This paper employs a new empirical approach to estimate the impact of subsidies on growth and productivity. Our key innovation is to use local political leader geographic rotation as a source of exogenous variation. By using Chinese Industrial Census data from 1999 to 2013, we find that more subsidies have a positive effect on growth but not on productivity. Further firm-level results suggest that the size expansion of firms to win subsidies might be the mechanism, a new spillover channel, in explaining our main finding.