The causes of the long stagnation in Japan

Abstract

The paper investigates whether the Japanese bank lending causes the long stagnation in the 1990s and if so whether this effect on the growth is more persistent than in the 1980s. Applying a VAR model for the annual prefecture panel data, the former can be verified by Granger causality test and the latter by impulse response function. There exists only one way causality from the loan to the GDP in the slump periods, while two way causalities exist in the 1980s. The shock in the loan equation is less persistent than the shock in GDP in the 1980s, but the persistence is reversed in the 1990s.

Acknowledgement

An earlier version of this paper was presented at the 2001 Applied Regional Science Conference, at the 2002 seminar in Honam University, and at the 2002 Western Economic Association International Conference. The authors are grateful for helpful comments from Sangho Kim, Komei Sasaki Toshihiko Miyagi, Benjamin N. Dennis, and many participants. The research of Miyakoshi was supported by Grant-in-Aid 13630104 of Ministry of Education, Culture, Sport, Science and Technology of Japan.

Notes

¹ In return for the injection, the banks issued the government with preferred stocks and subordinated bonds. See Deposit Insurance Corporation of Japan (2001, pp. 64–66).

² The Domestically Licensed Banks lent 139 trillions Japanese yen in the 1980 fiscal year, 448 in 1990 and 446 in 2000, while the Shinkin banks made 37 trillion yen in 1980, 60 in 1990 and 66 in 2000 and the Credit Cooperatives did 1 trillion yen in 1980, 18 in 1990 and13 in 2000.

³ The mutual loan and savings banks began to be gradually converted to the regional banks II after the Bank Acts of Japan was enforced in February 1989. The conversion had been completed until April 1992. The data for the regional banks II are the ones of the mutual loan and savings banks until February 1989.

⁴ The missing of the Deflator for Fukushima, Saitama, Toyama, Hyogo, Okayama, and Okinawa prefectures in several years lead to omit these prefectures from the analysis.

⁵ Though Levin and Lin (1992) and Im et al. (1995) introduced the methods for treating the panel unit root test, the methods might not be established in the published articles yet. In this paper, it is supposed that there is no unit roots in the data set after removing the piecewise linear trend from the original panel data.

⁶ Following Aoki and Patrick (1994, pp. 3–4), a main bank relationship is a long-term relationship between a firm and a particular bank from which the firm obtains its largest share of borrowings. This bank is called a main bank. In Japan the main bank plays an important role of corporate monitoring, governance and protection against the take-over for the firms. In spite of rapid securitization of corporate financing, the main bank still remains to be the core stockholder for firms and so there is no sign of the development of an active take-over mechanism. The main bank system may be worthy of being considered as one of the reasons to explain the results. These tasks are left to future research.