This paper examines impacts of industrial agglomeration and foreign direct investment (FDI) on productivity of Taiwanese firms. Firms may be driven by essential motives of economic agglomeration originally found in Alfred Marshall (1920), such as neighboring with customers or suppliers, labor market pooling and knowledge spillovers, to locate in heavily inhabited clusters. Potential flows of employees are connected with flows of knowledge and rising labor productivity since knowledge may be labor-embodied. Alternatively, recent empirical studies indicate that through patent citation knowledge spillovers may leave paper trail and decay with distance. In the same vein, these firms tend to perform FDI to benefit from cross-country knowledge spillovers. A conceptual framework of domestic location choices and FDI decisions is provided in order to specify underlying links between agglomeration externalities, global knowledge sourcing and labor productivities. According to Holmes (1999), this paper constructs two indicators of industrial agglomeration to appraise domestic agglomeration externalities in terms of labor market pooling and technological spillovers since we are short of input-output data. Comprehensive firm-level data from the Taiwanese Survey and Census of Manufacturing Operation are used to measure the two industrial agglomeration indicators across different villages and 4-digit Standard Industrial Classification (SIC) industries. The Taiwan Economic Journal (TEJ) reports annual FDI amounts, sales, ages, SIC codes, employees, R&D expenditures, and other firm characteristics of companies listed on Taiwan Stock Exchange and Gre-Tai Securities Market. With the TEJ data of 1,362 firms and two agglomeration indicators, we setup a cross-sectional econometric model to empirically assess the impacts of industrial agglomeration and FDI on firms’ performances. We expect to identify them as major determinants of labor productivity of Taiwanese firms.

Keywords: industrial agglomeration, knowledge spillovers, labor productivity.

JEL Classification: L25, R30