Conference Coverage: Topics in Applied Microeconomics
Client: The Becker Friedman Institute for Research in Economics at the University of Chicago
What causes “brain drain,” the phenomenon where highly-skilled people migrate around the world in search of better opportunities? Though widely debated in the media and often attributed to the desire to avoid high taxes, surprisingly little evidence exists about what motivates talented people to move across international borders.
Stefanie Stantcheva of the Harvard Society of Fellows and her colleagues have begun to unravel the brain drain question by applying innovative methods to US Patent Office (USPTO) and European Patent Office (EPO) data. They matched it with marginal tax rate data and tracked the mobility of inventors over time. Her study provides new insights into what might be motivating migration of talented workers. Ufuk Akcigit, University of Chicago, and Salomé Baslandze, Einaudi Institute for Economics and Finance, coauthored the paper.
Stantcheva concluded that top-performing or “superstar” inventors, as measured by those who hold the top one percent of total patents in any given year, do react to top tax rates, but not to a large extent. She measured the elasticity of domestic superstar inventors to the top net-of-tax rate relatively small, around 0.03. In other words, for a 10 percentage point decrease in top tax rates from a 60 percent rate, the number of domestic superstar inventors inclined to stay at home jumps by one percent. The elasticity for the number of foreign superstar inventors is much higher, at about one. That translates into a 26 percent increase in superstar inventors migrating for the same decrease in the top tax rate.
Superstar inventors who worked for multinational corporations were the most sensitive to tax rates. They were more likely to move to take advantage of beneficial tax rates than similar superstar inventors who worked at universities, independently, or at smaller firms. This may be because employment at multinational corporations makes the transition across borders easier for the employee and benefits the inventor’s career with greater international visibility. However, for multinational corporation talent who lived in a country where their specialty was well established, they were more likely to stay put.
Stantcheva developed this analytical approach by drawing on the 4.2 million USPTO patent records from 1975 to 2010, representing more than three million inventors; the database comprises 26 percnet of all patents in the world. Stantcheva used this data and the smaller database from the EPO, which has a higher percentage of EU inventors. Her study focuses on eight of the top patent-producing nations: Canada, France, Germany, Great Britain, Italy, Japan, Switzerland, and the United States. Because the inventors must enter their location at the time of filing for their patents, the data allow for tracking inventors’ mobility.
Stantcheva isolated those inventors who could be deemed the highest quality by tracking the number of patents held and the citation rates for those patents. Patent quality is especially important, she notes, because this is what partly dictates long-term income and tax brackets.
The top 10 percent of these inventors were dubbed superstars and then broken down into the top five percent and one percent of performers. (For example, Shunpei Yamazaki was determined to be the top superstar inventor of the one percent up until 2008, with 3,780 patents to his name; he lives and works in Japan.) She tracked the location of these inventors and combined that data with marginal tax rates, using a location choice model and evidence drawn from three case studies on countries that had experienced large changes in migration or tax policies.
Stantcheva acknowledged that their study also seems to show that career concerns matter to inventors when they choose where to live and work. As Stantcheva wryly observed during her presentation, all three authors on the paper migrated for reasons unrelated to taxes, unless studying taxes counts, too.