ABSTRACT
With Japan’s declining population, improving productivity is important to achieving continuous regional economic growth. This study focuses on the regional effects of population agglomeration and accessibility on total factor productivity (TFP). Empirical analysis shows that population agglomeration contributed to TFP growth, most significantly in the Greater Tokyo area. The interaction (flow) of people was examined and the importance of high-speed transportation network for TFP growth was clarified as the effect of accessibility on TFP growth. Population agglomeration and the development of transportation network are thus important strategies for growing regional economies.
Acknowledgements
I would like to thank reviewers whose comments have improved the quality of this study.
Disclosure statement
No potential conflict of interest was reported by the authors.
ORCID
Akihiro Otsuka http://orcid.org/0000-0001-7852-4590
Notes
1 In Japan, the effect of interregional spillovers is regarded as one of the economic effect of the interaction (flow) of people. ‘Interaction’ is a term used in the 2015 edition of the Japanese National Spatial Strategy. Originally, the term meant ‘movements based on temperature differences within a fluid’. However, in the National Spatial Strategy, the term is used to mean ‘the active movement of information, money, goods, and people between regions produced by the cooperation of regions with diverse resources’. The basic concept of the National Spatial Strategy is the formation of ‘spaces promoting interaction’. These spaces facilitate the dynamic occurrence of interaction in every part of the country.
2 It is assumed that fluctuations in can be attributed to a change in capital utilization, as well as a time trend. Based on this assumption, a proxy of the capital utilization rate can be measured with a residual error term () in the regression , where is a time trend and is the time-invariant slope of .
3 We can calculate the interregional travel time by using NITAS (National Integrated Transport Analysis System). The NITAS specifies the departure and destination sites and calculates the ‘total time required’, including time spent in transport, time spent transferring between modes of transport, and time spent waiting.
4 For models estimated by GMM, we can compute the first-order and second-order serial correlation statistics proposed by Arellano and Bond (Citation1991) as one method testing for serial correlation. The test actually returns two separate statistics, one for the first-order correlation and one for the second-order correlation. If the innovations are independently and identically distributed (i.i.d.), we expect the first-order statistic to be significant (with a negative autocorrelation coefficient), and the second-order statistic to be insignificant. The m-statistics are calculated aswhere is the average jth order auto-covariance.