A large number of overseas elites were brought back to China by the policy in the past decade. However, name disambiguation defied investigations on the relationship between their mobility and research performance. By taking advantage of the ORCID website and applying causal inference strategies, we investigated 2489 China-connected scientists’ academic performance in the Web of Science database in terms of their job mobility, including 1388 scientists who moved to China the treatment group, and 1101 scientists with a possibility to move to China the control group. The results show that first, scientists moving to China have a new growth pattern where both their productivity and the rates of being corresponding authors in publications grew more rapidly than before; however, they made fewer contributions to the four top journals, Nature, Science, Cell, and PNAS. Second, the research performance of the scientists is affected by the time of mobility towards China, the countries from which they moved, and the disciplines of their publications. Last, China now maintains symmetrical inflow-outflow patterns with most countries, especially developed countries in Europe and North America, with only a few exceptions (e.g., Pakistan).