One of the most controversial aspects of U.S. immigration policy is the H-1B visa program. The program currently allocates 85,000 visas to high-skill workers, especially in tech industries. Silicon Valley executives at places such as Google and Facebook promote H-1B visas as a way they can attract some of the best programmers in the world to American companies and bemoan the limit on total visas. However, a large percentage of visas go to IT-services companies, whose H-1B employees make lower salaries and more often directly displace American IT workers. The Trump administration has proposed reforms to the program. One particular proposal suggests holding the total number of visas awarded constant, but assigning visas by a preference system based on skills rather than by lottery — disfavoring the IT companies, if they are using the program mostly for entry-level IT positions rather than, for example, software engineers with more specialized skill sets.
What has been the effect of the visa system? A conference paper by economists at University of Michigan and University of California–San Diego takes a direct look at this question with respect to the IT industry, considering impacts both in India and in the United States. They aim to measure the impact on different types of workers in each country, as well as the overall effect of the policy. Unsurprisingly, they find that the majority of the benefits from trade accrue to Indian natives, both from wages to high-skilled emigrants to the United States and because of the increased skill in the Indian workforce from the growth of the IT sector driven by those who stay in India or emigrate and eventually return.
The overall effect on the U.S. workforce is small. Even within computer science, the most affected field, the authors find a 1.5 percent wage reduction and a 9 percent employment decline as compared to a scenario with no H-1B visas awarded between 1995 and 2010. Most of the employment decline comes not from layoffs but from fewer students going into computer science, taking into account the increased competition in that industry from the program. The total wage effect on non-CS majors and non-college graduates, driven slightly down by increased total labor supply and slightly up by complementarities in skills, is negligible. And American consumers enjoy slightly lower prices because of improved and less expensive IT services.
These estimates provide an interesting insight into the economics and politics of the immigration debate. The authors find benefits to India are much larger than to the United States, which might make Americans feel as if they are getting the worse end of the deal even if there are gains on both sides. The overall effect on the U.S. economy is minimal, and the price and wage benefits are diffuse and tiny while the employment and wage effects in computer science are concentrated and relatively large. In my view, the paper’s results reinforce the belief that the visa program benefits society overall, but they do not directly address the debate over whether a fixed number of visas could be allocated differently to provide more benefits. Given the small overall economic effects found from the large number of visas in IT over the last 20 years, it is certainly possible that reforms designed to get more visas to other industries could spur more innovation and have larger spillover benefits to American consumers.