What it Takes to Improve Productivity: Labor input quality as a factor behind regional inequality

Faculty Fellow, RIETI

The role of productivity as a driving force of macroeconomic growth needs no elaboration. This article will stress the importance of productivity as a factor behind regional inequality. While the analysis of productivity as a driving force of economic growth focuses on productivity growth, the analysis of productivity as a factor behind the regional inequality will focus on cross-sectional comparison of regional productivity levels.

Usually, regional inequality is discussed on the basis of per-capita income including income transfer. However, the decisive factor of workers' income is productivity. In regions where productivity remains low, workers cannot climb the income ladder from a low bracket. As a result, youth population flows out in those regions, which is a situation known as a social decline in population. The social decline adds to the impact of the natural decline that is bound to occur in the future.

In order to accurately measure the productivity of an economy, it is necessary to use a consistent measurement method taking into account of the quality of capital and labor. It is not easy to apply such method for the measurement of productivity of regional economies given the data constraints. A research project at RIETI which I lead is undertaking this challenge.

The data developed for the analysis of regional-level productivity have been published on RIETI's website as the Regional-Level Japan Industrial Productivity Database (R-JIP). In this article, I will explain the findings obtained through the development and analysis of the database.

♦  ♦  ♦

The use of the database makes it possible to decompose prefectural differences in labor productivity into differences in the capital-to-labor ratio (calculated by dividing capital input by working hours), differences in labor input quality, and differences in total factor productivity (TFP), which reflect such factors as technological advances. Through the decomposition, we can obtain an overview of the factors behind regional differences in labor productivity. From the outcome of this decomposition, it is clear that the factors have dramatically changed compared with that from a half century ago.

Half of a century ago, in the 1970s, the period of high economic growth in Japan was approaching its end. Looking at prefectural labor productivity at that time, Kanagawa was on top, followed in order by Chiba, Osaka, and Tokyo. The main source of the prefectural differences in labor productivity was the differences in the capital-to-labor ratio. For example, Kanagawa's top position reflected the presence of the Keihin industrial area, where major heavy chemicals industrial complexes were concentrated.

Meanwhile, a similar decomposition of the prefectural differences using the data for 2010 shows a quite different picture. Tokyo ranks first with much higher labor productivity than other prefectures. The differences between the 46 prefectures excluding Tokyo have apparently narrowed, but because of this, Tokyo's outstanding strength in labor productivity is all the more pronounced.

The source of the prefectural differences has also changed dramatically. For example, Tokyo's capital-to-labor ratio is lower than the national average. Then, which factors have boosted Tokyo's labor productivity? They are the higher TFP and labor input quality.

TFP is calculated by subtracting from output the values obtained by multiplying factor inputs by the corresponding marginal productivity. In principle, a similar approach is used in looking at growth in time series, or in cross-sectional comparison of the productivity level as discussed above, although the calculation method may be different. This means that identifying TFP as one of the sources behind the differences in labor productivity is not sufficient to fully explain the differences.

♦  ♦  ♦

Therefore, taking advantage of the intrinsic property of the R-JIP database which enables industry-level analysis, we examined in which industries the TFP gap contributes significantly to the prefectural differences in labor productivity. As a result, we found that the regional differences in TFP in the private services industries and the wholesale and retail industries have recently had a significant impact.

Some readers may find this finding somewhat surprising given the conventional wisdom that while the manufacturing industry has achieved dramatic innovation and productivity improvement due to exposure to international competition, the services industries and the wholesale and retail industries have not made notable productivity improvement.

However, it should be noted that what matters here is the regional productivity gap, rather than productivity growth. It is also necessary to take into consideration of the fact that value added by the manufacturing industry account for 30% and its share of workers is even smaller. Moreover, the service industries covered by this analysis include areas where the quality of human resources is particularly important, such as research and development, and information services.

In calculating in which industries the differences in TFP contribute to the differences in overall labor productivity, not only the size of regional TFP differences in each industry, but also each industry's share in each regional economy matters. Even though the regional differences in TFP within individual industries such as services and wholesale and retail may not be very large, those industries' large shares in each regional economy may have a significant impact.

Another factor contributing to Tokyo's relatively high labor productivity is the high quality of workers. In order to look at this from a different angle, we prepared a figure comprised of a horizontal axis that represents the labor input quality indicator and a vertical axis that represents labor productivity, with the 47 prefectures distributed in positions relative to Tokyo, which is represented by the value 1 on each axis (See the chart). This figure is based on the R-JIP database 2017, which covers the data for 2010.

Figure: Correlation between the Level of Labor Productivity and the Relative Labor Input Quality Indicator by Prefecture (2010)
Figure: Correlation between the Level of Labor Productivity and the Relative Labor Input Quality Indicator by Prefecture (2010)
Note: The value for Tokyo is used as the base figure of 1 on each of the vertical and horizontal axes.
Source: R-JIP Database 2017

The labor input quality indicator was developed on the basis of data obtained from the national census in the corresponding years with respect to the mix of attributes (gender, age group, academic achievement and occupational position) of workers classified by prefecture and industry and on the assumption that the differences in labor productivity are closely related with relative wages across those attributes. For example, if the labor productivity quality indicator shows a value 20% higher than the average, it is assumed that there is a labor input in terms of efficiency unit that is 20% higher per unit of working time.

This figure indicates a positive correlation between the labor input quality of workers and the level of labor productivity in the same region. Of course, this alone does not prove a causal effect. It is possible either that highly productive industries are concentrated in a region because high-quality human resources are available there, or that high-quality human resources are attracted to a region because highly productive industries are concentrated there.

The analysis of the correlation between the labor input quality and the level of labor productivity over the years shows that the regional differences in the labor input quality have narrowed over the past half century, while the slope of positive correlation has become steeper upward more recently. This finding, which is consistent with the results of the above mentioned decomposition of the differences in labor productivity by factor, suggests that the location of human-capital-intensive industries, many of which are classified as services industries, has come to have an increasingly strong influence on the regional differences in productivity.

♦  ♦  ♦

What lessons can be drawn from the above mentioned findings in order to achieve regional revitalization? From my experience of living in a provincial region, it is not unusual for me to hear about "manufacturing promotion plans" and "industrial zones," but it is not often that the broad array of services industries whose share of the entire economy has grown over the past half century becomes a topic of discussion about local economic development.

It is true that as the simultaneity of consumption and production applies widely to service industries, the benefits of agglomeration due to population size work in many cases. However, in this internet era, there are chances to overcome a disadvantage in terms of population size, and if struggling regions are to achieve a dramatic turnaround, the key will be human resource development. In particular, now that the human-capital-intensive nature of industries has become prominent, the presence of personnel with advanced skills is critical for regions to become locations for highly productive industries. As it is not easy for many regions to attract personnel with advanced skills across regional boundaries, they must foster such personnel on their own.

>> Original text in Japanese

* Translated by RIETI.

August 24, 2017 Nihon Keizai Shimbun

September 22, 2017