R-JIP Database 2017

On the Regional-level Japan Industrial Productivity (R-JIP) Database 2017

Depopulation and aging progress more rapidly in rural areas in Japan. The outsourcing trend of manufacturing work has hit these areas more profoundly. All of these occurrences cause serious concern about the widening of economic inequality among regions and the sustainability of local communities. The basic element underlying these economic discrepancies is productivity disparity among regions.

The standard method of productivity measurement currently used is the KLEMS approach, which tries to measure the output and various inputs of production (such as capital stock services and quality of labor inputs) for each industry as accurately as possible, and generates total factor productivity (TFP). But this approach is rarely utilized in the analysis of regional productivity disparity mainly because of a lack of ready-made necessary data.

With this aim, we started the Regional-level Japan Industrial Productivity (R-JIP) project in 2011 in collaboration with the Institute of Economic Research, Hitotsubashi University. Our project released the first version of the R-JIP database in 2012. The R-JIP 2017 is the third revision of the database.

The R-JIP database compiles value-added output in current and constant prices, quality-adjusted labor input, and quality-adjusted capital input for all 23 industrial sectors of the 47 prefectures. These data are constructed to add up to the control total of corresponding data in the national-level JIP database. Using these data, we calculate both the differences in TFP level among prefectures for each industry and the rate of change in TFP for each industry of each prefecture. In the R-JIP 2017, these data are available for every year from 1970 to 2012.

In order to simplify the task of compiling regional-level data, we compress the number of industrial sectors from over 100 in the national-level JIP to 23 in the R-JIP. Regarding counting output, only value-added output is available in the R-JIP compared to both gross output and intermediate inputs in the JIP. In spite of this data squeeze, our R-JIP database is unique in providing a regional-level dataset which enables productivity comparison among prefectures and taking into account differences in input quality.

Major revisions in the estimation procedures of the R-JIP 2017

The data period is extended from 1970-2009 in the R-JIP 2014 to 1970-2012 in the R-JIP 2017. The newly extended period covers 2011, when the Great East Japan Earthquake hit Japan. This fact necessitates the estimation of the amount of damaged capital stock in the earthquake-hit regions. In addition, we undertake the following two major revisions to our estimation procedures.

Increased availability of order-made data from the population census permits us to undertake revised estimation of labor-input quality for each prefecture. In calculating labor-input quality, we need the number of employed cross-classified by gender, age, education, status in employment, industry, and prefecture on a place of employment basis. These detailed data on a place of employment basis are only available as order-made data from the population census. At the time when we estimated R-JIP 2014, we could only utilize order-made data from the 1990 census and the 2000 census. In preparing R-JIP 2017, the order-made data from the 1980 census and the 2010 census became available, and we used them to obtain a more accurate estimation.

The revision of estimation procedures is also done to the fixed capital investment series for each industry of each prefecture, which are the base data for calculating capital stock and capital services. Up to R-JIP 2014, we had heavily relied on the Cabinet Office estimation of fixed capital investment series of each prefecture. However, after running some comparisons between the Cabinet Office data and the Census of Manufacturing data, we judged the latter is more appropriate for manufacturing industries. Therefore, in the R-JIP 2017, we use the Census of Manufacturing data for compiling manufacturing investment data. For broadly-defined non-manufacturing industries, we continue to utilize the Building Starts data from the Ministry of Land, Infrastructure, Transport and Tourism as a dividing share among prefectures excluding the electricity, gas and water industry, for which we can rely on its specific data. But the method of division to minute industry-classification is changed to avoid unnatural fluctuations in the R-JIP 2014 data.

Two trial calculations not reflected in the R-JIP 2017 data but still important

Our project not only conducts technical improvement of estimation procedures and renewal of the data mentioned above but also tackles more difficult problems associated with compilation of regional data. The System of National Accounts 2008 raises two challenges in the task of compiling regional accounts. One of the difficulties is related to the transactions of multiregional units between various regions. The other is the possibility that there may be significant variation in the prices of the same product across different regions. We undertake two trial calculations related to these challenges.

The first issue is on the input/output of headquarter services across different regions. The current Annual Report on Prefectural Accounts compiled by the Cabinet Office defines value-added without taking into account the input of headquarter services across different regions. The 46 prefectures except Tokyo abide by this definition. In contrast, Tokyo, where many large companies locate their headquarters, performs in its own way by calculating value-added output generated from providing various headquarter services to other prefectures. There is an obvious gap in the definition between Tokyo and other prefectures. To cope with this problem, we apply Tokyo's definition to all of the other prefectures by estimating the input of headquarter services across different prefectures [3].

The other issue is on the regional variation in service prices. Japan is not a geographically large country, and the prices of many easily-transported goods converge among regions by arbitrage transactions. However, many service prices do not hold the same property because they are produced and consumed at the same time in the same place. We apply the absolute purchasing power parity (PPP) method to the variety of service prices among prefectures to calculate price differences in service among prefectures. The result of this calculation and its effect on our estimate of productivity differences among prefectures are reported in another RIETI discussion paper [4].

Further information on R-JIP 2017

We are preparing a book (in Japanese, titled Regional Inequality and Productivity in Japan: An Analysis based on the R-JIP 2017 Database) that gives detailed explanation of the construction of R-JIP 2017 and its application to the Japanese regional productivity differences. The data construction part of the book will be summarized into a forthcoming RIETI discussion paper.

Until then, the below RIETI discussion papers [1] and [2] (their revised version can also be found in The Economic Review, Vol. 64, No. 3, July 2013, the Institute of Economic Research, Hitotsubashi University) provide basic information of the R-JIP 2012 database construction. There is no major change in the estimation procedures from R-JIP 2012 to R-JIP 2014. Furthermore, the R-JIP 2017 takes a step to major revision of data compiling procedures, which are briefly explained above (Major Revisions in the Estimation Procedures of the R-JIP 2017).

Our works related to R-JIP data that are written in English are as follows. [5] is a book on long-term trends in prefectural income inequality and its determinants in Japan. Chapter 6 of this book is the analysis based on R-JIP 2012, and Appendix 4 provides a detailed explanation of the data construction of the R-JIP 2012. We also have related RIETI Discussion Papers in English [6] and [7].

On the comparison between capital stock in the R-JIP and the social infrastructure

Capital stock in the R-JIP database, as well as the JIP database, is broadly defined in terms of whether its capital service is measureable as input to the production process of some specific industries. It does not depend upon its provider, namely, whether private sector or public sector. For example, there are many roads and irrigation canals used by the agricultural sector. Revenue collected from toll roads is the output of the transportation industry. The same holds true with water pipes and equipment of the water service industry. The public service sector uses such infrastructure as school buildings, cultural facilities, airports, and harbor facilities. All of these are included in capital stock in the R-JIP definition and are provided by the public sectors.

On the other hand, measurement of social infrastructure often includes all publicly provided capital stock, a typical example of which is the measurement by the Cabinet Office. When you analyze productivity of social infrastructure by region and use both R-JIP defined capital stock and social infrastructure measured by the Cabinet Office as inputs, double counting of inputs occurs, and you get false productivity.

To draw attention to this point, we define supplementary social infrastructure that is not included in the capital stock counted by the R-JIP database. The supplementary social infrastructure covers such infrastructure as general purpose roads without tolls, city parks, embankments for flood control, afforestation for disaster control, and breakwaters at seashore. We provide these data as a part of the R-JIP 2017 database.


Current members of the R-JIP project

Current members of the R-JIP Project are as follows.

  • Joji Tokui (Shinshu University, RIETI)
  • Kyoji Fukao (Hitotsubashi University, RIETI)
  • Sonoe Arai (RIETI)
  • Tatsuji Makino (Hitotsubashi University)
  • Takeshi Mizuta (Hitotsubashi University)
  • YongGak Kim (Senshu University)
  • Kazuyasu Kawasaki (Toyo University)
  • HyeogUg Kwon (Nihon University)
  • Kenta Ikeuchi (RIETI)

To our users

Construction of regional accounts such as the R-JIP database faces not only limited availability of data but also several difficult problems specific to regional accounting. In the R-JIP project, we address two of them and undertake trial calculations as above mentioned; the input of headquarter services across different prefectures is one issue and differences in service price among prefectures is another. Either of the trial calculations proves to confirm the significance of these issues. But after considering the additional burden to our task in the event that we reflect these issues to the R-JIP data regularly, we put these trial calculations aside of regular R-JIP data compilation. A detailed report on the effect of taking these two issues into account will be prepared as a separate paper. We will try our best to improve and extend our database henceforward.

You can freely download and use the R-JIP 2017 database and its supplementary data shown below. Please specify the data source as R-JIP 2017 in your paper and presentation and send one copy to us.

Contact us:
Service Productivity Project Room in the Institute of Economic Research, Hitotsubashi University

Download Data

R-JIP Database 2017 [XLSX: 7.1 MB]

Uploaded on April 24, 2018

This database compiles the following data series from 1970 through 2012 (calendar year) annually, for each industry of each prefecture.

  1. value added
    1. 1) real (million yen in 2000 price)
    2. 2) nominal (million yen)
  2. capital input
    1. 1) real capital stock (million yen in 2000 price)
    2. 2) cost of capital (million yen)
    3. 3) quality of capital (time-series index, 1 in 2000, same quality for each industry is applied nationwide)
  3. labor input
    1. 1) man-hour (number of employed persons multiplied by average yearly working hours, divided by 1000)
    2. 2) cost of labor (million yen)
    3. 3) quality of labor (time-series index, 1 in 2000, same quality for each industry is applied nationwide)
    4. 4) quality of labor (cross-sectional comparison index)
    5. 5) number of employed persons (persons)
  4. reference data series
    1. 1) population (persons)
    2. 2) per capita prefectural income (thousand yen)

Worksheet for the Analysis of Prefectural Differences in Labor Productivity [XLSX: 3.5 MB]

Uploaded on Sept 7, 2017

This worksheet uses the R-JIP 2017 data and analyzes prefectural differences in labor productivity. It decomposes prefectural differences in labor productivity into capital-labor ratio, labor quality, and total factor productivity (TFP).

Worksheet for Prefectural Growth Accounting [XLSX: 7.2 MB]

Uploaded on April. 24, 2018

This worksheet uses the R-JIP 2017 data and conducts prefectural-level growth accounting. It decomposes the annual rate of change in labor productivity from 1970 through 2012 into the rate of change in capital-labor ratio, labor quality, and total factor productivity (TFP).

Supplementary Social Infrastructure Data [XLSX: 266KB]

Uploaded on March 16, 2018

We provide supplementary social infrastructure data that are not included in the capital stock counted by the R-JIP database. It covers such infrastructure as general purpose roads without tolls, city parks, embankments for flood control, afforestation for disaster control, and breakwaters at seashore by prefecture. All data are in 2000 prices. The period of the data is from 1970 through 2012.

Long-term Prefectural Dataset for Productivity Analysis [XLSX: 145KB]

Uploaded on May 18, 2015

The Long-term Prefectural Dataset covers 1955, 1970, 1990, and 2008 data, which enables us to conduct longer period analysis. On the other hand, it lacks industry-classified data, and all data are aggregated to the total industry of each prefecture. Since this dataset is compiled based on the R-JIP 2012 database, it is not exactly consistent with the R-JIP 2014 data.