Seeing economic data through the fog of immigration estimates

The US immigration rate in 2025 is unknown and will continue to be for a while. Policymakers and economic forecasters need to accept significant uncertainty about the rate and the resulting pace of overall population growth. They need to be skeptical about immigration estimates, and they should be explicit about the population-growth assumptions they incorporate into their interpretations and projections of economic data.

While there has surely been a dramatic decline in the immigration rate since 2022, 2023, and the first half of 2024, there are no solid data yet on whether net immigration has slowed, stopped, turned slightly negative, or turned sharply negative. The recent Current Population Survey (CPS) data showing that there were 2.2 million fewer immigrants in the US in July 2025 than in January are implausibly large and uncorroborated, and imply absurd gyrations in the labor market. At the same time, Census Bureau population estimates for 2025, which were produced in late 2024 and are the basis for population figures used by the Bureau of Labor Statistics (BLS) and the Bureau of Economic Analysis (BEA), almost certainly assume an unrealistically high rate of immigration this year.

This blog post extends previous work to show that survey response rates could account for the big swings that CPS reported in the foreign-born and native-born population—confirming that immigration estimates are highly uncertain. This post also shows how uncertainty about immigration leads to wildly varying estimates of the breakeven rate of monthly US nonfarm payroll growth needed to keep up with labor force growth.

Declining response rates could account for millions fewer reported immigrants

The CPS—or the “household survey,” conducted monthly by the Census Bureau—has seen its response rates decline for years, from 90.3 percent in January 2013 to 67.1 percent in July 2025. The decline has been fairly steady, aside from a huge temporary drop during the pandemic.

It is plausible that response rates have fallen more among immigrants than for the native-born in 2025. Immigrants might be increasingly wary of responding to government surveys as the Trump administration is trying to use IRS taxpayer data and insurance claim data for immigration enforcement. Some immigration researchers have pushed back against response rates as an explanation for the reported drop in the foreign-born population. But the decline in response rates, along with some reasonable assumptions, could account for the drop in the foreign-born population reported by the CPS.

CPS response rates have declined faster in 2025. Response rates fell from 73.3 percent in January 2022 to 68.8 percent in January 2025, or 1.5 percentage points per year. In 2025, response rates fell from 68.8 percent in January to 67.1 percent in July, an annualized pace of 3.4 percentage points. There is no direct evidence of changes in response rates for native-born versus foreign-born US residents because citizenship and nativity (i.e. birthplace) are unknown for survey non-respondents. So, some assumptions are necessary.

Let’s assume that (1) the response rates for the native-born and foreign-born populations were the same until January 2025; (2) the native-born response rate then continued to decline in 2025 at the same rate as the prior three years, from 68.8 percent in January to 68.1 percent in July; (3) and the foreign-born response rate accounted for the accelerated decline in 2025. For the overall response rate to be 67.1 percent in July 2025 as reported, the foreign-born response rate would have had to have fallen to 62.9 percent. Since the foreign-born account for only 18 percent of the adult population, even a small drop in the overall response rate could imply a meaningful drop in the foreign-born response rate.

Those assumed declines in response rates, combined with the population controls that require the total US population in July 2025 to equal the predetermined number estimated by the Census in late 2024, would imply a foreign-born adult population of 47.5 million in July 2025 and a native-born adult population of 226.3 million. Compared with what the January 2025 CPS reported, that’s a 3.0 million drop in the foreign-born population and a 4.1 million jump in the native-born population. That’s an even more dramatic swing than what the CPS actually reported, which was a 1.9 million drop in foreign-born adults and a 3.0 million jump in native-born adults between January and July 2025. The 2025 decline in CPS response rates, combined with reasonable assumptions about the relative decline in response rates for the foreign-born, could therefore explain the entire reported decline in the foreign-born population.

The point is not to take this simple calculation and conclude that the foreign-born population has not declined (or has even increased). Response rates are volatile month-to-month; my assumption about the relative changes to native-born and foreign-born response rates is reasonable but unverifiable and possibly wrong; and CPS is a small sample that Census warns against using to estimate the size of the foreign-born and native-born populations. The point is that declines in survey response rates could plausibly explain much or all of the decline in the reported CPS decline in the foreign-born population, reinforcing the warning that CPS data on native-born and foreign-born counts should be viewed skeptically.

If the immigration rate is uncertain, so is the breakeven rate

As the rate of immigration slows or turns negative, the overall population and labor force grow more slowly—so fewer new jobs are needed to keep the labor market steady. In a recent Policy Brief, I estimated the breakeven rate at 86,000 jobs in June 2025: This is the level of monthly nonfarm payroll growth that keeps the percentage of employed adults constant, after accounting for aging. That breakeven rate assumed population growth in line with Census population estimates for 2025, which were produced in late 2024 and almost certainly assume too high an immigration rate for 2025.

Here, I estimate the monthly breakeven rate for the rest of 2025 under three different scenarios about net immigration:

1. Net immigration at the level implied by the Census population estimates, which are the official population controls used in the CPS. Census estimated that the adult population would grow by 1.2 million between June and December 2025. Previously I estimated that “natural increase”—that is, births minus deaths—would boost the adult population by 686,000 over six months, implying net adult immigration of 545,000 over six months. That rate of immigration is lower than what Census estimated in 2022, 2023, and 2024, but it is the highest of these three scenarios.
2. Net immigration of zero, such that new arrivals equal the sum of voluntary departures and involuntary deportations. This is within the range of net immigration recently estimated by immigration experts at the American Enterprise Institute and the Brookings Institution.
3. Negative net immigration—that is, more foreign-born people leave the US than enter—of 1.9 million adults over six months, equal to what CPS reports for the six months from January to July.

More immigration means a higher breakeven rate. I estimate the breakeven rate under the original Census estimates to be 88,000 jobs monthly for the rest of 2025. In the second scenario, where net immigration is zero, the breakeven rate falls to 27,000 jobs monthly. In the third scenario, where net immigration is sharply negative, the breakeven rate is a loss of 184,000 jobs monthly.

It's hard to make sense of the jobs report with so much uncertainty about the breakeven rate. Over the past three months, nonfarm payrolls increased 19,000 in May, 14,000 in June, and 73,000 in July. If the Census population estimates (scenario one) are correct, then job growth is below the breakeven rate, the labor market is weakening, and the Federal Reserve should be more inclined to cut interest rates. But if the reported CPS decline in immigration (scenario three) is correct, then job growth is far above the breakeven rate, the labor market is tightening, and the Fed should be less inclined to cut interest rates and should watch for overheating. In all likelihood, policymakers will focus less on the nonfarm payroll numbers and instead rely more on economic indicators whose interpretation does not depend on assumptions about immigration or population trends, like the unemployment rate and the employment-population ratio.

Notes

1. An indirect way to compare response rates for foreign-born and native-born populations is to estimate the correlation across geographic areas between survey response rates and immigrant share of the local population.

2. Mathematically, the overall response rate is the weighted average of the native-born and foreign-born response rates, weighted by each group’s share of the total population.

3. These population controls mean that the reported native-born population will be adjusted to offset reported changes in the foreign-born population, through the CPS sampling and weighting procedures. See this post for extensive details. See accompanying spreadsheet for calculations.