The idea that the United States needs more scientists in order to produce more science has become a mantra among government, industry, and academic leaders seeking to energize the lagging economy. The major role of innovation in the nation’s prosperity gives the notion intuitive appeal.

“Knowledge produced by one researcher is both an output and an input into another researcher’s [work],” write economists George Borjas of Harvard University and Kirk Doran of the University of Notre Dame in a recently published study. “Not surprisingly, there is a consensus among policy makers that increases in the supply of a highly skilled workforce will increase the productivity of the preexisting workforce, and lead to a substantial increase in national wealth.”

But is this idea actually true? “The Collapse of the Soviet Union and the Productivity of American Mathematicians,” the intriguing and important article from which I borrowed that quotation, raises serious doubts. Borjas and Doran have cleverly turned the major geo-political event of the 1990s into a natural experiment to find out what can happen when immigration sharply boosts the number of researchers in a scientific field. By tracking the large influx of Russian mathematicians into the United States following the Communist regime’s demise, they found a number of significant effects.

An overall increase in research production, however, was not among them. Instead, the output of American mathematicians whose research interests coincided with the Russians’ actually “shrank,” the paper states. The recent arrivals’ work did “fill the gap” in numbers of publications, but “there is no evidence that they greatly increased the size of the ‘mathematics pie.’ ” “The point of the paper,” Borjas tells Science Careers in an interview, is that “it’s certainly true that high-skill immigrants will bring in new ideas and some of these ideas will rub off on us, and be innovative and productive. … But it’s also true that supply and demand work.”

The Russians are coming

The fall of the Iron Curtain in late 1991 ended nearly 70 years of isolation of Soviet mathematicians from the world mathematical community. Suddenly free to travel and emigrate, approximately 1000 Soviet mathematicians, mostly highly productive researchers, relocated to other countries. Three-hundred-thirty-six scientists came to the United States.

During the decades of scant contact with foreign colleagues, Russian mathematicians, for political reasons that Borjas and Doran explain, concentrated on certain fields that tended to get much less attention in the West. American mathematicians, meanwhile, had moved ahead in areas that Russians largely ignored. When the émigrés arrived in America, therefore, they had a great impact on certain mathematical fields but much less on others.

These differential impacts allow Borjas and Doran to compare what happened in fields heavily and lightly affected by the influx and to analyze what a large infusion of new talent and ideas does to a field of research. Combining information from several large databases, they track the productivity and affiliations of the Russian and American mathematicians. Their inquiry concentrates on two major effects; the “knowledge shock” from all of the new approaches and insights suddenly available to American mathematics, and the “labor market shock” from all the new people suddenly on the American mathematics job market.

What they found does not support what Borjas calls the conventional “rubbing-off” theory, which holds that if “we get all these highly skilled immigrants, … somehow there’s a rubbing-off effect that makes you and me more innovative.”

Since the fields with heavy Russian influence had “incredibly bright new mathematicians coming in, with all these new theorems, all these new techniques flooding the market, you would expect that people working in those areas are going to learn quite a lot from them,” Borjas says. “At the same time, the number of academic jobs where mathematical research is actually done is really not increasing all that much. So something has to give.”

That something, the study found, was the career prospects and productivity of many of the mathematicians already here. “When you increase the number of very smart people in a field by a substantial amount,” Borjas says, “… not everybody benefits. The typical pre-existing American mathematician actually lost out.”

That’s because “a generation of American mathematicians at the very peak of their mathematical efficiency by luck just happened to graduate at the same time these Russians [were] coming in,” Borjas explains. The people of that young generation lost the most. Faculty members protected by tenure kept their jobs, but mathematicians who had not yet attained it found themselves facing sharply heightened competition, which produced an “unprecedented 12 percent unemployment rate for new American mathematics PhDs” and “a dramatic decrease in the probability of obtaining a position in research universities,” the article says. The overall unemployment rate for college graduates, meanwhile, was dropping rapidly, from 3.2% to 2.2% between 1992 and 1996. Many of the young mathematicians in heavily affected fields ended up moving to lower-ranking institutions or leaving academic mathematics altogether.

The situation “drove a lot of American mathematicians into Wall Street,” Borjas says. There, in their new role as “quants” (quantitative analysts), they turned their talents to inventing intricate financial instruments. Ironically, one of Borjas’s colleagues has joked, the migration from the formerly communist country probably helped to fuel the 2008 economic crisis that nearly brought down American capitalism.

The market in ideas

The mundane workings of labor market supply and demand may be the stuff of Economics 101, but didn’t all the ideas the Russians brought boost research? The “rubbing-off” theory argues that they should have done just that. But even there, Borjas says, “The constraints of the market were way too strong.”

“The actual post-1992 output of mathematicians whose work most overlapped with that of the Soviets and hence could have benefited more from the influx of Soviet ideas is far below what would have been expected” compared with former productivity and Americans’ productivity in other fields, the paper states. Many in the heavily affected fields “ceased publishing relatively early in their career. [Also] they became much less likely to publish a ‘home run’ paper after the arrival of the Soviet émigrés.” As in the job market, these Americans often found their papers crowded out of the limited number of available journal publication slots, especially in the premier journals, and their ideas crowded out of the center of attention in their field.

“Even if the ideas of a highly qualified single worker spill over to other workers with whom they interact, the overall effect of the interaction can still be deleterious to the productivity of other workers,” the article concludes. “In particular, in a world with constraints on the funding and dissemination of ideas (e.g., a limit on the number of faculty slots, or, more abstractly, a limit on the attention span of the potential audience), large and sudden increases in the population of producers of knowledge can result in diminishing marginal productivity for a pre-existing worker,” it continues.

“Even if the supply of jobs wasn’t fixed and the tenure system wasn’t in place, the fact of the matter is that the leaders of the field couldn’t really have consumed all this new stuff. … There are only so many hours in a day,” Borjas says. What’s more, “the loss of a generation of American mathematicians led to the loss of a lot of theorems that perhaps could have been important. We’ll never know that.”

But we do know that “constraints … limit the amount of rubbing-off effect that you can actually enjoy,” he goes on. “You could fill up the world with new ideas, but they all can’t really be consumed” in the available time and attention. In terms of supply and demand, Borjas says, “Producing theorems is no different than producing widgets.”

Implications

Borjas and Doran staunchly decline to discuss their work’s implications for immigration policy. And Borjas emphasizes that “you can’t take this paper and say it will apply in every single context in the world” because the data reveal different effects from the Russian mathematical influx depending on local factors.

At least one conclusion, however, seems clear. It is not the raw numbers of researchers but the institutional arrangements in which they work that shape the course of productivity and innovation. The American academic system, whose main features have changed little since the 1990s, could not take optimal advantage of either the new ideas or the new people. Instead, the system irreversibly wasted both talent and possibility and altered the incentives that influenced talented young Americans’ career decisions.

This does not, however, in itself argue against welcoming highly skilled immigrants and their ideas. Instead, it argues for crafting science labor force policies based on the real effects that they are likely to produce, not on the ones that people hope for or imagine.

Beryl Lieff Benderly writes from Washington, D.C.

Beryl Lieff Benderly writes from Washington, D.C.

10.1126/science.caredit.a1200024