Immigration and Wages — Digest
TL;DR
The surge-replication dispute is now a formal, staged adjudication: the Hunt-Orrenius-Zavodny BPEA draft (+0.9% native wages) was discussed at the conference by Michael Clemens himself, while Wilson & Zhou (−0.7%) presented the next week at the Dallas Fed — but no direct replication of the sign flip has appeared. Meanwhile the $100,000 H-1B fee is unenforced after two court blocks, a new NBER paper finds H-1B exposure raised native incomes in the 1999–2003 window, and the 1920s quota literature delivered its mirror-image result: restriction raised low-skilled wages through labor scarcity.
The surge wage dispute moves from footnotes to a conference room
The native-wage sign fight is being adjudicated in person now, not just in footnotes. Michael Clemens and Wendy Edelberg served as formal discussants of the Hunt, Orrenius & Zavodny BPEA conference draft presented September 24 — meaning the author of the competing +1.3% estimate engaged the +0.9% paper directly (post-pandemic-surge-labor-housing-market-impacts
). Days later, Wilson & Zhou presented their −0.7% estimate (Dallas Fed WP 2607, significant only at the 10% level) at the Dallas Fed migration workshop, with no revision posted as of October 12 (post-pandemic-surge-labor-housing-market-impacts).
"Immigrants filled low-wage jobs leading wages for all workers to potentially decline by up to 1.5% but overall wages for U.S. natives rose by 0.9%."
— Brookings summary of the BPEA conference draft via post-pandemic-surge-labor-housing-market-impacts
What's settled and what isn't is now cleanly separated. Established: no detectable native employment harm, metro GDP up 1.5%, rents up ~1.4–1.6%. Contested: only the sign of a small native wage effect — and the published Clemens/Edelberg comments accompanying the journal issue are where the specification choices (weighting, commuting zones vs metros, EOIR vs CBP instruments, net vs gross migration) should finally be pinned down (post-pandemic-surge-labor-housing-market-impacts).
What to watch: the formal BPEA comments when the journal issue lands — the first real venue where the sign-flip specifications get adjudicated on the record.
High-skilled immigration: the fee is a legal ghost, and the complementarity evidence is strong
The H-1B "treatment" has effectively dissolved while the evidence of native gains firms up. The $100,000 fee is currently not being enforced: the District of Massachusetts vacated the implementing guidance on June 8 (California v. Mullin), the First Circuit declined a stay on July 24, and on September 30 Judge Haywood S. Gilliam Jr. issued a preliminary injunction in Global Nurse Force v. Trump on APA notice-and-comment grounds (high-skilled-immigration-h1b-lottery-and-firm-performance
).
"As a result, the $100,000 H-1B fee is currently not being enforced."
— Visa Lawyer Blog, Oct 8, 2026 via high-skilled-immigration-h1b-lottery-and-firm-performance
The empirical counterpoint landed at the same Dallas Fed workshop: NBER WP 35560 (Abramitzky, Boustan, Gulek & Hainmueller) uses the 1999–2003 cap expansion and finds "H-1B exposure raised incomes for natives and pre-existing immigrants, with gains concentrated in non-STEM occupations," propagating downstream through supply chains — a productivity-shock signature, not substitution (high-skilled-immigration-h1b-lottery-and-firm-performance). The caveat matters: that window had a different skill mix than today, so generalizing to the current regime outruns the design. Note the irony — the paper's complementarity finding is exactly what the surge literature's education breakdown failed to confirm.
What to watch: the DHS proposed ~$103,265 cap-fee rulemaking (FR 2026-17324) — the notice-and-comment path the courts demanded — and any merits ruling from the First Circuit.
The 1920s quotas: the sharpest historical test says restriction did raise low-skilled wages
The historical record now supplies the cleanest pro-restriction wage result in the literature — with the mechanism doing the adjudicating work. Biddle & Cohen's revised KC Fed paper (RWP 22-12, presented October 2 with Santiago Pérez discussing) finds that after the quotas "abruptly and unevenly curtailed immigration from key source countries," wages rose faster in markets dependent on restricted-origin immigrants (1920s-quotas-low-skilled-wages
).
"Evidence on wage dynamics and mechanisms indicates that these gains reflect sustained labor scarcity rather than mechanical compositional changes."
— Biddle & Cohen, KC Fed RWP 22-12 via 1920s-quotas-low-skilled-wages
This is not a contradiction of the modern enforcement literature — it's a boundary condition. The quota shock was far larger, permanent, and landed in a pre-globalized economy with little scope for offshoring or capital adjustment; the open question is whether modern enforcement can ever replicate quota-scale scarcity (1920s-quotas-low-skilled-wages). Anyone citing the 1920s as proof that restriction raises wages today is generalizing beyond what the design supports.
What to watch: whether the scarcity-mechanism framing gets tested against the modern enforcement-shock studies (Cox & East, Escobari et al., Hernandez) in a single comparative framework.
The data infrastructure break is now dated, and the Bureau's own research argues against it
The 2030 census rule moved from proposal to contested process. The comment period, originally ending October 13, was extended to November 2, with the docket showing 47,113 comments (census-2030-residence-criteria-rule
). The rule would exclude non-citizens without green cards — including H-1B holders, students, and refugees — from the apportionment base, drop race/ethnicity questions, and infer status from administrative records across six agencies (census-2030-residence-criteria-rule).
"It's not only folks who are non-citizens who won't respond. We have found in our research that there's a subpopulation of white citizens who don't respond, kind of out of protest."
— Misty Heggeness, co-author of Census CES WP 24-31 via census-2030-residence-criteria-rule
For this topic, the stakes are definitional: the decennial census and ACS are the denominator for every skill-group wage study tracked here, and a citizenship question the Bureau's own working paper shows suppresses response would corrupt the measurement exactly when the post-surge retrenchment cohort needs tracking — while the 2025 ACS remains indefinitely delayed (census-2030-residence-criteria-rule).
What to watch: whether immigrant-rights groups' planned lawsuits materialize before the rule is finalized, and whether the Kansas-led suit is mooted or merged into the new litigation.
What surprised us
- The discussant assignments are doing the adjudication. Clemens discussing Hunt et al., Hunt discussing the Gulek H-1B paper, Pérez discussing Biddle & Cohen — the field's rivals are reviewing each other in formal venues, which is how this literature usually converges (post-pandemic-surge-labor-housing-market-impacts).
- The H-1B fee's litigation timeline makes it a bad natural experiment. Unenforced since June 8, with DHS planning to collect "if this order is later lifted" — any study treating the September 2025 announcement as binding will misestimate the treatment (high-skilled-immigration-h1b-lottery-and-firm-performance).
- The BPEA headline "wages fell up to 1.5%" is a composition effect, not a native harm — the same wage bill over more lower-wage workers. The all-workers vs natives decomposition (+0.9% for natives) is the number that answers the actual question (post-pandemic-surge-labor-housing-market-impacts).
- A softening macro backdrop raises the stakes for the retrenchment era. Unemployment ticked to 4.2% in September 2026 — the short-run/long-run distinction matters more as the surge cohort is partially unwound into a weaker market (post-pandemic-surge-labor-housing-market-impacts).
Immigration and Wages — Digest (Δ from last issue)
Since last time
- Promoted (entirely new this issue): High-skilled immigration is now a full section — the $100,000 H-1B fee is unenforced after two court blocks, and a new NBER paper (WP 35560) finds H-1B exposure raised native incomes in 1999–2003. The 1920s quotas also get their own section: Biddle & Cohen's revised KC Fed paper finds restriction raised low-skilled wages via labor scarcity — the mirror image of the growth-counterfactual framing you saw last time.
- Escalated: The surge wage dispute moved from footnote archaeology to formal, in-person adjudication — Clemens discussed the Hunt-Orrenius-Zavodny draft at the BPEA conference; Wilson & Zhou presented at the Dallas Fed the next week. The 2030 census rule moved from proposal to contested process: comment deadline extended to November 2, docket at 47,113 comments, and the Bureau's own working paper now supplies evidence against the rule's design.
- Demoted: The 2025 ACS delay (was a full section; now a single clause — "remains indefinitely delayed"). The surge literature's complementarity failure (was the headline negative disclosure; now a one-line callback in the H-1B section). The enforcement-shock literature (Cox & East, Escobari et al., Hernandez — was a surprise bullet; now a passing mention in the 1920s what-to-watch). The aggregate wage puzzle (was explicitly open; now partially addressed — see the third surprise bullet).
- Disappeared: The Arkolakis, Lee & Peters long-run growth paper (NBER WP 35392 — the 8.2%-by-1940 / 9.6%-by-2000 results, the skill-selection decomposition, and the ~1.7% quota-acts counterfactual) is entirely absent. So are two related surprise bullets: construction employment not rising with the surge, and "the quota acts look cheap in the growth counterfactual." If you wanted those on your radar, they're not tracked here anymore.
- Unchanged: The settled findings — no detectable native employment harm, metro GDP +1.5%, rents up ~1.4–1.6% — and the footnote-39 list of specification choices (weighting, commuting zones vs metros, EOIR vs CBP instruments, net vs gross migration), which is unchanged in substance but now expected to be pinned down in the published Clemens/Edelberg comments.
The surge wage dispute moves from footnotes to a conference room (Escalated)
The sign fight is being adjudicated in person now. Michael Clemens and Wendy Edelberg served as formal discussants of the Hunt, Orrenius & Zavodny BPEA conference draft presented September 24 — meaning the author of the competing +1.3% estimate engaged the +0.9% paper directly (post-pandemic-surge-labor-housing-market-impacts). Days later, Wilson & Zhou presented their −0.7% estimate (Dallas Fed WP 2607, significant only at the 10% level) at the Dallas Fed migration workshop, with no revision posted as of October 12 (post-pandemic-surge-labor-housing-market-impacts).
"Immigrants filled low-wage jobs leading wages for all workers to potentially decline by up to 1.5% but overall wages for U.S. natives rose by 0.9%."
— Brookings summary of the BPEA conference draft via post-pandemic-surge-labor-housing-market-impacts
What's settled and what isn't is now cleanly separated. Established: no detectable native employment harm, metro GDP up 1.5%, rents up ~1.4–1.6%. Contested: only the sign of a small native wage effect — and the published Clemens/Edelberg comments accompanying the journal issue are where the specification choices (weighting, commuting zones vs metros, EOIR vs CBP instruments, net vs gross migration) should finally be pinned down (post-pandemic-surge-labor-housing-market-impacts).
What to watch: the formal BPEA comments when the journal issue lands — the first real venue where the sign-flip specifications get adjudicated on the record.
High-skilled immigration: the fee is a legal ghost, and the complementarity evidence is strong (Promoted — new section)
The $100,000 fee is currently not being enforced: the District of Massachusetts vacated the implementing guidance on June 8 (California v. Mullin), the First Circuit declined a stay on July 24, and on September 30 Judge Haywood S. Gilliam Jr. issued a preliminary injunction in Global Nurse Force v. Trump on APA notice-and-comment grounds (high-skilled-immigration-h1b-lottery-and-firm-performance).
"As a result, the $100,000 H-1B fee is currently not being enforced."
— Visa Lawyer Blog, Oct 8, 2026 via high-skilled-immigration-h1b-lottery-and-firm-performance
The empirical counterpoint landed at the same Dallas Fed workshop: NBER WP 35560 (Abramitzky, Boustan, Gulek & Hainmueller) uses the 1999–2003 cap expansion and finds "H-1B exposure raised incomes for natives and pre-existing immigrants, with gains concentrated in non-STEM occupations," propagating downstream through supply chains — a productivity-shock signature, not substitution (high-skilled-immigration-h1b-lottery-and-firm-performance). The caveat matters: that window had a different skill mix than today, so generalizing to the current regime outruns the design. Note the irony — the paper's complementarity finding is exactly what the surge literature's education breakdown failed to confirm.
What to watch: the DHS proposed ~$103,265 cap-fee rulemaking (FR 2026-17324) — the notice-and-comment path the courts demanded — and any merits ruling from the First Circuit.
The 1920s quotas: the sharpest historical test says restriction did raise low-skilled wages (Promoted — new section)
Biddle & Cohen's revised KC Fed paper (RWP 22-12, presented October 2 with Santiago Pérez discussing) finds that after the quotas "abruptly and unevenly curtailed immigration from key source countries," wages rose faster in markets dependent on restricted-origin immigrants (1920s-quotas-low-skilled-wages).
"Evidence on wage dynamics and mechanisms indicates that these gains reflect sustained labor scarcity rather than mechanical compositional changes."
— Biddle & Cohen, KC Fed RWP 22-12 via 1920s-quotas-low-skilled-wages
This is not a contradiction of the modern enforcement literature — it's a boundary condition. The quota shock was far larger, permanent, and landed in a pre-globalized economy with little scope for offshoring or capital adjustment; the open question is whether modern enforcement can ever replicate quota-scale scarcity (1920s-quotas-low-skilled-wages). Anyone citing the 1920s as proof that restriction raises wages today is generalizing beyond what the design supports.
What to watch: whether the scarcity-mechanism framing gets tested against the modern enforcement-shock studies (Cox & East, Escobari et al., Hernandez) in a single comparative framework.
The data infrastructure break is now dated, and the Bureau's own research argues against it (Escalated)
The 2030 census rule moved from proposal to contested process. The comment period, originally ending October 13, was extended to November 2, with the docket showing 47,113 comments (census-2030-residence-criteria-rule). The rule would exclude non-citizens without green cards — including H-1B holders, students, and refugees — from the apportionment base, drop race/ethnicity questions, and infer status from administrative records across six agencies (census-2030-residence-criteria-rule).
"It's not only folks who are non-citizens who won't respond. We have found in our research that there's a subpopulation of white citizens who don't respond, kind of out of protest."
— Misty Heggeness, co-author of Census CES WP 24-31 via census-2030-residence-criteria-rule
For this topic, the stakes are definitional: the decennial census and ACS are the denominator for every skill-group wage study tracked here, and a citizenship question the Bureau's own working paper shows suppresses response would corrupt the measurement exactly when the post-surge retrenchment cohort needs tracking — while the 2025 ACS remains indefinitely delayed (census-2030-residence-criteria-rule).
What to watch: whether immigrant-rights groups' planned lawsuits materialize before the rule is finalized, and whether the Kansas-led suit is mooted or merged into the new litigation.
What surprised us
- [NEW] The discussant assignments are doing the adjudication. Clemens discussing Hunt et al., Hunt discussing the Gulek H-1B paper, Pérez discussing Biddle & Cohen — the field's rivals are reviewing each other in formal venues, which is how this literature usually converges (post-pandemic-surge-labor-housing-market-impacts).
- [NEW] The H-1B fee's litigation timeline makes it a bad natural experiment. Unenforced since June 8, with DHS planning to collect "if this order is later lifted" — any study treating the September 2025 announcement as binding will misestimate the treatment (high-skilled-immigration-h1b-lottery-and-firm-performance).
- [UPDATED] The BPEA headline "wages fell up to 1.5%" is a composition effect, not a native harm — the same wage bill over more lower-wage workers. This refines last issue's puzzle that the aggregate decline seemed "too big to be composition"; the all-workers vs natives decomposition (+0.9% for natives) is the number that answers the actual question (post-pandemic-surge-labor-housing-market-impacts).
- [NEW] A softening macro backdrop raises the stakes for the retrenchment era. Unemployment ticked to 4.2% in September 2026 — the short-run/long-run distinction matters more as the surge cohort is partially unwound into a weaker market (post-pandemic-surge-labor-housing-market-impacts).
Open threads
- Still open, updated: 2030 Census Residence-Criteria Proposed Rule: Finalization or Litigation — the thread continues with new specifics: comment deadline now November 2, 47,113 comments on the docket, and the watch item sharpened to whether immigrant-rights suits materialize pre-finalization and whether the Kansas-led suit is mooted or merged.
- No threads closed. The previous briefing's other watch items were absorbed into the body: the call for a replication isolating the footnote-39 sign flip is now the BPEA-comments watch (section 1), and the enforcement-shock pipeline is folded into the 1920s what-to-watch (section 3).