Trump’s Anti-immigration Policies are Based on Flawed Data

This article first appeared on the Cato Institute site.

Senior Trump administration aide Stephen Miller gave a press briefing on Thursday defending the RAISE Act—a bill introduced by Sens. Cotton (R-AR) and Perdue (R-GA) that would slash the number of legal immigrants without increasing skilled or merit-based immigration.

The purpose of the RAISE Act is to restrict low-skilled immigration in order to raise the wages of American workers.   

When asked by a reporter for evidence that restricting low-skilled immigration would raise wages, Miller cited research by Harvard economist George Borjas on the Mariel Boatlift.

The Mariel Boatlift produced an unexpected surge of 125,000 Cubans (henceforth Marielitos ) to Miami in 1980. Because at least 60 percent of the Marielitos were high school dropouts, Borjas tested whether they lowered the wages of American dropouts.

Since Borjas published his Mariel paper, there have been many rebuttals, criticisms, and additional research on it that should substantially diminish confidence in his findings. Below I will briefly summarize these results.

The first such criticism is by economists Michael Clemens and Jennifer Hunt. They conclude that the entirety of the wage decline observed by Borjas can be explained in how the wage survey in Miami increased the proportion of black workers surveyed, far in excess of their proportion of the population, when the Boatlift occurred (the CPS change was unrelated to the Boatlift).

Black American workers with less than a high school degree have lower wages than similarly skilled non-black Americans for myriad reasons that have nothing to do with immigration.

By including more of them in the survey at the same time the Marielitos were arriving made it look like there was a drastic wage decline when the observed effect was entirely due to shifting the demographics of the surveyed population.

That survey shift entirely explains the negative wage effect observed by Borjas. Borjas’s response to Clemens and Hunt is weak.

GettyImages-825652944 Senior Advisor to the President for Policy Stephen Miller talks to reporters about Donald Trump's support for creating a 'merit-based immigration system' in the James Brady Press Briefing Room at the White House August 2, 2017 in Washington, DC. Chip Somodevilla/Getty

The second criticism of Borjas’s Mariel Boatlift research is by economists Giovanni Peri and Vasil Yasenov.

They note that wages in Miami must be compared to wages in similar cities at the same time to measure how wages changed. By selecting a set of comparison cities using the Synthetic Control Method, a different method than Borjas used, they found no statistically significant deviation in Miami’s wages compared to similar cities that did not absorb the Marielitos .

Furthermore, Borjas relied on smaller surveys with few relevant observations for Miami and other cities. Peri and Yasenov used the larger to get even more data on wages. Including the additional data also showed that the Marielitos did not lower wages.

Cole Blondin and I wrote the third response, which is not a rebuttal to Borjas’s work but shows other implications.

We assumed that Borjas’s methods and data were beyond reproach so we copied them. We found that Americans in Miami with only a high school degree saw a significant increase in wages that coincided with Mariel.

Because Miamians with only a high school degree outnumbered dropouts, the wage losses reported by Borjas were more than offset by wage gains made by Miamians with only a high school education in most of our estimates.

We also found that wages for dropout Hispanic workers in Miami actually increased right after the Boatlift, relatively, compared to Hispanic workers in other cities as the Marielitos arrived. Since Hispanic workers are the most substitutable with the Marielitos , this result is wholly inconsistent with Borjas’s explanation.

The fourth response by David Roodman points out substantial methodological critiques of Borjas’s paper that should raise serious concerns such as the fact that his analysis excludes women entirely and that the wage drop observed by Borjas took years to develop when it should have happened more rapidly.

There is no good methodological reason for excluding the wages of working women from the analysis.

The other big problem with Miller’s response was that he responded to a question about how the RAISE Act would boost wages by citing research that argues an increase in immigration lowers wages under the exceptional circumstances of the Mariel Boatlift.

The actual research that examines wage effects from immigration restriction finds that wage growth for American workers slows after Congress passes restrictions for that purpose.

Economists Michael Clemens, Ethan Lewis, and Hannah Postel recently wrote a groundbreaking paper that took advantage of Congress’ 1964 cancellation of the Bracero program to see how a cut in legal migration affect American wages.

The Bracero program allowed hundreds of thousands of Mexican workers to work legally on American farms from 1942 to 1964. Congress canceled Bracero due to lobbying from labor unions and others that argued such a cancellation would raise wages.

Clemens, Lewis, and Postel found that farm worker wages rose more slowly because farmers mechanized production and planted crops that required less labor. Not only did the supposed wage gains from cutting legal migration not occur, the rate of wage increase actually slowed.

The details and justifications for the RAISE Act and Congress’s 1964 cancellation of Bracero are eerily similar. The Bracero program allowed in half a million workers a year before it was eliminated – which is roughly the number of green cards that RAISE would cut.

Bracero workers have a relatively similar skills level, compared to Americans, of the workers who currently enter of the family-based green cards that the RAISE Act intends to cut. Braceros were also concentrated in some states just like new immigrants are.

George Borjas’s research on the Mariel Boatlift is not the empirical slam-dunk that Stephen Miller thinks it is. At a minimum, the work I summarize above should raise doubts and diminish his confidence in the RAISE Act.

Supporters of the RAISE Act need to deal with the totality of the economics research on the effect of cutting legal immigration. The fact that they are ignoring most of it and cherry-picking findings should raise alarm bells that are louder than a debate over the history of the Statue of Liberty.

Alex Nowrasteh is an immigration policy analyst at the Cato Institute’s Center for Global Liberty and Prosperity.