Productivity Growth and Potential Economic Growth of Japan: Analysis Based on the JIP2006 Database

The JIP2006 is comprised of various kinds of annual data sets necessary for the estimation of sectoral TFP covering the entire Japanese economy for each of 108 industries. These data sets include annual data on capital stock and capital costs by industry, labor inputs by attribute (gender, education, age, etc.), gross output and intermediate input, and so forth for years between 1970 and 2002, as well as accompanying tables such as those showing indices for trade and deregulation.

Among the members of our project team, those primarily engaged in the development of JIP2006 include: Tsutomu Miyagawa, Professor at Gakushuin University, and Tomohiko Inui, Professor at Nihon University, in charge of capital-related data; Hyeog Ug Kwon, Associate Professor at Nihon University, and Kyoji Fukao in charge of input-output tables; Joji Tokui, Professor at Shinshu University, in charge of labor-related data; Yasuo Nakanishi, Professor at Senshu University, and Keiko Ito, Lecturer at Senshu University, in charge of accompanying tables; a number of other researchers plus about 10 graduate school students. It is quite rare both in Japan and elsewhere to have such a substantial lineup at work on the estimation of industrial productivity. I believe that the generous input of resources - enviable even in the eyes of researchers in the United States and Europe - has enabled us to make very precise estimations.

The estimation work received full support from the Economic and Social Research Institute (ESRI) of the Cabinet Office, under the auspices of which we had carried out estimations for the preceding version of the JIP database (JIP2003), as well as from the Hitotsubashi University 21st Century COE (centers for excellence) Program on "Statistical Analysis in Social Sciences." I would like to express our deep appreciation for their support.

The key characteristics of the JIP2006 include the following:

1. Both the database itself and its underlying base data are defined as public property and thus, in principle, will be made available to the general public (forthcoming on the RIETI website);
2. The data contained in the JIP2006 are consistent with the 93SNA and national economic accounts are used as a control total; and
3. The JIP2006 will participate in the EU KLEMS project, which has been estimating TFP separated into 72 industries for EU major member states, the U.S., South Korea and so forth, thereby enabling international comparison of productivity levels in absolute terms including with those of Japan.

The following table and chart were assembled using data contained in the preliminary version of the JIP2006.

Table 1 shows the growth accounting of the Japanese economy as calculated based on data contained in the preliminary version of the JIP2006. In this table, the real growth rates of past gross domestic product (GDP) were calculated using Laspeyres-type chain-linking methodology. Thus, the resulting estimates of growth rates of real GDP and TFP growth in the more distant past, for instance the 1970s, are higher than those resulting from calculation using fixed-base methodology in which prices of relatively recent years are fixed as base-year prices and long-term retrospective estimates are based on such prices. This is because the fixed-base methodology underestimates the importance of the electronics machinery industry in the 1970s in terms of its rapid growth and remarkable role in improving TFP; the current prices of electronics parts, for instance, are far cheaper than in the 1970s and thus, the industry's contribution to TFP improvement in the past is substantially diluted when calculated based on the current price levels. We estimated the time series of chain-linked real GDP on our own because the Cabinet Office provides such data only for years from 1994. The results of this growth accounting show that the sharp fall in the TFP growth rate began in 1990. The average TFP growth rate for years from 1970-1990 was reaching a level of 1.69% per annum. A rise in TFP would accelerate economic growth by pushing up a return on capital (ROE) and thus promoting capital accumulation. I will not go into the details of our analysis here but our findings suggest that the recovery of TFP growth to an annual 1.5% would raise Japan's potential economic growth rate to around 2% per annum.

Chart 1 shows the average annual growth rates of TFP for 106 industries (excluding housing and activities not elsewhere classified). From this chart, we can see that TFP growth rates differ considerably from one industry to another. A notable rise in productivity was observed in industries producing information and communication devices such as semiconductor devices and integrated circuits (ICs), computers and computer accessories, and electronic parts and components, as well as in the pharmaceutical industry. Meanwhile, within the service sector, the insurance and telecommunications industries posted high growth. These facts imply that Japan will be able to achieve and enjoy high TFP growth rates and the resulting steady economic growth if the country successfully promotes a further shift to high-tech industries. (For more details, see the JIP2006 database soon to be posted on the website.)

Other than the JIP2006 database, our project team has also been working on the development of a JIP micro-database of individual company- or establishment-level information covering almost the whole Japanese economy, whereby we are undertaking research to find out what factors caused the stagnation of TFP growth in the 1990s that was observed particularly in the manufacturing sectors, and what factors have been hindering the productivity growth of certain non-manufacturing industries whose productivity is said to be extremely low compared to that of their overseas counterparts.