Working Paper: NBER ID: w21096
Authors: George J. Borjas; Kirk B. Doran; Ying Shen
Abstract: The largest and most important flow of scientific talent in the world is the migration of international students to the doctoral programs offered by universities in industrialized countries. This paper uses the opening of China in 1978 to estimate the causal effect of this flow on the productivity of their professors in mathematics departments across the United States. Our identification strategy relies on both the suddenness of the opening of China and on a key feature of scientific production: intra-ethnic collaboration. The new Chinese students were more likely to be mentored by American professors with Chinese heritage. The increased access that the Chinese-American advisors had to a new pool of considerable talent led to a substantial increase in their productivity. Despite these sizable intra-ethnic knowledge spillovers, the relatively fixed size of doctoral mathematics programs (and the resulting crowdout of American students) implied that comparable non-Chinese advisors experienced a decline in the number of students they mentored and a concurrent decline in their research productivity. In fact, the productivity gains accruing to Chinese-American advisors were almost exactly offset by the losses suffered by the non-Chinese advisors. Finally, it is unlikely that the gains from the supply shock will be more evident in the next generation, as the Chinese students begin to contribute to mathematical knowledge. The rate of publication and the quality of the output of the Chinese students is comparable to that of the American students in their cohort.
Keywords: Chinese graduate students; productivity; mentoring; ethnic complementarities; knowledge spillovers
JEL Codes: D83; J24; J61; O31
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| influx of Chinese graduate students (F29) | productivity of Chinese-American advisors (D29) |
| mentoring more Chinese students (A29) | productivity of Chinese-American advisors (D29) |
| Chinese-American advisors mentoring more Chinese students (C92) | number of papers published by Chinese-American advisors (A14) |
| increase in productivity for Chinese-American advisors (O49) | decline in productivity for non-Chinese advisors (F29) |
| quality of output from Chinese students comparable to American students (D29) | benefits of the supply shock may not extend into future generations of mathematicians (D29) |