Reassessing the Firm Selection Hypothesis: New Evidence from Chinese Highways

This paper re-investigates whether larger cities eliminate more low-productivity firms, the so-called firm selection hypothesis. We exploit a huge boom of infrastructure construction in China during 1998-2007. We find that in coastal provinces, when we compare large cities to small cities which are not connected by controlled-access highways, the firm selection is apparent; however, comparing large cities to small ones which are connected by highways, the firm selection disappears. This result suggests that market size is dictated not so much by geography (city boundaries), but rather by transportation costs. The estimated effects of firm selection are robust to the potential endogeneity of highway construction. Evidence for firm selection is absent in inland provinces, perhaps because the market economy was relatively poorly developed in those areas during our study period.