WHICH AGGLOMERATION EXTERNALITIES MATTER MOST AND WHY?
基于73篇科学文章，重新审视专业化、多样性和竞争等集聚外部性对创新、生产率和城市就业增长的影响，发现研究异质性很大且存在发表偏倚，建议未来采用微观数据和动态模型等新方法。
Abstract This paper revisits the ongoing discussion on the importance of agglomeration externalities – specifically specialization, diversity and competition effects – that may contribute to innovation, productivity and urban employment growth. Previous meta‐analyses suggested that the evidence on agglomeration externalities is strongly context‐specific. Expanding an earlier analysis of 31 articles, we seek to draw in this paper more robust conclusions by means of the statistical evidence for agglomeration externalities presented in 73 scientific articles, all building on the seminal work of Glaeser et al . (1992). Our results confirm that the heterogeneity among studies is huge and can only be partially accounted for by means of an ordered probit analysis. Additionally, some evidence of publication bias is found. We conclude that the conventional lines of inquiry in this literature may now have reached strongly diminishing returns. New lines of inquiry, using rich micro‐level data on firms and workers, dynamic general equilibrium models at the macro level, more attention for spatial and temporal variation in the impacts of agglomeration, and further investigations into the spatial scope of externalities are promising avenues for further research that can enhance our understanding of how agglomeration externalities continue to fuel our increasingly urbanized world.