Public Policy Review, 2005, Vol.1, No.1
Structural changes in business fluctuations have been gathering attention in Europe and the US in recent years. It has become clear that business fluctuations in the US began to stabilize from the middle of the 1980s, and similar structural changes have been observed in Europe. On the other hand, there have been only a few studies concerning structural changes in Japanese business fluctuations. With this background, this paper presents an analysis as to whether or not there has been a structural change in Japanese business fluctuations in recent years, and if so, when and what kind of change. There are various econometric models for business fluctuations; a common one is the Markov switching model proposed by Hamilton (1989). In this model, it is understood that average growth rates differ between periods of expansion and periods of recession and the shifts between the period of expansion and the period of recession are formulated in accordance with the Markov process. Kim and Nelson (1999) expanded the Markov switching model of Hamilton (1989) considering structural changes and estimated the expanded model by using Bayesian estimation based on Markov chain Monte Carlo. In this paper, maximum likelihood estimation was performed by changing the points of structural change for each period in the model of Kim and Nelson (1999) and the period for which the likelihood becomes the highest is estimated as the point of structural change. The variables for Japanese business fluctuations used in this paper are composite index (CI) and the index of industrial production (IIP) published by the Economic and Social Research Institute of the Cabinet Office of the government of Japan. To focus on structural change in recent years, the sample period was March 1980 to November 2003. It was estimated that the points in time of structural change in business fluctuations were April 1989, based on CI, and January 1992, based on IIP. The analysis also revealed that structural changes were statistically significant for both of the variables. In more detail, average growth rates showed significant reductions for the recession period and for the expansion period, and increases in the amplitudes of business fluctuations and dispersion of short-term deviation from business fluctuations. The estimated points in time of the structural changes, April 1989 and January 1992, are almost the same as, or just after, the time of commencement of the Nikkei 225 futures transactions. Nikkei 225 futures transactions were blamed for lowering stock prices in Japan, thereby increasing stock price volatility, because stock prices started to drop significantly at the beginning of the 1990s. This paper also presents an analysis of the effect of Nikkei 225 futures transactions on stock price fluctuations and business fluctuations in Japan. More concretely, an analysis was performed to determine whether or not there are any Granger causalities between trading volume or open interest of Nikkei 225 futures transactions and the levels of the Nikkei 225 stock index, CI, IIP or their volatility. The analysis revealed no significant interactions. Therefore, Nikkei 225 futures transactions were not the cause of the instability of stock price fluctuations or business fluctuations in Japan.