are other, possibly larger, factors affecting homeownership, such as the changing face of America. Although homeownership rates by family types and racial groups are a bit above the level seen in 1994, the average person in 2015 was about as likely to live in a home that is owned or being bought. Thus, the shift in the distribution of the population toward racial groups and family types (and likely other factors) that tend to have lower homeownership rates is likely exerting an important influence on the overall homeownership rate.
June 06, 2016
After the Conference, Another Look at Liquidity
When it comes to assessing the impact of central bank asset purchase programs (often called quantitative easing or QE), economists tend to focus their attention on the potential effects on the real economy and inflation. After all, the Federal Reserve's dual mandate for monetary policy is price stability and full employment. But there is another aspect of QE that may also be quite important in assessing its usefulness as a policy tool: the potential effect of asset purchases on financial markets through the collateral channel.
Asset purchase programs involve central bank purchases of large quantities of high-quality, highly liquid assets. Postcrisis, the Fed has purchased more than $3 trillion of U.S. Treasury securities and agency mortgage-backed securities, the European Central Bank (ECB) has purchased roughly 727 billion euros' worth of public-sector bonds (issued by central governments and agencies), and the Bank of Japan is maintaining an annual purchase target of 80 trillion yen. These bonds are not merely assets held by investors to realize a return; they are also securities highly valued for their use as collateral in financial transactions. The Atlanta Fed's 21st annual Financial Markets Conference explored the potential consequences of these asset purchase programs in the context of financial market liquidity.
The collateral channel effect focuses on the role that these low-risk securities play in the plumbing of U.S. financial markets. Financial firms fund a large fraction of their securities holdings in the repurchase (or repo) markets. Repurchase agreements are legally structured as the sale of a security with a promise to repurchase the security at a fixed price at a given point in the future. The economics of this transaction are essentially similar to those of a collateralized loan.
The sold and repurchased securities are often termed "pledged collateral." In these transactions, which are typically overnight, the lender will ordinarily lend cash equal to only a fraction of the securities value, with the remaining unfunded part called the "haircut." The size of the haircut is inversely related to the safety and liquidity of the security, with Treasury securities requiring the smallest haircuts. When the securities are repurchased the following day, the borrower will pay back the initial cash plus an additional amount known as the repo rate. The repo rate is essentially an overnight interest rate paid on a collateralized loan.
Central bank purchases of Treasury securities may have a multiplicative effect on the potential efficiency of the repo market because these securities are often used in a chain of transactions before reaching a final holder for the evening. Here's a great diagram presented by Phil Prince of Pine River Capital Management illustrating the role that bonds and U.S. Treasuries play in facilitating a variety of transactions. In this example, the UST (U.S. Treasury) securities are first used as collateral in an exchange between the UST securities lender and the globally systemically important financial institution (GSIFI bank/broker dealer), then between the GSIFI bank and the cash provider, a money market mutual fund (MMMF), corporation, or sovereign wealth fund (SWF). The reuse of the UST collateral reduces the funding cost of the GSIFI bank and, hence, the cost to the levered investor/hedge fund who is trying to exploit discrepancies in the pricing of a corporate bond and stock.
Just how important or large is this pool of reusable collateral? Manmohan Singh of the International Monetary Fund presented the following charts, depicting the pledged collateral at major U.S. and European financial institutions that can be reused in other transactions.
So how do central bank purchases of high-quality, liquid assets affect the repo market—and why should macroeconomists care? In his presentation, Marvin Goodfriend of Carnegie Mellon University concluded that central bank asset purchases, which he terms "pure monetary policy," lower short-term interest rates (especially bank-to-bank lending) but increase the cost of funding illiquid assets through the repo market. And Singh noted that repo rates are an important part of the constellation of short-term interest rates and directly link overnight markets with the longer-term collateral being pledged. Thus, the interaction between a central bank's interest-rate policy and its balance sheet policy is an important aspect of the transmission of monetary policy to longer-term interest rates and real economic activity.
Ulrich Bindseil, director of general market operations at the ECB, discussed a variety of ways in which central bank actions may affect, or be affected by, bond market liquidity. One way that central banks may mitigate any adverse impact on market liquidity is through their securities lending programs, according to Bindseil. Central banks use such programs to lend particular bonds back out to the market to "provide a secondary and temporary source of securities to the financing market...to promote smooth clearing of Treasury and Agency securities."
On June 2, for example, the New York Fed lent $17.8 billion of UST securities from the Fed's portfolio. These operations are structured as collateral swaps—dealers pledge other U.S. Treasury bonds as collateral with the Fed. During the financial crisis, the Federal Reserve used an expanded version of its securities lending program called the Term Securities Lending Facility to allow firms to replace lower-quality collateral that was difficult to use in repo transactions with Treasury securities.
Finally, the Fed currently releases some bonds to the market each day in return for cash, through its overnight reverse repo operations, a supplementary facility used to support control of the federal funds rate as the Federal Open Market Committee proceeds with normalization. However, this release has an important limitation: these operations are conducted in the triparty repo market, and the bonds released through these operations can be reused only within that market. In contrast, if the Fed were to sell its U.S. Treasuries, the securities could not only be used in the triparty repo market but also as collateral in other transactions including ones in the bilateral repo market (you can read more on these markets here). As long as central bank portfolios remain large and continue to grow as in Europe and Japan, policymakers are integrally linked to the financial plumbing at its most basic level.
To see a video of the full discussion of these issues as well as other conference presentations on bond market liquidity, market infrastructure, and the management of liquidity within financial institutions, please visit Getting a Grip on Liquidity: Markets, Institutions, and Central Banks. My colleague Larry Wall's conference takeaways on the elusive definition of liquidity, along with the impact of innovation and regulation on liquidity, are here.
June 02, 2016
Moving On Up
People who move from one job to another tend to experience greater proportionate wage gains than those who stay in their job, except when the labor market is weak and there are relatively few employment options. This point was illustrated using the Atlanta Fed's Wage Growth Tracker in this macroblog post from last year.Given that the Wage Growth Tracker ticked higher in April, it is interesting to see how much of that increase can be attributed to job switching. Here's what I found:
A note about the chart: In the chart, a "job stayer" is defined as someone who is in the same occupation and industry as he or she was 12 months ago and has been with the same employer for at least the last three months. A "job switcher" is everyone else.
The overall Wage Growth Tracker for April was 3.4 percent (up from 3.2 percent in March). For job stayers, the Tracker was 3.0 percent (up from 2.9 percent), and for job switchers it was 3.9 percent (up from 3.7 percent). So the wage gains of job switchers do appear to have helped pull up our overall wage growth measure.
Moreover, unlike the wage growth of job stayers, job switchers are now tending to see wage growth of a similar magnitude to that experienced before the recession. This observation is broadly consistent with the improvement seen during the last year in the quits rate (the number of workers who quit their jobs as a percent of total employment) from the Job Openings and Labor Turnover Survey.
I think it will be interesting to continue to monitor the influence of job switching on wage growth as a further indicator of improving labor market dynamism. An update that includes the May data should be available in a few weeks.
June 01, 2016
Putting the Wage Growth Tracker to Work
The April pop in the Atlanta Fed's Wage Growth Tracker has attracted some attention in recent weeks, resulting in some interesting analysis. What is the tracker telling us about the tightness of the labor market and the risks to the inflation outlook?
We had earlier noted the strong correlation between the Wage Growth Tracker and the unemployment rate. Tim Duy took the correlation a step further and estimated a wage Phillips curve. Here's what he found:
The chart shows that lower unemployment generally coincides with higher wage growth (as measured by the Wage Growth Tracker), but wage growth varies a lot by unemployment rate. In the past, an unemployment rate around 5 percent has often been associated with higher wage growth than we currently have.
If the Wage Growth Tracker increased further, would that necessarily lead to an increase in inflation? Jared Bernstein suggests that there isn't much of an inflation signal coming from the Wage Growth Tracker. His primary evidence is the insignificant response of core personal consumption expenditure (PCE) inflation to an increase in the Wage Growth Tracker in a model that relates inflation to lags of inflation, wage growth, and the exchange rate.
However, I don't think the absence of a wage-push inflation connection using the Wage Growth Tracker is really that surprising. The Wage Growth Tracker better captures the wage dynamics associated with improving labor market conditions than rising labor cost pressures per se. For example, if firms are replacing departing workers with relatively low-wage hires, then the wages of incumbent workers could rise faster than do total wage costs (as this analysis by our colleagues at the San Francisco Fed shows). That said, as Bernstein also pointed out in the Washington Post, it's also pretty hard to find evidence of wage pass-through pushing up inflation in his model using more direct measures of labor costs.
I look forward to seeing more commentary about Atlanta Fed tools like the Wage Growth Tracker and how they can be part of the broader discussion of economic policy.
May 23, 2016
Can Two Wrongs Make a Right?
In a recent macroblog post, I showed that forecasts from the Atlanta Fed's real gross domestic product (GDP) nowcasting model—GDPNow—have been about as accurate a forecast of the U.S. Bureau of Economic Analysis's (BEA) first estimate of real GDP growth as the consensus from the Wall Street Journal Economic Forecasting Survey. Because GDPNow essentially uses a "bean-counting" approach that tallies the forecasts of the various main subcomponents of GDP, the total GDP forecast error can be broken up into the forecast errors coming from each piece of GDP. For most of the subcomponents of GDP, the contribution to total GDP growth is approximately its real growth rate multiplied by its expenditure share of nominal GDP (the exact formulas are in the working paper for GDPNow). The following chart shows the subcomponent contributions to the GDPNow forecast errors since the third quarter of 2011. (I want to note that the forecast errors are based on the final GDPNow forecasts formed before the BEA's first estimates of GDP are released.)
The forecast errors for the subcomponents can sometimes be quite large. For example, for the fourth quarter of 2013, GDPNow underestimated the combined contributions of net exports and inventory investment by nearly 2 percentage points. However, these misses were nearly offset by overestimates of the other contributions to growth (consumption, business and residential fixed investment, and government spending).
The pattern of large but largely offsetting GDP subcomponent errors has been attributed to the work of a fictional "Saint Offset," as former Fed Governor Laurence Meyer noted in a 1998 speech. Unfortunately, "Saint Offset" doesn't always come to the forecaster's aid. For example, in the fourth quarter of 2011, GDPNow predicted 5.2 percent growth—well above the BEA's first estimate of 2.8 percent—and the subcomponent errors were predominantly on the high side.
A closer look at the chart also reveals that GDPNow has had a tendency to overestimate the contribution of business fixed investment to growth and underestimate the growth contribution of inventory investment. Although these subcomponent biases have nearly offset one another on average, we really don't want to have to rely on "Saint Offset." We would like the subcomponent forecasts to be reasonably accurate because the subcomponents of GDP are of interest in their own right.Have the subcomponent biases been a unique feature of GDPNow forecasts? It appears not. Both the Survey of Professional of Forecasters (SPF), conducted about 11 weeks prior to the first GDP release, and Blue Chip Economic Indicators, conducted as close as three weeks prior to the first release, provide consensus forecasts for some GDP subcomponents. The following table provides an average forecast error (as a measure of bias) and average absolute forecast error (as a measure of accuracy) of the subcomponent growth contributions for the two surveys and comparably timed GDPNow forecasts.
We see that the biases in GDPNow's subcomponents have been fairly similar to those in the two surveys. For example, all three sources have underestimated the average inventory investment contribution to growth by fairly similar magnitudes.
The relative accuracy of GDPNow's subcomponent and overall GDP forecasts has also been similar to the accuracy of the two surveys. "Saint Offset" has helped all three forecasters; the standard errors of the real GDP forecasts are 20 percent to 40 percent lower than they would be if the forecast errors of the subcomponents did not cancel each other out.
Finally, notice that some GDP subcomponents appear to be much more difficult to forecast than others. For instance, the bias and accuracy metrics for consumer spending are smaller than they are for inventory investment. This differential is not really that surprising, because more monthly source data are available prior to the first GDP release for consumer spending than for inventory investment.
Can we take any comfort in knowing that private forecasters have mirrored the biases in GDPNow's subcomponent forecasts? An optimistic interpretation is that the string of one-sided misses are the result of bad luck—an atypical sequence of shocks that neither GDPNow nor private forecasters could account for. A more troubling interpretation is that there have been structural changes in the economy that neither GDPNow nor the consensus of private forecasters have identified. Irrespective of the reason, though, optimal forecasts should be unbiased. If biases in some of the subcomponents continue, then forecasters will need to look for a robust way to eliminate them.
May 19, 2016
Are People in Middle-Wage Jobs Getting Bigger Raises?
As observed in this Bloomberg article and elsewhere, the Atlanta Fed's Wage Growth Tracker (WGT) reached its highest postrecession level in April. This related piece from Yahoo Finance suggests that the uptick in the WGT represents good news for middle-wage workers. That might be so.
Technically, though, the WGT is the median change in the wages of all continuously employed workers, not the change in wages among middle-income earners. However, we can create versions of the WGT by occupation group that roughly correspond to low-, middle-, and high-wage jobs, which allows us to assess whether middle-wage workers really are experiencing better wage growth. Chart 1 shows median wage growth experienced by each group over time. (Note that the chart shows a 12-month moving average instead of a three-month average, as depicted in the overall WGT on our website.)
Wage growth for all three categories has risen during the past few years. However, the timing of the trough and the speed of recovery vary somewhat. For example, wage growth among low-wage earners stayed low for longer and then recovered relatively more quickly. Wage growth of those in high-wage jobs fell by less but also has recovered by relatively less. In fact, while the median wage growth of low-wage jobs is back to its 2003–07 average, wage growth for those in high-wage jobs sits at about 75 percent of its prerecession average.
Are middle-wage earners experiencing good wage growth? In a relative sense, yes. The 12-month WGT for high-wage earners was 3.1 percent in April compared with 3.2 percent and 3.0 percent for middle- and low-wage workers, respectively. So the typical wage growth of those in middle-wage jobs is trending slightly higher than for high-wage earners, a deviation from the historical picture.
Interestingly, this pattern of wage growth doesn't quite jibe with the relative tightness of the labor market for different types of jobs. As was shown here, the overall WGT appears to broadly reflect the tightness of the labor market (possibly with some lag).
In theory, as the pool of unemployed shrinks, employers will face pressure to increase wages to attract and retain talent. Chart 2 shows the 12-month average unemployment rates for people who were previously working in one of the three wage groups.
Like the relationship between overall WGT and the unemployment rate, wage growth and the unemployment rate within these wage groups are negatively correlated (in other words, when the unemployment rate is high, wage growth is sluggish). The correlation ranges from minus 0.81 for low-wage occupations to minus 0.88 for middle-wage occupations.
However, notice that although the current gap between unemployment rates across the wage spectrum is similar to prerecession averages, the current relative gap in median wage growth is different than in the past. In particular, the wage growth for those in higher-wage jobs has been sluggish compared to middle- and lower-wage occupations.
Nonetheless, it's clear that the labor market is getting tighter. Wage growth overall has moved higher over the past year, driven primarily by those working in low- and middle-wage jobs. Is firming wage growth starting to show up in price inflation? Perhaps.
The consumer price index inflation numbers moved higher again in April, and Atlanta Fed President Dennis Lockhart said on Tuesday that—from a monetary policy perspective—recent inflation readings and signs of better growth in economic activity during the second quarter (as indicated by the Atlanta Fed's GDPNow tracker) are encouraging signs.
May 16, 2016
GDPNow and Then
Real-time forecasts from the Atlanta Fed’s real gross domestic product (GDP) nowcasting model—GDPNow—have been regularly updated since August 2011 (the model was introduced online in July 2014). So we now have a nearly five-year history to allow us to evaluate the accuracy of the model’s forecasts. The chart below shows forecasts from GDPNow (red dots) alongside actual first estimates of real GDP growth (gray bars) from the U.S. Bureau of Economic Analysis (BEA). For comparison, the blue dots in the chart are the consensus (average) forecasts from the Wall Street Journal Economic Forecasting Survey (WSJ Survey).
The initial estimate of real GDP growth for a particular quarter is usually published at the end of the subsequent month. The WSJ Survey consensus forecasts plotted above were released about two weeks before these estimates. To maintain comparable timing with the WSJ Survey, the GDPNow forecasts shown in the chart are those constructed on or before the 12th day of the same month.
Occasionally, there has been relatively large disagreement between GDPNow and the WSJ consensus. For example, GDPNow predicted that GDP growth would be below 0.5 percent for five out of 19 quarters between 2011 and 2016, and the lowest WSJ Survey consensus forecast for any of those quarters was 1.3 percent. Nonetheless, the average accuracy of the GDPNow and WSJ Survey consensus forecasts has been similar: the average absolute forecast error (average error without regard to sign) for GDPNow was 0.56 versus 0.60 for the WSJ Survey consensus.
Studies have shown that the average or median of a set of professional forecasts tends to be more accurate than an individual forecaster (see, for example, here and here). Therefore, it’s surprising that GDPNow has been about as accurate on average as the WSJ Survey consensus. To see just how surprising this result is, I used the fact that the WSJ Survey provides both the names and forecasts of its respondents. From these, I constructed a panel dataset with each respondent’s absolute forecast errors and their absolute disagreement (difference) from the consensus forecast. Using a standard econometric technique (a two-way fixed-effects regression), we can then calculate each panelist’s average absolute GDP forecast error and their average absolute disagreement with the WSJ Survey consensus. These points are shown in the scatterplot below.
There is a clear inverse relationship between average forecast accuracy and average disagreement with the WSJ Survey consensus. However, GDPNow’s accuracy and disagreement statistics do not fit the general pattern. GDPNow (the orange diamond in the chart) was more accurate on average than all but six out of 49 WSJ panelists, though at the same time it differed from the consensus by more on average than all but four of the panelists.
What should one infer from all of this? Differences in forecasting method could be part of the explanation. GDPNow differs from many other approaches to nowcasting in that it is essentially a “bean counting” exercise. It doesn’t use historical correlations of GDP with other economic series in the way that commonly used dynamic factor models do, and it also doesn’t incorporate judgmental adjustments (see here for more discussion of these differences). During a period when the economy has been giving very mixed signals, perhaps it doesn’t come as a surprise that GDPNow’s forecasts occasionally deviate quite a bit from the WSJ Survey consensus. Time will tell if GDPNow continues to perform at least as well as the consensus.
May 04, 2016
What's behind the Recent Uptick in Labor Force Participation?
The labor force participation rate had been generally declining since around 2007. However, that trend has partially reversed in recent months. As noted in the minutes of the March meeting of the Federal Open Market Committee, this rise was interpreted as further strengthening of the labor market. But will the increase persist?
As shown in a previous macroblog post, the dominant contributor to the decline in participation during the last several years has been the aging of the population. To see what's behind the increase in participation during the last few months, the following chart breaks the participation rate change between the first quarters of 2015 and 2016 into a part that is the result of shifts in the age distribution (holding behavior within age groups fixed), and the parts that are the result of changes in behavior (holding the age distribution fixed).
During the last year, the negative effect on participation attributable to an aging population (0.22 percentage points) has been offset by a 0.23 percentage point decline in the share of people who want a job but are not counted as unemployed (including people who are marginally attached). This decline is an encouraging sign, and consistent with a tightening labor market.
How much more can the want-a-job category improve? We don't really know. But that category's share of the population is currently about 0.3 percentage points above the prerecession trough of 2.0 percent. So at the current pace we would be at prerecession levels in about a year.
Despite the recent uptick, projections over the next decade or so have the labor force participation rate moving lower, chiefly because of an aging population. But how much farther participation actually declines will also depend on the evolution of various behavioral factors. The employment report for April will be released this Friday by the U.S. Bureau of Labor Statistics, and it will be interesting to see whether the number of people on the margin of the labor force continues to shrink.
April 29, 2016
Is the Number of Stay-at-Home Dads Going Up or Down?
A recent Wall Street Journal post observed that most of the recession's "stay-at-home dads" are going back to work. Specifically, data from the U.S. Labor Department shows that the share of married men with children under 18 who are not employed (but their spouse is) rose during the recession and has since given back much of that increase, as the Journal's chart below indicates.
Of course, being a stay-at-home dad in the sense defined in the previous chart (that is, not employed) can be either involuntary because of unemployment, or it can be the result of a voluntary decision to not be in the workforce. Most of the variation in the previous chart is cyclical, suggesting that it is related to the rise and fall in unemployment. But it also looks like the share of stay-at-home dads is higher now than it was a decade or so ago. So perhaps there is also an increasing trend in the propensity to voluntarily be a stay-at-home dad.
To explore this possibility, the next chart shows the annual average share of married men ages 25–54 who have children and who say the main reason they do not currently want a job is because of family or household responsibilities. (This reason doesn't necessarily imply that they are looking after children, but it is likely to be the leading reason.) The fraction is very small—about 1.3 percent in 2015, or 285,000 men—but the share has more than doubled during the last 15 years and would account for about half of the elevated level of the stay-at-home rate in 2015 relative to 2000.
So although large numbers of unemployed stay-at-home dads have been going back to work, it also appears that there's a small but growing group of men who are choosing to take on household and family responsibilities instead.
April 15, 2016
Labor Force Participation: Aging Is Only Half of the Story
The labor force participation rate (LFPR) is an important ingredient in projecting employment growth and the unemployment rate. However, predicting the LFPR has proven difficult. For example, in 2011 the Congressional Budget Office (CBO) projected that the LFPR in 2015 would be about 64.3 percent. In reality, the LFPR turned out to be 62.6 percent. Based on the CBO projection, the economy would have needed to create about 4 million more jobs to reach the 2015 unemployment rate of 5.3 percent.
Why is the LFPR so hard to predict? Leaving aside the challenge of projecting the size of the population, movements in LFPR primarily reflect shifts in the age distribution of the population as well as a number of behavioral factors. Although the aging trends are largely baked in, the behavioral factors vary over time. According to our estimates, about half of the 3.4 percentage-point decline in the LFPR between 2007 and 2015 is the result of the aging of the population, while behavioral factors account for the rest.
The complication is that the specific behaviors can change. The following chart shows a decomposition of the change in LFPR from 2007 to 2011 and from 2011 to 2015. Though the aging of the population contributed about the same amount to the decline in LFPR in both periods, the contributions from other factors have varied a lot. (We delve into the changes in the factors following the chart.)
Aging: The single largest factor contributing to the decline in the overall LFPR has been the rising share of older Americans in the population. In 2007, about one in five Americans were over 60 years old. In 2015, almost one in four were over 60. Moreover, this demographic force will continue to suppress the overall LFPR as the share of older Americans increases further in coming years. (For an in-depth discussion of the economic implications of an aging population—including changes in the labor market—please read the Atlanta Fed's 2015 annual report.)
Later retirement: One countervailing factor to an aging population has been the rising LFPR of older individuals. The retirement rate of those over 60 declined between 2007 and 2011 by a similar amount as it had before the recession. However, the trend toward later retirement has slowed considerably in recent years. The reason for this slowing is a puzzle and has important implications for the future course of overall LFPR.
Schooling among the young: The enrollment in educational programs by American youth has been generally rising over several decades, and this trend has put downward pressure on the overall LFPR. During the 2007–11 period, the share of 16- to 24-year-olds who do not want a job because they were in school or college accelerated relative to the prerecession trend. However, enrollment rates have since flattened out. The slowing may reflect enrollment rates catching up to the longer-term trend or may be a result of changes in the opportunity cost of education.
Not in the labor force but want a job: The share of the population saying they want a job but are not classified as unemployed by the U.S. Bureau of Labor Statistics definition is countercyclical—it tends to go up during bad times and down during good times. The relative size of this group increased between 2007 and 2011 and has since retraced about half of that increase as the economy has strengthened. We expect that this category will continue to shrink some more as the economy continues to expand.
Health: The share of individuals who do not want a job because they were too ill or disabled to work has increased over time. The relative size of this group increased between 2007 and 2011. Since 2011, the rate of increase has slowed, and it actually declined in 2015. It is not clear what drove the larger increase during the 2007–11 period, but there is some literature linking weak labor market conditions to poor health outcomes.
Prime-age reasons for not wanting a job (other than health): During the recession, the share of prime-age (ages 25 to 54) women not wanting a job because of household or family responsibilities decreased. One explanation is that some women entered the labor force to help make ends meet. At the same time, there was an offsetting effect from a rise in educational enrollment. Since the recession, nonparticipation because of household or family responsibilities has returned to near prerecession levels, and educational enrollment has leveled off.
To summarize, we find that relative to the 2007–11 period there has been a:
- flattening in the rate of retirement by older individuals,
- flattening in the rate of educational program enrollment by younger individuals,
- declining share of the population saying they want a job but not officially counted as unemployed,
- smaller drag from nonparticipation because of health, and
- larger drag for reasons other than health among prime-age individuals.
Where will LFPR be by the end of 2016? What about five years from now?
During the first three months of 2016, there has been an increase in the overall LFPR. This was largely the result of a decline in the share of prime-age people citing health reasons for nonparticipation, with some contribution from a decline in the share who want a job but are not "unemployed."
Though these boosts to participation may offset the effect of an aging population in the short term, most forecasts have the LFPR declining over the next several years. How much participation will actually decline depends on the answers to several difficult questions, such as: Will older individuals push retirement out even farther? Will school enrollment rates rise more rapidly again? Will the health status of the population improve? The difficulty of answering these questions helps explain why making accurate labor force projections is challenging.
- Introducing the Atlanta Fed's Taylor Rule Utility
- Payroll Employment Growth: Strong Enough?
- Forecasting Loan Losses for Stress Tests
- Men at Work: Are We Seeing a Turnaround in Male Labor Force Participation?
- What’s Moving the Market’s Views on the Path of Short-Term Rates?
- Lockhart Casts a Line into the Murky Waters of Uncertainty
- How Will Employers Respond to New Overtime Regulations?
- How Good Is The Employment Trend? Decide for Yourself
- Is the Labor Market Tossing a Fair Coin?
- When It Rains, It Pours
- September 2016
- August 2016
- July 2016
- June 2016
- May 2016
- April 2016
- March 2016
- February 2016
- January 2016
- November 2015
- Business Cycles
- Business Inflation Expectations
- Capital and Investment
- Capital Markets
- Data Releases
- Economic conditions
- Economic Growth and Development
- Exchange Rates and the Dollar
- Fed Funds Futures
- Federal Debt and Deficits
- Federal Reserve and Monetary Policy
- Financial System
- Fiscal Policy
- Health Care
- Inflation Expectations
- Interest Rates
- Labor Markets
- Latin America/South America
- Monetary Policy
- Money Markets
- Real Estate
- Saving, Capital, and Investment
- Small Business
- Social Security
- This, That, and the Other
- Trade Deficit
- Wage Growth