In the past few months we have learned that hundreds Americans at Southern California Edison have been replaced by H-1B workers. A few hundred more Americans at Walt Disney World have been replaced by H-1B workers. Ditto at Northeast Utilities (now Eversource).
Such replacements have been going on since at least 1994. But a paper from a new Washington, D.C. libertarian think tank claims they could not have happened.
Usually, the Washington Examiner is a better source of immigration-related news than most news outlets. That is why I was surprised to read this article about the paper, by David Bier of the Niskanen Center, entitled "H-1Bs Don't Replace U.S. Workers".
I am puzzled by some of his logic. He bases one of his chief findings on the assertion that:
If H-1Bs were primarily cheaper substitutes for American labor, the pace of H-1B requests—measured by the length of time before the cap on visas is reached—should rise when unemployment rises, as employers look to cut labor costs by laying off workers.
That is the exact opposite of what a thinking person would expect applying basic economics. When unemployment rises, the supply of available labor increases; under the law of supply-and-demand the cost of labor decreases, creating less of a cost advantage from using cheap foreign labor. When unemployment declines, the supply of available labor decreases; causing the cost of labor to increase; creating greater demand for cheap foreign labor—which is exactly what we see.
The rest of Bier's report is no better. He attributes this conclusion:
Using cross-city panel regressions, they found that a one percent increase in foreign STEM workers as a share of total employment raises the wages of college-educated natives by 7 to 8 percent in a given city
to this study. But the actual results are:
We find that H-1B-driven increases in STEM workers in a city were associated with significant increases in wages paid to college educated natives.
There is a big difference between "associated with" and "raises". Regression analysis can establish relationships; it does not show causation. When regression shows an association between H-1B numbers and wages, it is just as possible (and more likely) the data shows that employers resort to more H-1B workers when wages are higher than that H-1B workers drive wages up.
Bier provides his own charts showing correlation between H-1B numbers and employment. He attributes H-1B as the cause of employment wage growth. Interestingly, in his own charts he uses the more descriptive "H-1Bs Associated with Greater Employment" – not "causes".
Bier also provides suggestions for "Responding to Occasional H-1B Abuse". The problem is H-1B abuse is not "occasional" but rather "normal". A USCIS audit in 2008 found that violation rate in the H-1B program was 20.7 percent. Bloomberg called that a "High Rate of H-1B visa fraud".
Of course, maybe Bier is correct and the workers at SCE, Northeast Utilities, and elsewhere were not, in fact, replaced by H-1B workers and they can go back to work on Monday.
In reality, Bier's data shows the exact opposite of what he claims. His conclusions rely on spinning the results to defy the law of supply and demand.
A more rational explanation of Bier's data is that when market forces drive wages up, employers turn to using more H-1B visas.
For another take on Bier's report, see computer science professor Norman Matloff's blog. For a debunking of industry lobbyists' claims of a tech labor shortage, see this report by Steven Camarota.