Explaining the productivity advantages of manufacturing firms in Russian urban agglomerations

This paper empirically analyzes the agglomeration-related productivity premium at the enterprise level of the manufacturing industry in Russia. A settlement is counted as part of an urban agglomeration in two cases: that of a large, central city and that of a town located within 50 kilometers of the central city. Data obtained from a 2009 manufacturing enterprise survey conducted by the Higher School of economics are used, along with linked data on hosting regions and cities. We employ a multilevel model, which allows us to consider firm, urban and regional heterogeneity and test two possible explanations of the productivity advantages of firms in urban agglomerations - own-sector and all economic activity concentration in the city and the surrounding region. In addition we check if agglomeration productivity premium increases if the city is located within the region opened for international trade. The results suggest that Russian plants in urban agglomerations enjoy 17-21% higher labor productivity. This gain arises as a result of urbanization and external scale economy - the agglomeration of firms belonging to different industries at both the urban and the regional levels of analysis. We also found that productivity gained from urban agglomeration is the highest in towns with populations of 100,000 to 250,000 people. Localization and clustering - the own-sector concentration of plants in the city - is not associated with higher labor productivity. The structure and size of the surrounding economy always matter: in contrast to urban clusters, regional own-industry clustering satisfactorily explains the productivity premium, suggesting that efficient clustering requires a scale economy larger than only a city. If the region is open to trade, has a good road network, attracts migrant workers from other regions and has a mature market with relatively high living standards, urban agglomeration effects will be much stronger. All of our results are robust to changes in estimation technique, sample structure and choice of spatial objects. We conclude with the assumptions for the policy framework. Our results do not imply that the maximization of agglomeration numbers can resolve the issue of low productivity in manufacturing. Rather, we have demonstrated that agglomeration mechanisms rely on urban diversity, scale and openness in the economy. Indeed, these dimensions deserve stimulation through the use of regional policy measures.