Key takeaways

• The depressed overall hires rate is a key driver of new labor market weakness for young college graduates, as it makes it harder for them to break into the labor market. This is true across industries, not just in those that disproportionately employ young college graduates—suggesting the culprit is not a structural change in the economy like AI but a labor market in which employers are hiring less and workers are holding onto the jobs they have.
• The information sector—posited to be more AI-exposed—has experienced recent job losses but employs only 2.3% of young college graduates.
• High-tech industries, which employ about 1 in 10 college workers, expanded at a historically rapid pace in the early 2020s but have shown signs of softening over the last three years.

The early 2020s labor market for young college graduates was strong. But, as we showed in this series’ first blog post, the Class of 2026 is graduating college into a labor market that has notably weakened in the past two years. A growing share of young college graduates are looking for jobs, but their employment rates have not kept pace—meaning unemployment is rising faster for young graduates than for the overall workforce. While their outcomes remain better than those of their noncollege counterparts, the uptick in unemployment has been a rising concern.

In this blog post, we delve deeper into the industries where young college graduates are likely to work,1 examining whether it has been relatively more difficult to secure employment in these fields as the labor market has weakened. Our analysis first examines employment changes, then turns to labor market flows, including hires and separations rates. We also scour the data for signs of contraction in the tech sector that may disproportionately affect the prospects for young college graduates as they enter the labor market.

In the third blog post in the series, we will examine the occupations where young college graduates work with particular attention to occupations that may have grown or shrunk, as well as to those most exposed to AI.

Young college graduates work in industries with strong growth in this business cycle

Over half of young college graduates work in private education and health services, professional and business services, or public-sector jobs. Figure A displays the share of employment in each industrial sector or sector grouping for young college graduates ages 22 to 27, all college graduates, and young workers without a four-year college degree. Industries in the figure appear in order of the share of young college graduates they employ, from largest to smallest.

The types of jobs where young college graduates work look similar to those of college graduates generally. Young workers without a college degree (i.e., noncollege) are far more likely to work in trade, transportation, and utilities; mining, construction, and manufacturing; or leisure and hospitality. All groups of workers are least likely to work in the information sector, closely watched for signs of AI-induced displacement.

Figure A

Figure B shows the change in employment in each sector between 2019 and 2026 and between 2023 and 2026, arranged in the same order as Figure A for comparison. The two fastest growing sectors since the last business cycle peak occurred in the two largest sectors for young college graduates: private education and health services and professional and business services.

Since the rollout of ChatGPT, many have looked at industries and occupations likely to be exposed to AI to see whether this has led to weaker job growth. Among the most closely watched of these industries is the information sector, which has seen an 8.5% employment decline since 2023.2 While these losses are notable—and especially relevant to understanding AI’s fingerprints on the labor market for young college-educated workers—it cannot be overemphasized just how small this sector is in the overall economy. Less than 2% of overall employment is in the information sector, including only 2.3% of young college graduates. Further, the sector saw a rapid employment expansion between 2019 and 2023—the employment loss between 2019 and today is just 2.0%.

Figure B

High-tech sectors have relatively more college graduates but haven’t experienced large AI-induced employment losses

In recent years, the Census Bureau has created an experimental data series on high-tech industries to better understand business dynamics. These include both manufacturing and service sector components of high tech.3 We translate their NAICS classification for high-tech industries into Census Industry Classifications used by the Current Population Survey (CPS) to determine the likelihood of young college graduates working in these sectors.

In the 2026 economy, about 5.6% of workers were employed in what the Census considers high-tech industries. Just about 1 in 10 (9.9%) of the college-educated workforce works in the tech industry. Young college graduates are similarly represented: 10.3% work in tech.

Overall, the Current Establishment Survey tells us that tech industry employment tracked changes in overall private employment in the prior business cycle (between 2007 and 2019) but expanded sharply in the early 2020s and has softened a bit in the last three years. Since 2023, the tech sector has fallen by 0.7%. While overall employment using CPS does show modest growth, neither shows large swings that suggest a large impact for young college graduates.4

Weak hires may be the biggest culprit to labor market weakness for young college graduates

The Job Openings and Labor Turnover Survey (JOLTS) can shed light on the question of whether entry-level workers—with or without college degrees—are facing a harder labor market to break into. While JOLTS doesn’t include demographic characteristics, it presents jobs openings as well as rates of hiring, layoffs, quits, and other separations. Today’s economy has substantially less churn than during the recovery from the pandemic, when millions of workers reentered the labor market after mass layoffs—many quit soon after as they searched for, and generally found, better opportunities.

Figure C shows the hires and separation rates. The lighter colors represent the monthly seasonally adjusted data for each series while the darker colors represent a 12-month moving average that provides a better overall picture of recent trends, smoothing out some data volatility. Over the last five years, the hires rate has steadily fallen and now sits at levels last seen in 2013 and 2014, when the labor market was still struggling to recover from the Great Recession.

The total separations rate includes quits, layoffs and discharges, and other separations.5 As with the hires rate, the separations rate has been declining over the last few years and now sits about where it was in 2014. Much of this is due to reductions in quits. Quits are higher when workers feel confident that they will find better job opportunities. Right now, workers are sitting tight, more so than any point in the past 10 years. Taken together, there is simply less churn in the labor market. But reduced churn is not inherently bad. If the frantic labor market of the early 2020s led to many workers and employers finding satisfactory matches, it could make sense that the following years would see less churn than normal. But for young workers looking to enter the job market, a reduction in hiring can make it harder to find a job.

Figure C

Table 1 breaks down the change in the hires and separations rate over the last three years, again using 12-month moving averages to smooth some volatility in the data. The industries are listed in order of the share of young college graduates they employ, similarly to Figure A. Overall, the hires rate fell 0.8 percentage points and the separations rate fell 0.6 percentage points between 2023 and 2026. The industries where young college graduates are more likely to work saw smaller reductions in both hires and quits than the overall. Industries where young workers without a college degree are more often found—over a quarter are in trade, transportation, and utilities—saw greater losses. Finally, leisure and hospitality, where young noncollege are more than twice as likely to work as young college graduates, saw the largest declines in hiring.

It does not appear that the industries where young college graduates tend to work are experiencing more weakness than other industries. Job gains are just as strong, if not stronger, and hiring hasn’t fallen as far in other industries. In short, there does not seem to be any profound structural change in the economy affecting the industry composition of employment—AI or anything else—that would easily explain the softening of the labor market for young college graduates in recent years. What it does appear to be is a harder labor market for young workers to break into when employers are less likely to hire and workers are more likely to sit tight in the job they have.

1. Throughout this blog post, we define young college graduates as people between the ages of 22 and 27 with only a four-year college degree. Unlike similar analyses of young workers, we do not exclude young college graduates that are currently enrolled in school, but the results here are robust either way. Unless otherwise noted, data for 2026 represent a 12-month average from April 2025 through March 2026 for the most up to date and reliable estimates, which removes seasonality and increases sample sizes.

2. ChatGPT was first introduced in November 2022 but took several months for more widespread usage.

3. High tech industries: Computer and Peripheral Equipment Manufacturing, Communications Equipment Manufacturing, Semiconductor and Other Electronic Component Manufacturing, Navigational, Measuring, Electromedical, and Control Instruments Manufacturing, Aerospace Product and Parts Manufacturing, Software Publishers, Data Processing, Hosting, and Related Services, Other Information Services, Architectural, Engineering, and Related Services, Computer Systems Design and Related Services, and Scientific Research and Development Services.

4.  It’s not unusual for the CPS and CES to display small differences in employment levels or trends considering nontrivial differences in their methodologies. 

5. Other separations include separations due to retirement, death, disability, and transfers to other locations of the same firm.

Job openings have generally been decreasing over the last year, slowly but steadily moving closer to their pre-pandemic levels, though clearly not there yet. Much of the elevated rates we’ve seen may have been because of the increased labor market churn and not net new demand. pic.twitter.com/XAYqdMNZUy

— Elise Gould (@eliselgould) May 31, 2023

Last month, it appeared that the labor market might be cooling given the rising layoffs in March, but job growth in April and the drop in layoffs in today’s report suggest the labor market remains steady and is not rapidly cooling. pic.twitter.com/vQCQhsX05u

— Elise Gould (@eliselgould) May 31, 2023

In November, job openings softened mildly (from 6.6 to 6.5 million) while hires were essentially unchanged, rising slightly (from 5.91 to 5.98 million). Another troubling sign was the large increase in layoffs (from 1.7 to 2.0 million). The number of layoffs and discharges increased notably in accommodation and food services, health care and social assistance, and state and local government.

On the whole, the U.S. economy is seeing a significantly slower hiring pace than we experienced in May or June—roughly where it was before the recession, which is a big problem given that we have only recovered just over half of the job losses from this spring. And job openings are now substantially below where they were before the recession began (6.5 million at the end of November, compared to 7.1 million on average in the year prior to the recession). And today’s data release only covers through November, so it doesn’t even capture December’s job losses, which were substantial. With hiring and job openings at these levels, the economy is facing a long, slow recovery without additional action from Congress.

One of the most striking indicators from today’s report is the job seekers ratio, that is, the ratio of unemployed workers (averaged for mid-November and mid-December) to job openings (at the end of October). On average, there were 10.7 million unemployed workers while there were only 6.5 million job openings. This translates into a job seeker ratio of about 1.6 unemployed workers to every job opening. Another way to think about this: for every 16 workers who were officially counted as unemployed, there were only available jobs for 10 of them. That means, no matter what they did, there were no jobs for 4.2 million unemployed workers. And this misses the fact that many more weren’t counted among the unemployed. The economic pain remains widespread with 26.8 million workers hurt by the coronavirus downturn.

With growing COIVD-19 cases and falling employment, the incoming Biden administration will be facing a mounting, not waning, crisis. The latest congressional relief bill is an important step toward addressing some of this pain, but it is not at the scale of the problem. I’m hopeful that more relief measures are on the horizon for increasingly desperate workers and their families. Senate Republicans forced the December bill to be far too small. Fortunately, with the Democratic majority in the Senate given the results of the Georgia runoffs, Democrats will now be able to get more relief measures through reconciliation. Their top priorities must be aid to state and local governments and further extensions of unemployment insurance.

The revisions will also provide details on changes in the initial payroll employment estimates by sector. For instance, in the preliminary release, the revisions were located primarily in “leisure and hospitality”, “professional and business services”, and “retail trade” with downward revisions of -175,000, -163,000, and -146,400, respectively. On Friday, the historical data will reflect the final benchmarks overall and by sector.

Tracking trends in nominal wage growth

Turning to nominal wage growth, the most important economic indicator to watch in 2020, last month there was a large drop for production/nonsupervisory workers. The figure below charts year-over-year changes in private-sector nominal average hourly earnings for “all nonfarm employees” as well as “production/nonsupervisory workers.” After remaining consistently higher than “all nonfarm” for nearly a year and at or above 3.4% for much of that time, it fell to 3.0% in December, its lowest point since September 2018. This begs the question of whether this is simply a blip and production/nonsupervisory workers will continue to pull away or if the separation in growth rates between the two over the last year was mostly statistical noise.

At this point in the recovery—with unemployment at or below 4.0% for 22 months—wage growth remains lower than expected. As employment growth consistently remains higher than working-age population growth, more and more workers are pulled into the labor force and finding jobs. As this slack gets absorbed, workers should be getting scarcer and scarcer. Therefore, employers would typically have to pay more to attract and retain the workers they want. After increasing in 2018, wage growth for all nonfarm employees has slowed for much of 2019 and remains below target levels.

Nominal Wage Tracker

On Friday, the BLS will also be employing new population controls in the Current Population Survey (CPS) starting in January 2020. Unlike the establishment survey, these changes to the CPS are not updated historically so caution should be exercised when making comparisons with data for December 2019 or earlier periods. The BLS is also making some changes to their methodology in terms of providing new seasonally adjusted series for measures of labor market underutilization as well as beginning to include both those in opposite-sex and same-sex marriages in estimates of married persons.

The new benchmarks to the establishment survey as well as revisions to the household survey will provide much fodder for thought on Friday morning. And, wage growth continues to be the most important indicator to watch as it lags behind overall improvements in the labor market.

The unemployment rate remained at 3.7% in August. The labor force participation rate edged up 0.2 percentage points to where it was when the year started, at 63.2%. The prime-working age labor force participation rate (25–54 years old) saw a significant gain in August, rising 0.6 percentage points. Like the overall labor force participation rate, it is now sitting exactly where it was at the start of 2019 (82.6%). The prime-age employment to population ratio rose as well (+0.5 percentage points) in August and is now slightly above its level in January (80.0% versus 79.9%). While the August increases are welcome, they are only just making up for the losses in participation that occurred earlier in this year.

Nominal wages grew 3.2% year-over-year in August, which is slower than expected in an economy that has had historically low unemployment—the unemployment rate has been at (or below) 4.0 % for the past 18 months. Wage growth has been particularly disappointing because it provides further evidence that the deceleration experienced in the first half seems to be holding.

Next week, when the annual Census data on poverty and income are released, we will have a fuller picture of how working people have fared this far into the economic recovery.

The unemployment rate remained stable at 3.7 percent. The labor force participation rate and the share of the population with a job each increased 0.1 percentage points in July, a sign that slowly more and more would-be workers are drawn back into the labor force in search of jobs and many of them are securing them. The share of workers who want full-time work but can only find part-time work (involuntarily part-time) also fell significantly in July. Unfortunately, the prime-working age population (25–54 years old) saw some losses in July; both the prime-age participation rate and the employment-to-population ratio fell 0.2 percentage points. Neither series has seen sustained improvement since last summer.

Wage growth continues to fall short of what we’d expect in an economy that has had historically low unemployment—the unemployment rate has been at (or below) 4.0 percent for the past 17 months. Nominal wages grew 3.2 percent in July over last year, remaining below levels consistent with Federal Reserve’s target for inflation and long-term potential productivity growth. Recent data releases have provided ample evidence of a slowdown in the U.S. economy, as well as a slowdown in nominal wage growth. This week’s decision by the Federal Reserve to cut short-term interest rates by a quarter-point makes lots of sense.

Unfortunately, disappointing wage growth is not a new feature of the U.S. economy. According to new data released this week, hourly pay has continued to diverge from economy-wide productivity. Since 1979, productivity has risen six times faster than hourly compensation for the typical U.S. worker.

Skeptical readers might argue that as much as the two states are similar, they are sufficiently different such that the diverging economic outcomes observed in our report are the result of fundamental differences in the two states’ economies and that state policy decisions were largely irrelevant. I think there’s ample evidence to indicate that such readers are wrong. In the paper, I discuss some of the policy decisions—such as those around Medicaid expansion, investment in infrastructure, and worker organizing—where one can draw a fairly straight line from the policy decision to the observed economic result. I also note that wage growth was actually stronger in Wisconsin than Minnesota in the seven years prior to Great Recession.

It’s also instructive to compare job growth in the two states in the economic expansion prior to the Great Recession. The data suggest that whatever their differences, prior to the recession, Wisconsin and Minnesota followed a very similar trajectory for employment growth.

Figure A

Figure A shows the number of jobs in Wisconsin, Minnesota, and the United States from November 2001 to December 2017, relative to the number of jobs in each geography in January 2011, the month that Governors Walker and Dayton took office. As you can see from the figure, changes in the level of jobs throughout the business cycle leading up to the Great Recession were remarkably similar between the two states. Both Minnesota and Wisconsin had modest job losses in the beginning of the period in the wake of the early 2000s recession, followed by modest job growth that tracked the U.S. average for a while and then flattened out for roughly the last two years prior to the onset of the Great Recession. In that earlier business cycle from November 2001 to December 2007, cumulative job growth was 3.7 percent in Minnesota and 3.3 percent in Wisconsin. Subsequently, the two states suffered losses in the recession that were similar, albeit slightly more severe in Wisconsin—with losses of 4.3 percent and 4.9 percent in Minnesota and Wisconsin, respectively, from December 2007 to December 2010.

The period from January 2011 to December 2017, after Governors Walker and Dayton assumed office, shows a starkly different picture. From early on in the recovery, Minnesota’s job growth accelerated noticeably more quickly than Wisconsin’s and the gap between the two states has increased fairly steadily ever since.

As I note in the paper, it would take a much more complicated and sophisticated analysis than I have done to pinpoint the precise jobs impact of particular policymakers’ decisions in each state. Nevertheless, I would be willing to bet that Minnesota’s decision to make large public investments in infrastructure, education, and healthcare—including the Affordable Care Act (ACA) Medicaid expansion—early in the latter period likely provided some extra fuel for the state’s economy. Conversely, Wisconsin lawmakers’ decisions to reject federal funding for the Milwaukee-Madison high-speed rail, to reject the ACA Medicaid expansion, and to take an ax to their public sector very likely stifled potential job growth.

The H-2B visa program is a guestworker program that allows employers to temporarily hire migrant workers in low-wage nonagricultural occupations, such as landscaping, forestry, hospitality, construction, and seafood processing, when no U.S. workers are available. (Table 1 below shows the top 10 H-2B occupations in fiscal 2016, according to labor certification disclosure data from the government.) Since May 2017, the U.S. Department of Homeland Security (DHS), in consultation with the U.S. Department of Labor (DOL), had been considering temporarily increasing the annual numerical limit of visas that can be issued to new workers (commonly referred to as the “cap”) through the H-2B visa program, which is currently set at 66,000. (Note that although the annual cap is set at 66,000, there were approximately 136,000 H-2B workers employed in the United States in fiscal 2016; this is because some H-2B workers remain employed for longer than one fiscal year [Costa 2017b]).

Table 1

DHS has the authority to increase the number of H-2B visas because in May 2017 Congress included language in the fiscal 2017 omnibus spending bill (which funds the federal government) allowing DHS to roughly double the number of H-2B visas that can be issued if the agency sees fit. On June 27, 2017, ProPublica reported that a DHS Interim Final Rule was likely to be published soon to increase the number of H-2B visas (Grabell and Elliott 2017) and Bloomberg BNA reported about a week later that the rule was under review at the Office of Management and Budget (Francis 2017). On July 17, 2017, a number of news outlets, including Politico (Hesson 2017) and The Washington Post (Jan 2017), reported that DHS had decided to increase the H-2B cap by 15,000, making the additional visas available to employers who apply for them during fiscal 2017. On the same day DHS and DOL jointly posted an unpublished rule to the Federal Register for public inspection, laying out the procedures and legal basis for the cap increase and noting that the final rule will be published on July 19 (U.S. DHS and U.S. DOL 2017).

Below is the actual language from Section 543 of the omnibus bill authorizing a cap increase:

Notwithstanding the numerical limitation set forth in section 214(g)(1)(B) of the Immigration and Nationality Act (8 U.S.C. 1184(g)(1)(B)), the Secretary of Homeland Security, after consultation with the Secretary of Labor, and upon the determination that the needs of American businesses cannot be satisfied in fiscal year 2017 with United States workers who are willing, qualified, and able to perform temporary nonagricultural labor, may increase the total number of aliens who may receive a visa under section 101(a)(15)(H)(ii)(b) of such Act (8 U.S.C. 1101(a)(15)(H)(ii)(b)) in such fiscal year above such limitation by not more than the highest number of H–2B nonimmigrants who participated in the H–2B returning worker program in any fiscal year in which returning workers were exempt from such numerical limitation.1 [emphasis added]

The language explicitly states that DHS’s decision on whether to raise the cap should be made after consulting with DOL to determine whether “the needs of American businesses cannot be satisfied in fiscal year 2017 with United States workers who are willing, qualified, and able to perform temporary nonagricultural labor.” If not, then DHS “may increase the total number of aliens who may receive a visa” in the H-2B program, up to about a total of 130,000 (which is approximately the number of visas issued in 2007, the year that “the highest number of H–2B nonimmigrants who participated in the H–2B returning worker program” were admitted [Bureau of Consular Affairs 2017]).

There are a number of reasons why the H-2B program should not be expanded. There are numerous cases of litigation, reports in the media, and government audits documenting how migrants employed through the H-2B program are often exploited and robbed by employers, and even become victims of human trafficking.2 While these most-egregious examples are clear legal violations, much of the abuse and discrimination in the H-2B program is perfectly legal because employers have been allowed to underpay H-2B workers for years thanks to loopholes in wage regulations (Costa 2016b, 2017a). And because U.S. workers are forced to compete with vulnerable H-2B workers, wages and working conditions for all workers in major H-2B occupations are degraded. As a result, there’s no question that the H-2B program needs major reforms to protect both migrant and American workers—but any push to expand H-2B should be made through the regular legislative process and after a full debate in Congress, so that members are accountable for their votes—not as a fly-by-night provision on a must-pass spending bill that has no identifiable author. In essence, expanding the program with no reform to improve worker protections will allow unscrupulous employers to carve out an even larger rights-free zone in the low-wage labor market that will drag on wage growth.

Nevertheless, we can engage with the actual language in the fiscal 2017 omnibus bill that authorizes DHS to increase the cap and consider whether the increase to the H-2B cap was justified. Section 543 of the bill indicates that DHS may increase the cap after consulting with the Secretary of Labor to determine whether there are any “United States workers who are willing, qualified, and able to perform temporary nonagricultural labor.” Leading up to the cap increase announced on July 17, 2017, employers that use the H-2B program and the corporate lobby groups that represent them argued that they are currently experiencing a severe shortage of available U.S. workers to fill H-2B jobs, and that, for this reason, DHS should raise the H-2B cap. But other than anecdotes of shortages—many of which have been reported by the media—very little labor market data or metrics have been offered to prove the existence of labor shortages in H-2B jobs. What do the data have to say?

At the national level, it is worth pointing out that there is no evidence at all of labor shortages in the labor market that are significant enough to move national data trends. Despite much-welcome progress in reducing the national unemployment rate in recent years, and despite the fact that the unemployment rate has now reached pre-Great Recession levels, many other labor market indicators signal an economy that still has some way to go before genuine full employment is attained. The share of adults between the ages of 25 and 54 (“prime-age” adults who traditionally have very strong labor force attachments), for example, has just recently clawed back half of the peak-to-trough decline it suffered during the Great Recession (EPI 2017a). Most persuasively, there is no evidence at all of a durable acceleration in the growth of either wages or prices. Full employment as a concept is precisely the lowest unemployment rate consistent with inflation that is stable at the Federal Reserve’s preferred target (2 percent). The U.S. economy has spent most of the past eight years with inflation below this target, and it remains significantly below it even as of May 2017. Crucially, inflation has not been rising in recent months. The story is similar for wage growth. Nominal wage growth has picked up a bit from its absolute nadir in the immediate aftermath of the Great Recession (EPI 2017b), but it remains far below any reasonable target of what it should be in a healthy economy (Bivens 2015), and there is little sign of any durable acceleration.

While there is no evidence that labor shortages are a significant feature of the national labor market, could there be shortages in particular occupations? It’s always possible. Determining whether a labor shortage exists in a particular occupation can be a difficult and inexact science, but those claiming shortages exist should at least provide the most rudimentary data consistent with such a claim. For example, if an occupation has seen flat or declining wages coupled with consistently high unemployment rates over a prolonged period of time, those are strong indicators that the occupation is not experiencing a shortage. With that in mind, let’s turn to the latest national-level data on wages and unemployment in H-2B jobs. (I have previously written about the change in wages and unemployment rates in H-2B occupations through 2014 and 2015; see Costa 2016b and Costa 2016a, respectively.)

Average hourly wages in the top 10 H-2B occupations, 2004–2016

Table 2 shows the change in average hourly wages for all workers in the United States (in all occupations) compared with the changes for all workers in the United States in the top 10 H-2B occupations, from 2004 to 2016, adjusted to 2016 dollars.3

Table 2

As Table 2 shows, there was little wage growth for all workers nationwide in all occupations—only 5.49 percent or $1.23 over the 2004–2016 period—but wage growth was even slower for workers in nine out of the top 10 H-2B occupations. Wages were stagnant—growing less than one-half of 1 percent annually in all but one of the top 10 H-2B occupations, and wages even declined for workers in two of the top 10 H-2B occupations (“Cooks, Restaurant” and “Helpers—Production Workers”) while one had zero growth (“Nonfarm Animal Caretakers”). The only one of the top 10 H-2B occupations that saw faster-than-average wage growth was “Waiters and Waitresses,” which grew at 1.57 percent per year, for a total of 20.52 percent or $2.00 in 2016 dollars, over the 12-year period from 2004 to 2016.

The economy overall, however, has seen wage growth tick upward slightly since 2014, and, consistent with overall trends, wages in the top H-2B occupations rose faster as well since 2014. Table 3 below shows the change in average hourly wages for all workers in the United States (in all occupations) and for all workers in the United States in the top 10 H-2B occupations, from 2014 to 2016, adjusted to 2016 dollars.

Table 3

As Table 3 shows, average hourly wages for all occupations grew 3.62 percent over the 2014–2016 period, at an average rate of 1.79 percent (roughly 2 percent) per year. Wages in all but one of the top 10 H-2B occupations grew in the 2 to 3 percent range per year, except for “Waiters and Waitresses,” which had above average wage growth at 5.81 percent per year.

While it is clearly an encouraging sign to finally see some improved wage growth in the top 10 H-2B occupations over the past three years, there are a few important caveats to consider. First, wage growth during 2015 and 2016 cannot be solely attributed to excess labor market tightness (i.e., labor shortages) pushing up nominal wages. Instead, inflation-adjusted wage growth in those years was helped quite a lot by a very rapid deceleration of inflation. Overall CPI inflation averaged less than 0.7 percent in those two years, after averaging around 2 percent in the four preceding years. Decelerating inflation will, all else being equal, push up growth in inflation-adjusted wages (Bivens 2016). If we thought this decelerating inflation would continue forever, then maybe we could assume that the 2014–2016 wage growth would be sustainable going forward. But we know that the deceleration of inflation in those years was driven by falling commodity prices (mostly oil). Commodity prices are notoriously volatile, and future years will almost surely see a recovery of these prices, and hence downward pressure on real wages. So, while the 2–3 percent real wage growth in H-2B occupations from 2014 to 2016 was clearly good for those workers, it is far from clear that it stemmed from a tight labor market or that it will reliably persist going forward.

Further, even if we ignore these mitigating factors and assume that the improved wage growth during 2014–2016 was due to labor market tightness and will continue, the decent gains made from 2014 to 2016 do not even come close to compensating workers in the top H-2B occupations for the lost decade of wage growth that the economy is just beginning to recover from. It will take at least a few more years of 2 to 3 percent annual wage growth for these occupations to start approaching the average hourly wage levels they would have reached if the Great Recession had not occurred. Heeding calls to raise the H-2B cap to stem the upward pressure of 2014–2016 would essentially be stomping out the only encouraging sign of wage growth these workers have seen over the past decade, and would lock in previous years’ wage declines.

Unemployment rates in the top 10 H-2B occupations, 2004–2016

Unemployment rates in H-2B occupations are calculated from Current Population Survey basic monthly microdata, which are jointly maintained by the U.S. Census Bureau and the Bureau of Labor Statistics. These data are not classified by SOC code, but instead use Census codes. However, the H-2B occupations with the 10 largest numbers of certifications in fiscal 2016 match up reasonably well with the same or similar occupations found in the CPS data using the government’s crosswalks between the occupations, even though the occupational titles may differ slightly in some cases.

Figure A shows the average unemployment rates in each of the top 10 H-2B occupations listed, for the periods 2004–2005, 2006–2007, and 2015–2016. Two years of data were pooled together to increase sample sizes. The first two periods listed were chosen because they exclude the years of the recession that began in 2008 and its aftermath, and 2015–2016 was chosen because those years represented the most recent data available.

Figure A

Figure A shows that while the unemployment rates in all but one of the top 10 H-2B occupations were high during the 2004–2005 and 2006–2007 periods, the unemployment rates did in fact decline in all but one of the top 10 H-2B occupations between 2004 and 2016, consistent with an improving economy after the Great Recession. Nevertheless, the unemployment rates in the top 10 H-2B occupations are far from being extremely low: The unemployment rates in all of the top 10 H-2B occupations in 2015–2016 were above the national unemployment rate in 2016 of 4.9 percent. Seven of the 10 occupations had very high unemployment rates of between of 7.78 percent and 10.64 percent. In two of the occupations—“Butchers and other meat, poultry, and fish processing workers” and “Nonfarm animal caretakers”—the 2015–2016 unemployment rate was very near the national unemployment rate (at 5.06 and 5.40 percent, respectively), but still slightly above it.

Conclusion: Scant evidence of labor shortages does not justify increasing the number of H-2B visas

The fact that, from 2004 to 2016, wages were stagnant or declined in nine of the top 10 H-2B occupations and only increased significantly in one of those occupations, combined with persistently high unemployment rates over the same period, makes it highly unlikely that labor shortages exist at the national level in the top H-2B occupations. Even considering the uptick in wage growth since 2014, the available evidence suggests that the labor market for H-2B occupations still has a lot of slack, which points to the high likelihood that there are many United States workers who are willing, qualified, and able to perform H-2B jobs. This does not mean that no labor shortages exist anywhere in the United States in these occupations—it is entirely possible that shortages may exist in some states or localities—but the relatively high national unemployment rates in H-2B occupations during the 2004–2016 period suggest that even the employers experiencing a local labor shortage might find available U.S. workers if they recruited outside their city, region, or state, and if they offered more attractive wages and benefits (including offering transportation and housing to workers from outside the local area). Given these realities, DHS’s fiscal 2017 expansion—by 15,000 visas—of a program as flawed as the H-2B program is ill-advised. In addition, the H-2B program still has a legal framework that allows employers to underpay and exploit migrant workers4 and to overlook unemployed U.S. workers5 seeking jobs in H-2B occupations. DHS and DOL should be focusing their resources and staff time on program oversight, on rooting out bad-actor employers who violate H-2B rules, and on protecting the migrant workers who come to the United States in search of better opportunities through the H-2B program—not on processing more visas for H-2B employers.

Acknowledgments

The author would like to thank Josh Bivens for his valuable insights and suggestions, and Jessica Schieder and Zane Mokhiber for technical and research assistance.

Endnotes

1.  H.R.244 – Consolidated Appropriations Act, 2017. 115th Congress (2017–2018).

2. See, for example, SPLC 2012; Garrison, Bensinger, and Singer-Vine 2015b; U.S. GAO 2015; and Stack 2015.

3. Wage data from the Bureau of Labor Statistics’ Occupational Employment Statistics (OES) survey data are used for this purpose because OES data are utilized by DOL to set H-2B prevailing wage rates and are therefore superior to other wage data sets for analyzing H-2B occupations or making comparisons between the wages of all workers and H-2B workers.

4. See, for example, Costa 2017a, New York Times Editorial Board 2016, and Bauer and Stewart 2013.

5. See, for example, Garrison, Bensinger, and Singer-Vine 2015a.

References

Bureau of Consular Affairs. 2017. Nonimmigrant Visa Statistics. U.S. Department of State.

Bauer, Mary, and Meredith Stewart. 2013. Close to Slavery: Guestworker Programs in the United States. Southern Poverty Law Center, February 18.

Bivens, Josh. 2015. A Vital Dashboard Indicator For Monetary Policy: Nominal Wage Targets. Center on Budget and Policy Priorities, June 15.

Bivens, Josh. 2016. Should We Care about Slow Nominal Wage Growth When Price Inflation Is Slow? YES. Economic Policy Institute, February 6.

Bureau of Labor Statistics. Various years. Current Population Survey. U.S. Department of Labor.

Bureau of Labor Statistics. Various years. Occupational Employment Statistics. U.S. Department of Labor.

Costa, Daniel. 2016a. “The H-2B Temporary Foreign Worker Program: Examining the Effects on Americans’ Job Opportunities and Wages.” Testimony before the U.S. Senate Subcommittee on Immigration and the National Interest, June 8.

Costa, Daniel. 2016b. The H-2B Temporary Foreign Worker Program: For Labor Shortages or Cheap, Temporary Labor? Economic Policy Institute, January 19.

Costa, Daniel. 2017a. “H-2B Crabpickers Are So Important to the Maryland Seafood Industry That They Get Paid $3 Less per Hour Than the State or Local Average Wage.” Working Economics (Economic Policy Institute blog), May 26.

Costa, Daniel. 2017b. “Modern-day Braceros: The United States Has 450,000 Guestworkers in Low-Wage Jobs and Doesn’t Need More.” Working Economics (Economic Policy Institute blog), March 31.

Economic Policy Institute (EPI). 2017a. “Employment-to-Population Ratio of Workers Ages 25–54, 2006–2017.” In The State of Working America.

Economic Policy Institute (EPI). 2017b. Nominal Wage Tracker. Updated July 7.

Francis, Laura. 2017. “More Visas for Seasonal Workers in Pipeline, DHS Says.” Bloomberg BNA, July 5.

Garrison, Jessica, Ken Bensinger, and Jeremy Singer-Vine. 2015a. “‘All You Americans Are Fired.’” Buzzfeed News, December 1.

Garrison, Jessica, Ken Bensinger, and Jeremy Singer-Vine. 2015b. “The New American Slavery: Invited to the U.S., Foreign Workers Find a Nightmare.” Buzzfeed News, July 24.

Grabell, Michael, and Justin Elliott. 2017. “Is Trump Administration’s Visa Push a Way to Win Health Care Votes?” ProPublica, June 27.

Hesson, Ted. 2017. “DHS to Add 15,000 H-2B Guest Worker Visas This Year.” Politico Pro, July 17.

Jan, Tracy. 2017. “Trump Officials Open Border to 15,000 More Foreign Workers.” Washington Post, July 17.

New York Times Editorial Board. 2016. “Why Guest Workers Are Easily Exploited.” New York Times, July 1.

Office of Foreign Labor Certification (OFLC). 2017. OFLC Performance Data. U.S. Department of Labor.

Southern Poverty Law Center (SPLC). 2012. “SPLC Wins Record $11.8 Million Judgment for Guestworkers in Suit against Forestry Company.” October 30.

Stack, Liam. 2015. “Indian Guest Workers Awarded $14 Million.” New York Times, February 18.

U.S. Department of Homeland Security and U.S. Department of Labor, Wage and Hour Division (U.S. DHS and U.S. DOL). 2017. Exercise of Time-Limited Authority to Increase the Fiscal Year 2017 Numerical Limitation for the H-2B Temporary Nonagricultural Worker Program (an unpublished Rule by the Wage and Hour Division and the Homeland Security Department, scheduled to be published on July 19). Document number 2017-15208. July 17.

United States Government Accountability Office (U.S. GAO). 2015. H-2A and H-2B Visa Programs: Increased Protections Needed for Foreign Workers. March 6 [reissued on May 30, 2017].

The positive news on layoffs is clear. The layoffs rate—the number of layoffs as a share of total employment—is now down to 1.1, better than the rate over the last entire business cycle, 2000-2007, when its trough was 1.2. That’s great news. Unfortunately, hires and quits remain below full employment levels—hires peaked at 4.0 in 2006 and 4.4 in 2001, while quits peaked at 2.2 in 2006 and 2.6 in 2001. It is particularly troubling that the hires rate has fallen for two months running, from 3.8 in February down to 3.5 in April, but it’s not surprising given the weak Employment Report for April. Unfortunately, the even weaker Employment Report for May suggests that hires may fall even further when the JOLTS report comes out in July. While increasing job openings is a good thing, it needs to translate into hires for workers to see the effects of that increase.

The quits rate fell slightly in April from 2.1 to 2.0. In a stronger economy, workers would feel more confidence to quit their job in search of a better one. For many years now, workers have continually stayed in their job rather than finding what might be a better match.

JOLTS

Some workers are clearly having an easier time of the recovery than others. As the chart below shows, unemployed workers in some sectors are seeing more job openings than others. Job openings exceed the number of unemployed workers in professional and business services, health care and social assistance, finance and insurance, and in wholesale trade. On the other hand, unemployed construction workers continue to face a weak job market. There are 38 unemployed construction workers for every 10 construction job openings. While some employers claim skills shortfalls, increasing wages, offering on-the-job training, and more intensive search techniques would most likely close those gaps.

JOLTS

Figure A

There is still a substantial gap between the number of job openings and the number of unemployed people, illustrating just how far we are from full employment. As shown in the figure below, this gap is far larger today than it would be in a tight labor market.

Figure B

The job-seekers-to-job openings ratio differs by industrial sector. Workers in some sectors (e.g., health care and social assistance, and professional and business services) are faring relatively well, but there is still a large gap between job seekers and job openings in other industries. For example, there are four unemployed construction workers for every one construction job opening.

Figure C

As shown in the figure below, the quits rate remains depressed, meaning that workers still aren’t quitting at the rate they would be if the economy was stronger. The best news in the data released today is that hires were up. Hopefully this is a new trend and not a short term blip in the data. We need job openings to keep translating into hires if we want to reach full employment.

Figure D