The Recent Surge in Immigration and Its Impact on Measured Productivity Growth

Labor productivity—real nonfarm business output divided by total hours worked—is a critical measure of economic health, and since 2023, it has been increasing at a brisk pace. Some observers, however, worry that mismeasurement in the denominator, particularly as a result of an undercount in recent immigration, might paint a rosier picture than reality.

In a recent blog post, I explored how the Census Bureau’s Current Population Survey (CPS), a key source for understanding U.S. labor market trends, significantly undercounted recent immigrants compared with estimates from the Congressional Budget Office (CBO). Between January 2022 and October 2024, the CPS reported a net increase of 3.94 million immigrants, while the CBO, using additional data sources, estimated that the true number was more than double, totaling 8.65 million during that same period.The net change reflects new immigrants entering the U.S. less immigrant deaths and immigrants leaving the country. For the CBO estimate, I prorated the CBO projection for full-year 2024, which can be found in “The Demographic Outlook: 2024 to 2054,” January 2024. Despite this undercount, the blog post showed that the impact of this discrepancy on the aggregate unemployment rate is very modest.

As my earlier post explained, the CPS definition of a recent immigrant depends on when the survey was conducted. In even survey years (as in 2024), recent immigrants are defined as those who reported immigrating in that year or the previous two calendar years. In odd survey years, recent immigrants are defined as those who reported immigrating in that survey year or the previous three calendar years.

Why is it important to identify recent immigrants? Recent immigrants tend to have higher unemployment rates than U.S. natives and earlier immigrants. Yet they constitute a small share of the labor force (1.9% in the CPS data for October 2024). Assuming the “missing” recent immigrants per the CBO’s estimates had similar labor market outcomes as those recorded in the CPS, the share of recent immigrants in the labor force would instead be 3.7%, as my earlier blog pointed out. Still, the aggregate unemployment rate would be only a modest 0.10 percentage points higher—4.24% instead of 4.14%.

Productivity Measurement: Why Undercounting Might Matter

While the undercount of recent immigrants has minimal effects on the unemployment rate, its implications for productivity measurement are potentially more significant. Productivity growth depends critically on the accurate measurement of total hours worked—the denominator in the calculation. If growth in total hours is understated because of missing recent immigrants, productivity growth will appear artificially high. It is important to note that this blog post assumes there is no mismeasurement of the numerator—namely output.

To get a sense of how an undercount of recent immigrants affects estimates of hours worked, I construct an alternative measure of total hours, assuming (1) a net increase of 8.65 million recent immigrants (as per the CBO) from January 2022 to October 2024 rather than 3.94 million (as per the CPS) and (2) that the “missing” recent immigrants had the same employment rates and average hours worked as those recorded in the CPS.As of October 2024, the CPS recorded 4.91 million people who had immigrated to the U.S. since January 2022 and were still residing in the country; this represents the total number of recent immigrants living in the country in that specific month. In this exercise, I took the difference between the CPS’ net change in immigrants (both recent and older) and the CBO’s net change in immigrants and then added it to 4.91 million. This provides an approximate upper bound for the number of recent immigrants.

The figure below compares the annualized quarterly growth rate in total hours between this alternative measure and the original CPS estimate. For example, total measured hours in the CPS decreased by 0.17% in the second quarter of 2024 but would have increased by 1.18% under this adjusted scenario. In the third quarter of 2024, the corresponding numbers were 0.79% and 1.96%, respectively.

Annualized Growth Rate in Total Hours Worked

SOURCES: IPUMS, Congressional Budget Office and author’s calculations.

NOTE: In the 2024 survey, “recent immigrants” are defined as those who reported immigrating in that year or the previous two calendar years.

BLS and CPS: Comparing Estimates

Because labor productivity growth can be approximated as the growth in output minus the growth in hours worked, the inclusion of the missing recent immigrants would reduce labor productivity growth by more than 1 percentage point in the second and third quarters of 2024, if adjusted CPS hours were used to calculate productivity. More broadly, the difference between the adjusted and unadjusted hours in the figure above serves as an indicator of the potential for upward bias in productivity growth caused by the undercounting of recent immigrants.

However, the Bureau of Labor Statistics (BLS), which calculates labor productivity, uses a different source for hourly data. The BLS primarily uses data from the Current Employment Statistics (CES) survey to calculate total hours worked, with the CPS providing supplemental data for unpaid family workers and self-employed proprietors. The two surveys are distinct: The CES is conducted monthly as an establishment survey of about 119,000 businesses and government agencies, while the CPS is conducted monthly as a sample survey of about 60,000 American households. And unlike the CPS, the CES does not collect information on immigration status, making it unclear whether the BLS total hours estimates are affected by similar undercounts.

To investigate this, the next figure compares total hours worked based on three sources: CPS data, adjusted CPS data (accounting for the missing recent immigrants) and BLS data—all normalized to 100 in the first quarter of 2022. While the evolution of CPS and BLS estimates tended to closely align before 2022, a gap emerged afterward. Notably, adjusting the CPS data for the missing immigrants eliminated this gap from 2023 onward, suggesting that the CES likely records recent immigrants’ contributions accurately.

Total Hours Worked by Different Data Sources

SOURCES: IPUMS, Congressional Budget Office, Haver Analytics and author’s calculations.

NOTES: In surveys conducted in even years by the CPS, recent immigrants are defined as those who reported immigrating in that year or the previous two calendar years; in odd survey years, recent immigrants are defined as those who reported immigrating in the survey year or the previous three calendar years. The figure’s data are available for download.

Conclusion

The undercounting of recent immigrants in the CPS raises important questions about the reliability of key economic indicators like labor productivity. If productivity growth is overstated due to missing growth in hours worked, policymakers and the public might overestimate the strength of the economy. However, the findings in this analysis suggest that the total hours worked used to measure labor productivity growth are likely not understated, given the alignment between the adjusted CPS and BLS data.

As immigration continues to shape the labor market, ensuring accurate and comprehensive data collection is essential for guiding policy decisions and understanding the true state of the economy.

Notes

1. The net change reflects new immigrants entering the U.S. less immigrant deaths and immigrants leaving the country. For the CBO estimate, I prorated the CBO projection for full-year 2024, which can be found in “The Demographic Outlook: 2024 to 2054,” January 2024.
2. As of October 2024, the CPS recorded 4.91 million people who had immigrated to the U.S. since January 2022 and were still residing in the country; this represents the total number of recent immigrants living in the country in that specific month. In this exercise, I took the difference between the CPS’ net change in immigrants (both recent and older) and the CBO’s net change in immigrants and then added it to 4.91 million. This provides an approximate upper bound for the number of recent immigrants.