Analysis of the Characteristics of the Innovation System for the JSIC Service Sector: A measurement of innovation factor sensitivities using a text mining method

The vitalization of the service sector is one of the major economic issues in Japan, where the service sector accounts for more than 70% of GDP and concerns exist of a reduction in potential growth through population decline and the weakening of global competitiveness. However, empirical studies on the innovation factors and productivity in the service sector have faced significant difficulties due to limited data availability (e.g. poor statistical data) and problems of the sectoral classification (e.g. existence of establishments with varying characteristics of the innovation system in each sector). This paper challenges the latter problem by making categories based on the characteristics of the innovation system for the Japan Standard Industrial Classification (JSIC) service sectors (4-digit). Applying a text mining method, we quantitatively measure sensitivities of three innovation factors—(1) knowledge and technologies, (2) actors and networks, and (3) institutions following the concept of "sectoral systems of innovation" by F. Malerva—and reveal the characteristics of the innovation system which each service sector has. We also classify the JSIC service sectors (4-digit) into eight categories that have specific innovation potential according to the proper innovation system obtained from the revealed characteristics. In addition, our study finds a positive correlation between promoting innovation factors ((1) and (2)) and long-term performance (production and total factor productivity (TFP) calculated by the Japan Industrial Productivity (JIP) database), and that innovation factor (3) controls the performance.