株価指数の時差共和分分析と予測可能性

Intertemporal Cointegration Analysis and Predictability of Stock Indices

大鋸, 崇

OGA, Takashi

[SUMMARY] The aim of this study is to analyse leading and lagging characteristics in multivariate time series of asset prices using the Intertemporal Cointegration Model constructed by Oga（2021）and to construct an investment strategy utilizing the leading index. Using the multivariate time series of the Tokyo Stock Price Index（TOPIX）and the Tokyo Stock Index by industry（TOPIX33）, we clarified the leading and lagging characteristics between the paired two series. As a result, a large number of set of series following intertemporal cointegrations is found; this indicates the existence of pairs of a stock price index series that has the property of co-varying with lead-lags in the long-term trend of stock index fluctuations. We constructed an investment strategy using a set of time series with intertemporal cointegration relationship and conducted a two-year simulation. Specifically, when the backward moving average of the optimal forecast given by the model changes to upward trend, we consider a simple investment strategy of“buying” the lagging index, and“selling”when it changes to downward trend. As a result, it was shown that a stable profit can be obtained by using our investment strategy that uses this leading and lagging property. Although It is difficult to obtain a positive excess return from prior information due to market efficiency, the predictability of asset prices was demonstrated by using the intertemporal cointegration model of Oga（2021）.

https://opac.ll.chiba-u.jp/da/curator/900121538/

https://opac.ll.chiba-u.jp/da/curator/900121538/S24368873-37-3-P001-OGA.pdf

千葉大学経済学会

Economics Association for Chiba University

紀要論文

24368873

AN10005358

千葉大学経済研究 = Economic journal of Chiba University

37

3-4

1

47

2023-03-15

10.20776/S24368873-37-3-P1

publisher

経済研究 (ONLINE ISSN 2436-8873)