A recently released National Bureau of Economic Research working paper is garnering a lot of comment among those who promote immigration of science and technology workers into the U.S. The paper, by economists Giovanni Peri, Kevin Shih, and Chad Sparber, concludes that a one percentage point increase in the foreign STEM share of a city’s total employment increased wages of native college educated labor by about 7-8 percentage points and the wages of non-college educated natives by 3-4 percentage points.
What is perhaps most interesting about this result is that it seems to refute several previous studies that have concluded the opposite. One in particular, published by the Economic Policy Institute, found that current U.S. immigration policies that facilitate large flows of guest workers (on H-1B and other kinds of work visas) appear to provide firms with access to plentiful STEM (science, technology, engineering, and mathematics) workers at wages that are too low to induce a significantly increased supply from the domestic workforce.
The NBER working paper just published, and I, for one, will be curious to hear reactions from other economists about its methodology. Here, the term "working paper" is important. A working paper is a preliminary scientific document that authors lease to share ideas about a topic or to elicit feedback typically before submitting to a peer reviewed conference or academic journal. So the final version of the piece could be quite different, or it may not publish at all, depending on the critiques other economists provide.
It is worth exploring how the authors on the two sides of the issue arrive at their result. For the EPI study, authors used State Dept. H-1B visa statistics and estimated the impact of this immigration on the IT labor market. Among other things, they found that guest workers had grown to represent a third of all the U.S. IT workers with a BS degree and above who started jobs in the recent past. They also noted a lack of wage growth in IT jobs since the year 2000. In a nutshell, they think the two trends are linked. They also note that in some cases, STEM employment trends are highly regional. "The Silicon Valley pattern in employment and wage levels is not representative of the industry as a whole," they say. "For example, the sharp increase in computer system design employment in Silicon Valley in 2010..... is primarily the result of the dramatic growth of employment at Apple in response to the success of the iPhone and the launch of the iPad.....Even in the dynamic technology regions of Silicon Valley, Route 128, Dallas, and Austin, the local earnings data.... reveal no signs of the rising wages that would be expected to occur if employers were hiring in a market that had a limited supply of workers."
These last comments are worth keeping in mind as you review the methods used by the three authors of the recent NBER paper. The economists looked only at data from 219 U.S. metropolitan areas. They then looked at those having the biggest populations of H-1B STEM workers. They then look at inflation-adjusted wage gains for native-born workers in those cities. The authors say these grew at between 17 and 28%. Thus their claim is that the influx of foreign workers actually raised wages for other STEM workers.
Conversely, the authors also looked at the metropolitan areas with a declining foreign STEM workforce. Those areas also saw a decline in wages for native college degree holders. The two declines are linked, claim the economists.
They come to these results through use of fairly involved mathematical modeling. But there is an unsettling point one can raise about their conclusions that doesn't require any kind of math model: If it were true that a given number of H-1B STEM workers raise wages by 7 to 8% as the authors claim, would doubling or tripling the number of foreign workers be expected to raise them proportionally more?
Regardless of what a math model said, it would be hard to buy this idea.