George Borjas on Immigration and We Wanted Workers
What this covers
Russ Roberts interviews George Borjas of Harvard on the economic and cultural effects of immigration, with Borjas arguing that standard economic models obscure real distributional harm. The conversation turns on a fundamental disagreement about method: economists typically present immigration as creating aggregate gains that wash out in the long run, but Borjas insists this conclusion is built into their assumptions rather than discovered in data. He contends that treating immigrants as interchangeable units—the Workers vs People Distinction—allows policymakers to dismiss genuine costs borne by specific groups, chiefly low-skilled native workers and taxpayers. Roberts presses back at several points, defending the historical record of assimilation and arguing that the remedy for immigration-driven wage pressure is skill upgrading, not restriction.
The discussion ranges across labor economics, history, and cultural change. Borjas makes three interconnected claims about recent decades: the constant-returns assumption in economic models guarantees the zero-wage-impact finding rather than proving it; low-skilled native wages have fallen 3–5% from recent immigration, a concentrated loss on those least equipped to absorb it; and past assimilation cannot be assumed to repeat because the conditions that enabled it—unionized manufacturing, wartime pressure, explicit pro-melting-pot culture—have disappeared. He illustrates Cohort Quality vs Assimilation with the Cuban case, where earlier and later waves were fundamentally different populations, not proof of immigrant advancement over time. On Supply and Demand, Borjas contends that claiming "immigrants do jobs Americans won't do" misses market mechanics; wages rise when supply is removed, natives step in. He also flags the welfare-state angle: unlike trade, low-skilled immigration may shrink the economic pie if welfare costs exceed labor gains. The cultural dimension surfaces when he notes that policies rewarding ethnic identity retention have weakened the Incentives Shape Behavior for assimilation itself.
Borjas argues that immigrants are people, not interchangeable workers, so immigration creates real distributional costs—especially for low-skilled natives and through welfare-state and cultural effects—that economists and policymakers wrongly dismiss by hiding behind aggregate models and assumptions.
- The finding that immigration has zero average long-run wage impact is built into the constant-returns assumption, not discovered in data
- Low-skilled native wages have fallen 3-5% from recent immigration, a real cost concentrated on those least able to bear it
- Past assimilation cannot be extrapolated because conditions (unionized manufacturing, wartime pressure, pro-melting-pot culture) no longer exist
Creative destruction from technology and trade harms displaced workers while raising overall living standards across generations.
- Over the past century, technological change, globalization of goods and capital, and immigration destroyed hundreds of millions of jobs while raising US material living standards 10-30 fold; although the displaced were harmed at the time, their children and grandchildren live far better material lives through this process of creative destruction and growth.
“a farmer living in 1900, when 40% of the workers were on the farm--today it's 2% in the United States--that farmer's children and grandchildren are much better off even though the changes that caused that to happen were very tough on farmers in the meanwhile”
- Low-skilled American workers have had a very challenging 25 years from three converging forces: increased immigration of low-skilled workers, technological change including artificial intelligence, and trade with countries that have abundant low-skilled labor—together explaining why many in certain parts of the country are doing poorly.
“low-skilled workers in America, the last 25 years, have been very challenging, for three reasons... One is the increase in immigration of low-skilled workers. The second is technology... And the third, of course, is trade.”
The policy solution is improving worker skills and education, not blocking immigration, innovation, or imports.
- The right policy response to immigration harming low-skilled natives is not to keep out low-skilled workers, reduce innovation, or block imports, but to improve the skill set of native workers—encouraging them to finish high school and improving the school system and culture.
“It seems to me the right policy response to that is not to keep out low-skilled workers. It's not to reduce innovation... It's to try to improve the skill-set of the workers who are here.”
Labor-market policy effects must be measured only on directly affected low-skill workers, not aggregate employment.
- When measuring the impact of a labor-market intervention like a minimum wage, you must look only at the affected workers at the low end of the skill distribution, not aggregate employment, because theory does not predict overall employment to change much when only a small proportion of workers are affected and the effect gets lost in growth-driven noise.
“you certainly don't want to look at, say, employment of all workers. Because the theory doesn't expect that to change very much, if only a small proportion of workers are affected by the minimum wage.”
Larger markets enable greater specialization and productivity, and technology creates genuine welfare improvements.
- Forty years after meeting, the two can now converse via Skype, recorded and available to anyone around the world at zero direct cost, reflecting the many good things technology has produced.
“we'd be talking not via phone but via Skype; it would be recorded and it would be available to anyone around the world at zero direct cost”
- Doubling a population need not leave per-capita income unchanged even under capital adjustment, because, following Adam Smith, the division of labor is limited by the extent of the market—so a larger market can enable greater specialization and higher productivity rather than mere replication.
“I actually think there's, via Adam Smith, the division of labor is limited by the extent of the market.”