The Extract and Usage of Common Factors of Macroeconomic Panel Data by Principal Component Analysis(in Japanese)

In this paper, we extract common factors from nearly 120 series of Japanese macroeconomic Panel Data as the diffusion index of business cycles, using the estimation method proposed by Stock and Watson (1998, 2002a,b). This model referred to as Approximated Dynamic Factor Model, can be estimated from rather easy methods based on principal component analysis. In addition, using these common factors, forecasting for principal macroeconomic time series and estimating forward looking Taylor rule are implemented. The estimation result suggests the common factors have useful infomation on forward-looking economic activities. This paper is also the survey dealing with advanced studies of Approximated Dynamic Factor Model in the Western countries.