research question about unemployment rate

What does the unemployment rate measure?

Subscribe to the hutchins roundup and newsletter, stephanie aaronson stephanie aaronson senior associate director, division of research and statistics - federal reserve board.

February 18, 2021

• 13 min read

This explainer builds off of a May 2020 post, “ Making Sense of the Monthly Jobs Report During the COVID-19 Pandemic .”

The unemployment rate soared from a 50-year low of 3.5 percent to 14.8 percent in April 2020 at the beginning of the COVID-19 pandemic, and then fell faster than many forecasters anticipated, to 6.3 percent in January 2021. But the labor market is far from healthy: for instance, the Bureau of Labor Statistics (BLS) counted 4.5 million more people as unemployed in January than were unemployed before the pandemic—and many more people weren’t counted as unemployed because they’d stopped looking for work. Here’s a guide to various measures of the health of the labor market.

What does the unemployment rate measure?   

The headline unemployment rate (known as U-3) measures the percentage of people over the age of 16 who aren’t working but are available and actively looking for work.

Where do the data on unemployment come from?  

Data on unemployment are collected every month in the Current Population Survey (CPS), a survey of about 60,000 households , conducted by the Census and the BLS every month, which includes roughly 105,000 people ages 16 and older . The questions about unemployment refer to what people were doing during the week that includes the 12th of the month, known as the “reference week”—so the survey to be released on Friday, March 4, 2021 will cover the week of February 8, 2021. The CPS is referred to as the household survey, to distinguish it from the establishment survey, which counts the number of people on employer payrolls. (The latest tally of payrolls showed that as of January 2021, we have recovered 9.9 million of the 22.4 million jobs lost since the beginning of 2020.)

How are individuals in the monthly survey identified as being unemployed ?   

Respondents to the survey are first asked whether they worked during the week that includes the 12th of the month. Individuals are counted as employed if they did any work at all as a paid employee, if they worked in their own business, or worked without pay for at least 15 hours in a family business. People are also counted as employed if they were temporarily absent from work as a result of sickness, bad weather, vacation, a strike, or personal reasons. Such workers are classified as employed but absent from work.

Respondents who are not employed then are asked if they have looked for work in the previous four weeks and are available to work. If so, they are counted as unemployed. Respondents who did not work but are on temporary layoff from a job with the expectation that they will be recalled—as many furloughed employees are today—are counted as unemployed whether they looked for a job or not.

What’s the difference between being unemployed and being out of the labor force?  

People who are not working and who don’t meet the criteria to be counted as unemployed are said to be out of the labor force. This category includes students, retirees, and those who stay at home to take care of family members. In addition, people who report wanting a job but who have not looked for work in the most recent four weeks are also considered out of the labor force. Between February 2020 and January 2021, 5.5 million people dropped out of the labor force, on net. At the same time, the number of people out of the labor force who said they wanted a job rose by 1.9 million.

Why is it  hard  to determine if people are unemployed in the COVID-19 pandemic?  

While the survey questions are the same as always, the nature of the COVID-19 economy means that people’s behavior, and hence the data, may not follow the same pattern that we usually see when the economy is turning down. For instance, early in the pandemic, the share of workers reporting themselves employed but not at work skyrocketed. While some of this increase was due to a rise in illness or childcare responsibilities, the number of people reporting themselves “not at work for other reasons” rose from around 600,000 prior to the pandemic to over 8 million. In conducting the survey, the BLS has attempted to categorize those employed but absent from work due to “pandemic-related business closures or cutbacks” as unemployed.  Despite this effort, an unusually large number of workers in this situation have still been counted as “not being at work for other reasons,” which the BLS views as misclassification . Following its typical procedure, the BLS has categorized individuals reported as “not being at work for other reasons” as employed rather than unemployed. While this classification is in keeping with their standard practice, it has the effect of depressing the unemployment rate relative to a case where these workers are categorized as unemployed. That said, this misclassification is certainly less severe now than it was early in the recession: in January 2021, the number was 1.7 million. In addition, the BLS has publicly stated its estimate of the size of this mismeasurement. In January, the BLS reported that if misclassified workers had been counted as unemployed, the unemployment rate would have been 0.6 percentage points higher, or 6.9 percent. It also noted that this was an upper-bound estimate.

Among those who have lost jobs, the typical behavior would be to transition from employment into unemployment rather than to transition out of the labor force. However, early in the pandemic, with stay-at-home orders in place and nonessential businesses closed in many communities, people who left employment were much less likely to seek work than would typically be the case. In addition, schools closed in many places, which meant that many people who lost their jobs had child-care responsibilities that prevented them from seeking or accepting a new job. Even now, nearly a year into the pandemic, many of these same dynamics are in place. As a result, relative to a typical downturn, we expect the headline unemployment rate to be relatively lower than in a typical recession and the percent of those out of the labor force to be relatively higher, especially the percent of those who say they want a job but aren’t looking.

Indeed, we saw evidence of this in March 2020 , when, relative to the prior trend, an additional 1.2 million people moved from employment to out of the labor force, and the number of people categorized as out of the labor force but wanting a job rose by 500,000. The resulting decline in the labor force participation rate was much larger than would be expected given the rise in the unemployment rate, and it remains unusually low.

Given the limitations of the unemployment rate as a measure of labor market slack, what are some alternatives?  

The BLS releases six measures of labor market slack in the monthly jobs report. These include the official unemployment rate (U-3), discussed above, as well as more narrow definitions, called U-1 and U-2, which respectively include only those unemployed at least 15 weeks (the long-term unemployed) or for less than a month (the short-term unemployed). The BLS also publishes broader definitions of slack. The U-6 rate, for instance, counts all those who are technically unemployed plus those are who are working part-time but would prefer full time work, and those “marginally attached to the labor force,” that is, people who say they want either a full-time or part-time job, have not looked for work in the most recent four weeks, but have looked for a job sometime in the past 12 months. When adults classified as “marginally attached” report that they did not recently seek work because they do not believe jobs are available for them, they are classified as “discouraged workers.” The U-4 counts the unemployed and discouraged workers, while U-5 adds in other marginally attached workers. In January, the broadest of these measures, U-6, stood at 11.1 percent, 4.8 percentage points higher than the official unemployment rate.

Given the measurement problems during the pandemic, are there alternatives to the BLS indicators of slack to gauge the health of the labor market?

Some economists have offered their own estimates of labor market slack trying to account for the misclassification and unusual movements in labor force participation during the pandemic. For example, Jason Furman and Wilson Powell III at the Peterson Institute for International Economics calculate what they call the “ realistic unemployment rate .” Their realistic unemployment rate was 8.3 percent in January , two percentage points higher than the official unemployment rate. Furman and Powell’s realistic unemployment rate differs from the official in two ways. First, they estimate the number of workers misclassified as being “not at work for other reasons” and count them as unemployed. Second, they try to estimate the excess decline in labor force participation beyond what would be expected given the rise in unemployment, and add those people to the unemployment rate as well. While this estimate is dependent on the specific modeling assumptions, it is nonetheless a useful attempt to reveal the extent to which we underestimate the true disruption to people’s livelihoods if we fail to account for the unusually large drop in labor force participation.

In a recent speech , Fed Chair Jerome Powell outlined his own method for adjusting the unemployment rate to capture the unusual features of the pandemic labor market. First, like Furman and Powell, he adds to the count of the unemployed an estimate of the misclassified workers. In addition, he adds in the number of people who have left the labor force since last February. This measure does not attempt to account for the fact that more people dropped out of the labor force than usual, and it does not account for structural reasons that individuals may have dropped out of the labor force, such as retirement. At the same time, it also isn’t sensitive to the specific modeling assumptions regarding the behavior of the participation rate. His calculations boost the unemployment rate to close to 10 percent in January .

One shortcoming of both these approaches is that they implicitly or explicitly make an assumption about what share of the individuals who are out of the labor force would be unemployed in a more normal recession. In addition, by counting individuals who are out of the labor force as unemployed, these measures would seem to assume that such individuals will act like the unemployed once the economy recovers. But typically, people who are out of the labor force are less likely to become employed than are those who are unemployed. One measure of the unemployment rate that includes individuals out of the labor force but also accounts for this variation in the propensity to return to work is the Hornstein-Kudlyak-Lange non-employment index , which was 9.3 percent in January.

A simpler measure is the employment-to-population-ratio (EPOP), a ratio of the number of people employed to the number of people in the population. The EPOP , which stood at 61.1 percent on the eve of the pandemic, declined by 9.8 percentage points between February and April—the largest decline since the series began in January 1948. Although the EPOP has recovered somewhat, to 57.5 percent, it still stands at its lowest level since the early 1980s, a time when far fewer women were in the labor force. This large drop is direct evidence of the unprecedented toll that the pandemic has taken on the labor market and people’s livelihoods.

What are “initial claims” for unemployment insurance?   

When people first file for unemployment insurance (UI), they are counted as an “initial claim.” So when unemployment increases, initial claims tend to rise. Because initial claims are reported weekly , they are often used as an early indicator of the overall unemployment rate.  

What is the relationship between initial claims and the unemployment rate — and why might it be different now?  

The number of people receiving UI and the number counted as unemployed do tend to move in the same direction, but there is no formal link between the two. The only criteria for being counted as unemployed (and hence included in the unemployment rate) are that you are without a job and that you have actively searched for work or are on temporary layoff. You don’t need to be collecting unemployment insurance to be counted as unemployed. And some people are eligible to collect partial unemployment insurance benefits if they are working but have been assigned a schedule that is far below their usual weekly hours.

Many people who become unemployed do not apply for UI benefits, either because they are not eligible or because they choose not to apply. So initial claims typically understate the number of people becoming unemployed in a given week. That said, there are people who file an initial claim and are not counted as unemployed in the CPS. This could happen if a person doesn’t meet the CPS criteria for being unemployed—for instance, if they file for UI because their work schedule was reduced, or if the person has a very short spell of unemployment which is not captured in the CPS (for example, a person who becomes unemployed and finds a job in  between survey reference weeks).  

Furthermore, many people who are unemployed and do file an initial claim do not end up receiving unemployment insurance benefits, either because they are not covered by the program, because they have not accumulated enough working hours to be eligible for benefits, or because they don’t satisfy the job search requirements. In February 2020, before the pandemic, the number of people unemployed was about 5.8 million while the number of people receiving UI benefits averaged only about 1.7 million.

What is the payroll survey? And why is it likely to be less useful than usual?

The payroll (or establishment) survey is a survey of 145,000 businesses—employing about one third of all workers on nonfarm payrolls. The payroll survey tends to have difficulty when the economy is at a turning point, as is the case now. To create the sample to be surveyed, the BLS picks firms from the universe of firms that have unemployment insurance tax accounts. However, new firms do not enter the BLS sample universe right away, and the BLS can have difficulty distinguishing non-response from a firm closure in real time. Since the net contribution of jobs created at new firms and jobs destroyed at closing firms is typically small, the BLS assumes that nonresponding firms have the same change in employment as occurred at firms that responded. It then uses a model, called the net birth-death model, to forecast the residual between that imputation and the actual data. This model tends to overestimate employment growth when the economy is weakening and underestimate it when the economy is improving. And while the model error is typically small, it can, on occasion, be large.

We know about these forecast errors because the BLS revises the data based on more complete information. In most years the benchmark is small, with the level of employment revising up or down by less than 0.2 percentage points. However, strikingly, when the establishment survey data for March 2009—the depths of the Great Recession—were benchmarked, the level of payroll employment was revised down by over 900,000 jobs, or 0.7 percentage points—meaning that employers had shed 75,000 more jobs each month between April 2008 and March 2009 than previously estimated.

If, as a result of the pandemic, an unusually large number of firms are closing and few are opening, it seems possible that even the dramatic decline in employment that we are likely to see will underestimate the true extent of job loss.

Why might the historical link between unemployment and poverty be a misleading way to look at today’s numbers?  

When people become unemployed, they lose an important (and sometimes their only) source of income and are at risk of falling into poverty. Of course, the more generous unemployment insurance is, the less likely it is for someone who loses a job to become poor. But unemployment insurance has typically replaced only about 40 percent of lost wages, on average, over the past 20 years, with a lot of variation in generosity across the states.

The federal response to the pandemic changed that. The CARES Act, for instance, added $600 a week to weekly unemployment insurance benefits through the end of July 2020, preventing many families from falling into poverty, and the December extension provided for an additional $300. And, of course, Congress provided two rounds of one-time payments for most families—$1,200 per adult and $500 per dependent child in the spring of 2020 and another $600 per individual in December, with payments phasing out for higher earners. Estimates suggest that about 13 million people were prevented from falling into poverty by these efforts.

The author thanks Francisca Alba for research assistance and Becca Portman for graphic design.  

Labor & Unemployment

Economic Studies

The Hutchins Center on Fiscal and Monetary Policy

Wendy Edelberg, Tara Watson

August 13, 2024

Anthony F. Pipa, Zoe Swarzenski

August 5, 2024

Wendy Edelberg, Noadia Steinmetz-Silber

July 31, 2024

• Main navigation
• Main content

Recessions and the Trend in the US Unemployment Rate

• Kurt G. Lunsford

The unemployment rate in the United States falls slowly in expansions, and it may not reach its previous low point before the next recession begins. Based on this feature, I document that the frequent recessions prior to 1983 are associated with an upward trend in the unemployment rate. In contrast, the long expansions beginning in 1983 are associated with a downward trend. I then estimate a two-variable vector autoregression (VAR) that includes the unemployment rate and a recession indicator. Long-horizon forecasts from this VAR conditioned on no future recessions project that the unemployment rate will go to 3.6 percent after a long period with no recessions.

The views authors express in Economic Commentary are theirs and not necessarily those of the Federal Reserve Bank of Cleveland or the Board of Governors of the Federal Reserve System. The series editor is Tasia Hane. This paper and its data are subject to revision; please visit clevelandfed.org  for updates.

When it comes to analyzing economic indicators to predict where the US economy is headed, the unemployment rate is arguably the variable familiar to most people. It receives attention from academics, policymakers, business economists, and politicians, but also the public at large. An appealing feature of the unemployment rate is its perceived ease of interpretation. A high or rising unemployment rate is a signal of macroeconomic slack or contraction, and a low or falling unemployment rate is a signal of macroeconomic health or expansion.

One issue that can confound this simple interpretation is that the unemployment rate may have a slow-moving trend that changes over time. If the trend is not static, then it is hard to know how far the current or forecasted unemployment rates are from the underlying trend. Figure 1 highlights this issue. It shows the monthly unemployment rate from January 1948 to October 2020 along with a line intended to estimate the unemployment rate trend. I compute this line with the statistical technique in Hodrick and Prescott (1997) (HP). 1  The trend line shows substantial variation, falling below 5 percent in the 1960s and 1990s and rising above 7 percent in the 1980s and 2010s. Because of this changing trend, an unemployment rate of 6 percent may be viewed as indicating macroeconomic slack in some periods but macroeconomic health in other periods, making it difficult for economists, policymakers, and the public at large to know where the economy stands. 2

Notes: Trend computed using a Hodrick and Prescott (1997) filter. Gray bars indicate recession periods. Sources: US Bureau of Labor Statistics, Unemployment Rate [UNRATE], retrieved from FRED, Federal Reserve Bank of St. Louis ( https://fred.stlouisfed.org/series/UNRATE ), and author’s calculations.

Researchers and policymakers often acknowledge the trend in the unemployment rate. Researchers typically remove a time-varying trend from the unemployment rate before studying its business cycle properties. 3  Policymakers on the Federal Reserve’s Federal Open Market Committee (FOMC) note in their Statement on Longer-Run Goals and Monetary Policy Strategy that the maximum level of employment “changes over time.” 4  In fact, the longer-run projections of the unemployment rate in the FOMC’s Summary of Economic Projections have drifted down since 2012. 5

Research has attributed much of the trend in the unemployment rate to demographic changes. 6  In this Commentary , I suggest an additional, previously unrecognized source of the trend: the frequency of recessions. Because the unemployment rate rises quickly in recessions but falls slowly in expansions, it may not fall to its previous low point if a recession cuts an expansion short. 7  Hence, frequent recessions can cause the unemployment rate to trend up over time. Figure 1 shows that this happened in the 1950s and the 1970s. Since 1983, recessions have been less frequent and expansions have been longer, causing the unemployment rate to regularly fall below its previous low point and generating a downward trend in the unemployment rate. 8  In February 2020, the unemployment rate fell to 3.5 percent, its lowest level since 1969.

I also estimate the relationship between recessions and the unemployment rate with a statistical model called a vector autoregression (VAR). I use the VAR to make forecasts of the unemployment rate under the hypothetical scenario that there will be no recessions in the future. I intend for this hypothetical scenario to match the spirit of the FOMC’s longer-run projections of the unemployment rate, which are made “in the absence of further shocks to the economy.” 9  My forecasts project that the unemployment rate will go to 3.6 percent after a long period with no recessions.

Recessions and Unemployment Rate Trends

Figure 2 depicts a series of computations that result in a view of the alignment between recessions and the unemployment rate. This view of the alignment (panel D) highlights the intuition that frequent recessions, separated by short expansions, are associated with upward drift in the unemployment rate, while infrequent recessions, separated by long expansions, are associated with downward drift.

Note: Gray bars indicate recession periods. Sources: US Bureau of Labor Statistics, Unemployment Rate [UNRATE], retrieved from FRED, Federal Reserve Bank of St. Louis ( https://fred.stlouisfed.org/series/UNRATE ), NBER-based Recession Indicators for the United States from the Period following the Peak through the Trough [USREC], retrieved from FRED, Federal Reserve Bank of St. Louis ( https://fred.stlouisfed.org/series/USREC ), and author’s calculations.

Panel A of figure 2 shows the cumulative sum of the National Bureau of Economic Research’s (NBER’s) recession months from January 1948 to October 2020. I define a recession as starting in the month following the NBER peak and ending in the month of an NBER trough. For the current period, the NBER announced a business cycle peak in February 2020 but has not announced a subsequent trough. In figure 2, I treat March and April 2020 as recession months. 10

In panel B of figure 2, I fit a linear time trend to the cumulative sum of the NBER recession months with ordinary least squares. This time trend gives an estimate of how quickly recessions accumulate on average. Then in panel C, I remove the linear time trend from the cumulative sum and show a detrended cumulative sum of NBER recession months. This detrended cumulative sum shows when recessions have accumulated more quickly and less quickly than average.

The detrended cumulative sum in panel C rises at a constant rate in every recessionary month and falls at a constant but slower rate in every expansionary month. This structure implies that this variable may not fall to its previous low point if a recession cuts an expansion short. As a result, frequent recessions, separated by short expansions, can cause this detrended cumulative sum to drift up over time. This upward drift occurs with the four recessions that begin in 1948, 1953, 1957, and 1960 and again with the four recessions that begin in 1970, 1973, 1980, and 1981. In other words, both 1948 to 1960 and 1970 to 1982 are 13-year periods where recessions accumulated more quickly than average. In contrast, recessions accumulated less quickly than average during the long expansions that occur mostly since 1983 and also in the 1960s. During these periods, the detrended cumulative sum falls below its low point from previous expansions, creating downward drifts in the series.

The periods of rapid recession accumulation, 1948 to 1960 and 1970 to 1982, are also periods when the unemployment rate trend rises in figure 1. In contrast, periods when recessions accumulate less quickly than average, the 1960s, 1983 to 2000, and the 2010s, are all periods when the unemployment rate trend falls in figure 1. To make this comparison between the accumulation of recessionary months and the unemployment rate more explicit, panel D shows the detrended cumulative sum of NBER recession months (left axis) along with the unemployment rate (right axis). The two series move closely together and have a correlation of about 0.7, even including the unusually large spike in the unemployment rate in April 2020.

A positive correlation between the frequency of recessionary months and the unemployment rate is not surprising. The NBER’s Business Cycle Dating Committee uses labor market variables when assigning business cycle peaks and troughs. 11  However, what is surprising about panel D is how closely the unemployment rate follows the detrended cumulative sum of recessionary months for such a long time—from 1948 to 2020. 12  This is surprising because the US labor market has been driven by a variety of economic shocks along with changing government policies, labor market regulations, and demographics; yet, the unemployment rate closely tracks the stable and linear structure of the detrended cumulative sum of recessionary months. As with the detrended cumulative sum of recessionary months, the unemployment rate rises quickly in recessions but falls slowly in expansions, and these features cause the unemployment rate to trend up with frequent recessions and trend down with infrequent recessions.

Longer-Run Unemployment Rate Projections

The results in the previous section show that the unemployment rate trend is aligned closely with how quickly recessionary months accumulate. Consequently, the unemployment rate trend may not be easily separated from the business cycle with statistical techniques that estimate slow-moving trends, such as in Hodrick and Prescott (1997), to offer just one example. This is because the unemployment rate’s trend is itself related to business cycles. 13  Instead, I model the unemployment rate and the NBER recession indicator, which has a value of zero in expansion months and a value of one in recession months, together with a statistical tool known as a VAR.

Using this VAR, I can produce longer-run projections of the unemployment rate in the spirit of the FOMC’s Summary of Economic Projections, which assumes that there will be no shocks to the economy in the future. I do this by producing forecasts of the unemployment rate while imposing that the recession indicator has a value of zero in all future periods. 14

There are two important steps for computing the forecasts. First, I use data from January 1948 to February 2020 to estimate the parameters of the VAR. These parameters establish the statistical relationship between the unemployment rate and the recession indicator, allowing me to predict how the unemployment rate will move in the future under the hypothetical scenario of no future recessions. Second, I choose the initial conditions as a starting point for my forecasts. 15  For example, I need to decide if I want to start my forecasts from a high unemployment rate or a low unemployment rate. Forecasters often use the most recent data as their starting points. However, they may also choose older data to check how accurate their projections would have been in the past.

Figure 3 uses this latter approach. Each panel shows the unemployment rate for the 7 most recent NBER expansions. 16  In addition, it shows the forecasts from the VAR under the hypothetical scenario of no recessions. I use the months before each expansion started as the starting point for the forecasts.

Note: Forecasts are computed under the hypothetical scenario of no future recessions. Sources: US Bureau of Labor Statistics, Unemployment Rate [UNRATE], retrieved from FRED, Federal Reserve Bank of St. Louis ( https://fred.stlouisfed.org/series/UNRATE ), and author’s calculations.

While these forecasts do not perfectly track the unemployment rate over the course of an expansion, they generally match the downward trend of the unemployment rate in expansions. These forecasts also demonstrate relative accuracy in predicting where the unemployment rate will fall at the end of expansions. For the long expansion from 1991 to early 2001, the forecast predicts almost perfectly where the unemployment rate fell. For the other two longest expansions—1961 through 1969 and 2009 to early 2020—the forecast overpredicts where the unemployment rate fell to by about 0.6 percentage points.

The longest expansion shown in figure 3 (panel G)—July 2009 to February 2020—lasted 10 years and 8 months. In order to compute unemployment rate forecasts in the spirit of the FOMC’s Summary of Economic Projections, which assumes that there will be no shocks to the economy in the future, I next consider how low the unemployment rate could fall in expansions that last much longer than 10 years and 8 months. Specifically, I produce forecasts by imposing no recessions for 20 years. 17  In addition, I study the importance of the starting points for the forecasts by considering three different starting points. The first two are November 1982 and June 2009. These are the same starting points that I used in panels D and G of figure 3 and they coincide with the ends of the deepest recessions since 1948. The third starting point is February 2020, which is the last month in my estimation sample.

Figure 4 shows the 20-year forecasts. When using the November 1982 and June 2009 starting points, the forecasts start with the unemployment rate at high levels. This is natural as both of these starting points coincide with the end of recessions. In contrast, the forecasts generated with the February 2020 starting point start with a low level of the unemployment rate, a measure which is consistent with the healthy labor market at the start of 2020.

Note: The three lines correspond to three starting points: November 1982, June 2009, and December 2019. Source: Author’s calculations.

The forecasts generated with the November 1982 and June 2009 starting points move down over time. The forecasts generated from the February 2020 starting point rise very slightly before falling again. The forecasts with all three initial conditions become very similar at about 20 years, showing that the starting point does not affect how low the VAR projects the unemployment rate will fall as long as expansions are sufficiently long. For each of the three starting points, the VAR projects that the unemployment rate will be about 3.6 percent after 20 years without a recession. 18  That is, if one were to use this VAR to make a longer-run projection of the unemployment rate in the absence of further shocks to the economy as done in the FOMC’s Summary of Economic Projections, then one would project 3.6 percent. 19

The value of the projections in figure 4 is that they provide an answer to where the unemployment rate could fall in a hypothetical world in which recessions do not occur, which I approximate with 20-year forecasts with no recessions. Of course, other choices of forecast length are possible. With shorter forecasts, the starting point of the forecast still matters. As shown in figure 4, the forecasts generated with the November 1982 and June 2009 starting points are not all the way down to 3.6 percent until about 20 years. I have also considered longer forecasts; however, I do not show these forecasts here because they also yield unemployment rates at 3.6 percent.

To check the robustness of my results, I drop early portions of my estimation sample and recompute the long-run forecasts under the assumption of no future recessions. Using samples of January 1958 to February 2020, January 1968 to February 2020, and January 1978 to February 2020, I compute the long-run forecasts of the unemployment rate to be 3.6 percent, 3.8 percent, and 3.6 percent, respectively. These values indicate that the early portions of my sample do not have a big impact on the results. Alternatively, if I use January 1948 to June 2009 as my sample, the long-run forecast is 3.9 percent. This sample choice shows that the 2010s, which were part of the longest expansion in US history, have only a small impact on the results. Overall, the findings suggest that how low the unemployment rate can fall in an expansion appears to be quite stable over a variety of sample periods.

Finally, to highlight how the assumption of no future recessions affects the forecast of the unemployment rate, I also compute the 20-year forecast of the unemployment rate without this assumption. This unconditional forecast of the unemployment rate is 5.7 percent, more than 2 percentage points above the forecast that assumes no future recessions. Clearly, when making longer-run projections of the unemployment rate, such as those in the FOMC’s Summary of Economic Projections, conditioning upon future recessions can make large changes in the projection.

Conclusions

The unemployment rate in the United States falls slowly in expansions, and it may not reach its previous low point before the next recession begins. This feature suggests that the unemployment rate trends up with frequent recessions and trends down when recessions are infrequent. In this Commentary , I show that the US unemployment rate indeed trended up with the rapid accumulation of recessions prior to 1983 and then trended down again with the slow accumulation of recessions after 1983. In addition, I estimate the relationship between recessions and the unemployment rate with a VAR. Long-run forecasts from this VAR under the scenario of no future recessions can be used to produce longer-run projections of the unemployment rate in the spirit of the FOMC’s Summary of Economic Projections. I find that the unemployment rate moves to 3.6 percent in the absence of future recessions. 

• I use a tuning parameter of 106 for the Hodrick and Prescott (1997) filter. This parameter is higher than what academic researchers often use for monthly data. I choose this higher value to highlight the lower frequency variation in the data. Using the filter from Müller and Watson (2015) with frequencies of 12 years or longer produces a similar picture. Return to 1
• For example, Weiner (1993) argues that the natural rate of unemployment was about 6.25 percent in 1993 but about 6.7 percent in 1980. An important point is that Weiner (1993) accounts for inflation in his estimate of the natural rate of unemployment. However, the HP filter in figure 1 does not account for inflation, nor do I throughout this Commentary . Return to 2
• For example, see the handbook chapter of Rogerson and Shimer (2011), which like many other studies, uses the Hodrick and Prescott (1997) filter to separate the trend and cycle components of the unemployment rate. Return to 3
• See https://www.federalreserve.gov/monetarypolicy/files/FOMC_LongerRunGoals.pdf . Return to 4
• For the April 2012 FOMC meeting, the range of longer-run unemployment rate projections was 4.9 percent to 6.0 percent. For the September 2020 FOMC meeting, this longer-run unemployment rate range had fallen to 3.5 percent to 4.7 percent. Return to 5
• Weiner (1993), also discussed in footnote 2, emphasizes demographic change. See Crump, Eusepi, Giannoni, and Șahin (2019) for more recent analysis and discussion of demographics and the unemployment rate. Return to 6
• Neftçi (1984) and Sichel (1993) have previously documented that the unemployment rate changes asymmetrically over the business cycle, rising quickly in recessions and falling slowly in expansions. This Commentary draws out the implication that this asymmetry can affect the longer-run trend in the unemployment rate. Return to 7
• Consistent with the slower accumulation of recession months, 1983 roughly corresponds to beginning of the “Great Moderation,” a period in US history in which many economic variables became less volatile (Kim and Nelson, 1999; McConnell and Perez-Quirós, 2000; Stock and Watson, 2002). The results in this Commentary link the Great Moderation to theoretical models of unemployment rate asymmetry, such as Dupraz, Nakamura, and Steinsson (2019) and Lepetit (2020), who find that more stable economic environments imply lower average unemployment rates. That is, the general downward trend in the unemployment rate after 1983 is consistent with theoretical models that show that the average unemployment rate can fall with a reduction in economic volatility. Return to 8
• See the notes to table 1 of https://www.federalreserve.gov/monetarypolicy/files/fomcprojtabl20200916.pdf . Return to 9
• See https://www.nber.org/research/data/us-business-cycle-expansions-and-contractions for a list NBER peaks and troughs. Return to 10
• The NBER’s Business Cycle Dating Committee also uses gross domestic product, gross domestic income, personal consumption expenditures, and personal income less transfers when assigning business cycle peaks and troughs. For an example of the Business Cycle Dating Committee’s reasoning, see https://www.nber.org/news/business-cycle-dating-committee-announcement-june-8-2020 . Return to 11
• Consistent with this finding, Hall and Kudlyak (2020) document that the pace of reduction of the unemployment rate in expansions has been roughly stable for 70 years. Additionally, figure 2 implies that the pace of unemployment rate increases in recessions has also been roughly stable for 70 years. Return to 12
• Separating trends and business cycles is challenging for a wide variety of economic indicators. For example, Coibion, Gorodnichenko, and Ulate (2018) show that the Congressional Budget Office has historically made large changes to its measure of potential output around recessions. Return to 13
• I use the conditional forecasting approach in Doan, Litterman, and Sims (1983). I provide details of the VAR and the conditional forecasting exercise in a supplemental appendix. Return to 14
• Mathematically, these initial conditions are the values of the right-hand side variables in the VAR that are needed to produce the forecasts. See the supplemental appendix for additional details. Return to 15
• I exclude the expansion that began in August 1980 because this expansion lasted only one year. Return to 16
• I choose 20 years because that is the horizon at which forecasts appear to converge for all of the initial conditions in figure 4. If I only use a 15-year forecast horizon, the November 1982 and June 2009 initial conditions yield forecasts of the unemployment that fall to 3.8 percent, but it does not fall all the way to 3.6 percent as shown in figure 4. If I use horizons of 25 or 30 years, then the unemployment rate falls to 3.6 percent for all the initial conditions shown in figure 4, but it does not fall further. Return to 17
• This finding uses a VAR with 6 lags and is sensitive to the number of lags included in the VAR. I have checked lag lengths from 1 to 13 and found that the 20-year unemployment rate projection varies from a low of about 2.9 percent (with 13 lags) to a high of about 3.7 percent (with 1 lag). Hence, the results that I provide are conservative in the sense that different lag choices may yield materially lower unemployment rate projections, but none yields materially higher unemployment rate projections. Return to 18
• This long-run projection of 3.6 percent is very similar to Hall and Kudlyak’s (2020) steady-state unemployment rate of 3.5 percent. Hence, my results provide empirical support for Hall and Kudlyak’s choice of a steady state. In contrast, my long-run projection is materially above the steady-state of Dupraz, Nakamura, and Steinsson (2019), which is 4.6 percent. In addition, the conditional forecasts that I produce can provide data moments that may help researchers estimate labor market congestion functions as in Section 5.3 of Hall and Kudlyak (2020). Return to 19
• Coibion, Olivier, Yuriy Gorodnichenko, and Mauricio Ulate. 2018. “The Cyclical Sensitivity in Estimates of Potential Output.” Brookings Papers on Economic Activity , Fall: 343–411. https://www.doi.org/10.1353/eca.2018.0020 .
• Crump, Richard K., Stefano Eusepi, Marc Giannoni, and Ayşegül Șahin. 2019. “A Unified Approach to Measuring u*.” Brookings Papers on Economic Activity , Spring: 143–214. https://www.doi.org/10.1353/eca.2019.0002 .
• Doan, Thomas, Robert Litterman, and Christopher A. Sims. 1983. “Forecasting and Conditional Projection Using Realistic Prior Distributions.” NBER Working Paper, No. 1202. https://www.doi.org/10.3386/w1202 .
• Dupraz, Stéphane, Emi Nakamura, and Jón Steinsson. 2019. “A Plucking Model of Business Cycles.” NBER Working Paper, No. 26351. https://www.doi.org/10.3386/w26351 .
• Hall, Robert E., and Marianna Kudlyak. 2020. “Why Has the US Economy Recovered So Consistently from Every Recession in the Past 70 Years?” NBER Working Paper No. 27234. https://www.doi.org/10.3386/w27234 .
• Hodrick, Robert J., and Edward C. Prescott. 1997. “Postwar US Business Cycles: An Empirical Investigation.” Journal of Money, Credit, and Banking , 29(1): 1–16. https://www.doi.org/10.2307/2953682 .
• Kim, Chang-Jin, and Charles R. Nelson. 1999. “Has the US Economy Become More Stable? A Bayesian Approach Based on a Markov
• Switching Model of the Business Cycle.” Review of Economics and Statistics , 81(4): 608–616. https://www.doi.org/10.1162/003465399558472 .
• Lepetit, Antoine, 2020. “Asymmetric Unemployment Fluctuations and Monetary Policy Trade-Offs.” Review of Economic Dynamic s, 36: 29–45. https://doi.org/10.1016/j.red.2019.07.005 .
• McConnell, Margaret M., and Gabriel Perez-Quirós. 2000. “Output Fluctuations in the United States: What Has Changed since the Early 1980s?” American Economic Review , 90(5): 1464–1476. https://doi.org/10.1257/aer.90.5.1464 .
• Müller, Ulrich K. and Mark W. Watson, 2015. “Low-Frequency Econometrics.” NBER Working Paper, No. 21564. https://www.doi.org/10.3386/w21564 .
• Neftçi, Salih N. 1984. “Are Economic Time Series Asymmetric over the Business Cycle?” Journal of Political Economy , 92(2): 307–328. https://www.doi.org/10.1086/261226 .
• Rogerson, Richard, and Robert Shimer. 2011. “Search in Macroeconomic Models of the Labor Market.” Handbook of Labor Economics , Volume 4a, Chapter 7, 619-700. https://www.doi.org/10.1016/S0169-7218(11)00413-8 .
• Sichel, Daniel E. 1993. “Business Cycle Asymmetry: A Deeper Look.” Economic Inquiry , 31(2): 224–236. https://www.doi.org/10.1111/j.1465-7295.1993.tb00879.x .
• Stock, James H., and Mark W. Watson. 2002. “Has the Business Cycle Changed and Why?” NBER Macroeconomics Annual , 17: 159–218. https://www.doi.org/10.1086/ma.17.3585284 .
• Weiner, Stuart E. 1993. “New Estimates of the Natural Rate of Unemployment.” Federal Reserve Bank of Kansas City, Economic Review , Fourth Quarter: 53–63. https://ideas.repec.org/a/fip/fedker/y1993iqivp53-69nv.78no.4.html .

Suggested Citation

Lunsford, Kurt G. 2021. “Recessions and the Trend in the US Unemployment Rate.” Federal Reserve Bank of Cleveland,  Economic Commentary  2021-01. https://doi.org/10.26509/frbc-ec-202101

The Pandemic's Impact on Unemployment and Labor Force Participation Trends

Following early 2020 responses to the pandemic, labor force participation declined dramatically and has remained below its 2019 level, whereas the unemployment rate recovered briskly. We estimate the trend of labor force participation and unemployment and find a substantial impact of the pandemic on estimates of trend. It turns out that levels of labor force participation and unemployment in 2021 were approaching their estimated trends. A return to 2019 levels would then represent a tight labor market, especially relative to long-run demographic trends that suggest further declines in the participation rate.

At the end of 2019, the labor market was hotter than it had been in years. Unemployment was at a historic low, and participation in the labor market was finally increasing after a prolonged decline. That tight labor market came to an abrupt halt with the COVID-19 pandemic in the spring of 2020.

Now, two years later, the labor market has mostly recovered from the depths of the pandemic recession. The unemployment rate is close to pre-pandemic lows, and job openings are at record highs. Yet, participation and employment rates have remained persistently below pre-pandemic levels. This suggests the possibility that the pandemic has permanently reduced participation in the economy and that current participation rates reflect a new normal. In this article, we explore how the pandemic has affected labor markets and whether a new normal is emerging.

What Is "Normal"?

One way to define the normal level of a variable is to estimate its trend and compare the observed data with the estimated trend values. Constructing a trend essentially means drawing a smooth line through the variations in the actual data.

But this means that constructing the trend for a point in time typically involves considering what happened both before and after that point in time. Thus, constructing the trend at the end of a sample is especially hard, since we do not yet know how the data will evolve.

We construct trends for three aggregate labor market ratios — the labor force participation (LFP) rate, the unemployment rate and the employment-population ratio (EPOP) — using methods described in our 2019 article " Projecting Unemployment and Demographic Trends ."

First, we estimate statistical models for LFP and unemployment rates of demographic groups defined by age, gender and education. For each gender and education, we decompose its unemployment and LFP into cyclical components common to all age groups and smooth local trends for age and cohort effects.

Second, we aggregate trends from the estimates of the group-specific trends. Specifically, we construct the trend for the aggregate LFP rate as the population-share-weighted sum of the corresponding estimated trends for demographic groups. We construct the aggregate unemployment rate and EPOP trends from the group-specific LFP and unemployment trends and the groups' population shares.

In our previous work, we estimated the trends for the unemployment rate and LFP rate of a gender-education group separately using maximum likelihood methods. The estimates reported in this article are based on the joint estimation of LFP and unemployment rate trends using Bayesian methods.

We separately estimate the trends using data from 1976 to 2019 (pre-pandemic) and from 1976 to 2021 (including the pandemic period). Figures 1, 2 and 3 display annual averages for the three aggregate labor market ratios — the LFP rate, the unemployment rate and EPOP, respectively — from 1976 to 2021.

In each figure, the solid black line denotes the observed values, and the blue and pink lines denote the estimated trend using data from 1976 up to and including 2019 and 2021, respectively. The estimated trends are subject to uncertainty, and the plotted trends represent the median estimate of the trend.

For the estimates based on data up to 2021, we also include the 90 percent coverage area shown as the shaded pink area. According to the statistical model, there is a 90 percent probability that the trend is contained in the coverage area. The blue and pink dotted lines represent our projections on how the labor market ratios will evolve until 2031, again based on the estimated trend up to and including 2019 and 2021. The shaded gray vertical areas highlight recessions as defined by the National Bureau of Economic Research (NBER).

Pre-Pandemic Trends: 1976-2019

We start with the pre-pandemic trends for the LFP rate and unemployment rate estimated for data from 1976 through 2019. After a long recovery from the 2007-09 recession, the LFP rate was 63.1 percent in 2019 (slightly above the estimated trend value of 62.8 percent), and the unemployment rate was 3.7 percent (noticeably below its estimated trend value of 4.7 percent).

The LFP rate being above trend and the unemployment rate being below trend reflects the characterization of the 2019 labor market as "hot." But note that even though the LFP rate exceeded its trend value in 2019, it was still lower than during the 2007-09 period. This difference is accounted for by the declining trend in the LFP rate.

As noted in our 2019 article , LFP rates and unemployment rates differ systematically across demographic groups. Participation rates tend to be higher for younger, more-educated workers and for men. Unemployment rates tend to be lower for men and for the older and more-educated population.

Thus, changes in the population composition over time — that is, the relative size of demographic groups — will affect the aggregate LFP and unemployment rates, in addition to changes in the LFP and unemployment rate trends of the demographic groups.

As also noted in our 2019 article, the hump-shaped trend of the aggregate LFP rate reflects a variety of forces:

• Prior to 1990, the aggregate LFP rate was boosted by an upward trend in the LFP rate of women. But after 1990, the LFP rate of women began declining. Combining this with declining trend LFP rates for other demographic groups has reduced the aggregate LFP rate.
• Changes in the age distribution had a limited impact prior to 2005, but the aging population since then has lowered the aggregate LFP rate substantially.
• Increasing educational attainment has contributed positively to aggregate LFP throughout the period.

The steady decline of the unemployment rate trend reflects mostly the contributions from an older and more-educated population and, to some extent, a decline in the trend unemployment rates of demographic groups.

Pre-Pandemic Expectations of Future LFP and Unemployment Trends

Our statistical model of smooth local trends for the LFP and unemployment rates of demographic groups has the property that the best forecast for future trend values of demographic groups is their last estimated trend value. Thus, the model will only predict a change in the trend of aggregate ratios if the population shares of its constituent groups are changing.

We combine the U.S. Census Bureau population forecasts for the gender-age groups with an estimated statistical model of education shares for gender-age groups to forecast population shares of our demographic groups from 2020 to 2031 (the dotted blue lines in Figures 1 and 2).

As we can see, the changing demographics alone imply further reductions of 1 percentage point and 0.2 percentage points in the trend LFP rate and unemployment rate, respectively. This projection is driven by the forecasted aging of the population, which is only partially offset by the forecasted higher educational attainment.

Based on data up to 2019, the same aggregate LFP rates in 2021 as in 2019 would have represented a substantial cyclical deviation upward from the pre-pandemic trends.

It is notable that the unemployment rate is much more volatile relative to its trend than the LFP rate is. In other words, cyclical deviations from trend are much more pronounced for the unemployment rate than for the LFP rate.

In fact, in our estimation, the behavior of the unemployment rate determines the common cyclical component of both the unemployment rate and the LFP rate. Whereas the unemployment rate spikes in recessions, the LFP rate response is more muted and tends to lag recessions. This feature will be important for interpreting how the estimated trend LFP rate changed with the pandemic.

Finally, Figure 3 combines the information from the LFP rate and unemployment rate and plots actual and trend rates for EPOP. On the one hand, given the relatively small trend decline of the unemployment rate, the trend for EPOP mainly reflects the trend for the LFP rate and inherits its hump-shaped path and the projected decline over the next 10 years. On the other hand, EPOP inherits the volatility from the unemployment rate. In 2019, EPOP is notably above trend, by about 1 percentage point.

Unemployment and Labor Force Participation During the Pandemic

The behavior of unemployment resulting from the pandemic-induced recession was different from past recessions:

• The entire increase in unemployment between February and April 2020 was accounted for by the increase in unemployment from temporary layoffs. This differed from previous recessions, when a spike in permanent layoffs led the bulge of unemployment in the trough.
• The recovery started in May 2020, and the speed of recovery was also much faster than in previous recessions. After only seven months, unemployment declined by 8 percentage points.
• The behavior of the unemployment rate is reflected in the 2020 recession being the shortest NBER recession on record: It lasted for two months (March to April 2020).

To summarize, the runup and decline of the unemployment rate during the pandemic were unusually rapid, but the qualitative features were not that different from previous recessions after properly accounting for temporary layoffs, as noted in the 2020 working paper " The Unemployed With Jobs and Without Jobs . "

The decline in the LFP rate was sharp and persistent. The LFP rate dropped from 63.4 percent in February 2020 to 60.2 percent in April 2020, an unprecedented drop during such a short period of time. After a rebound to 61.7 percent in August 2020, the LFP rate essentially moved sideways and remained below 62 percent until the end of 2021.

The large drop in the aggregate LFP rate has been attributed to, among others:

• More people — especially women — leaving the labor force to care for children because of school closings or to care for relatives at increased health risk, as noted in the 2021 work " Assessing Five Statements About the Economic Impact of COVID-19 on Women (PDF) " and the 2021 article " Caregiving for Children and Parental Labor Force Participation During the Pandemic "
• An increase in retirement due to health concerns, as noted in the 2021 working paper " How Has COVID-19 Affected the Labor Force Participation of Older Workers? "
• Generous pandemic income transfers and unemployment insurance programs, as noted in the 2021 article " COVID Transfers Dampening Employment Growth, but Not Necessarily a Bad Thing "

All of these factors might impact the participation trend, but by how much?

The Pandemic's Effect on Trend Estimates for LFP and Unemployment

The aggregate trend assessment for the LFP and unemployment rates has changed considerably as a result of 2020 and 2021. Repeating the estimation of trend and cycle for our demographic groups using data from 1976 up to 2021 yields the pink trend lines in Figures 1 and 2.

The updated trend estimates now put the positive cyclical gap in 2019 for LFP at 0.5 percentage points (rather than 0.3 percentage points) and the negative cyclical gap for the unemployment rate at 1.4 percentage points (rather than 1 percentage point). That is, by this estimate of the trend, the labor market in 2019 was even hotter than by the estimates from the 1976-2019 period.

In 2021, the actual LFP rate is essentially at trend, and the unemployment rate is only slightly above trend. That is, by this estimate of the trend, the labor market is relatively tight.

Notice that even though the new 2021 trend estimates for both the LFP and the unemployment rates differ noticeably from the trend values predicted for 2021 based on data up to 2019, the trend revisions for the LFP rate are limited to more recent years, whereas the trend revisions for the unemployment rate apply to the whole sample.  

The difference in revisions is related to how confident we can be about the estimated trends. The 90 percent coverage area is quite narrow for the LFP rate for the entire sample up to the last four years. Thus, there is no need to drastically revise the estimated trend prior to 2017.

On the other hand, the 90 percent coverage area for the trend unemployment rate is quite broad throughout the sample. That is, a wide range of values for trend unemployment is potentially consistent with observed unemployment values. Consequently, the last two observations lead to a wholesale reassessment of the level of the trend unemployment rate.

Another way to frame the 2020-21 trend revisions is as follows. The unemployment rate is very cyclical, deviations from trend are large, and though the sharp increase and decline of the unemployment rate in 2020-21 is unusual, an upward level shift of the trend unemployment rate best reflects the additional pandemic data.

The LFP rate, however, is usually not very cyclical, and it is only weakly related to the unemployment rate. Since the model assumes that the cyclical response does not change over the sample, it then attributes the large 2020-21 drop of the LFP rate to a decline in its trend and ultimately to a decline of the trend LFP rates of most demographic groups.

Finally, the EPOP trend is again mainly determined by the LFP trend, seen in Figure 3. Including the pandemic years noticeably lowers the estimated trend for the years from 2017 onwards. The cyclical gap in 2019 is now estimated to be 1.4 percentage points, and 2021 EPOP is close to its estimated trend.

What Does the Future Hold?

In our framework, current estimates of trend LFP and the unemployment rate for demographic groups are the best forecasts of future rates. Combined with projected demographic changes, this implies a continued noticeable downward trend for the LFP rate and a slight downward trend for the unemployment rate.

The trend unemployment rate is low, independent of how we estimate the trend. But given the highly unusual circumstances of the pandemic, the model may well overstate the decline in the trend LFP rate. Therefore, it is likely that the "true" trend lies somewhere between the trends estimated using data up to 2019 and data up to 2021.

That being a possibility, it remains that labor markets as of now have been unusually tight by most other measures, such as nominal wage growth and posted job openings relative to hires. This suggests that the true trend is closer to the revised 2021 trend than to the 2019 trend. In other words, the LFP rate and unemployment rate at the end of 2021 relative to the 2021 estimate of trend LFP and unemployment rate are consistent with a tight labor market.

Andreas Hornstein is a senior advisor in the Research Department at the Federal Reserve Bank of Richmond. Marianna Kudlyak is a research advisor in the Research Department at the Federal Reserve Bank of San Francisco.

To cite this Economic Brief, please use the following format: Hornstein, Andreas; and Kudlyak, Marianna. (April 2022) "The Pandemic's Impact on Unemployment and Labor Force Participation Trends." Federal Reserve Bank of Richmond Economic Brief , No. 22-12.

This article may be photocopied or reprinted in its entirety. Please credit the authors, source, and the Federal Reserve Bank of Richmond and include the italicized statement below.

V iews expressed in this article are those of the authors and not necessarily those of the Federal Reserve Bank of Richmond or the Federal Reserve System.

Subscribe to Economic Brief

Receive a notification when Economic Brief is posted online.

By submitting this form you agree to the Bank's Terms & Conditions and Privacy Notice.

Thank you for signing up!

As a new subscriber, you will need to confirm your request to receive email notifications from the Richmond Fed. Please click the  confirm subscription  link in the email to activate your request.  

If you do not receive a confirmation email, check your junk or spam folder as the email may have been diverted.

Phone Icon Contact Us

• Original Article
• Open access
• Published: 08 March 2018

Unemployment among younger and older individuals: does conventional data about unemployment tell us the whole story?

• Hila Axelrad 1 , 2 ,
• Miki Malul 3 &
• Israel Luski 4  

Journal for Labour Market Research volume  52 , Article number:  3 ( 2018 ) Cite this article

71k Accesses

31 Citations

13 Altmetric

Metrics details

In this research we show that workers aged 30–44 were significantly more likely than those aged 45–59 to find a job a year after being unemployed. The main contribution is demonstrating empirically that since older workers’ difficulties are related to their age, while for younger individuals the difficulties are more related to the business cycle, policy makers must devise different programs to address unemployment among young and older individuals. The solution to youth unemployment is the creation of more jobs, and combining differential minimum wage levels and earned income tax credits might improve the rate of employment for older individuals.

1 Introduction

Literature about unemployment references both the unemployment of older workers (ages 45 or 50 and over) and youth unemployment (15–24). These two phenomena differ from one another in their characteristics, scope and solutions.

Unemployment among young people begins when they are eligible to work. According to the International Labor Office (ILO), young people are increasingly having trouble when looking for their first job (ILO 2011 ). The sharp increase in youth unemployment and underemployment is rooted in long-standing structural obstacles that prevent many youngsters in both OECD countries and emerging economies from making a successful transition from school to work. Not all young people face the same difficulties in gaining access to productive and rewarding jobs, and the extent of these difficulties varies across countries. Nevertheless, in all countries, there is a core group of young people facing various combinations of high and persistent unemployment, poor quality jobs when they do find work and a high risk of social exclusion (Keese et al. 2013 ). The rate of youth unemployment is much higher than that of adults in most countries of the world (ILO 2011 ; Keese et al. 2013 ; O’Higgins 1997 ; Morsy 2012 ). Official youth unemployment rates in the early decade of the 2010s ranged from under 10% in Germany to around 50% in Spain ( http://www.indexmundi.com/g/r.aspx?v=2229 ; Pasquali 2012 ). The youngest employees, typically the newest, are more likely to be let go compared to older employees who have been in their jobs for a long time and have more job experience and job security (Furlong et al. 2012 ). However, although unemployment rates among young workers are relatively higher than those of older people, the period of time they spend unemployed is generally shorter than that of older adults (O’Higgins 2001 ).

We would like to argue that one of the most important determinants of youth unemployment is the economy’s rate of growth. When the aggregate level of economic activity and the level of adult employment are high, youth employment is also high. Footnote 1 Quantitatively, the employment of young people appears to be one of the most sensitive variables in the labor market, rising substantially during boom periods and falling substantially during less active periods (Freeman and Wise 1982 ; Bell and Blanchflower 2011 ; Dietrich and Möller 2016 ). Several explanations have been offered for this phenomenon. First, youth unemployment might be caused by insufficient skills of young workers. Another reason is a fall in aggregate demand, which leads to a decline in the demand for labor in general. Young workers are affected more strongly than older workers by such changes in aggregate demand (O’Higgins 2001 ). Thus, our first research question is whether young adults are more vulnerable to economic shocks compared to their older counterparts.

Older workers’ unemployment is mainly characterized by difficulties in finding a new job for those who have lost their jobs (Axelrad et al. et al. 2013 ). This fact seems counter-intuitive because older workers have the experience and accumulated knowledge that the younger working population lacks. The losses to society and the individuals are substantial because life expectancy is increasing, the retirement age is rising in many countries, and people are generally in good health (Axelrad et al. 2013 ; Vodopivec and Dolenc 2008 ).

The difficulty that adults have in reintegrating into the labor market after losing their jobs is more severe than that of the younger unemployed. Studies show that as workers get older, the duration of their unemployment lengthens and the chances of finding a job decline (Böheim et al. 2011 ; De Coen et al. 2010 ). Therefore, our second research question is whether older workers’ unemployment stems from their age.

In this paper, we argue that the unemployment rates of young people and older workers are often misinterpreted. Even if the data show that unemployment rates are higher among young people, such statistics do not necessarily imply that it is harder for them to find a job compared to older individuals. We maintain that youth unemployment stems mainly from the characteristics of the labor market, not from specific attributes of young people. In contrast, the unemployment of older individuals is more related to their specific characteristics, such as higher salary expectations, higher labor costs and stereotypes about being less productive (Henkens and Schippers 2008 ; Keese et al. 2006 ). To test these hypotheses, we conduct an empirical analysis using statistics from the Israeli labor market and data published by the OECD. We also discuss some policy implications stemming from our results, specifically, a differential policy of minimum wages and earned income tax credits depending on the worker’s age.

Following the introduction and literary review, the next part of our paper presents the existing data about the unemployment rates of young people and adults in the OECD countries in general and Israel in particular. Than we present the research hypotheses and theoretical model, we describe the data, variables and methods used to test our hypotheses. The regression results are presented in Sect.  4 , the model of Business Cycle is presented in Sect.  5 , and the paper concludes with some policy implications, a summary and conclusions in Sect.  6 .

2 Literature review

Over the past 30 years, unemployment in general and youth unemployment in particular has been a major problem in many industrial societies (Isengard 2003 ). The transition from school to work is a rather complex and turbulent period. The risk of unemployment is greater for young people than for adults, and first jobs are often unstable and rather short-lived (Jacob 2008 ). Many young people have short spells of unemployment during their transition from school to work; however, some often get trapped in unemployment and risk becoming unemployed in the long term (Kelly et al. 2012 ).

Youth unemployment leads to social problems such as a lack of orientation and hostility towards foreigners, which in turn lead to increased social expenditures. At the societal level, high youth unemployment endangers the functioning of social security systems, which depend on a sufficient number of compulsory payments from workers in order to operate (Isengard 2003 ).

Workers 45 and older who have lost their jobs often encounter difficulties in finding a new job (Axelrad et al. 2013 ; Marmora and Ritter 2015 ) although today they are more able to work longer than in years past (Johnson 2004 ). In addition to the monetary rewards, work also offers mental and psychological benefits (Axelrad et al. 2016 ; Jahoda 1982 ; Winkelmann and Winkelmann 1998 ). Working at an older age may contribute to an individual’s mental acuity and provide a sense of usefulness.

On average, throughout the OECD, the hiring rate of workers aged 50 and over is less than half the rate for workers aged 25–49. The low re-employment rates among older job seekers reflect, among other things, the reluctance of employers to hire older workers. Lahey ( 2005 ) found evidence of age discrimination against older workers in labor markets. Older job applicants (aged 50 or older), are treated differently than younger applicants. A younger worker is more than 40% more likely to be called back for an interview compared to an older worker. Age discrimination is also reflected in the time it takes for older adults to find a job. Many workers aged 45 or 50 and older who have lost their jobs often encounter difficulties in finding a new job, even if they are physically and intellectually fit (Hendels 2008 ; Malul 2009 ). Despite the fact that older workers are considered to be more reliable (McGregor and Gray 2002 ) and to have better business ethics, they are perceived as less flexible or adaptable, less productive and having higher salary expectations (Henkens and Schippers 2008 ). Employers who hesitated in hiring older workers also mentioned factors such as wages and non-wage labor costs that rise more steeply with age and the difficulties firms may face in adjusting working conditions to meet the requirements of employment protection rules (Keese et al. 2006 ).

Thus, we have a paradox. On one hand, people live longer, the retirement age is rising, and older people in good health want or need to keep working. At the same time, employers seek more and more young workers all the time. This phenomenon might marginalize skilled and experience workers, and take away their ability to make a living and accrue pension rights. Thus, employers’ reluctance to hire older workers creates a cycle of poverty and distress, burdening the already overcrowded social institutions and negatively affecting the economy’s productivity and GDP (Axelrad et al. 2013 ).

2.1 OECD countries during the post 2008 crisis

The recent global economic crisis took an outsized toll on young workers across the globe, especially in advanced economies, which were hit harder and recovered more slowly than emerging markets and developing economies. Does this fact imply that the labor market in Spain and Portugal (with relatively high youth unemployment rates) is less “friendly” toward younger individuals than the labor market in Israel and Germany (with a relatively low youth unemployment rate)? Has the market in Spain and Portugal become less “friendly” toward young people during the last 4 years? We argue that the main factor causing the increasing youth unemployment rates in Spain and Portugal is the poor state of the economy in the last 4 years in these countries rather than a change in attitudes toward hiring young people.

OECD data indicate that adult unemployment is significantly lower than youth unemployment. The global economic crisis has hit young people very hard. In 2010, there were nearly 15 million unemployed youngsters in the OECD area, about four million more than at the end of 2007 (Scarpetta et al. 2010 ).

From an international perspective, and unlike other developed countries, Israel has a young age structure, with a high birthrate and a small fraction of elderly population. Israel has a mandatory retirement age, which differs for men (67) and women (62), and the labor force participation of older workers is relatively high (Stier and Endeweld 2015 ), therefore, we believe that Israel is an interesting case for studying.

The Israeli labor market is extremely flexible (e.g. hiring and firing are relatively easy), and mobile (workers can easily move between jobs) (Peretz 2016 ). Focusing on Israel’s labor market, we want to check whether this is true for older Israeli workers as well, and whether there is a difference between young and older workers.

The problem of unemployment among young people in Israel is less severe than in most other developed countries. This low unemployment rate is a result of long-term processes that have enabled the labor market to respond relatively quickly to changes in the economic environment and have reduced structural unemployment. Footnote 2 Furthermore, responsible fiscal and monetary policies, and strong integration into the global market have also promoted employment at all ages. With regard to the differences between younger and older workers in Israel, Stier and Endeweld ( 2015 ) determined that older workers, men and women alike, are indeed less likely to leave their jobs. This finding is similar to other studies showing that older workers are less likely to move from one employer to another. According to the U.S. Bureau of Labor Statistics, the median employee tenure is generally higher among older workers than younger ones (BLS 2014 ). Movement in and out of the labor market is highest among the youngest workers. However, these young people are re-employed quickly, while older workers have the hardest time finding jobs once they become unemployed. The Bank of Israel calculated the chances of unemployed people finding work between two consecutive quarters using a panel of the Labor Force Survey for the years 1996–2011. Their calculations show that since the middle of the last decade the chances of unemployed people finding a job between two consecutive quarters increased. Footnote 3 However, as noted earlier, as workers age, the duration of their unemployment lengthens. Prolonged unemployment erodes the human capital of the unemployed (Addison et al. 2004 ), which has a particularly deleterious effect on older workers. Thus, the longer the period of unemployment of older workers, the less likely they will find a job (Axelrad and Luski 2017 ). Nevertheless, as Fig.  1 shows, the rates of youth unemployment in Israel are higher than those of older workers.

(Source: Calculated by the authors by using data from the Labor Force survey of the Israeli CBS, 2011)

Unemployed persons and discouraged workers as percentages of the civilian labor force, by age group (Bank of Israel 2011 ). We excluded those living outside settled communities or in institutions. The percentages of discouraged workers are calculated from the civilian labor force after including them in it

We argue that the main reason for this situation is the status quo in the labor market, which is general and not specific to Israel. It applies both to older workers and young workers who have a job. The status quo is evident in the situation in which adults (and young people) already in the labor market manage to keep their jobs, making the entrance of new young people into the labor market more difficult. What we are witnessing is not evidence of a preference for the old over the young, but the maintaining of the status quo.

The rate of employed Israelis covered by collective bargaining agreements increases with age: up to age 35, the rate is less than one-quarter, and between 50 and 64 the rate reaches about one-half. In effect, in each age group between 25 and 60, there are about 100,000 covered employees, and the lower coverage rate among the younger ages derives from the natural growth in the cohorts over time (Bank of Israel 2013 ). The wave of unionization in recent years is likely to change only the age profile of the unionization rate and the decline in the share of covered people over the years, to the extent that it strengthens and includes tens of thousands more employees from the younger age groups. Footnote 4

The fact that the percentage of employees covered by collective agreement increases with age implies that there is a status quo effect. Older workers are protected by collective agreements, and it is hard to dismiss them (Culpepper 2002 ; Palier and Thelen 2010 ). However, young workers enter the workforce with individual contracts and are not protected, making it is easier to change their working conditions and dismiss them.

To complete the picture, Fig.  2 shows that the number of layoffs among adults is lower, possibly due to their protection under collective bargaining agreements.

(Source: Israeli Central Bureau of Statistics, 2008, data processed by the authors)

Dismissal of employees in Israel, by age. Percentage of total employed persons ages 20–75 and over including those dismissed

In order to determine the real difference between the difficulties of older versus younger individuals in finding work, we have to eliminate the effect of the status quo in the labor market. For example, if we removed all of the workers from the labor market, what would be the difference between the difficulties of older people versus younger individuals in finding work? In the next section we will analyze the probability of younger and older individuals moving from unemployment to employment when we control for the status quo. We will do so by considering only individuals who have not been employed at least part of the previous year.

3 Estimating the chances of finding a job and research hypotheses

Based on the literature and the classic premise that young workers are more vulnerable to economic shocks (ILO 2011 ), we posit that:

H 1 : The unemployment rate of young people stems mainly from the characteristics of the labor market and less from their personal attributes.

Based on the low hiring rate of older workers (OECD 2006 ) and the literature about age discrimination against older workers in labor markets (Axelrad et al. 2013 ; Lahey 2005 ), we hypothesis that:

H 2 : The difficulty face by unemployed older workers searching for a job stems mainly from their age and less from the characteristics of the labor market.

To assess the chances of younger and older workers finding a job, we used a logit regression model that has been validated in previous studies (Brander et al. 2002 ; Flug and Kassir 2001 ). Being employed was the dependent variable, and the characteristics of the respondents (age, gender, ethnicity and education) were the independent variables. The dependent variable was nominal and dichotomous with two categories: 0 or 1. We defined the unemployed as those who did not work at all during the last year or worked less than 9 months last year. The dependent variable was a dummy variable of the current employment situation, which received the value of 1 if the individual worked last week and 0 otherwise.

3.1 The model

i—individual i, P i —the chances that individual i will have a full or part time job (at the time of the survey). \(\underset{\raise0.3em\hbox{$\smash{\scriptscriptstyle-}$}}{\text{X}}_{\text{i}}\) —vector of explanatory variables of individual i. Each of the variables in vector \(\underset{\raise0.3em\hbox{$\smash{\scriptscriptstyle-}$}}{X}_{i}\) was defined as a dummy variable with the value of 1 or 0. β—vector of marginal addition to the log of the odds ratio. For example, if the explanatory variable was the log of 13 years or more of schooling, then the log odds ratio refers to the marginal addition of 13 years of education to the chances of being employed, compared with 12 years of education or less.

The regression allowed us to predict the probability of an individual finding a job. The dependent variable was the natural base log of the probability ratio P divided by (1 − P) that a particular individual would find a job. The odds ratio from the regression answers the question of how much more likely it is that an individual will find a job if he or she has certain characteristics. The importance of the probability analysis is the consideration of the marginal contribution of each feature to the probability of finding a job.

3.2 The sample

We used data gathered from the 2011 Labor Force Survey Footnote 5 of the Israeli Central Bureau of Statistics (CBS), Footnote 6 which is a major survey conducted annually among households. The survey follows the development of the labor force in Israel, its size and characteristics, as well as the extent of unemployment and other trends. Given our focus on working age individuals, we excluded all of the respondents under the age of 18 or over the age of 59. The data sample includes only the Jewish population, because structural problems in the non-Jewish sector made it difficult to estimate this sector using the existing data only. The sample does not include the ultra-Orthodox population because of their special characteristics, particularly the limited involvement of men in this population in the labor market.

The base population is individuals who did not work at all during the past year or worked less than 9 months last year (meaning that they worked but were unemployed at least part of last year). To determine whether they managed to find work after 1 year of unemployment, we used the question on the ICBS questionnaire, “Did you work last week?” We used the answer to this question to distinguish between those who had succeeded in finding a job and those who did not. The data include individuals who were out of the labor force Footnote 7 at the time of the survey, but exclude those who were not working for medical reasons (illness, disability or other medical restrictions) or due to their mandatory military service. Footnote 8

3.3 Data and variables

The survey contains 104,055 respondents, but after omitting all of the respondents under the age of 18 or above 59, those who were outside the labor force for medical reasons or due to mandatory military service, non-Jews, the ultra-Orthodox, and those who worked more than 9 months last year, the sample includes 13,494 individuals (the base population). Of these, 9409 are individuals who had not managed to find work, and 4085 are individuals who were employed when the survey was conducted.

The participants’ ages range between 18 and 59, with the average age being 33.07 (SD 12.88) and the median age being 29. 40.8% are males; 43.5% have an academic education; 52.5% are single, and 53.5% of the respondents have no children under 17.

3.4 Dependent and independent variables

While previous studies have assessed the probability of being unemployed in the general population, our study examines a more specific case: the probability of unemployed individuals finding a job. Therefore, we use the same explanatory variables that have been used in similar studies conducted in Israel (Brander et al. 2002 ; Flug and Kassir 2001 ), which were also based on an income survey and the Labor Force Survey of the Central Bureau of Statistics.

3.5 The dependent variable—being employed

According to the definition of the CBS, employed persons are those who worked at least 1 h during a given week for pay, profit or other compensation.

3.6 Independent variables

We divided the population into sub-groups of age intervals: 18–24, 25–29, 30–44, 45–54 and 55–59, according to the sub-groups provided by the CBS. We then assigned a special dummy variable to each group—except the 30–44 sub-group, which is considered as the base group. Age is measured as a dummy variable, and is codded as 1 if the individual belongs to the age group, and 0 otherwise. Age appears in the regression results as a variable in and of itself. Its significance is the marginal contribution of each age group to the probability of finding work relative to the base group (ages 30–44), and also as an interaction variable.

3.6.2 Gender

This variable is codded as 1 if the individual is female and 0 otherwise. Gender also appears in the interaction with age.

3.6.3 Marital status

Two dummy variables are used: one for married respondents and one for those who are divorced or widowed. In accordance with the practice of the CBS, we combined the divorced and the widowed into one variable. This variable is a dummy variable that is codded as 1 if the individual belongs to the appropriate group (divorced/widowed or married) and 0 otherwise. The base group is those who are single.

3.6.4 Education

This variable is codded as 1 if the individual has 13 or more years of schooling, and 0 otherwise. The variable also appears in interactions between it and the age variable.

3.6.5 Vocational education

This variable is codded as 1 if the individual has a secondary school diploma that is not an academic degree or another diploma, and 0 otherwise.

3.6.6 Academic education

This variable is codded as 1 if the individual has any university degree (bachelors, masters or Ph.D.) and 0 otherwise.

3.6.7 Children

In accordance with similar studies that examined the probability of employment in Israel (Brander et al. 2002 ), we define children as those up to age 17. This variable is a dummy variable that is codded as 1 if the respondents have children under the age of 17, and 0 otherwise.

3.6.8 Ethnicity

This variable is codded as 1 if the individual was born in an Arabic-speaking country, in an African country other than South Africa, or in an Asian country, or was born in Israel but had a father who was born in one of these countries. Israel generally refers to such individuals as Mizrahim. Respondents who were not Mizrahim received a value of 0. The base group in our study are men aged 30–44 who are not Mizrahim.

We also assessed the interactions between the variables. For example, the interaction between age and the number of years of schooling is the contribution of education (i.e., 13 years of schooling) to the probability of finding a job for every age group separately relative to the situation of having less education (i.e., 12 years of education). The interaction between age and gender is the contribution of gender (i.e., being a female respondent) to the probability of finding a job for each age group separately relative to being a man.

To demonstrate the differences between old and young individuals in their chances of finding a job, we computed the rates of those who managed to find a job relative to all of the respondents in the sample. Table  1 shows that the rate of those who found a job declines with age. For example, 36% of the men age 30–44 found a job, but those rates drop to 29% at the age of 45–54 and decline again to 17% at the age of 55–59. As for women, 31% of them aged 30–44 found a job, but those rates drop to 20% at the age of 45–54 and decline again to 9% at the age of 55–59.

In an attempt to determine the role of education in finding employment, we created Model 1 and Model 2, which differ only in terms of how we defined education. In Model 1 the sample is divided into two groups: those with up to 12 years of schooling (the base group) and those with 13 or more years of schooling. In Model 2 there are three sub-groups: those with a university degree, those who have a vocational education, and the base group that has only a high school degree.

Table  2 shows that the probability of a young person (age 18–24) getting a job is larger than that of an individual aged 30–44 who belongs to the base group (the coefficient of the dummy variable “age 18–24” is significant and positive). Similarly, individuals who are older than 45 are less likely than those in the base group to find work.

Women aged 30–44 are less likely to be employed than men in the same age group. Additionally, when we compare women aged 18–24 to women aged 30–44, we see that the chances of the latter being employed are lower. Older women (45+) are much less likely than men of the same age group to find work. Additionally, having children under the age of 17 at home reduces the probability of finding a job.

A university education increases the probability of being employed for both men and women aged 30–44. Furthermore, for older people (55+) an academic education reduces the negative effect of age on the probability of being employed. While a vocational education increases the likelihood of finding a job for those aged 30–44, such a qualification has no significant impact on the prospects of older people.

Interestingly, being a Mizrahi Jew increases the probability of being employed.

In addition, we estimated the models separately twice—for the male and for the female population. For male and female, the probability of an unemployed individual finding a job declines with age.

Analyzing the male population (Table  3 ) reveals that those aged 18–24 are more likely than the base group (ages 30–44) to find a job. However, the significance level is relatively low, and in Model 2, this variable is not significant at all. Those 45 and older are less likely than the base group (ages 30–44) to find a job. Married men are more likely than single men to be employed. However, divorced and widowed men are less likely than single men to find a job. For men, the presence in their household of children under the age of 17 further reduces the probability of their being employed. Mizrahi men aged 18–24 are more likely to be employed than men of the same age who are from other regions.

Table  3 illustrates that educated men are more likely to find work than those who are not. However, in Model 1, at the ages 18–29 and 45–54, the probability of finding a job for educated men is less than that of uneducated males. Among younger workers, this might be due to excess supply—the share of academic degree owners has risen, in contrast to almost no change in the overall share of individuals receiving some other post-secondary certificate (Fuchs 2015 ). Among older job seeking men, this might be due to the fact that the increase in employment among men during 2002–2010 occurred mainly in part-time jobs (Bank of Israel 2011 ). In Model 2, men with an academic or vocational education have a better chance of finding a job, but at the group age of 18–24, those with a vocational education are less likely to find a job compared to those without a vocational education. The reason might be the lack of experience of young workers (18–24), experience that is particularly needed in jobs that require vocational education (Salvisberg and Sacchi 2014 ).

Analyzing the female population (Table  3 ) reveals that women between 18 and 24 are more likely to be employed than those who are 30–44, and those who are 45–59 are less likely to be employed than those who are 30–44. The probability of finding a job for women at the age of 25 to 29 is not significantly different from the probability of the base group (women ages 30–44).

Married women are less likely than single women to be employed. Women who have children under the age of 17 are less likely to be employed than women who do not have dependents that age. According to Model 2, Mizrahi women are more likely to be employed compared to women from other regions. According to both models, women originally from Asia or Africa ages 25–29 have a better chance of being employed than women the same age from other regions. Future research should examine this finding in depth to understand it.

With regard to education, in Model 1 (Table  3 ), where we divided the respondents simply on the question of whether they had a post-high school education, women who were educated were more likely to find work than those who were not. However, in the 18–29 age categories, educated women were less likely to find a job compared to uneducated women, probably due to the same reason cited above for men in the same age group—the inflation of academic degrees (Fuchs 2015 ). These findings become more nuanced when we consider the results of Model 2. There, women with an academic or vocational education have a better chance of finding a job, but at the ages of 18–24 those with an academic education are less likely to find a job than those without an academic education. Finally, at the ages of 25–29, those with a vocational education have a better chance of finding a job than those without a vocational education, due to the stagnation in the overall share of individuals receiving post-secondary certificate (Fuchs 2015 ).

Thus, based on the results in Table  3 , we can draw several conclusions. First, the effect of aging on women is more severe than the impact on men. In addition, the “marriage premium” is positive for men and negative for women. Divorced or widowed men lose their “marriage premium”. Finally, having children at home has a negative effect on both men and women—almost at the same magnitude.

5 Unemployment as a function of the business cycle

To determine whether unemployment of young workers is caused by the business cycle, we examined the unemployment figures in 34 OECD countries in 2007–2009, years of economic crisis, and in 2009–2011, years of recovery and economic growth. For each country, we considered the data on unemployment among young workers (15–24) and older adults (55–64) and calculated the difference between 2009 and 2007 and between 2011 and 2009 for both groups. The data were taken from OECD publications and included information about the growth rates from 2007 to 2011. Our assessment of unemployment rates in 34 OECD countries reveals that the average rate of youth unemployment in 2007 was 13.4%, compared to 18.9% in 2011, so the delta of youth unemployment before and after the economic crisis was 5.55. The average rate of adult unemployment in 2007 was 4% compared to 5.8% in 2011, so the delta for adults was 1.88. Both of the differences are significantly different from zero, and the delta for young people is significantly larger than the delta for adults. These results indicate that among young people (15–24), the increase in unemployment due to the crisis was very large.

An OLS model of the reduced form was estimated to determine whether unemployment is a function of the business cycle, which is represented by the growth rate. The variables GR2007, GR2009 and GR2011 are the rate of GDP growth in 2007, 2009 and 2011 respectively ( Appendix ). The explanatory variable is either GR2009 minus GR2007 or GR2011 minus GR2009. In both periods, 2007–2009 and 2009–2011, the coefficient of the change in growth rates is negative and significant for young people, but insignificant for adults. Thus, it seems that the unemployment rates of young people are affected by the business cycle, but those of older workers are not. In a time of recession (2007–2009), unemployment among young individuals increases whereas for older individuals the increase in unemployment is not significant. In recovery periods (2009–2011), unemployment among young individuals declines, whereas the drop in unemployment among older individuals is not significant (Table  4 ).

6 Summary and conclusions

The purpose of this paper was to show that while the unemployment rates of young workers are higher than those of older workers, the data alone do not necessarily tell the whole story. Our findings confirm our first hypothesis, that the high unemployment rate of young people stems mainly from the characteristics of the labor market and less from their personal attributes. Using data from Israel and 34 OECD countries, we demonstrated that a country’s growth rate is the main factor that determines youth unemployment. However, the GDP rate of growth cannot explain adult unemployment. Our results also support our second hypothesis, that the difficulties faced by unemployed older workers when searching for a job are more a function of their age than the overall business environment.

Indeed, one limitation of the study is the fact that we could not follow individuals over time and capture individual changes. We analyze a sample of those who have been unemployed in the previous year and then analyze the probability of being employed in the subsequent year but cannot take into account people could have found a job in between which they already lost again. Yet, in this sample we could isolate and analyze those who did not work last year and look at their employment status in the present. By doing so, we found out that the rate of those who found a job declines with age, and that the difficulties faced by unemployed older workers stems mainly from their age.

To solve both of these problems, youth unemployment and older workers unemployment, countries need to adopt different methods. Creating more jobs will help young people enter the labor market. Creating differential levels for the minimum wage and supplementing the income of older workers with earned income tax credits will help older people re-enter the job market.

Further research may explore the effect of structural and institutional differences which can also determine individual unemployment vs. employment among different age groups.

In addition to presenting a theory about the factors that affect the differences in employment opportunities for young people and those over 45, the main contribution of this paper is demonstrating the validity of our contention that it is age specifically that works to keep older people out of the job market, whereas it is the business cycle that has a deleterious effect on the job prospects of younger people. Given these differences, these two sectors of unemployment require different approaches for solving their employment problems. The common wisdom maintains that the high level of youth unemployment requires policy makers to focus on programs targeting younger unemployed individuals. However, we argue that given the results of our study, policy makers must adopt two different strategies to dealing with unemployment in these two groups.

6.1 Policy implications

In order to cope with the problem of youth unemployment, we must create more jobs. When the recession ends in Portugal and Spain, the problem of youth unemployment should be alleviated. Since there is no discrimination against young people—evidenced by the fact that when the aggregate level of economic activity and the level of adult employment are high, youth employment is also high—creating more jobs in general by enhancing economic growth should improve the employment rates of young workers.

In contrast, the issue of adult unemployment requires a different solution due to the fact that their chances of finding a job are related specifically to their age. One solution might be a differential minimum wage for older and younger individuals and earned income tax credits (EITC) Footnote 9 for older individuals, as Malul and Luski ( 2009 ) suggested.

According to this solution, the government should reduce the minimum wage for older individuals. As a complementary policy and in order to avoid differences in wages between older and younger individuals, the former would receive an earned income tax credit so that their minimum wage together with their EITC would be equal to the minimum wage of younger individuals. Earned income tax credits could increase employment among older workers while increasing their income. For older workers, EITCs are more effective than a minimum wage both in terms of employment and income. Such policies of a differential minimum wage plus an EITC can help older adults and constitute a kind of social safety net for them. Imposing a higher minimum wage exclusively for younger individuals may be beneficial in encouraging them to seek more education.

Young workers who face layoffs as a result of their high minimum wage (Kalenkoski and Lacombe 2008 ) may choose to increase their investment in their human capital (Nawakitphaitoon 2014 ). The ability of young workers to improve their professional level protects them against the unemployment that might result from a higher minimum wage (Malul and Luski 2009 ). For older workers, if the minimum wage is higher than their productivity, they will be unemployed. This will be true even if their productivity is higher than the value of their leisure. Such a situation might result in an inefficient allocation between work and leisure for this group. One way to fix this inefficient allocation without reducing the wages of older individuals is to use the EITC, which is actually a subsidy for this group. This social policy might prompt employers to substitute older workers with a lower minimum wage for more expensive younger workers, making it possible for traditional factories to continue their domestic production. However, a necessary condition for this suggestion to work is the availability of efficient systems of training and learning. Axelrad et al. ( 2013 ) provided another justification for subsidizing the work of older individuals. They found that stereotypes about older workers might lead to a distorted allocation of the labor force. Subsidizing the work of older workers might correct this distortion. Ultimately, however, policy makers must understand that they must implement two different approaches to dealing with the problems of unemployment among young people and in the older population.

For example, in the US, the UK and Portugal, we witnessed higher rates of growth during late 1990 s and lower rates of youth unemployment compared to 2011.

Bank of Israel Annual Report—2013, http://www.boi.org.il/en/NewsAndPublications/RegularPublications/Research%20Department%20Publications/BankIsraelAnnualReport/Annual%20Report-2013/p5-2013e.pdf .

http://www.boi.org.il/en/NewsAndPublications/RegularPublications/Research%20Department%20Publications/RecentEconomicDevelopments/develop136e.pdf .

The Labor Force Survey is a major survey conducted by the Israeli Central Bureau of Statistics among households nationwide. The survey follows the development of the labor force in Israel, its size and characteristics, as well as the extent of unemployment and other trends. The publication contains detailed data on labor force characteristics such as their age, years of schooling, type of school last attended, and immigration status. It is also a source of information on living conditions, mobility in employment, and many other topics.

The survey population is the permanent (de jure) population of Israel aged 15 and over. For more details see: http://www.cbs.gov.il/publications13/1504/pdf/intro04_e.pdf .

When we looked at those who had not managed to find a job at the time of the survey, we included all individuals who were not working, regardless of whether they were discouraged workers, volunteers or had other reasons. As long as they are not out of the labor force due to medical reasons or their mandatory military service, we classified them as "did not manage to find a job."

Until 2012, active soldiers were considered outside the labor force in the samples of the CBS.

EITC is a refundable tax credit for low to moderate income working individuals and couples.

Addison, J.T., Centeno, M., Portugal, P.: Reservation wages, search duration, and accepted wages in Europe (No. 1252). IZA Discussion paper series (2004)

Axelrad, H., Luski, I., Malul, M.: Difficulties of integrating older workers into the labor market: exploring the Israeli labor market. Int. J. Soc. Econ. 40 (12), 1058–1076 (2013). https://doi.org/10.1108/ijse-12-2011-0098

Article   Google Scholar  

Axelrad, H., Luski, I., Malul, M.: Behavioral biases in the labor market differences between older and younger individuals. J. Behav. Exp. Econ. 60 , 23–28 (2016). https://doi.org/10.1016/j.socec.2015.11.003

Axelrad, H., Luski, I., Malul, M.: Reservation wages and unemployment among older workers. J. Labor Res. 38 (2), 206–227 (2017). https://doi.org/10.1007/s12122-017-9247-6

Bank of Israel.: Bank of Israel report, Bank of Israel Research Department Jerusalem, March 2011. https://www.boi.org.il/he/NewsAndPublications/RegularPublications/Doch2010/p5.pdf (2011). Retrieved 17 Jan 2018

Bank of Israel.: Recent economic developments 136, Bank of Israel Research Department Jerusalem, December 2013. http://www.boi.org.il/en/NewsAndPublications/RegularPublications/Research%20Department%20Publications/BankIsraelAnnualReport/Annual%20Report-2013/p5-2013e.pdf (2013). Retrieved 16 July 2014

Bell, D.N., Blanchflower, D.G.: Young people and the great recession. Oxford Rev. Econ. Pol. 27 (2), 241–267 (2011). https://doi.org/10.1093/oxrep/grr011

Böheim, R., Horvath, G.T., Winter-Ebmer, R.: Great expectations: past wages and unemployment durations. Labour Econ. 18 (6), 778–785 (2011). https://doi.org/10.1016/j.labeco.2011.06.009

Brander, A., Peled Levy, O., Kassir, N.: Government policy and labor force participation rates of the working age population Israel and the OECD countries in the 90s. Bank of Israel Survey, August 2002 (2002), 7–61 ISSN 0552-2761 (Hebrew)

Bureau of Labor Statistics.: Employee tenure summary employee tenure in 2014, Bureau of Labor Statistics, U.S. Department of Labor. http://www.bls.gov/news.release/tenure.nr0.htm (2014). Retrieved 22 June 2015

Culpepper, P.D.: Powering, puzzling, and ‘pacting’: the informational logic of negotiated reforms. J. Eur. Public Policy 9 (5), 774–790 (2002)

De Coen, A., Forrier, A., Sels, L.: The impact of age on the reservation wage: the role of employability. FBE Research Report MO_1001, pp. 1–36. (2010).‏ https://ssrn.com/abstract=1620368 or http://dx.doi.org/10.2139/ssrn.1620368

Dietrich, H., Möller, J.: Youth unemployment in Europe—business cycle and institutional effects. Int. Econ. Econ. Policy 13 (1), 5–25 (2016). https://doi.org/10.1007/s10368-015-0331-1

Flug, K., Kassir, N.: On poverty, work and everything in between. Research Department Bank of Israel (2001) (Hebrow)

Freeman, R.B., Wise, D.A.: The youth labor market problem: its nature causes and consequences. In: The youth labor market problem: Its nature, causes, and consequences, pp. 1–16. University of Chicago Press, Chicago (1982)

Fuchs, H.: The socioeconomic situation of young adults in Israel. State of the nation report: society, economy and policy in Israel, pp. 139–181 (2015)

Furlong, A.: Youth studies: an introduction, pp. 72–97. Routledge, New York (2012)

Hendeles, S.: The Center for Adult Employment. National Insurance Institute, Research and Planning, Development Services Division (2008) (Hebrew)

Henkens, K., Schippers, J.: Labor market policies regarding older workers in the Netherlands. In: Taylor, P. (ed.) The ageing labor force: promises and prospects, pp. 141–157. Edward Elgar, Cheltenham (2008)

Google Scholar  

International Labor Office: Global employment trends for youth: 2011 update. International Labor Office, Geneva (2011)

Isengard, B.: Youth unemployment: Individual risk factors and institutional determinants. A case study of Germany and the United Kingdom. J. Youth Stud. 6 (4), 357–376 (2003). https://doi.org/10.1080/1367626032000162096

Jacob, M.: Unemployment benefits and parental resources: what helps the young unemployed with labor market integration? J. Youth Stud. 11 (2), 147–163 (2008). https://doi.org/10.1080/13676260701863413

Jahoda, M. (ed.): Employment and unemployment”. Cambridge University Press, Cambridge (1982)

Johnson, R.W.: Trends in job demands among older workers, 1992–2002. Mon. Labor Rev. 127 , 48 (2004)

Kalenkoski, C.M., Lacombe, D.J.: Effects of minimum wages on youth employment: the importance of accounting for spatial correlation. J. Labor Res. 29 (4), 303–317 (2008). https://doi.org/10.1007/s12122-007-9038-6

Keese, M., Queisser, M., Whitehouse, E.: Older workers living longer, working longer, DELSA Newsletter 2, OECD (2006)

Keese, M., Roseveare, D., Giguere, S.: The OECD action plan for youth, giving youth a better start in the labor market. OECD, Paris (2013)

Kelly, E., McGuinness, S., O'Connell, P.: J: Transitions to long-term unemployment risk among young people: evidence from Ireland. J Youth Stud 15 (6), 780–801 (2012). https://doi.org/10.1080/13676261.2012.678047

Lahey, J.N.: Do older workers face discrimination? Center for Retirement Research at Boston College (2005)

Malul, M.: Older workers’ employment in dynamic technology changes. J. Socio Econ. 38 (5), 809–813 (2009). https://doi.org/10.1016/j.socec.2009.05.005

Malul, M., Luski, I.: The optimal policy combination of the minimum wage and the earned income tax credit. BE J. Econ. Anal. Policy 9 , 1 (2009). https://doi.org/10.2202/1935-1682.1953

Marmora, P., Ritter, M.: Unemployment and the retirement decisions of older workers. J. Labor Res. 36 (3), 274–290 (2015). https://doi.org/10.1007/s12122-015-9207-y

McGregor, J., Gray, L.: Stereotypes and older workers: The New Zealand experience. Soc. Policy J. NZ. 163–177 (2002).

Morsy, H.: Scarred generation. Financ. Dev. 49 , 1 (2012)

Nawakitphaitoon, K.: Occupational human capital and wages: the role of skills transferability across occupations. J. Labor Res. 35 (1), 63–87 (2014). https://doi.org/10.1007/s12122-013-9172-2

OECD.: Live longer, work longer—ISBN-92-64-035877 (2006)

O’Higgins, N.: The challenge of youth unemployment. Int. Soc. Secur. Rev. 50 (4), 63–93 (1997). https://doi.org/10.1111/j.1468-246X.1997.tb01084.x

O’Higgins, N.: Youth unemployment and employment policy: a global perspective. International Labor Office, Geneva (2001)

Palier, B., Thelen, K.: Institutionalizing dualism: complementarities and change in France and Germany. Politics Soc. 38 (1), 119–148 (2010)

Pasquali, V.: Unemployment rates in countries around the world. Global Finance. https://www.gfmag.com/global-data/economic-data/worlds-unemployment-ratescom (2012). Retrieved 4 Nov 2013

Peretz, S.: The secret of the Israeli labor market’s flexibility. Haaretz. https://www.haaretz.com/israel-news/business/opinion/.premium-1.698583 (2016). Accessed 7 Jan 2018

Salvisberg, A., Sacchi, S.: Labour market prospects of Swiss career entrants after completion of vocational education and training. Eur. Soc. 16 (2), 255–274 (2014). https://doi.org/10.1080/14616696.2013.821623

Scarpetta, S., Sonnet, A., Manfredi, T.: Rising youth unemployment during the crisis: how to prevent negative long-term consequences on a generation?, OECD Social, Employment and Migration Working Papers, No. 106, OECD Publishing (2010). https://doi.org/10.1787/5kmh79zb2mmv-en

Stier, H., Endeweld, M.: Employment transitions and labor market exits: age and gender in the Israeli labor market. Res. Soc. Stratif. Mobil. 30 (41), 93–103 (2015). https://doi.org/10.1016/j.rssm.2015.01.002

Vodopivec, M., Dolenc, P.: Live longer, work longer: making it happen in the labor market. Financ. Theory Pract. 32 (1), 65–81 (2008)

Winkelmann, L., Winkelmann, R.: Why are the unemployed so unhappy? Evidence from panel data. Economica 65 (257), 1–15 (1998)

Download references

Authors’ contributions

HA, MM and IL conceptualized and designed the study. HA collected and managed study data, HA and IL carried out statistical analyses. HA drafted the initial manuscript. MM and IL reviewed and revised the manuscript. All authors read and approved the final manuscript.

Competing interests

The authors declare that they have any no competing interests.

Ethics approval and consent to participate

Not applicable.

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Author information

Authors and affiliations.

Center on Aging & Work, Boston College, Chestnut Hill, MA, 02467, USA

Hila Axelrad

The School of Social and Policy Studies, The Faculty of Social Sciences, Tel Aviv University, P.O. Box 39040, 6997801, Tel Aviv, Israel

Department of Public Policy & Administration, Guilford Glazer Faculty of Business & Management, Ben-Gurion University of the Negev, Beer Sheva, Israel

Department of Economics, The Western Galilee College, Akko, Israel

Israel Luski

You can also search for this author in PubMed   Google Scholar

Corresponding author

Correspondence to Hila Axelrad .

See Table  5 .

Rights and permissions

Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License ( http://creativecommons.org/licenses/by/4.0/ ), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.

Reprints and permissions

About this article

Cite this article.

Axelrad, H., Malul, M. & Luski, I. Unemployment among younger and older individuals: does conventional data about unemployment tell us the whole story?. J Labour Market Res 52 , 3 (2018). https://doi.org/10.1186/s12651-018-0237-9

Download citation

Received : 22 September 2017

Accepted : 25 February 2018

Published : 08 March 2018

DOI : https://doi.org/10.1186/s12651-018-0237-9

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Unemployment
• Older workers

JEL Classification

Numbers, Facts and Trends Shaping Your World

Read our research on:

Full Topic List

Regions & Countries

• Publications
• Our Methods
• Short Reads
• Tools & Resources

Read Our Research On:

What the unemployment rate does – and doesn’t – say about the economy

Every month, the federal Bureau of Labor Statistics releases a flood of data about employment and unemployment in the U.S. And every month, the lion’s share of the attention goes to one figure – the unemployment rate, which was a seasonally adjusted 4.8% in January. (The February report comes out on Friday.)

But the unemployment rate is just one indicator of how the U.S. economy is doing, and it’s not always the best one. Simply being out of work isn’t enough for a person to be counted as unemployed; he or she also has to be available to work and actively looking for work  (or on temporary layoff). In any given month, the unemployment rate can rise or fall based not just on how many people find or lose jobs, but on how many join or leave the active labor force.

There are, in fact, five other monthly measures of what the BLS calls “labor underutilization” besides the official unemployment rate, as well as scores of other measurements – labor force participation rates, employment-population ratios, average weekly wages, average hours worked and more. Knowing what those other data points are, where they come from and how they’re calculated is critical in understanding what they do – and don’t – tell us about the nation’s workers.

Take the concept of unemployment. Since U.S. economists first began trying to systematically measure unemployment in the 1870s , one of the main issues has been defining exactly what “being unemployed” means  – since many people who don’t have jobs, such as retirees and students, may not actually want paying work. (As the BLS itself noted  once upon a time , “Being employed is an observable experience, while being unemployed often lacks that same concreteness.”)

Since 1945, the official definition has been that to be considered unemployed, you must not only not have a job but be available for work (i.e., not too sick to work) and have actively looked for a job in the past four weeks. If you’re neither employed nor, according to the official definition, unemployed, you’re not considered part of the labor force.

The BLS derives its unemployment data from the Census Bureau’s Current Population Survey , which interviews about 60,000 people each month (and not, as is sometimes supposed, by counting how many people drew unemployment benefits). The CPS covers the entire civilian non-institutional population ages 16 and older, including self-employed people; prison inmates, residents of mental facilities and homes for the aged, and active-duty military personnel are excluded. (A separate survey of 146,000 private- and public-sector employers produces the monthly nonfarm-payroll numbers.)

Since 1994, no major changes have been made in how unemployment is measured, though there have been some modest  updates to the CPS  over time. For example, a 2010 change  raised the upper limit on reporting how long someone has been jobless from “99 weeks and over” to “260 weeks and over” in order to better track long-term unemployment.

As many observers have pointed out, the official unemployment definition leaves out some significant groups. The underemployed – part-time workers who would prefer to work full-time – are counted among the employed. And discouraged workers – people who’d like a job but have stopped looking because they don’t believe any work is available – aren’t counted as part of the labor force at all.

But the CPS asks people a lot more than whether they are working or searching for work. Questions include how long jobless people have been out of work, how recently they looked for work, why part-timers aren’t working full time, why people choose not to look for work, and even why people with jobs may not have been working during the survey period (for example, if they were sick, on vacation, temporarily laid off or snowed in).

All that extra data enables the BLS to calculate six different measures of labor underutilization, labeled U-1 through U-6, with broader or narrower parameters than the official unemployment rate (which is known as U-3). The broadest, U-6, includes all “marginally attached” workers (including discouraged workers) and involuntary part-time workers. The seasonally adjusted U-6 rate stood at 9.4% in January; since 1994 it has ranged from 6.8% (in October 2000) to 17.1% (most recently in April 2010). While the U-6 typically runs anywhere from 3 to 7 percentage points higher than the regular unemployment rate, with the gap wider during recessions and narrower in good economic times, it tends to follow the same pattern as the official unemployment rate.

Beyond the unemployment rate, a key metric in the monthly jobs report is the labor force participation rate – the share of the 16-and-over civilian non-institutional population either working or looking for work. The participation rate rose for several decades, peaked in early 2000 at 67.3%, then began falling; in January it was 62.9%, about where it was in the late 1970s. Labor economists generally agree that waves of retiring Baby Boomers explain part (but not all) of the decline, which has been especially steep for men in their prime working years .

There’s also the employment-population ratio, which measures employed people as a percentage of the 16-and-over civilian non-institutional population. Though the ratio has some quirks, it’s less affected by seasonal variations or short-term fluctuations in labor-market behavior than the unemployment rate. According to the January jobs report, the seasonally adjusted employment-population ratio was 59.9%, three-tenths of a percentage point higher than it was a year earlier.

Like the labor force participation rate, the employment-population ratio can be affected by more people retiring or deciding to go back to school. That’s why many labor-market economists focus on the 25-to-54 age group, which strips out most students and retirees. In January, the employment-population ratio for that subgroup was a seasonally adjusted 78.2%, a figure that’s been constant for the past four months. While that’s better than it was during the long hangover from the Great Recession (only 75.2% of 25- to 54-year-olds were employed in January 2011), it’s still below the indicator’s pre-recession high (80.3% in January 2007).

It’s also important to remember that not all employment is created equal. Before the Great Recession, fewer than 20% of all part-time workers said they were working less than 35 hours a week for economic reasons, such as slack demand or inability to find full-time work. During the slump, that share jumped to a third of all part-timers; the “involuntary part-time” share has fallen since, to 22.2% of all part-timers in January, but still is above typical pre-recession levels.

Note: This post has been updated with January 2017 unemployment data. It was originally published in June 2013.

• Business & Workplace
• Economic Conditions
• Economic Policy
• Economy & Work
• Unemployment

Drew DeSilver is a senior writer at Pew Research Center .

A look at small businesses in the U.S.

A look at black-owned businesses in the u.s., 2023 saw some of the biggest, hardest-fought labor disputes in recent decades, do you tip more or less often than the average american, diversity, equity and inclusion in the workplace, most popular.

901 E St. NW, Suite 300 Washington, DC 20004 USA (+1) 202-419-4300 | Main (+1) 202-857-8562 | Fax (+1) 202-419-4372 |  Media Inquiries

Research Topics

• Email Newsletters

ABOUT PEW RESEARCH CENTER  Pew Research Center is a nonpartisan fact tank that informs the public about the issues, attitudes and trends shaping the world. It conducts public opinion polling, demographic research, media content analysis and other empirical social science research. Pew Research Center does not take policy positions. It is a subsidiary of  The Pew Charitable Trusts .

© 2024 Pew Research Center

ReviewEcon.com

AP, IB, and College Microeconomic and Macroeconomic Principles 

6 questions about unemployment and the labor force

Unemployment.

Updated 10/29/2021 Jacob Reed 1. What does it mean to be unemployed?

The unemployment rate is one of the most watched and publicized labor force statistics, but many people are confused about what it actually measures. The unemployment rate is the percentage of people in the labor force who are not working, but are actively looking for work.  As of the writing of this article, the official (U-3) unemployment rate was 4.0% (see the  BLS for the current unemployment rate ).

The formula for the unemployment rate is: Unemployed/Labor Force x 100 = Unemployment Rate

2. Who is not counted in the unemployment rate?

The official unemployment rate (U-3) does not count people who are not actively looking for work. As a result, there may be some workers who recently lost their jobs and want jobs, but aren’t actively looking. These people are considered out of the labor force. There are also some people who lost their jobs a long time ago but have given up looking for work within the last 12 months. These are called discouraged workers and they are also considered out of the labor force so they are not reflected in the official unemployment rate.  Also, underemployed members of the labor force who have part-time jobs but are looking for full time jobs are counted as employed. For all of these reasons, the official unemployment rate may not always reflect an accurate picture of the overall labor market. The bureau of labor statistics does track all of these groups and includes them in the U-6 measure of unemployment. As of the writing of this article, the U-6 rate was 8.1% (see the  BLS for the current rate ).

3. Who is in the labor force? The labor force includes civilian citizens who are at least 16 years of age and are either employed or actively looking for work. To be considered employed, a person must work for pay or profit for one or more hours in the given week,  work without pay in a family business for 15 or more hours, or have a job but didn’t work due to vacation, illness, labor dispute, etc. After a recession, the unemployment rate may fall if unemployed workers leave the workforce. This lowers the unemployment rate and can give a false sense that the labor market has improved.

The formula for the labor force is:

Working + Looking for Work = Labor Force

4. What is the labor force participation rate?

The labor force participation rate is the percentage the working age population that is either working or looking for work. The formula is:

[(Working + Looking for Work)/Working Age Civilian Population] x 100 = Labor Force Participation Rate

The labor force participation rate in the United States fell during the last recession as some citizens lost their jobs and gave up looking for work. As the economy has improved the labor force participation rate is rising again, but it has not recovered back to pre recession levels.  As of the writing of this article, the labor force participation rate in the US was 63.2% (see the  BLS for the current rate ). This statistic gives economists a sense of how many people are choosing to be part of an economy’s labor force. A higher participation rate will shift the  production possibilities curve  outward. A lower participation rate will shift it inward. 

5. What types of unemployment are there?

Seasonal Unemployment:  This type of unemployment is often not discussed on many macroeconomics exams because the official unemployment rate is seasonally adjusted; meaning seasonal unemployment has been deleted out of the statistic. Seasonal unemployment occurs when workers lose their jobs due to the time of year. Lifeguards getting laid off in the winter and temporary store retail clerks getting laid off after the holiday shopping season are two examples. Seasonal unemployment is a natural part of a healthy economy.

Frictional Unemployment:  This type of unemployment is characterized by movement between jobs. When a college graduate is looking for her first job, a cook quits his restaurant job, or a brick mason is fired from construction company, all three of these people are now frictionally unemployed. Frictional unemployment is a natural part of a healthy economy.

Structural Unemployment:  This type of unemployment is most often characterized by a skills mismatch; meaning the skills unemployed workers have do not match the skills needed for the jobs available. These workers must go back to school or be retrained to get the skills they need. This type of unemployment can be caused by technological changes like ATM machines replacing banking tellers.  Structural unemployment is also a natural part of a healthy economy as well. As the economy changes, some structural unemployment is inevitable.

Cyclical Unemployment:  This is unemployment caused by the business cycle. People unemployed as a result of the great depression of the 1930’s and the recent great recession were cyclically unemployed. Cyclical unemployment is characterized by an overall downturn in the economy. A recessionary gap in the  AS/AD model  is an indication of cyclical unemployment. 

6. What is full employment (or the natural rate of unemployment)? Full employment is defined as zero cyclical unemployment; or when the unemployment rate equals frictional unemployment plus structural unemployment (seasonal unemployment is already deleted from official numbers). When the economy is at full employment the unemployment rate will equal what is called the natural rate of unemployment (NRU). This occurs when an economy is at long-run equilibrium in the  AS/AD model  and there is no inflationary or recessionary gap. As of the writing of this article, the long run natural rate of unemployment, as estimated by the Congressional Budget Office, was 4.55% (see the  St. Louis Fed for current estimates ).

Up Next:  Content Review Page:   Production Possibilities Curve

 Outside Resources:  mjmfoodie  (definition of unemployment),  mjmfoodie  (types of unemployment)

**AP©, Advanced Placement Program©, and College Board© are registered trademarks of the College Board, which was not involved in the production of, and does not endorse, this material. IB is a registered trade mark of International Baccalaureate Organization which was also not involved in the production of and does not endorse this material.**

I would like to acknowledge the work of Dick Brunelle and Steven Reff from Reffonomics.com whose work inspired many of the review games on this site. I would also like to thank Francis McMann, James Chasey, and Steven Reff who taught me how to be an effective AP Economics teacher at AP summer institutes; as well as the countless high school teachers, and college professors from the AP readings, economics Facebook groups, and #econtwitter. 

Have A Question?

Site Search

• Effective Communication
• Job Qualifications
• Medical Exams
• Physical Access
• Reasonable Accommodations
• Reasonable Modifications
• Service Animals
• Tax Incentives
• ADA Coordinator
• Education Access
• Emergency Management
• Historic Exemptions
• Program Access
• Transportation
• Individuals
• Discrimination & Enforcement
• The Americans with Disabilities Act
• South Dakota Network of ADA Coordinators
• Utah Network of ADA Coordinators
• News & Blog
• Blog - Access Granted
• In The Press
• Research & Resources

High Rates of Unemployment for People with Disabilities

Compiled by Erin Moser 1

Edited by Jill L. Bezyak 2

Background Information

Misperceptions and stigma toward people with disabilities are factors that contribute to high unemployment rates. Fear of individuals losing their social security disability is yet another. Regardless of the reason, people with disabilities continue to struggle finding and maintaining suitable employment, and high rates of unemployment for individuals with disabilities continue to be a problem in the United States, leading to poor physical and mental health, social isolation, low self-esteem, and low life satisfaction 3 .

In 2021, the unemployment rate for persons with a disability was 10.1%, approximately twice as high as those without a disability 4 . With attention to psychiatric disabilities, more than 11 million adults in the United States are diagnosed with a serious mental illness (i.e., schizophrenia, anxiety, bipolar disorder), and 90% of this population are unemployed 5 . Additionally, 29% of workers with a disability were employed part-time, while 16% of individuals without disabilities were employed similarly 3 . In fact, in most developed countries, the unemployment rate for individuals with disabilities who are of working age is at least twice as high as individuals without disabilities 6 . If workers with disabilities reached the same employment rate as those without a disability, nearly 14 million more individuals would have been employed in 2021 7 . These statistics support the ongoing notion that individuals with disabilities continue to not be hired, contributing to significant consequences of unemployment and poverty.

The Americans with Disabilities Act (ADA) of 1990, is the most comprehensive disability rights legislation in history. Signed by President George H.W. Bush, the ADA prohibits discrimination in job application procedures, promotions, termination, compensation, job training, and many other employment-related factors 8 . Despite the implementation of this law, it is not self-enforcing, and research has shown that employers are not only ill-informed about the responsibilities and requirements under the law, but they are also skeptical of hiring individuals with disabilities due to concerns (unfounded or otherwise) that these individuals will not be able to work efficiently 9 . Despite employers who have reported satisfaction with the performance of their employees with disabilities, including the minimal costs to hire and accommodate such workers, individuals with disabilities continue to be underrepresented and unsupported in the U.S. workforce.

Research Question What leads to the high rate of unemployment among individuals with disabilities?

Employer Perception

Often, employers lack awareness on how to hire and/or accommodate workers with disabilities. As a result, many employers feel that employing a worker who may require accommodations will increase the burden on supervisors, co-workers, and human resource staff 10 . Employers have concerns about learning their responsibilities under legislation, evaluating costs and benefits of hiring, safety, potential legal liability, reactions of coworkers and customers, need for additional supervision, and potential unforeseen issues that may arise 6 8 . This lack of awareness or lack of exposure to successfully employed and accommodated workers with disabilities perpetuates this fear of the unknown and removes opportunities for individuals with disabilities to gain competitive employment. Many employers who are unfamiliar with the hiring process of individuals with disabilities rely on stereotypes of workers with disabilities that have manifested over decades. Employers erroneously believe that people with disabilities are often late or miss work, that they are poor job performers, and that there is a general discomfort working with a person with a disability 6 .

Previous research in this area suggests these types of negative attitudes toward employers and coworkers represent some of the most significant barriers to employment for people with disabilities. Negative attitudes toward workers with disabilities are multifaceted, with both workplace prejudice and the rejection of workplace initiatives intended to increase diversity influencing hiring authorities 11 . Workplace prejudice is exemplified by unreasonable and unrealistic beliefs that workers, such as individuals with disabilities, belong to a particular social group that is less capable, less motivated, or more problematic than their non-disabled coworkers 11 . These beliefs are based, in part, by stereotypes employers have developed about workers with disabilities and how poorly their attributes align with behaviors and requirements for particular roles in the workplace 11 . Despite legislation, more work needs to be done to educate employers and decrease the perpetual stigma that continues in hiring processes.

Another common barrier regarding the hiring process of individuals with disabilities is that employers have concerns over the potential expenses of providing accommodations for a person with a disability. Many employers fear exorbitant costs for accommodations, when, in fact, most workplace accommodation costs are low or cost nothing. For example, the Job Accommodation Network (JAN), a service of the U.S. Department of Labor’s Office of Disability Employment Policy (ODEP), has shown that workplace accommodations are low cost and positively impact the workplace in several ways 12 . The JAN survey has been ongoing since 2004, with over 3,300 employers completing the survey 12 . Survey results have consistently shown that employers find the benefits of providing workplace accommodations far outweigh the accompanying costs 12 . Many employers reported results of providing such accommodations included, but were not limited to, retaining valuable employees, improvement in productivity and workplace morale, reduction of workers’ compensation and training costs, and an increase in company diversity 12 . Employers who participated in the survey reported a high percentage (56%) of accommodations cost the company nothing, whereas the remaining accommodations cost less than $500 12 .

Additional research has found that employers who have hired individuals with disabilities in the past have typically done so if the employee is “easier” and less costly to accommodate 12 . Employers also have a propensity for hiring and retaining workers with physical disabilities compared to other types of disabilities such as psychiatric illnesses 12 . Employers have also expressed concerns about resentment toward workers with disabilities, even though resentment may come from the same prejudice and stereotypes that the ADA was designed to combat 12 .

Social Security Disability

There are two primary federal programs that offer benefits and entitlements to individuals with disabilities who are unable to work. Supplemental Security Income (SSI) is an income assistance program for individuals with disabilities who have little or no income and are unable to work to sustain themselves 13 . SSI provides monthly payments to adults and children with a disability who have income and resources below specific financial limits. These payments assist with meeting basic needs such as food, clothing, and shelter. Social Security Disability Insurance (SSDI) is a federal program that pays benefits to individuals if they are “insured” 13 . To be insured, individuals must work long enough and recently enough to pay Social Security taxes on their earnings 13 . While both programs were developed and are meant to aid individuals in need, there are also disincentives that have unintentionally contributed to the persistent problem of under and unemployment for individuals with disabilities.

In 1999, the Ticket to Work and Work Incentives Improvement Act became law. At the time, less than one-half of one percent of SSDI and SSI recipients stopped receiving benefits and returned to work 14 . The financial disincentives to work and earn income instead of receiving benefits coupled with the lack of adequate training and placement services were, and continue to be, significant barriers to employment for people with disabilities. The Ticket to Work Program was implemented to help overcome these significant barriers to employment. Services offered through this program include benefits planning, assistance, and outreach; expedited reinstatement (to address the fear of losing benefits); demonstration projects (designed to test the impact of various changes to the law on the ability of recipients to work and earn more); state partnership initiative (to help states develop innovative statewide programs that provide support to increase job opportunities and earnings), to name a few 14 .

Although the Ticket to Work Program was created to assist individuals with returning to work, there continues to be disincentives that make this transition difficult. One disincentive for individuals with disabilities who are receiving social security benefits (SSI or SSDI) is that to return to work, federal regulations mandate an administrative review of their disability status (i.e., continuing disability review) 15 . This review discourages many people from seeking employment. Secondly, once individuals do begin to work, their cash payments decrease as their earnings increase. SSDI beneficiaries can earn up to Social Security Administration’s (SSA) substantial gainful activity (SGA) level each month with no loss of their benefits. In 2022, the monthly SGA amount is $1,350 and $2,260 for individuals who are blind 16 . There is an incentive in place to help support individuals receiving SSDI who want to return to work. Individuals earning under SGA will continue to receive their benefits; however, once earnings exceed that amount (i.e., $1,350 in 2022) for nine nonconsecutive months (plus a three-month grace period), all SSDI case benefits cease, which is called the “earnings cliff” 17 . Although this incentive is meant to help support return to work, the fear and stress that arises when an individual begins to lose their benefits overshadows the incentive to work 18 . This fear is partly due to the lengthy approval process for SSDI, making it understandable that SSDI recipients are hesitant to return to work and lose their benefits. In most cases, if an individual returns to work but is later unable to continue working (due to the same disability), they won’t need to “re-qualify” for disability benefits. They are simply placed back on SSDI, or whatever disability programs they qualify for. While this is an incentive that can be extremely beneficial for people who would like to return to work, it is often overlooked or even unknown to SSDI recipients. Although this incentive is in place to support return to work, it continues to overshadow the complexity of the process of applying for social security and the fear of losing those benefits.

While many incentives for individuals with disabilities to return to work have been put into place over the last few decades, there continues to be a significant gap in the number of people without disabilities who are unemployed versus individuals with disabilities. Stigma regarding workers with disabilities continues to hinder employers from hiring an individual, despite their qualifications. These negative attitudes contribute to the lack of hiring of individuals with disabilities, despite an absence of research that supports these beliefs.

Additionally, these beliefs are fueled by a lack of information and education regarding individuals with disabilities and their positive contributions to the workplace. The lack of understanding or education regarding appropriate accommodations and concern for increased cost or undue hardship to their organization continues to be a significant barrier to hiring as well. Additionally, individuals receiving SSDI fear their loss of benefits when returning to work, which disincentivizes their return. Although these barriers continue to be addressed, further education and awareness can contribute greatly to assisting not only the individuals seeking employment but also the employers that are hiring.

• 1 Erin N. Moser, PhD, Associate Professor, University of Northern Colorado.
• 2 Jill L. Bezyak, PhD, Professor and PI of the Rocky Mountain ADA Center, University of Northern Colorado.
• 3 a b Chan, F., Tansey, T. N., Iwanaga, K., Bezyak, J., Wehman, P., Phillips, B. N., ... & Anderson, C. (2021). Company characteristics, disability inclusion practices, and employment of people with disabilities in the post COVID-19 job economy: A cross sectional survey study. Journal of Occupational Rehabilitation, 31(3), 463-473.
• 4 https://www.bls.gov/news.release/pdf/disabl.pdf
• 5 https://www.rutgers.edu/news/poor-physical-health-barrier-job-seekers-serious-mental-illness#:~:text=Research%20%26%20Innovation-,Poor%20Physical%20Health%20a%20Barrier%20for%20Job%20Seekers%20with%20Serious,to%2090%20percen
• 6 a b c https://www.un.org/development/desa/disabilities/resources/factsheet-on-persons-with-disabilities/disability-and-employment.html
• 7 https://www.americanprogress.org/article/removing-obstacles-for-disabled-workers-would-strengthen-the-u-s-labor-market/
• 8 a b https://www.dol.gov/agencies/odep/ada30/timeline
• 9 Lee, B. A. (1996). Legal requirements and employer responses to accommodating employees with disabilities. Human Resource Management Review, 6(4), 231-251.
• 10 Kaye, H. S., Jans, L. H., & Jones, E. C. (2011). Why don’t employers hire and retain workers with disabilities? Journal of Occupational Rehabilitation, 21(4), 526-536.
• 11 a b c Anglim, J., Sojo, V., Ashford, L. J., Newman, A., & Marty, A. (2019). Predicting employee attitudes to workplace diversity from personality, values, and cognitive ability. Journal of Research in Personality, 83, 103865.
• 12 a b c d e f g h https://askjan.org/topics/costs.cfm
• 13 a b c https://www.ssa.gov/benefits/ssi/
• 14 a b https://www.ssa.gov/policy/docs/ssb/v66n3/v66n3p29.html
• 15 Cook, J. A. (2006). Employment barriers for persons with psychiatric disabilities: Update of a report for the President's Commission. Psychiatric Services, 57(10), 1391-1405.
• 16 https://www.ssa.gov/oact/cola/sgadet.html
• 17 https://choosework.ssa.gov/library/fact-sheet-trial-work-period-twp
• 18 https://www.disability-benefits-help.org/faq/go-back-work-when-receiving-ssd

Federal Reserve Economic Data

Data in this graph are copyrighted. Please review the copyright information in the series notes before sharing.

• Page short URL
• Embed in website
• Image short URL

Units:   Percent , Seasonally Adjusted

Frequency:   Monthly

The unemployment rate represents the number of unemployed as a percentage of the labor force. Labor force data are restricted to people 16 years of age and older, who currently reside in 1 of the 50 states or the District of Columbia, who do not reside in institutions (e.g., penal and mental facilities, homes for the aged), and who are not on active duty in the Armed Forces. This rate is also defined as the U-3 measure of labor underutilization. The series comes from the 'Current Population Survey (Household Survey)' The source code is: LNS14000000

Suggested Citation:

U.S. Bureau of Labor Statistics, Unemployment Rate [UNRATE], retrieved from FRED, Federal Reserve Bank of St. Louis; https://fred.stlouisfed.org/series/UNRATE, .

Units:   Thousands of Persons , Seasonally Adjusted

Persons 16 years of age and older. The series comes from the 'Current Population Survey (Household Survey)' The source code is: LNS11000000

U.S. Bureau of Labor Statistics, Civilian Labor Force Level [CLF16OV], retrieved from FRED, Federal Reserve Bank of St. Louis; https://fred.stlouisfed.org/series/CLF16OV, .

The civilian noninstitutional population is defined as: persons 16 years of age and older residing in the 50 states and the District of Columbia, who are not inmates of institutions (e.g., penal and mental facilities, homes for the aged), and who are not on active duty in the Armed Forces. The series comes from the 'Current Population Survey (Household Survey)' The source code is: LNS12000000

U.S. Bureau of Labor Statistics, Employment Level [CE16OV], retrieved from FRED, Federal Reserve Bank of St. Louis; https://fred.stlouisfed.org/series/CE16OV, .

RELEASE TABLES

• Labor Force Status Flows by Sex: Seasonally Adjusted
• Monthly, Seasonally Adjusted (population data is not adjusted for seasonal variation; not seasonally adjusted version used)
• Table A-9. Selected employment indicators: Monthly, Seasonally Adjusted
• Table A-10. Selected unemployment indicators, Seasonally adjusted: Monthly, Unemployment Rates
• Table A-15. Alternative measures of labor underutilization: Monthly, Seasonally Adjusted

RELATED DATA AND CONTENT

Data suggestions based on your search.

Content Suggestions

Other Formats

Unemployment Rate

Civilian Labor Force Level

Employment Level

Related Categories

Permalink/embed, subscribe to the fred newsletter.

Third Quarter 2024 Survey of Professional Forecasters

Upward Revisions to Current Year Growth Expectations, but Downward Revisions to Employment

The outlook for the U.S. economy is more mixed than three months ago, according to 36 forecasters surveyed by the Federal Reserve Bank of Philadelphia. The panelists predict GDP will grow at an annual rate of 1.9 percent this quarter, down from 2.0 percent in the previous survey. However, the panelists revised upward their expectations for GDP growth in the fourth quarter of 2024 from 1.5 percent in the previous survey to 1.7 percent in the current survey. Overall, the forecasters revised upward their expectations for 2024 GDP growth on an annual-average over annual-average basis from 2.5 percent to 2.6 percent.

The forecasters see higher unemployment rates across all horizons compared with the previous survey. On an annual-average basis, the forecasters revised upward their expectations by 0.2 percentage point for both 2024 and 2025 to 4.1 percent and 4.3 percent, respectively. The forecasters expect the unemployment rate to be 4.2 percent in both 2026 and 2027, each up 0.1 percentage point from the last survey.

On the employment front, the panelists have slightly lowered their expectations for the next three quarters compared with the previous survey. The forecasters see job gains at a monthly rate steadily decreasing from 143,900 in the current quarter to 116,200 in the second quarter of 2025. However, in the third quarter of 2025, the forecasters expect monthly job gains to bounce back to 145,800. Overall, annual-average projections for nonfarm payroll employment decreased from a monthly rate of 210,100 in 2024 to 130,000 in 2025, both lower than those in the previous survey. (These annual-average projections are computed as the year-to-year change in the annual-average level of nonfarm payroll employment, converted to a monthly rate.) 

Median Forecasts for Selected Variables in the Current and Previous Surveys

The charts below provide some insight into the degree of uncertainty the forecasters have about their projections for the rate of growth in the annual-average level of real GDP. Each chart presents the forecasters’ current and previous estimates of the probability that growth will fall into each of 11 ranges.

The forecasters deem the most likely growth rate in all four years to be in the range of 1.5 percent to 2.4 percent, which matches the expectations of the previous survey.

• Mean Probabilities for Real GDP Growth in 2024  (chart)
• Mean Probabilities for Real GDP Growth in 2025  (chart)
• Mean Probabilities for Real GDP Growth in 2026  (chart)
• Mean Probabilities for Real GDP Growth in 2027  (chart)

The forecasters’ density projections for unemployment, shown below, shed light on uncertainty about the labor market over the next four years. Each chart presents the forecasters’ current and previous estimates of the probability that unemployment will fall into each of 10 ranges.

The forecasters predict the unemployment rate will most likely be in the range of 3.7 percent to 4.2 percent for 2024 and 2027, matching the expectations for the highest probability of the previous survey. However, for 2025 and 2026, the forecasters have shifted their highest unemployment rate expectations from a range of 3.7 percent to 4.2 percent in the previous survey to 4.3 percent to 4.8 percent in the current survey.

• Mean Probabilities for Unemployment Rate in 2024  (chart)
• Mean Probabilities for Unemployment Rate in 2025  (chart)
• Mean Probabilities for Unemployment Rate in 2026  (chart)
• Mean Probabilities for Unemployment Rate in 2027  (chart)

Forecasters Revise Down Inflation Expectations at Most Horizons

The forecasters have lowered their expectations for all current-quarter measures of inflation. Current-quarter headline CPI and PCE inflation are now expected to be 2.3 percent and 2.1 percent, respectively, down from 2.8 percent and 2.4 percent in the previous survey. Similarly, current-quarter core CPI and PCE inflation expectations were also revised downward to 2.6 percent and 2.4 percent, respectively.

Predictions at all other horizons for both headline and core CPI and PCE inflation have either been revised downward or remain the same when compared with the previous survey. The only two exceptions to this are the second quarter of 2025 projections for headline CPI and PCE inflation, both of which were revised upward by 0.1 percentage point to 2.4 percent and 2.2 percent, respectively.

Over the next 10 years, 2024 to 2033, the forecasters predict headline CPI inflation will be an annual-average rate of 2.30 percent, down from 2.33 percent in the previous survey. The corresponding estimate for 10-year annual-average PCE inflation is 2.10 percent, unchanged from the previous survey.

Median Short-Run and Long-Run Projections for Inflation (Annualized Percentage Points)

The charts below show the median projections (the red line) and the associated interquartile ranges (gray areas around the red line) for 10-year annual-average CPI and PCE inflation. The charts provide historical perspective on the 10-year inflation expectations in the current survey.

• Projections for the 10-Year Annual-Average Rate of CPI Inflation (chart)
• Projections for the 10-Year Annual-Average Rate of PCE Inflation (chart)

The figures below show the probabilities that the forecasters are assigning to each of 10 possible ranges for fourth-quarter over fourth-quarter core PCE inflation in 2024 and 2025. In 2024, the forecasters expect the most likely range for core PCE inflation to be in the range of 2.5 percent and 2.9 percent. In 2025, the forecasters see the most likely range to be lower than that in 2024, placing expectations in the range of 2.0 percent and 2.4 percent.

• Mean Probabilities for Core PCE Inflation in 2024  (chart)
• Mean Probabilities for Core PCE Inflation in 2025  (chart)

Lower Risk of Negative Quarter-over-Quarter Growth in 2024, but Higher Risk in 2025

The forecasters see a lower risk of a contraction in real GDP for both quarters of 2024 than in the previous survey but have revised upward their risk expectations for 2025. They see the least amount of risk of a contraction in the current quarter, which they estimate to be 16.2 percent, and the highest amount of risk in the second quarter of 2025, which they estimate to be 27.6 percent.

Risk of a Negative Quarter (%) Survey Means

Natural Rate of Unemployment Estimated at 4.40 Percent

In third-quarter surveys, we ask the forecasters to provide their estimates of the natural rate of unemployment — the rate of unemployment that occurs when the economy reaches equilibrium. The forecasters estimate this rate at 4.40 percent. The table below shows, for each third-quarter survey since 1996, the percentage of respondents who use the natural rate in their forecasts and, for those who use it, the median estimate and the lowest and highest estimates. Forty-four percent of the 27 forecasters who answered the question report that they use the natural rate in their forecasts. The lowest estimate is 3.50 percent, and the highest estimate is 5.16 percent.

Median Estimates of the Natural Rate of Unemployment

Technical Note Moody's Aaa and Baa Historical Rates The historical values of Moody's Aaa and Baa rates are proprietary and, therefore, not available in the data files on the Bank’s website or on the tables that accompany the survey’s complete write-up in the PDF.

The Federal Reserve Bank of Philadelphia thanks the following forecasters for their participation in recent surveys:

William Adams , Comerica Bank; Ed Al-Hussainy and Alexander Spitz , Columbia Threadneedle Investments; Scott Anderson and Doug Porter, BMO Capital Markets; Robert J. Barbera , Johns Hopkins University Center for Financial Economics; Peter Bernstein , RCF Economic and Financial Consulting, Inc.; Wayne Best and Michael Brown , Visa, Inc.; Jay Bryson , Wells Fargo; Seth Carpenter , Morgan Stanley; Christine Chmura , Ph.D. , and Xiaobing Shuai , Ph.D. , Chmura Economics & Analytics; Gary Ciminero , CFA , GLC Financial Economics; Grant Collins , AIM Research, LLC; Rajeev Dhawan , Georgia State University; Bill Diviney , ABN AMRO Bank NV; Gabriel Ehrlich , Daniil Manaenkov , and Yinuo Zhang , RSQE, University of Michigan; Michael R. Englund , Action Economics, LLC; Michael Feroli , J.P. Morgan; Tani Fukui and Shan Ahmed , MetLife Investment Management; Sacha Gelfer , Bentley University; James Glassman , Independent Economist; Jan Hatzius , Goldman Sachs; Steve Kihm , Citizens Utility Board of Wisconsin; Yaniv Konchitchki , University of California, Berkeley; Thomas Lam , Independent Economist (Singapore); Brian Martin , Australia New Zealand Bank (ANZ); Robert McNab , Old Dominion University; R. Anthony Metz , Pareto Optimal Economics, LLC; R. M. Monaco , TitanRM; Joel L. Naroff , Naroff Economics, LLC; Nomura Securities International ; Brendon Ogmundson , BC Real Estate Association; Perc Pineda, Ph.D. , Plastics Industry Association; Joel Prakken and Chris Varvares , S&P Global Market Intelligence; Jason Prole , Capital Risk Management; Michael Roberts , Dan Roberts , and Jeffrey Baldwin , Roberts Capital Advisors, LLC; Parker Ross , Arch Capital Group; Philip Rothman , East Carolina University; Allen Sinai and Lu Yu , Decision Economics, Inc.; Sean Snaith , University of Central Florida; Stephen Stanley , Santander US Capital Markets; Charles Steindel , Editor, NABE Business Economics ; Susan M. Sterne , Economic Analysis Associates, Inc.; Edward Sullivan , Portland Cement Association; Ryan Sweet , Oxford Economics USA, Inc.; Jordan Vickers and Maira Trimble , Eaton Corporation; Lawrence Werther , Daiwa Capital Markets America; Mark Zandi , Moody’s Analytics.

This is a partial list of participants. We also thank those who wish to remain anonymous.

Return to the main page for the Survey of Professional Forecasters .

• Topics ›
• Bangladesh ›

Youth Unemployment High in South Asia

Youth unemployment is being cited as one of the core drivers of the unrest in Bangladesh , which led to weeks of protests and Prime Minister Sheikh Hasina stepping down from office.

The following chart, based on ILO data , shows how labor force unemployment for people aged 15-24 years in Bangladesh stood at 15.7 percent in 2023, above the world average for youth unemployment of 13.8 percent and the low and middle income average of 14.1 percent. Youth unemployment is a regional issue, with India having hit a similar level in 2023, while Nepal and Sri Lanka’s rates last year were worse, both surpassing the 20-percent mark. Of this selection of countries, Pakistan fared better in 2023 at around 9.7 percent.

According to a report by the Japan Times, the latest figures indicate that in 2024, roughly 40 percent of Bangladeshi youth are not in education, employment or training, including those no longer looking for work or registered unemployed. The authors write that stagnant job growth in the private sector as well as a cooling economy has made public sector jobs more attractive. Protests started weeks ago over a quota for such civil service jobs which reserved 30 percent of government roles to relatives of veterans of the 1971 war of independence from Pakistan.

While all of the countries' unemployment rates have fallen from a pandemic-induced peak, they have in all five cases risen in the past decade.

Description

This chart shows the share of youth labor force who are unemployed in a selection of South Asian countries, by year.

Can I integrate infographics into my blog or website?

Yes, Statista allows the easy integration of many infographics on other websites. Simply copy the HTML code that is shown for the relevant statistic in order to integrate it. Our standard is 660 pixels, but you can customize how the statistic is displayed to suit your site by setting the width and the display size. Please note that the code must be integrated into the HTML code (not only the text) for WordPress pages and other CMS sites.

Infographic Newsletter

Statista offers daily infographics about trending topics, covering: Economy & Finance , Politics & Society , Tech & Media , Health & Environment , Consumer , Sports and many more.

Related Infographics

Indian budget, how big is india's new jobs plan, international politics, u.s. image deteriorates in parts of asia, heavy monsoon rains hit india one month into season, forced labor & human trafficking, which region has the highest number of child laborers, indian worker satisfaction tops regional average, what the u.s. is most worried about, child poverty, 1 in 7 children live in poverty in the u.s., where most people were internally displaced in 2023, world’s most pessimistic nation, gloomy greeks remain the world’s most pessimistic nation, inflation in the global economy, stubborn inflation fuels fears of 'no landing' scenario, the employment gap for people with disabilities, citizenship amendment act, india sees relatively few refugees from caa countries.

• Who may use the "Chart of the Day"? The Statista "Chart of the Day", made available under the Creative Commons License CC BY-ND 3.0, may be used and displayed without charge by all commercial and non-commercial websites. Use is, however, only permitted with proper attribution to Statista. When publishing one of these graphics, please include a backlink to the respective infographic URL. More Information
• Which topics are covered by the "Chart of the Day"? The Statista "Chart of the Day" currently focuses on two sectors: "Media and Technology", updated daily and featuring the latest statistics from the media, internet, telecommunications and consumer electronics industries; and "Economy and Society", which current data from the United States and around the world relating to economic and political issues as well as sports and entertainment.
• Does Statista also create infographics in a customized design? For individual content and infographics in your Corporate Design, please visit our agency website www.statista.design

Any more questions?

Get in touch with us quickly and easily. we are happy to help.

Feel free to contact us anytime using our contact form or visit our FAQ page .

Statista Content & Design

Need infographics, animated videos, presentations, data research or social media charts?

More Information

The Statista Infographic Newsletter

Receive a new up-to-date issue every day for free.

• Our infographics team prepares current information in a clear and understandable format
• Relevant facts covering media, economy, e-commerce, and FMCG topics
• Use our newsletter overview to manage the topics that you have subscribed to

Skip to content

Labour Market Report August 2024

Date published: 13 August 2024

The latest labour market statistics were published today (13th August 2024) by the Northern Ireland Statistics & Research Agency.

Payrolled employees increased over the month and median earnings decreased

• The number of employees receiving pay through HMRC PAYE in NI in July 2024 was 807,700, an increase of 0.1% over the month and 2.3% over the year.
• Earnings data from HMRC PAYE indicated that NI employees had a median monthly pay of £2,249 in July 2024, a decrease of £82 (3.5%) over the month and an increase of £164 (7.9%) over the year.
• When considering the annual change in median monthly pay by industry sector, the largest percentage increases over the year were recorded in the ‘Education’ (11.9%), ‘Other service activities’ (10.9%), ‘Agriculture, forestry and fishing’ (9.6%) and ‘Arts, entertainment and recreation’ (9.2%) sectors. These sectors all had median monthly earnings below the NI average.
• The estimates from HMRC PAYE for the latest period, are based on early data and, therefore, are more likely to be subject to larger revisions.

Increase in the seasonally adjusted claimant count rate over the month

• In July 2024, the seasonally adjusted number of people on the claimant count was 41,000 (4.2% of the workforce), an increase of 6.1% from the previous month’s revised figure. The July 2024 claimant count remains 37.4% higher than the pre-pandemic count in March 2020. These increases are largely due to the increase in the administrative earnings threshold for Universal Credit in May 2024.

Proposed redundancies lower than previous year while Confirmed redundancies almost double

• NISRA, acting on behalf of the Department for the Economy, received confirmation that 40 redundancies occurred in July 2024. Over the year August 2023 to July 2024, 2,550 redundancies were confirmed, which was almost double the figure for the previous year (1,340).
• There were 2,820 redundancies proposed in the twelve months to July 2024, which was around three quarters of the figure for the previous year (3,940).

Labour Force Survey headline measures

• The latest NI seasonally adjusted unemployment rate (the proportion of economically active people aged 16 and over who were unemployed) for the period April-June 2024 was estimated from the Labour Force Survey at 1.9%. This was a decrease of 0.2 percentage points (pps) over the quarter and a decrease of 0.7pps over the year.
• The proportion of people aged 16 to 64 in work (the employment rate) decreased by 0.3pps over the quarter and increased by 1.2pps over the year to 71.6%.
• The total number of weekly hours worked in NI was estimated at 29.0 million hours, an increase of 2.2% on the previous quarter and an increase of 2.7% on the equivalent period last year.
• The economic inactivity rate (the proportion of people aged 16 to 64 who were not working and not seeking or available to work) increased by 0.5pps over the quarter and decreased by 0.7pps over the year to 27.1%.

The statistical bulletin and associated tables are available on the Labour Market Report - August 2024 page.

Latest news

• Labour Market Report August 2024 13 August 2024
• Weekly deaths tables – week ending 2 August 2024 09 August 2024
• Census 2021 update 08 August 2024
• Weekly deaths tables – week ending 26 July 2024 02 August 2024

More news …

Skip to content

Northern Ireland Labour Market Statistics

Date published: 13 August 2024

The labour market statistics were published today by the Northern Ireland Statistics & Research Agency.

Payrolled employees increased over the month and median earnings decreased

• The number of employees receiving pay through HMRC PAYE in NI in July 2024 was 807,700, an increase of 0.1% over the month and 2.3% over the year.
• Earnings data from HMRC PAYE indicated that NI employees had a median monthly pay of £2,249 in July 2024, a decrease of £82 (3.5%) over the month and an increase of £164 (7.9%) over the year.
• When considering the annual change in median monthly pay by industry sector, the largest percentage increases over the year were recorded in the ‘Education’ (11.9%), ‘Other service activities’ (10.9%), ‘Agriculture, forestry and fishing’ (9.6%) and ‘Arts, entertainment and recreation’ (9.2%) sectors. These sectors all had median monthly earnings below the NI average.
• The estimates from HMRC PAYE for the latest period, are based on early data and, therefore, are more likely to be subject to larger revisions.

Increase in the seasonally adjusted claimant count rate over the month

• In July 2024, the seasonally adjusted number of people on the claimant count was 41,000 (4.2% of the workforce), an increase of 6.1% from the previous month’s revised figure. The July 2024 claimant count remains 37.4% higher than the pre-pandemic count in March 2020. These increases are largely due to the increase in the administrative earnings threshold for Universal Credit in May 2024.

Proposed redundancies lower than previous year while Confirmed redundancies almost double

• NISRA, acting on behalf of the Department for the Economy, received confirmation that 40 redundancies occurred in July 2024. Over the year August 2023 to July 2024, 2,550 redundancies were confirmed, which was almost double the figure for the previous year (1,340).
• There were 2,820 redundancies proposed in the twelve months to July 2024, which was around three quarters of the figure for the previous year (3,940).

Labour Force Survey headline measures

• The latest NI seasonally adjusted unemployment rate (the proportion of economically active people aged 16 and over who were unemployed) for the period April-June 2024 was estimated from the Labour Force Survey at 1.9%. This was a decrease of 0.2 percentage points (pps) over the quarter and a decrease of 0.7pps over the year.
• The proportion of people aged 16 to 64 in work (the employment rate) decreased by 0.3pps over the quarter and increased by 1.2pps over the year to 71.6%.
• The total number of weekly hours worked in NI was estimated at 29.0 million hours, an increase of 2.2% on the previous quarter and an increase of 2.7% on the equivalent period last year.
• The economic inactivity rate (the proportion of people aged 16 to 64 who were not working and not seeking or available to work) increased by 0.5pps over the quarter and decreased by 0.7pps over the year to 27.1%.
• The latest Labour Market release shows that over the year both payrolled employee numbers and earnings have increased. In addition, all the Labour Force Survey headline measures have improved over the year, with the unemployment and economic inactivity rates both decreasing and the employment rate increasing.
• The latest HMRC payroll data shows that payrolled employees increased by 0.1% over the month and by 2.3% over the year. Payrolled earnings decreased by 3.5% over the month and were 7.9% higher than July 2023.
• Households reported, via the Labour Force Survey (LFS), over the year to April-June 2024, a 1.2pps increase in the employment rate (to 71.6%) and decreases of 0.7pps in both the economic inactivity rate (to 27.1%) and the unemployment rate (to 1.9%). None of these annual changes were statistically significant.
• The total number of hours worked in April-June 2024 increased by 2.7% over the year, to 29.0 million hours per week. This is 0.6% below the pre-pandemic position recorded in October-December 2019.
• There was an increase of 6.1% in the claimant count estimate over the month to July 2024, from the revised figure for June 2024. The claimant count rate for July 2024 was 4.2%, an increase from the revised rate for June 2024 (3.9%). These increases are largely due to the increase in the administrative earnings threshold for Universal Credit in May 2024.
• Finally, in July 2024, the Department was notified of 40 confirmed redundancies, bringing the rolling twelve-month total of confirmed redundancies to 2,550, almost double the figure for the previous year (1,340). Although the rolling twelve-month total of confirmed redundancies is substantially higher than that of the previous year, it is similar to the levels seen in the decade preceding the pandemic. Over the year, there were 2,820 proposed redundancies reported to the Department, around three quarters of the figure for the previous year (3,940), and below the trend seen immediately before the pandemic.

Notes to editors: 

• The statistical report and associated tables are available at:

https://www.nisra.gov.uk/publications/labour-market-report-august-2024

• The Northern Ireland Statistics and Research Agency wishes to thank the participating households and businesses for their co-operation in agreeing to take part in the surveys and for facilitating the collection of the relevant data.
• ‘Over the quarter’ refer to comparisons between the latest quarterly estimates for the period April-June 2024 and the quarter preceding that (i.e. January-March 2024). ‘Over the year’ refer to comparisons between the latest quarterly estimates for the period April-June 2024 and those of the corresponding quarter one year previously (i.e. April-June 2023). Changes that are significant in a statistical sense (i.e. where the estimated change exceeded the variability expected from a sample survey of this size and was likely to reflect real change) are specifically highlighted.
• Estimates relating to April-June 2024 should be compared with the estimates for January-March 2024. This provides a more robust estimate than comparing with the estimates for March-May 2024, as the April and May data are included within both estimates.
• The official measure of unemployment is from the Labour Force Survey. This measure of unemployment relates to people without a job who were available for work and had either looked for work in the last four weeks or were waiting to start a job. This is the International Labour Organisation definition. Labour Force Survey estimates are subject to sampling error. This means that the exact figure is likely to be contained in a range surrounding the estimate quoted. For example, the unemployment rate is likely to fall within 0.6pps of the quoted estimate (i.e. between 1.3% and 2.5%).
• The claimant count is an administrative data source derived from Jobs and Benefits Offices systems, which records the number of people claiming unemployment-related benefits. In March 2018, the NI claimant count measure changed from one based solely on Jobseekers Allowance (JSA) to an experimental measure based on JSA claimants and out-of-work Universal Credit (UC) claimants who were claiming principally for the reason of being unemployed. Those claiming unemployment-related benefits (either UC or JSA) may be wholly unemployed and seeking work or may be employed but with low income and/or low hours, that make them eligible for unemployment-related benefit support. Under UC a broader span of claimants became eligible for unemployment-related benefit than under the previous benefit regime.
• Redundancies are provided by companies under the Employment Rights (Northern Ireland) Order 1996 (Amended 8 October 2006) whereby they are legally required to notify the Department of impending redundancies of 20 or more employees. Companies who propose fewer than 20 redundancies are not required to notify the Department, therefore the figures provided are likely to be an underestimate of total job losses, however, it is not possible to quantify the extent of the shortfall. All other things being equal we would expect more redundancies in sectors dominated by large businesses as they are the businesses that meet the 20 or more collective redundancy criteria.
• To prevent the potential identification of individual businesses, redundancy totals relating to fewer than three businesses are not disclosed. The Statistical Disclosure Control Policy is available here: https://www.nisra.gov.uk/publications/redundancies-background-information . Where the number of businesses does not meet the threshold for release (as detailed in the Statistical Disclosure Control Policy), individual monthly totals are not published.
• HMRC’s Pay As You Earn (PAYE) Real Time Information (RTI) system is an administrative data source. The PAYE RTI system is the system employers use to take Income Tax and National Insurance contributions before they pay wages to employees. These data relate to employees paid by employers only, and do not include self-employment income.
• Estimates of the number of paid employees and employee earnings from PAYE are classed as official statistics in development as they are still in their development phase. As a result, the data are subject to revisions. Early estimates (flash estimates) for July 2024 are based on around 85% of information and will be subject to revision in the next month’s release when between 98% and 99% of data will be available (main estimates). The size of revisions to main and flash estimates are similar for employees, while revisions to earnings flash estimates are typically larger than main estimate revisions. The HMRC PAYE covers the whole population rather than a sample of employees or companies. Data are based on where employees live and not the location of their place of work within the UK. Data are seasonally adjusted but not adjusted for inflation.
• The Labour Market Report will be of interest to policy makers, public bodies, the business community, banks, economic commentators, academics, and the general public with an interest in the local economy.
• The next scheduled release of the Labour Market Report will be published on the NISRA website on Tuesday 10th September 2024.
• Feedback is welcomed and should be addressed to:

Responsible statistician:

Mark McFetridge,

Economic & Labour Market Statistics (ELMS),

[email protected] or Tel: 028 902 55172.

Latest news

• Northern Ireland Labour Market Statistics 13 August 2024
• Murphy announces Assured Skills Welding Academy at South West College 12 August 2024
• Murphy announces Assured Skills Academy with EY 05 August 2024
• Department for the Economy opens consultation on Postgraduate Award Scheme 01 August 2024

More news …