One-two punch in the labor market

Stephen Stanley | July 18, 2025

This material is a Marketing Communication and does not constitute Independent Investment Research.

The labor market data weakened substantially last summer, sparking a near-panic from Fed officials that ultimately led to a 50 bp rate cut in September. That step, in retrospect, is widely regarded as an overreaction. So far, there is less evidence of another round of temporary weakness this summer, though some special factors may have flattered the May and June data.  There are two key forces likely to weigh on the monthly numbers this summer: a difficult job market for college graduates and students as well as heightened immigration enforcement, which may be cutting into labor supply.

Last year

Last year, the monthly labor data clearly underperformed during the summer.  Through May 2024, payroll gains had been 118,000 or higher for 14 straight months. But the June, July, and August increases were all below 100,000, leading to concerns that momentum was dissipating.  Even more notably, the unemployment rate backed up noticeably in May, June, and July, rising by a cumulative 36 bp over the three months from 3.86% to 4.22%, and holding at 4.2% in August.

These readings led to considerable worry regarding the health of the economy.  Yields on the 10-year Treasury slid from as high as 4.60% in late May to 3.70% by early September.  Similarly, Fed officials grew increasingly concerned about the health of the labor market and, in turn, the economy, which ultimately led to their aggressive 50 bp cut in September.

At the time last year, I laid out the case that at least a significant portion of the apparent weakness in the summer numbers reflected a temporary factor.  The job market for new college graduates and college students looking for summer work was unusually soft last year.  As a result, in May 2024, the level of employment for 20-to-24 year olds, as recorded in the household survey, sank by nearly half a million, driving the unemployment rate for this age group up in a single month from 6.7% to 7.9%.  Over the course of the summer, the jobless rate for this age segment stayed elevated through August (ranging from 7.5% in June to 7.9% in August).

I made the argument at that time that Fed officials and many other economists were misreading the aggregate numbers as evidence of broad underlying weakness, and I predicted that the labor market data would strengthen again beginning in September. Indeed, the September employment report, the next one released immediately after the September FOMC move, showed a whopping 254,000 rise in payroll employment (since revised to 240,000) and a drop in the unemployment rate of a full tenth.

Trending toward a repeat?

The monthly employment data have not been as weak in May and June as was the case last year, as least on their face.  Recall that last year, the payroll weakness began in June.  This year, the headline job gain for June was 147,000, in line with recent trends.  However, the June reading was flattered by an outsized seasonally-adjusted jump in government employment, which I suspect reflected unusually low end-of-school-year layoffs, probably more a timing issue than a fundamental development.  Private sector payrolls increased in June by only 74,000, the weakest reading since a hurricane-impacted October and only a little better than the average gain of 46,000 recorded in June, July, and August 2024.

As was the case a year ago, the unemployment rate increased noticeably in May.  The rounded reading held at 4.2%, but the unrounded figure rose by 6 bp to 4.244%.  However, unlike last year, the unemployment rate fell sharply in June – by 13 bp.

There are three factors to consider regarding the performance of the labor market this summer, two of which are new from last year and reflect government policy choices.

College-age job market

The one force that is a carryover from last year is the difficult job market for college graduates and college students looking for summer jobs.  My insight on this issue benefits from having a couple of kids in the age group.  As my daughter reminded me a few weeks back, it was two summers ago that she first brought this theme to my attention.  At the dinner table each night, she would frequently talk about how many of her friends were having difficulty finding jobs after graduating.  Being the supposed expert on the labor market, I expressed skepticism, noting that the national unemployment rate at the time was about 3.5%, close to a 50-year low.  So, how bad could it be for college grads?

In reality, her friends’ experience reflected a broader trend.  Once the post-pandemic labor shortage was resolved, openings for entry-level professional positions dried up.  The labor market overall was still robust a year ago, but the market for young adults, if anything, got even tougher.

The household survey data a year ago bears that out.  In May 2024, employment for 20-to-24 year olds fell by a whopping 466,000, or more than 3%.  To put that into perspective, if overall household employment had fallen by the same percentage as the drop for 20-to-24 year olds, the aggregate would have plunged by over 5 million in a single month.

Of that 466,000 collapse, 312,000 left the labor force, meaning that they were not actively looking during the month, while the other 154,000 were unemployed.  So, the labor force participation rate fell by a whopping 1.5 percentage points in May 2024, and the jobless rate climbed by 1.2 percentage points.  The unemployment rate for this group briefly recovered in September, when all the summer job seekers would have gone back to school and thus dropped out of the labor force, but then quickly reverted back to the higher level first reached in May.

This year, the job market for this age group has been soft again.  The Wall Street Journal ran a story on the topic on June 16.  The piece included several eye-popping statistics.  The 12-month average for the unemployment rate for 20-to-24 year olds with a bachelor’s degree rose to 6.6%, up by a full percentage point from two years earlier.  A New York Fed study found that the unemployment rate for college graduates ages 22-to-27 averaged 5.8% in the first three months of this year, the widest gap between this group and the general population in at least 35 years. I can once again corroborate anecdotally that finding a job straight out of college was tough for the Class of 2025.

The May 2025 household survey data points to discouragement. The fall in seasonally-adjusted employment fell by less than a year ago, “only” by 163,000, still equivalent on a percentage basis to around 2 million if spread over the aggregate employment level, but the labor force dropped by 178,000.  That is, more of the recent grads who were struggling just gave up and were not actively looking for work.  Thus, the unemployment rate for this age group held steady in May (and in June) at 8.2%, but the labor force participation rate slid by 0.8 percentage points in May and another three tenths in June.

Unlike in 2024, the 20-to-24 age group did not influence the aggregate unemployment rate in May and June.  It did, however, account for about one third of the decline in the labor force and in household employment over the past two months.

What are the implications for the overall economy?  Presuming that the market for recent college grads fails to magically resurrect in the coming months, joblessness for this group seems likely to remain elevated and labor force participation depressed.  However, the seasonal adjustments incorporate the influx of new grads into the workforce over the course of the summer.  Once we get to the fall, that seasonal force likely dissipates and the month-to-month changes in the key labor market metrics should no longer be heavily influenced by this age group, which may appear, as it did last year, as a seasonally-adjusted firming in job growth.

Wild cards for 2025

There are two additional drivers in the labor market at the moment that did not exist a year ago, both related to government policy.  First, businesses are reporting that they are slowing their new hiring amid the elevated degree of policy-related uncertainty, especially related to tariffs.  This is likely also weighing on job growth at the moment and may continue to do so throughout the summer.  The good news here is that businesses remain broadly optimistic about the economic outlook and are likely to accelerate their hiring activity once they have a clear picture of the end result for tariffs.

The second wild card is a ratcheting up in the aggressiveness of immigration enforcement.  In the first several months of the new administration, border crossings into the country fell sharply, which undoubtedly contributed to a slowing in the pace of net hiring, as labor supply has expanded at a slower pace.

The household survey is likely not reflecting this development accurately, given how the survey is put together.  Assumptions by the Bureau of Labor Statistics about population growth are only revised once a year, with the release of the January data each year.  The actual survey mechanics involve asking a sample of people two questions: 1) are you working? and 2) if not, are you looking?

The hard outputs of the survey are all ratios, the labor force participation rate, the employment-to-population ratio, and the unemployment rate.  Those ratios are then applied to a population estimate to derive levels for the labor force and employment.  The point is that the changes in household employment and the labor force will not be impacted by a general drop in population growth until the BLS incorporates a new population estimate that reflects the slowdown in net immigration.  Until early next year, the BLS population figures assume about a 0.9% annual rate of the working-age population, roughly the same as in 2022, 2023, and 2024.

A new development more recently could have a more concrete impact on the household survey data. The more aggressive enforcement, including more frequent workplace raids, by ICE may compel at-risk workers to stay home.  This dynamic may have contributed to the softness in the labor force (down a combined 735,000) and household employment (down a combined 603,000) in May and June.

The detailed household survey data offer some interesting nuance.  The BLS breaks the household survey results down to isolate a variety of age, sex, and ethnicity groups.  One table in the monthly report details the statistics for the “Hispanic or Latino population,” a group that has reportedly garnered disproportionate attention from immigration authorities.

The data show that the labor force participation for Hispanics has indeed fallen over the past two months, from 67.3% to 66.8%.  However, the employment level for the group has actually risen slightly over the past two months, and the employment-to-population ratio has only fallen by two tenths of a percentage point, less than the aggregate slide (three tenths).  The implication of these two numbers is that unemployment is down sharply for Hispanics, from 1.80 million in April to 1.63 million in June, yielding a plunge in the unemployment rate from 5.2% in April to 4.8% in June.

I should insert the standard caveat here. Household survey data tend to exhibit wild month-to-month swings that often reflect statistical noise rather than fundamental shifts.  This would be even more of an issue for subsets of the data (such as the statistics for Hispanics).  I would never recommend taking one month’s household survey data as evidence of anything.

Having said that, the literal interpretation of the data on Hispanics is interesting.  The numbers suggest that immigration enforcement is not having much of an impact on those who are already working, but it may be dissuading those who would ordinarily be looking for a job from actively searching.

There is another detail of these data that runs somewhat at odds with the prevailing narrative.  Most of the press reports regarding stiffer immigration enforcement have focused on ICE agents going to construction sites, Home Depot parking lots and so on, targeting mostly young males.  However, the household survey data show that for men over 20 years old, the labor force participation rate inched down by only two tenths of a percentage point from April to June and the employment-to-population ratio rose.  In fact, the weakening amongst Hispanics in the past two months came mainly among women 20 years and over.  For that cohort, the labor force participation rate plunged by over a full percentage point and the employment-to-population ratio sank by a full percentage point.  The story based on the data, to the extent they can be taken literally, points to a little different situation that what one might have expected from the press coverage.

We can strip out the numbers for Hispanics and calculate the labor market statistics for the non-Hispanic population.  Over the past two months, for that group, the labor force participation rate has fallen by 33 bp, about the same as the aggregate slide of 36 bp.  Similarly, the employment-to-population ratio fell by 33 bp, not far from the 31 bp drop in the aggregate figure.  The key impact is on the unemployment rate.  The overall unemployment rate is down by 7 bp on balance over the past two months, but the non-Hispanic jobless figure is up by 3 bp.

The data suggest that this immigration story is not having a major impact, at least not yet, on the household survey figures. It could be depressing payroll job gains, but we can only guess about that, as those data are not broken down by age, race and so on.  If it is having any impact at all, it is actually lowering the unemployment rate.  Moreover, the detailed household survey data on Hispanics does not seem to be precisely consistent with the prevailing narrative of how immigration enforcement is impacting the job market.

In any case, workplace immigration enforcement is proving controversial, as it pits two elements of President Trump’s political base, businesses in some industries who rely heavily on immigrant labor and those favoring very strict immigration policy, against each other.  As with tariff-related uncertainty, it is unclear how long this factor may persist, but it would not be surprising to see more clarity at some point soon on which immigrants may be at risk of deportation, which would presumably lessen the potential impact on the broad labor market.