The U.S. economy has long been powered in part by the nation's ability to attract the world's most educated and skilled people to its shores.
But a new study of the worldwide migration of professionals to the U.S. shows a sharp drop-off in its proportional share of those workers — raising the question of whether the nation will remain competitive in attracting top talent in an increasingly globalized economy.
The study, "Migration of Professionals to the U.S.," was presented at the recent SocInfo 2014 conference in Barcelona, Spain. It uses a novel method of tracking people through data from the social media site LinkedIn, is believed to be the first to monitor global migrations of professionals to the U.S., says co-author Emilio Zagheni, a University of Washington assistant professor of sociology and fellow of the UW eScience Institute.
"This is the first time a worldwide data set has been used to answer this question," Zagheni says. "That hasn't been done before."
The study found that:
"These other countries are attracting not only a higher share of migrants, but also migrants from the top universities in the world," Zagheni says. "That was surprising."
The study, which comes as the United States is mired in a divisive fight over immigration reform, counters conventional wisdom that the nation is the incontestable top choice for professionals migrating from other countries.
"The U.S. is still the top destination for migrations, but [the study] shows that this is something that should not be taken for granted," says co-author Bogdan State, who worked on the study as a Stanford University graduate student alongside co-author Mario Rodriguez, a LinkedIn senior data scientist.
The study suggests numerous possible reasons for the proportional migration decline — the U.S.'s complex visa system, greater demand for professionals in other countries, fewer opportunities for immigrants due to the dot-com collapse of the early 2000s, and the 2008 recession.
Charles Hirschman, a UW sociology professor and expert on immigration to the United States, says the findings underscore the emergence of other nations in sectors once dominated by the United States.
"The United States has to work really hard to stay competitive in this environment," Hirschman says. "Even if we're doing everything right, we're still going to face increasing competition."
Tracking international migration patterns is key to making fiscal projections and shaping effective policies. But migration data tend to be inconsistent across countries and expensive to gather. And there's no consensus between nations on what constitutes a migrant.
Zagheni and other researchers realized a few years ago that the borderless and ubiquitous nature of social media made it a virtual goldmine of migration data. In 2012, Zagheni, State, and Ingmar Weber, then a research scientist at Yahoo, analyzed the Internet Protocol addresses of 43 million Yahoo mail users to calculate migration rates to and from countries worldwide, producing the first-ever curve of U.S. emigration by age and gender. Their work was published as "Using IP Geolocation to Study Human Mobility."
Earlier this year, Zagheni, State and researchers from the Qatar Computing Research Institute published a study that estimated migration patterns among Organization for Economic Cooperation and Development countries by analyzing 500,000 Twitter users. Their work, "Inferring International and Internal Migration Patterns from Twitter Data," found that migration rates from Mexico to the U.S. had dropped, while migration increased from European countries hit hard by the economic crisis, such as Greece and Ireland.
Zagheni acknowledges that the LinkedIn research has limitations. For example, since the study did not include citizenship information, researchers couldn't distinguish between U.S. expatriates returning to the country and in-migration of foreign workers.
More significantly, LinkedIn users aren't a representative sample of the entire population of highly skilled migrants in the U.S. The researchers addressed that by dividing the data set into 10 groups, one for each new cluster of users annually since 2004, and confirmed that there was a significant downward trend in U.S. migration among all 10 groups.
Respecting the privacy of LinkedIn's members was a primary concern. LinkedIn researchers used algorithms to generate the dataset used in the study, which did not include any personal information, according to a LinkedIn representative. The company declined to specify how many LinkedIn users were involved in the study.
Zagheni thinks social media data can be valuable to demographers, geographers, and economists to help understand long-term migration patterns and develop improved theories. Hirschman calls the researchers' work "a major step forward" in effectively using data from social media.
"This is a very, very creative approach, and I think it's opening up a whole new world of data analysis that's going to enrich the field of demography and international migration studies," he says.
The other co-author is Dirk Helbing, a professor of sociology at ETH Zurich in Switzerland.
No entries found