OPINION: A study finds zero American jobs saved by ICE raids

Introduction

Zero. That is the number of American jobs a new academic study finds were protected by the mass immigration enforcement raids conducted under the second Trump administration. The National Bureau of Economic Research paper, authored by Elizabeth Cox and Chloe N. East, published in May 2026 and resurfaced by Raw Story on July 29, 2026, finds no positive employment effect for U.S.-born workers from ICE raids. A promise sold as protection, measured instead as damage.

The core message behind the raids has always been simple: remove undocumented workers, and the jobs they held will go to American citizens. This NBER working paper, referenced as number 35129, tests that premise directly against labor data, and finds the opposite of what the message promises. For every 1,000 immigrant men arrested, employment among immigrant men without a college degree falls by 2.5%, and employment among U.S.-born men also falls, by 0.54%, in areas with heavy ICE activity.

This is an opinion column, not a neutral report. It states plainly: a paper is not a verdict, and a working paper that has not necessarily undergone traditional peer review deserves scrutiny, not blind acceptance. It was also criticized as flawed by the Washington Times, a conservative outlet, on May 10, 2026. But the burden of proof now sits with the side that promised jobs, and so far, no counter-study refuting these specific figures has been identified in the sources consulted.

What the study actually measures

A precise number, not a vague impression

The NBER paper does not deal in vague impressions of an "economy that's improving" or "getting worse." It measures a specific ratio: for every 1,000 immigrant men arrested in a local labor market, employment among immigrant men without a college degree drops 2.5%, and employment among U.S.-born men drops 0.54%, specifically in areas experiencing heavy ICE activity. Both groups lose. Neither gains.

No source consulted specifies the exact geographic scope of "high-ICE-activity areas," nor the total number of workers covered by the underlying dataset. This is a real limitation of the study as relayed, and it should be stated rather than smoothed over. A number without full geography is still a number. It should just be handled carefully.

Two authors, one shared institution

Elizabeth Cox and Chloe N. East authored the paper under the NBER's institutional umbrella, a nonpartisan economic research organization widely cited across the political spectrum for decades. Their paper was published in May 2026 and only gained broader public attention after Raw Story's July 29, 2026 report revived it.

No detail on the authors' other published work, their institutional affiliation beyond NBER, or their funding sources appears in the sources consulted for this column. That absence of detail is worth flagging honestly rather than filling in with assumption. What is documented stays documented. What is missing stays missing.

The criticism that exists — and its limits

"Flawed," according to the Washington Times

On May 10, 2026, the Washington Times, a conservative-leaning outlet, published a piece describing the NBER's research as flawed. The available sources do not detail the specific methodological objections raised in that critique — no breakdown of disputed variables, no counter-calculation, no alternative dataset is described in the material consulted for this column.

A general critique without a documented methodological detail carries less weight than a critique that shows its work. Calling a number flawed is easy. Showing exactly where it breaks is harder, and it hasn't been shown here.

What "working paper" actually means

An NBER working paper is not automatically equivalent to a study published in a peer-reviewed academic journal after formal review by other specialists in the field. That distinction matters, and this column states it clearly rather than letting the NBER's institutional prestige stand in for a review process that may not yet be complete for this specific paper.

Working paper status is a real limitation. It is not, however, a reason to discard the numbers outright, absent a documented alternative calculation that contradicts them. The paper stands until someone shows the math wrong.

The gap between message and number

A rhetoric built entirely on scarcity

The political argument for large-scale immigration enforcement raids has rested, publicly, on a story of job scarcity: remove workers without legal status, and jobs open up for citizens waiting for them. That story requires a labor market where immigrant and native-born workers are direct substitutes competing for the same identical positions.

The NBER paper's finding — a negative effect on both groups simultaneously — undercuts that substitution model directly. If removing immigrant workers actually freed up jobs for American-born workers, the data would show U.S.-born employment rising in high-ICE areas, not falling. The theory promised addition. The data shows subtraction on both sides of the ledger.

An economic argument that owes an explanation

When a signature policy's central economic justification fails an empirical test, the burden shifts. It becomes the responsibility of those who made the promise to explain the gap, not the responsibility of critics to prove a negative. No official response addressing this specific NBER paper by name has been identified in the sources consulted for this column.

Silence, in this specific case, is itself a data point worth naming. A promise made loudly deserves a defense made just as loudly when challenged.

Immigration raids inside a broader legal storm

Two circuits, same week, same administration

This NBER paper does not exist in isolation. On July 30, 2026, the 9th Circuit and 7th Circuit federal appeals courts each ruled against the administration's policy of indefinite detention, without a bond hearing, for immigrants who entered the country illegally years earlier, according to the New York Times. At least six circuits have now ruled against that detention policy, against two in its favor, with three petitions pending before the Supreme Court.

No source consulted establishes a proven causal link between these judicial rulings and the NBER employment findings. They fall within the same window, on the same broad policy area, which is worth naming as context, not as connected proof. Two separate fronts, one shared subject: the actual cost of enforcement, measured against its promised benefit.

Courts blocking related enforcement tools

A Reuters report dated July 21, 2026 documents a federal judge blocking the administration from stripping work permits from certain immigrants. A separate Reuters report from July 16, 2026 covers a related ruling permitting the administration to re-revoke the status of migrants who had used a Biden-era humanitarian parole program. Multiple courts, multiple rulings, one recurring theme: enforcement tools facing sustained legal resistance.

None of these separate rulings directly addresses the NBER employment study. They document, together, a period where the administration's immigration enforcement approach is being contested on several fronts simultaneously, legal and empirical alike.

What this means for the cost-of-living argument

A second dossier undermining the same message

The NBER finding does not stand alone in raising questions about the administration's economic messaging. A separate CNN fact-check by Daniel Dale, published July 31, 2026, ruled Trump's claim that "everything else is down" in prices "absolutely, unequivocally false," after analyzing 332 products and CPI categories and finding roughly five times more increases than decreases since January 2025.

Two distinct claims — jobs protected by raids, and prices coming down broadly — both face documented empirical challenges within the same short window. One data point is an outlier. Two, on two different promises, start to look like a pattern.

The electoral backdrop this feeds into

An analysis by CNN's Harry Enten, relayed by Raw Story on July 30, 2026, found Trump's net approval among voters under 50 collapsed from minus-3 in the fall of 2024 to minus-41 in 2026. No source consulted proves the NBER study alone caused any specific portion of that shift. The economic gap between promise and outcome, documented across multiple separate issues, provides plausible context for a broader erosion of trust.

Context is not causation. But repeated context, across separate independent measurements, deserves to be named rather than ignored.

Who actually pays the measured cost

Immigrant workers absorb the largest measured hit

The NBER paper's numbers are asymmetric: immigrant men without a college degree see employment fall 2.5% per 1,000 arrests, nearly five times the 0.54% drop measured among U.S.-born men. The largest human cost, according to this specific study, falls on the very communities targeted directly by the raids, not on the workers the policy claims to protect.

That asymmetry does not erase the second finding. It sharpens it. A policy that hurts its stated target and its stated beneficiary at the same time has failed on its own terms.

What "high-ICE-activity areas" means for local economies

Local economies where enforcement activity concentrates do not operate as isolated labor pools. Immigrant and native-born workers frequently work in adjacent roles within the same businesses, supply chains, and local consumer markets. A contraction on one side of that chain plausibly ripples into the other, which is consistent with — though not proof of — the negative effect measured for U.S.-born workers in the same areas.

No source consulted models this ripple mechanism directly. It remains a plausible explanation for the finding, not a confirmed one. The number stands regardless of which specific mechanism produced it.

What "flawed" would need to show

The standard a real rebuttal has to meet

A serious rebuttal to an NBER working paper would need to identify a specific methodological error: a flawed control group, a mismeasured variable, an unrepresentative sample, or a statistical technique applied incorrectly. The Washington Times critique, as relayed in the sources consulted, does not provide that level of detail.

An institution's ideology is not, by itself, evidence its math is wrong. Nor is a study's institutional prestige, by itself, evidence its math is right. Both sides owe the same rigor.

The honest posture for this column

This column holds two things as true simultaneously: the NBER numbers, as currently available, show no employment benefit for U.S.-born workers from ICE raids, and a specific, unanswered methodological critique exists from a conservative outlet. Neither fact cancels the other. Both belong in the same paragraph, not in separate arguments.

The people around this policy, not just the policy itself

Political allies breaking over adjacent issues

Representative Marjorie Taylor Greene, a longtime ally of Trump since 2020, publicly broke with him in late 2025, citing among other reasons his hesitation on releasing more documents tied to Jeffrey Epstein. That break concerns a different policy area entirely, but it documents a broader erosion of unquestioning loyalty within the coalition that once defended every administration policy without exception.

No source consulted links Greene's break specifically to immigration enforcement. It is cited here only as evidence that internal political cohesion, more broadly, is fraying across multiple fronts at the same time as this study surfaced. When allies start asking questions out loud, studies asking the same questions get harder to dismiss.

The Senate confirmation fight running in parallel

On July 30, 2026, the confirmation of Todd Blanche as permanent Attorney General was delayed after Republican senators John Cornyn and Thom Tillis blocked it over a disputed $1.8 billion fund. This dispute is unrelated in substance to immigration enforcement, but it belongs to the same broader picture of an administration facing friction from within its own governing majority, not only from courts and researchers.

Three separate fronts — courts, Congress, and labor economics — converging in the same short window is worth naming plainly.

What responsible coverage of this study requires

Neither dismissal nor blind acceptance

Responsible coverage of this NBER paper requires stating its numbers precisely, stating the criticism against it precisely, and stating clearly what remains unknown: the exact geographic scope, the total sample size, and whether a formal peer-reviewed version will later confirm or revise these specific figures. Precision is not neutrality's opposite. It's neutrality's actual method.

This column does not claim the debate over immigration enforcement's economic effects is closed. It claims, specifically, that one identified study finds no jobs saved for American-born workers, and that no counter-study refuting those precise figures has surfaced in the sources reviewed.

What would change this column's conclusion

A documented, methodologically detailed rebuttal — showing a specific flawed variable, an unrepresentative dataset, or a superior alternative calculation — would change the weight this column assigns to the NBER findings. Absent that, the number holds: zero jobs saved, according to the best available evidence identified.

Evidence changes conclusions. Ideology alone does not. That distinction is the whole argument of this column.

What the study says about the promise, plainly

A promise measured against its own outcome

The political promise was specific: raids protect American jobs. The measured outcome, according to this specific NBER paper, is specific too: no positive effect for U.S.-born workers, and a documented negative effect for the immigrant workers targeted directly. The promise and the outcome do not match.

A policy is not judged by its stated intention. It is judged by what actually happens once it runs.

What accountability looks like here

Accountability, in this specific case, does not require abandoning immigration enforcement as a policy area entirely. It requires the administration to either produce a documented rebuttal of these specific figures, or to stop repeating a jobs-protection claim that its own field's best available research does not currently support.

A claim repeated without evidence does not become true through repetition. It becomes, instead, a claim waiting for its own fact-check.

What defenders of the raids still argue

The theory behind mass enforcement, stated fairly

Supporters of large-scale immigration enforcement argue that short-term dislocation in a local labor market can still produce long-term wage gains for native-born workers, even if immediate employment numbers dip. That argument deserves to be stated on its own terms rather than dismissed outright, because it does not, by itself, contradict the NBER's short-term employment findings.

No source consulted for this column provides a long-term wage study testing that specific claim against the same dataset used by Cox and East. A theory about the future does not answer a measurement about the present. The two can both be worth tracking, without one canceling the other.

Why the burden still falls on the promise-makers

A policy sold on an immediate jobs claim cannot retreat to a long-term theoretical argument only once the immediate data turns unfavorable. If the original promise was short-term relief for American-born workers, the short-term data is the correct yardstick, and on that yardstick, the NBER paper shows a documented shortfall.

Changing the terms of the promise after the data arrives is not a rebuttal. It is a retreat. The original claim, made loudly, deserves to be judged by its original terms.

What local businesses report, where documented

Labor shortages named directly by employers

Reporting connected to the broader immigration enforcement push has documented agricultural and construction employers describing labor shortages following enforcement surges in specific regions, though the sources reviewed for this column do not provide a single, consolidated national study quantifying that effect precisely alongside the NBER figures. Anecdote is not data, and this column treats it as exactly that: context, not proof.

What can be stated plainly is that the NBER's own dataset measures a negative employment effect for U.S.-born workers specifically in high-enforcement areas — the same areas where labor-shortage complaints would be expected to concentrate if the substitution theory held. The shortage and the study point the same direction. Neither, alone, is the full picture.

What a fuller accounting would require

A complete accounting of the raids' economic effect would require sector-by-sector data — agriculture, construction, hospitality, meatpacking — matched against local employment figures over a sustained period, not a single aggregate national number. No source consulted assembles that fuller sector breakdown for this specific policy window.

Until that fuller accounting exists, the NBER's aggregate figures remain the most specific, citable measurement available. A partial picture, honestly labeled, still beats an assumption dressed up as certainty.

What this study means heading into November

An economic argument entering an election year exposed

The 2026 midterms will test whether voters weigh a documented jobs shortfall against a policy sold explicitly as job protection. Democratic strategists, through a joint push nicknamed "Making America Affordable Again" led by Chuck Schumer and Hakeem Jeffries, have already signaled a strategy centered on cost-of-living and labor arguments rather than personal attacks on Trump, according to a Los Angeles Times report dated July 17, 2026.

No source consulted confirms the NBER study itself will feature directly in campaign messaging. Its underlying finding — a jobs promise that data does not support — fits squarely within the economic argument Democrats have already chosen to run on. A study does not need to be quoted on the campaign trail to shape the argument campaigns are already having.

What Republicans would need to counter it

An effective Republican counter would require either a documented methodological rebuttal of the NBER's specific figures, or an alternative dataset showing a different result under comparable conditions. Neither has surfaced in the sources reviewed for this column as of the period covered. An unanswered study does not disappear because it goes unmentioned. It simply sits, waiting for a rebuttal that has not yet arrived.

Conclusion

What is true, according to the best available evidence identified: the NBER's Cox-East working paper finds no positive employment effect for U.S.-born workers from ICE raids, and a documented negative effect for both immigrant and native-born men in high-enforcement areas. What remains unresolved: the Washington Times critique has not, as documented in available sources, produced a methodologically detailed rebuttal of these specific figures.

What the next research cycle will decide: whether a peer-reviewed version of this paper, or a competing study, confirms, revises, or overturns this number. Until then, the number as it stands says one thing plainly. Zero jobs saved is not a rounding error. It is the whole argument, reduced to its actual result.

Sources

Primary sources

• NBER — Working Paper 35129, effects of ICE raids on employment — May 2026
• Washington Times — Critique of the NBER study — May 10, 2026
• New York Times — 9th and 7th Circuit rulings on indefinite detention — July 30, 2026

Secondary sources

• Raw Story — Coverage of the NBER study on ICE raids — July 29, 2026
• Raw Story — Harry Enten's (CNN) analysis on under-50 voters — July 30, 2026
• The Guardian — Trump v. Slaughter and agency firings — July 14, 2026