Iris Claus: In this paper, Hong Gao, Ming Lu, and Hiroshi Sato quantify the impact of city scale, which is measured by cities’ population size, on employment in China using probit model estimations and individual level survey data from 2002 and 2007. To control for a potential endogeneity bias in the estimations, cities’ population growth rates between 1953 and 1982 are used as instrumental variables. The potential endogeneity bias arises from two main sources. First, employment could affect city scale because individuals may locate where they are more likely to find work. Second, unobservable shocks may simultaneously affect city scale and employment. Using cities’ population growth rates between 1953 and 1982 overcomes the endogeneity problem because voluntary relocations were prohibited during this period because in China's planned economy the government strictly controlled migration.
There are several reasons why city scale may affect employment outcomes. For instance, the economic geography literature...