Immigration Lottery Design: Engineered and Coincidental Consequences of H-1B Reforms

In response to increasing demand for high-skilled labor, the U.S. Congress legislated in 2005 that the H-1B visa program create 20,000 additional slots for advanced degree applicants on top of 65,000 slots open to all. Since then, the U.S. Customs and Immigration Service (USCIS) has implemented this policy through visa allocation rules that comply with this legislation. Following a directive in the April 2017 Buy American and Hire American Executive Order by President Trump, USCIS tweaked its H-1B visa allocation rule in 2019, in an explicit effort to increase the share of higher-skill beneficiaries, bypassing the need for Congressional approval to increase the number of advanced degree slots. The USCIS estimated that the rule change, engineered solely for this objective, would increase the number of higher-skill beneficiaries by more than 5,000 at the expense of lower-skill beneficiaries. In this paper, we characterize all visa allocation rules that comply with the legislation. Despite specifying rigid caps, we show that the legislation still allows for rules that can change the number of high-skill awards by as many as 14,000 in an average year. Of all rules that comply with the legislation, the 2019 rule adopted by the Trump administration produces the best possible outcome for higher-skill applicants and the worst possible outcome for lower-skill applicants. We also discover that each of the two previous and much less known changes to the H-1B visa allocation rule resulted in more substantial changes to the share of higher-skill beneficiaries than the 2019 reform. The distributional effects of these earlier reforms in 2006 and 2008, how- ever, were motivated by logistical considerations, potentially without understanding of their importance for the rate of higher-skill awards.