The Phantom Softness in the Labor Market
The Dallas Fed published an extraordinary update to its break-even employment framework on Tuesday, and the implications are even bigger than the authors realize.
The new analysis—by economists Anton Cheremukhin, Daniel Wilson, and Xiaoqing Zhou—extends their earlier work, which used immigration court microdata obtained through Freedom of Information Act requests. The data now run through December 2025, and they paint a picture that undercuts the idea that the labor market has been softening.
The break-even rate of employment growth—the number of net new jobs needed each month to keep the unemployment rate constant—has collapsed to approximately zero, the Dallas Fed economists find. It peaked at about 250,000 jobs per month in 2023, when the Biden border crisis was flooding the labor market with new workers. By July 2025, it had fallen to around 10,000. From August through December 2025, it averaged negative 3,000—meaning the economy could actually lose a few thousand jobs a month without the unemployment rate budging.
The primary driver is the sharp reversal of unauthorized immigration. Net unauthorized immigration averaged negative 55,000 per month in the second half of 2025, the Dallas Fedsters write. For the full year, it totaled negative 548,000—roughly 50 percent larger in magnitude than the Congressional Budget Office had projected (-365,000).
The Department of Homeland Security puts the total number of illegal aliens who have left the country even higher—nearly three million since January 2025, including over 700,000 deportations and an estimated 2.2 million self-deportations. Even cautious estimates combining the Dallas Fed’s unauthorized outflows with legal departures and visa expirations exceed one million.
The Participation Rate May Be Haunted by a Ghost Population
This has implications far beyond the break-even calculation.
Consider the labor force participation rate, which has declined from 62.5 percent a year ago to 62.0 percent in February. That drop has been cited by bears (people who think the economy is deteriorating and stocks are likely headed lower) and doves (people who think monetary policy is too tight) alike as evidence that the labor market is softening—that Americans are dropping out, demand is flagging, and the Fed needs to cut rates.
But the participation rate is simply the labor force divided by the civilian noninstitutional population aged 16 and over. That denominator is based on Census Bureau population estimates that are updated annually via the Vintage series and incorporated into BLS population controls. While the January 2026 adjustment reflects the Census Bureau’s Vintage 2025 estimates—which include monthly projections through December 2025 and note a decline in net international migration—the annual process can still lag or imperfectly capture the full scale and composition of outflows in the second half of 2025, particularly among working-age unauthorized immigrants with high labor force attachment.
Here’s where the arithmetic gets interesting. The people leaving are overwhelmingly working-age adults who came here to work, and most were working. Assume a conservative 75 percent participation rate for this group. If large numbers have departed but the population denominator hasn’t fully adjusted downward for the change in composition, the measured participation rate is being artificially dragged down by a phantom population—people the BLS is still counting who are no longer in the country. Even using conservative unauthorized outflow figures combined with other departures, a sizable chunk of the observed decline in participation could be statistical noise rather than genuine weakness. The entire reported drop over the past year may be exaggerated, if not largely illusory.
Let’s do the numbers, as they say on the radio. At the administration’s estimate of three million departures with our estimate of 75 percent participation, the true labor force participation rate would be approximately 61.9 percent—but the measured rate, distorted by the stale denominator, would read 61.2 percent. That is a measurement distortion of nearly 0.7 percentage points. In other words, the entire decline in the participation rate over the past year would be a statistical ghost.
Even using a lowball figure of one million total departures—well below both the DHS estimate and what you would get by combining the Dallas Fed’s unauthorized outflow data with legal departures—the distortion is still around 0.2 percentage points, which would account for roughly half the observed decline.
The Real Unemployment Rate May Be (a Bit) Lower
This implies that the labor market is tighter than the headline numbers suggest. It means the participation rate has not fallen as much as it appears—and may not have fallen meaningfully at all once adjusted. The break-even employment threshold has collapsed to near zero, meaning payroll prints that would have signaled weakness two years ago now signal stability or even modest tightening.
The February jobs report, which showed a decline of 92,000 payroll jobs and the unemployment rate ticking up to 4.4 percent, looks alarming only if you are still using the old 150,000-to-200,000 benchmark from the open-border era. It is worth remembering that even the CBO—the official scorekeeper—underestimated unauthorized outflows by 50 percent. The labor market models built on those assumptions are misfiring for a reason: the structural shift has been bigger and faster than the forecasters anticipated.
The distortion may also be nudging the unemployment rate higher than it should be. The Bureau of Labor Statistics uses population controls to weight its household survey responses by demographic cell—age, sex, race, ethnicity—and if those controls overcount prime-age workers who have left the country, the weighting process distorts the resulting estimates. The unemployment rate has risen only from 4.2 percent to 4.4 percent over the past year—a modest move that may be even more modest than it appears, once you account for a denominator that is overcounting workers who are no longer here.
While this might seem to flatter the Trump economy, correcting the impression that the labor market has softened, it also complicates the case for aggressive rate cuts. If the true labor supply is lower and the effective participation rate higher than the distorted data show, the disinflationary slack that the doves are counting on may not exist. The Fed would be easing into a snugger labor market than it appears.
The Cheremukhin anlaysis is careful not to overstate its conclusions, noting that “payroll gains that might historically have signaled economic slack are now consistent with a balanced labor market.” That’s the tweedy Fed economist version. The blunt version: the people telling you the labor market is cracking probably failed to update their models for the single biggest structural change in the U.S. labor force in a generation—the reversal of the Biden-era immigration surge.
Read the full article here