Compilation of the Regional-Level Japan Industrial Productivity Database (R-JIP)

This paper describes the construction of the Regional-Level Japan Industrial Productivity Database (R-JIP), which covers the period from 1970 through 2008 and comprises annual industrial output and factor input data of Japan's 47 prefectures classified by 23 industry. Output is measured in terms of value added, while the input data consist of labor input and capital service input, which take both (time series) quality changes and (cross-sectional) quality differences into account. This database makes it possible to calculate both changes in TFP over time and TFP differences across prefectures.

Using the data, we analyze Japanese prefectural labor productivity differences and changes therein from 1970 onward. The main findings of our analysis are as follows. (1) Not only differences in capital-labor ratios, but also differences in TFP and labor input quality play a part in explaining the differences in prefectural labor productivity. In 1970, both differences in capital-labor ratios and TFP were the main sources of prefectural labor productivity differences. However, thereafter, the contribution of differences in capital-labor ratios rapidly declined, while the contribution of differences in TFP continued to play a role. Differences in labor input quality account for only a small part of prefectural labor productivity differences throughout the period. (2) Looking at different industries, we find a marked contrast between the manufacturing and non-manufacturing sector. In the case of the manufacturing sector, both the expansion of human capital-intensive industries in the less-developed rural areas and decreases in TFP gaps across prefectures within each industry contributed to the observed decline in prefectural labor productivity differences. On the other hand, in the non-manufacturing sector, both capital and human capital-intensive industries - such as real estate, private services, and transportation and communication - are concentrated in more advanced areas like Tokyo, where labor productivity has been high from the start. This contributes to the prevailing prefectural labor productivity differences. (3) From the viewpoint of growth accounting, increases in the capital-labor ratio have generally been higher in the lagging prefectures, which have worked in the direction of reducing labor productivity differences. Generally speaking, the same can be said for changes in labor input quality. In sharp contrast, TFP has generally been growing more quickly in the leading prefectures, thus working in the direction of increasing labor productivity differences.