Immigration can’t explain declining employment growth

A current challenge for policymakers is interpreting recent softness in the labor market. Relative to 2024, the U.S. labor market produced 156,000 fewer jobs per month from May to August of 2025. Here we present four pieces of evidence suggesting that this slowdown reflects both weakening labor demand as well as a decline in net migration.

1. Around half of the decline in monthly employment payroll growth is attributable to declining net immigration. A range of different computations attribute between 40 and 60 percent of the decline in nonfarm employment growth to slower net migration.
2. The decline in employment growth is broad-based and less pronounced in states and sectors with larger shares of unauthorized employment.¹ The 10 states with nearly 70 percent of unauthorized workers employ more than half of American workers but account for less than half of the decline in employment growth. The six industry sectors that employ almost 80 percent of unauthorized workers also employ more than half of American workers but account for only 36 percent of the decline in employment growth.
3. Median real wage growth has halved, dropping by even more for workers who compete with unauthorized workers. A large decline in net migration might have muted effects on wages as labor demand and supply offset each other. However, median real wages in 2025 are growing at close to half the rate of 2023 and 2024 (1.7 times slower growth). This is more pronounced for low-wage workers (2.5 times slower), who might be expected to benefit the most from less competition from unauthorized workers.
4. Finally, we find that the share of workers who are employed has declined while the share of people who are not in the labor force has increased. If a decline in net immigration was largely behind the decline in the level of employment, then we might expect the shares of workers in employment, nonemployment, and unemployment to remain steady. Instead, there is a clear shift away from employment and toward nonparticipation in the labor force, a sign of a generally discouraging hiring market.

Throughout, we apply the preliminary revisions to benchmark employment estimates released by the U.S. Bureau of Labor Statistics in September 2025.² We take the revision, divide by 12 months, and push this revision through data from April 2024 to the most recent data (August 2025). This approach assumes that the 2025 data require the same revisions. Note that this only differentially affects computation of changes in employment because we do not revise data from January to March 2024. Since estimates of benchmark revisions are not available for the employment statistics by state, we use the published data.

Declining immigration accounts for roughly half of slower job growth

The data that we want to explain is the decline in employment growth between 2024 and 2025. We compare all of 2024 with May–August 2025. This allows changes in immigration policy to come into effect and to account for any lag in unauthorized workers finding jobs.

In 2024, average monthly nonfarm employment growth was 121,000 jobs. In May to August 2025, it was −35,000 jobs, a decline in job growth of 156,000 jobs per month.³ Our aim is to provide an estimate for the share of this 156,000 monthly decline that could be directly attributable to a decline in net migration of unauthorized individuals.

To do this requires three numbers: the decline in net immigration of unauthorized individuals, the share of these missing workers that would be 16 and over, and the share of these missing workers that could be expected to be working. We are interested in computing an upper bound, so throughout we will make assumptions that push toward a higher estimate and then return to consider how this estimate falls as more conservative assumptions are made.

1. We use the U.S. Congressional Budget Office’s (CBO) estimate for the 2024 to 2025 slowdown in net immigration of individuals who are not in lawful permanent resident, student visa, or employment visa categories. This is 2,079,000 individuals,⁴ or 173,000 individuals per month when spread out evenly over 12 months.⁵ (CBO’s estimate of the decline in net immigration is the largest we have found, and hence we use this as an upper bound and return to more conservative estimates below.)
2. These statistics pertain to individuals, not workers. To estimate employment, we need to compute the share of workers 16 and over and the share we may expect to be employed. In 2024, 99 percent of apprehensions by the U.S. Customs and Border Protection were either single adults (64 percent) or family units, which by definition include a child (35 percent). If all family units consist of two adults and a single child, then we obtain an upper bound on the share of individuals that are adults of 79 percent.⁶
3. We need to adjust for the share that we estimate is employed. The employment-to-population rate for individuals from the ages of 16 to 64 without a disability is around 75 percent (roughly midway between the latest reported rates for men and women).⁷ This is an upper bound, assuming zero disability and that all the change in net immigration for individuals 16 and over is for individuals younger than 65.⁸

If we combine these figures, then we obtain an estimate of a decline in 103,000 jobs per month due to changes in net migration (103,000 ≈ 173,000 × 0.79 × 0.75). This is 66 percent of the 156,000 decline in employment growth that we are trying to understand.

When we relax any of our strong assumptions, this share declines.

• If the median family entering is two adults and two children, the share of working-age immigrants declines, and hence the share of employment growth due to declining net immigration declines to 55 percent.
• If we use alternative estimates of the decline in net migration from the Federal Reserve Bank of San Francisco, the share declines to 51 percent.⁹
• If we use the U.S. Bureau of Economic Analysis’ lower estimates for changes in population growth rates rather than CBO’s estimates, the share declines to 29 percent.¹⁰

Taken together, and further supported by the evidence below, these calculations indicate around half of the decline in employment growth is due to slowing net migration.

Decline in employment growth is more pronounced in states and sectors with fewer unauthorized workers

The decline in payroll employment growth is remarkably broad across sectors and states. And we find that the states and sectors most exposed to unauthorized workers experienced disproportionately smaller declines.

Suppose the U.S. economy were divided into two sectors: One accounts for 50 percent of employment but 80 percent of unauthorized employment. If unauthorized employment were responsible for the entire decline in employment and contracted the same amount in both sectors, then we would expect this sector to account for more than 50 percent of the decline in employment growth. However, we find the opposite in the data.

Sectoral analysis

All 13 major sectors of the economy experienced declines in employment growth. We divide sectors into two groups with roughly equal shares of employment, but where one employs more unauthorized labor than the other. The sectors in this group are construction, transportation and warehousing, leisure and hospitality, manufacturing, business services, and retail.

Together, this group employs 78 percent of unauthorized workers, according to the 2023 American Community Survey.¹¹ These sectors account for 52 percent of total nonfarm employment. However, they account for only 36 percent of the decline in employment from 2024 to May–August 2025 (Figure 1).

State analysis

Employment growth declined in 33 states. California, Texas, Florida, Nevada, New Jersey, Illinois, Georgia, Massachusetts, New York, and Washington employ 69 percent of unauthorized workers, according to calculations from the Pew Research Center.¹² These states account for 51 percent of total nonfarm employment, but only 48 percent of the decline from 2024 to May–August 2025.

When we look across sectors and states, we see that the decline in employment growth is broad-based. If anything, states and sectors that account for the largest shares of unauthorized workers account for a disproportionately small share of the decline in employment growth.

Despite less competition from unauthorized labor, low-wage workers’ wage growth is slower

In 2025, real wage growth has slowed substantially. The slowdown is more pronounced for workers at the bottom of the income distribution. These are workers that we may think would most benefit from less competition as immigrant labor is reduced.

Since mid-2022, nominal wage growth for workers exceeded inflation, as workers made up the lost ground in real wages due to the surge in inflation in 2021–2022. During 2023 and 2024, real wages grew around 3.6 percent annually. Growth was higher for workers in the bottom quarter of the wage distribution (3.9 percent) than at the top (3.1 percent). Growth was remarkably stable for all groups over the period (Figure 2).

In 2025, real wage growth nearly halved for all workers (3.6 percent to 2.1 percent), reflecting a moderation in nominal wage growth and a pickup in inflation.¹³ The decline is more pronounced for low-wage workers (3.9 percent to 1.5 percent, a decline of nearly two-thirds) than for high-wage workers (3.1 percent to 2.4 percent, a decline of less than one-third). This differential decline entirely represents differences in nominal wage growth.

These patterns are clear from the two panels of Figure 3. Both show how real wage growth has slowed, and more so for low-wage workers.

We might expect that a decline in unauthorized immigration would put upward pressure on wages of other workers, especially in the lower parts of the income distribution. However, this has not been the case. The decline in growth in real wages points to a significant role of lower labor demand in reducing employment.

Decline in job growth is accompanied by growth in share of people not in the labor force

If there has been a marked decline in employment growth, then why hasn’t the unemployment rate increased?

Consider the explanation that the decline in employment growth entirely reflects a slowdown in net immigration. If this were the case, we might expect the shares of individuals who are employed, unemployed, or not in the labor force to remain constant.¹⁴ However, this is not the case.

In Figure 4 we add together unemployed individuals (U), those not in the labor force (N), and those that are employed (E). We then plot the share (for example, U / (U + E + N)). The trend lines are computed using data from 2022 to 2024.

There is a clear shift away from employment and toward nonparticipation in the labor force. Why is the share of individuals who are out of the labor force expanding?

One answer is that while layoff rates have remained low, unemployed workers have become discouraged by a market with lower real wage growth, a low hiring rate, fewer job openings, and longer unemployment durations. Consistent with this, the rate at which unemployed workers give up looking for a job and hence are classified as not in the labor force has ticked up in 2025.

Conclusion

Net migration has plummeted in 2025 to date, with CBO projecting a decline of nearly 2.1 million individuals this year compared with 2024. At the same time employment growth has slowed. In this note we have shown that the 2025 decline in employment growth is larger than what we would expect from the direct effect of declining net migration.

Further, there has been a significant decline in real wage growth, consistent with a slowing in demand beyond the direct effect of lower net migration. Intuitively, other features of the economy also point to slowing beyond the change in net immigration: Employment declines are balanced across sectors and states that employ many unauthorized workers, and sectors and states that employ relatively few. Meanwhile, we can understand the lack of an increase in unemployment through unemployed workers giving up searching for jobs, and people not in the labor force being less likely to begin looking for work; both are consistent with a broadly slowing labor market.

Taken together, the evidence points toward a slowdown in employment growth that goes beyond the direct effects of a slowdown in net migration.

Endnotes

1 Our calculations of “unauthorized” employment or workers follow the U.S. Congressional Budget Office (CBO) category of “other foreign nationals,” defined by the CBO as “people who entered the United States illegally and who have not obtained a permanent legal status, people who were permitted to enter the country lawfully through the use of parole authority and who may be awaiting proceedings in immigration court, and people who previously resided in the United States legally in a temporary status but who remained in the country after that legal status expired.” State-level estimates come from the Pew Research Center, following their definition of “unauthorized immigrants.”

2 U.S. Bureau of Labor Statistics (BLS), “Current Employment Statistics Preliminary Benchmark (National) Summary,” September 2025.

3 Recall that we are putting the benchmark revisions through the data. Absent these revisions, the decline would be slightly smaller: 168,000 (2024) − 27,000 (May–Aug. 2025) = 141,000 jobs per month.

4 Specifically, we are looking at the CBO category of “other foreign nationals.” See CBO report An Update to the Demographic Outlook, 2025 to 2055; total decline calculated using data underlying figures 5 and 6.

5 A note on front-loading: In our computation, we divide the 2025 slowdown in net immigration of unauthorized individuals evenly across the 12 months of 2025. Why not front-load the immigration decline into the first half of 2025? This is because CBO projections for the decline in net immigration are almost entirely due to reduced entry. Hence, the counterfactual we are interested in understanding is what employment growth would have been like if the flow throughout 2024 had continued into 2025. Front-loading would only be appropriate if the reduction in net migration was entirely obtained from removing individuals from the U.S., but so far this is not the case.

6 U.S. Customs and Border Patrol, Nationwide Encounters, accessed October 24, 2025.

7 BLS, “Table A-6. Employment status of the civilian population by sex, age, and disability status, not seasonally adjusted,” August 2025.

8 A reasonable assumption; see BLS, “Labor Force Characteristics of Foreign-Born Workers Summary,” May 20, 2025.

9 Duzhak and Schmidt, “Updated Estimates of Net International Migration,” Federal Reserve Bank of San Francisco, July 2025. They estimate net international migration to the U.S will be around 1 million in 2025, down 1.6 million from 2024 and 2.5 million less than 2023.

10 Cheremukhin, “Break-even employment declined after immigration changes,” Federal Reserve Bank of Dallas, October 2025.

11 Lisiecki, “The Role of Undocumented Workers in High-Growth Occupations and Industries Across the United States,” Center for Migration Studies, August 2025.

12 Passel and Krogstad, U.S. Unauthorized Immigrant Population Reached a Record 14 Million in 2023, Pew Research Center, August 2025.

13 Overall, slightly more than half (54 percent) of the decline in nominal wage growth is due to higher inflation, and slightly less than half (46 percent) is due to slowing real wage growth.

14 We deliberately focus on people “not in the labor force” (N) versus the broader category of “nonemployed” (unemployed + not in the labor force).