CHAMPAIGN, Ill. — Immigrants whittle into a broad earnings gap with American-born workers only about half as fast as long-accepted estimates suggest, according to new research by a University of Illinois economist.
Darren Lubotsky says immigrants' typically low starting wages grow just 10 to 15 percent faster than native-born workers over their first 20 years in the U.S., well short of the 26 percent catch-up rate in widely used, census-based projections.
Lubotsky says census estimates are flawed because they only compare wages of immigrants and native workers polled in the once-a-decade surveys, and fail to factor in the roughly third of immigrants – most low paid -- who come to the U.S. then leave.
"The reason it appears immigrants catch up quicker based on census data is because, in that data, we're mostly seeing the successful ones, those who decided to remain here," said Lubotsky, a member of the U. of I. Institute of Government and Public Affairs.
"Immigrants who have been in the U.S. for decades tend to have relatively high earnings, but the data don't include many low-earning immigrants who left and therefore weren't surveyed," he said. "The earnings of those who stay do grow, but only about half as fast as previous estimates indicated."
Census estimates are further skewed, he says, because seasonal workers and other low-earning immigrants who routinely go back and forth between the U.S. and their home countries are sometimes misclassified as new arrivals, inflating the pay averages for long-term immigrants.
Lubotsky's study, published in the Journal of Political Economy, examined annual earnings of about 14,000 natives and immigrants from 1951 to 1997, using data from two Bureau of Labor Statistics surveys in the early 1990s that was matched to Social Security records.
"The problem with census data is that the composition of the survey group keeps changing – it gets smaller because every 10 years you're losing people who leave the country," he said. "The data I used track the same people over time and show that the earnings of those who stayed were high to begin with, but tended to not grow very fast."
But even the broader data sample has shortcomings, Lubotsky said, such as missing information on immigrants who are paid in cash and providing no insight on why some choose to leave the country.
"The data aren't ideal, but it's still a vast improvement," he said. "I view the research as getting the facts straight, showing that earnings projections from census data are inaccurate. Whether discussing policy or labor markets, people need to realize that it's really not the case that immigrants arrive poor then end up closer to the average worker."
Lubotsky cautions against giving the new findings too much weight in the more than century-old debate over U.S. immigration policy.
"I tend to think that we should not use this research as a guide to what we do in the future," he said. "We should understand how immigrants assimilate into our labor markets, but there are broader costs and benefits to letting more immigrants in or keeping them out. Whether these new facts matter for policy is up to policymakers."
But he says the study could be useful for foreign workers who are thinking about packing up and crossing U.S. borders in quest of riches.
"For prospective immigrants, it's something they might want to consider," Lubotsky said. "If you're coming for money, have a contingency plan."