The Effect of Imported Intermediate Inputs on Firm Performance: Firm and Establishment Level Evidence from Japan

Author Name KIM YoungGak (Senshu University) / INUI Tomohiko (Faculty Fellow, RIETI)
Creation Date/NO. March 2019 19-J-016
Research Project East Asian Industrial Productivity
Download / Links


Under the recent rapid development of production fragmentation across countries, there is a growing number of studies that examines the effect of international outsourcing and offshoring decisions on the domestic operations of firms. According to the Global Value Chain Index (GVCI) developed by OECD, the participation rate of Japanese firms in global value chains increased from 29.3 % in 1995 to 47.7% in 2007. RIETI-TID database shows that Japanese parts and product imports from China increased from 5.8 billion US dollars in 2000 to 25.0 billion US dollars in 2016.

This paper examines the effect of increases in imported intermediate input use goods on firm outcomes, especially exports. First, we examine the effect on firm productivity, and we find a positive relationship between increases in imported intermediate input use goods and productivity level. When we separate the analysis by import source regions, a positive relationship is observed from imports from North America and Europe, but no relationship is found in the case of imports from China. These results indicate that firms benefit from the technology spillover effects from the advanced countries by importing from them. Our analysis also indicates that the increase in imports supports the improvement of resource allocation within firms.

Next, we analyze the effect of the increase of imported inputs on both extensive margin and intensive margin of firm exports, controlling for the productivity of firms and establishments, finding a positive relationship in both cases. These results imply that imported inputs have a positive impact on firm exports both through productivity improvement and reduced intermediate input price. We also examine the effect of imported inputs on the number of employees, but we find no significant effect.