صرفه جویی مقیاسی، تغییر فنی و رشد بهره وری در خدمات مخابراتی محلی ژاپنی

This paper aims to examine the cost structure of local telecommunications services in Japan to provide measurements of productivity growth and a decomposition of this growth. We estimate a variable cost function using panel data on eleven regional telecommunications sectors of NTT for the 1992–1997 period. Total factor productivity (TFP) is measured on the basis of the estimated parameters, and the TFP growth rate is then decomposed into three sources: scale economies, technical change and capital adjustment. Our results reveal that the contributions of technical change to productivity growth are greater than those of scale economies and capital adjustment.

This paper aims to examine the cost structure of local telecommunications services in Japan to provide measurements of productivity growth and a decomposition of this growth. We estimate a variable cost function using panel data on eleven regional telecommunications sectors of NTT for the period 1992–1997.1 Total factor productivity (TFP) is measured on the basis of the estimated parameters, and the TFP growth rate is then decomposed into three sources: scale economies, technical change and capital adjustment. Our results reveal that the contributions of technical change to productivity growth are greater than those of scale economies and capital adjustment. We also find out that sluggish adjustment of capital prevents the fast growth of productivity for some regional sectors. This paper has some policy implications for the price regulation of incumbent local exchange carriers. Like British Telecom in the UK, the regional sectors of NTT have been subject to the price-cap regulation in the form of RPI-X, where RPI and X are the inflation rate and the expected rate of productivity growth, respectively.2 Thus, the increase in the average price of local services must be kept to less than an increase in RPI-X. In implementing the price-cap regulation, the level of the X factor is debated because a high X factor may jeopardize the firm’s survival while a low X factor may fail to provide incentives for cost reduction. The analysis in the present paper suggests that the current X set by MPT is comparable to our estimates of TFP growth if its slowdown tendency is taken into consideration.3 Although the telecommunications industry has received increasing attention worldwide, there are only a few empirical studies on cost structures and productivity of Japanese telecommunications. To the best of our knowledge, these include Oniki et al. (1994) and Sueyoshi, 1996 and Sueyoshi, 1997. Oniki et al. (1994) investigated the productive performance of NTT during 1958–1987 using a framework similar to that of the present study. Their primary concern was the effects of the 1985 reform by which NTT was privatized and the telephony market was liberalized. Sueyoshi (1997) used the data envelopment analysis (DEA) to evaluate the 1985 reform by comparing NTT’s managerial efficiency before and after 1985. Using econometric techniques as well as DEA, Sueyoshi (1996) also examined the divestiture of NTT in 1999 by testing whether the cost structure exhibits subadditivity or economies of scope. The previous studies are mainly concerned with the managerial form of NTT and thereby deal with its entire cost structure. Unlike those studies, we focus on the local services, for which issues of regulating incumbents provoke considerable controversy. We intend to shed light on the possibilities of fostering competition in the local telecommunications market in Japan. The paper is organized as follows. Section 2 briefly reviews the variable cost function model to be estimated. Section 3 discusses the measurement of TFP and its decomposition analysis. Section 4 explains the data construction. Section 5 reports and discusses the empirical results. Section 6 provides concluding remarks.

We have examined the productivity growth of NTT regional sectors for the 1992–1997 period. The variable cost function is estimated to measure TFP and decomposition analysis is conducted. Technical change is revealed to be the dominant contributor to growth in TFP. This is contrary to a prevailing belief that technological progress does not readily lead to reduced costs in local services because a large part of their costs arise from the labor intensive activities of maintenance. We also find evidence for the constant returns to scale of technology in the long-run. The full equilibrium is supported by 20 observations while the fixity of capital is confirmed in the other observations. The sluggish adjustment of capital lessens the growth rate of TFP in some regional sectors servicing rural areas. Among these findings, both the possibilities of technical progress and non-increasing returns to scale favor a further promotion of competition in the local service market. However, our results also suggest that the technical progress tended to decline as digitization of the network reached completion. Thus, it becomes important for policymakers to comprehend the effects on productivity growth of introducing new technology such as fiber-optic cables and internet telephony. Future research is needed to investigate this issue with extended data after the divestiture of NTT in 1999. In addition, this paper has shown that measurement of TFP growth provides a useful reference for regulators to determine the required level of improvement in productivity. We again need further observations to explore the performance of incentive schemes introduced as major steps for the reform of the Japanese telecommunications industry.