A Method for Mining Information about Business Cycles from Stock Prices Data
Main Article Content
Abstract
In this paper, we propose a method for analyzing business cycles based on stock prices data. We take the Nikkei Stock Average (NSA) as a proxy of stock prices and use the Composite Index in Japan (CIJ) as an indicator for business cycles. We investigate the correlation of the CIJ with respect to a difference between two moving average series of the NSA, and introduce a lag parameter in the CIJ. Thus, we can determine an effective frequency interval for the differenced series by maximizing the correlation coefficient, and we can analyze the lead–lag relation between the NSA and the CIJ based on the estimate of the lag. Moreover, we analyze the dynamic relationship between the differenced series of the NSA with the most effective frequency interval and the CIJ using a regression model that is constructed using a time-varying coefficient. As an empirical example, we analyze the daily time series of the NSA for closing values from January 1, 1991 to September 28, 2018, together with the monthly CIJ data over the same period.