Measuring the Impacts of the H-1B Visa Program on U.S. Labor Markets: Two Recent Quasi-Experimental Studies

The H-1B visa program, created by the Immigration Act of 1990 (P.L. 101-649), allows U.S. employers to hire temporary foreign workers in specialty occupations requiring a college degree. Many H-1B workers eventually become lawful permanent residents through the employment-based immigration system.

Congress has a long-standing interest in the H-1B program, sometimes increasing and sometimes decreasing the number of new H-1B workers permitted each year. Under current law, cap-subject employers may hire up to 65,000 new H-1B workers each year, as well as an additional 20,000 workers with masters' degrees or greater education levels. There is no cap on H-1B workers extending their stay, or on the number of new H-1B workers hired by university and nonprofit research employers.

Proponents of the H-1B program say it helps U.S. employers hire workers with skills difficult to find in the domestic labor market, increasing these employers' overall production and employment. Critics of the H-1B program say it enables employers to displace domestic workers with cheaper foreign labor. Difficulties in measuring the impacts of the H-1B program contribute to these disagreements.

Measurement Issues

Several fundamental challenges make it difficult to measure the impacts of the H-1B program on domestic labor markets. These issues are discussed below, as are steps taken to address them in the studies profiled.

Selection Issues

Increasingly, H-1B workers are hired by a small group of very large firms in the scientific and business services, computer, and information technology staffing industries with expertise in navigating the H-1B application process. These large employers are not typical of U.S. employers; they are a selected group. This selection problem means that comparing outcomes for employers that use the H-1B program (and their employees) with employers that do not use the H-1B program (and their employees) will not yield accurate estimates of the impact of program.

Data Availability

Data availability issues affect the time horizon that can be studied, and require using proxy measures to examine certain outcomes. For example, U.S. Citizenship and Immigration Services (USCIS) has posted H-1B employer petition data on its website for petitions filed starting in FY2009. The researchers who conducted the studies profiled in this In Focus, examining long-term impacts of the H-1B program on employment and wages, filed Freedom of Information Act (FOIA) requests with USCIS to gain access to H-1B petition data for years before 2009. USCIS only retained data for FY2007 and FY2008 for approved petitions. To construct comparisons involving these years, researchers infer which companies entered the H-1B lottery and did not win by using data from the Labor Condition Applications (LCAs) employers must file with the U.S. Department of Labor before filing H-1B petitions.

Researchers who want to study the impact of the H-1B program on employment and wages have linked data on employers using the program with large confidential tax datasets that can be used to illuminate employer and employee outcomes. However, these tax datasets do not identify which employees have H-1B status; the closest available proxy measure is information on which employees of these firms are foreign-born. This proxy measure is limited because there are many ways employers can hire foreign-born workers other than via H-1B visas. Some foreign-born workers have permanent residency or citizenship. Employers may hire foreign graduates of U.S. colleges and universities through the Optional Practical Training (OPT) program, or use other visa programs.

Time Lag

Studying the impacts of the H-1B program can require many years of data, as impacts may take time to observe. Data access issues cause further delays. For example, gaining research access to federal corporate tax data or state unemployment insurance data is a lengthy process. Meanwhile, the U.S. labor market changes, as do the ways firms use the H-1B program. These delays affect the timeliness of study results.

Quasi-Experimental Studies

Because of the above-mentioned measurement issues, until recently, research on the H-1B visa program has not directly examined whether the program expands or displaces employment in the United States.

However, annual employer demand for new H-1B workers exceeded the number available to cap-subject firms in FY2006-FY2009. In these years, USCIS used lotteries to select which petitions to consider. Having H-1B petition acceptance decided by lottery makes it possible to avoid selection bias issues in studying the impacts of the H-1B program by comparing outcomes for firms whose petitions were accepted ("won the lottery") with outcomes for firms whose petitions were denied ("lost the lottery"). Studies using random variation that was not intended as an experiment are called quasi-experimental studies. They provide the best available evidence of how H-1B petition decisions affect outcomes for cap-subject firms that apply for the program and their incumbent employees.

There are two recent quasi-experimental studies of the impacts of the H-1B program on employment and wages in the United States. They are summarized in Table 1. These two studies use data on firms selected or not selected in the H-1B lottery by random chance. However, they differ in many ways. They focus on different years, with differences in how many (and which) firms were subject to the lottery; use different data; and rely on different assumptions.

Most findings on the employment and wage outcomes of H-1B workers differ across these two studies. One area of similarity is that the first study finds negative employment impacts of H-1B petitions that fall most heavily on other foreign-born workers. This complements the finding in the second study of no overall employment loss for native-born workers when their employers win the H-1B lottery.