Grey model as a tool in dynamic portfolio selection

Abstract

The goal of this chapter is to provide a simple approach to applying the grey model (GM) (1,1) of stock price prediction. Namely, many different quantitative methods and models are developed to facilitate the decision-making process within the dynamic portfolio selection. However, the grey methodology is still insufficiently examined, although it has approaches tailored towards financial data. Thus, the main contribution of this chapter is to focus on a GM of stock price prediction on a rolling-window basis and use the results in the portfolio rebalancing over time. Previous research does not give a detailed analysis of portfolio performance over time as this chapter is providing. Daily data have been collected for a Croatian stock market index from 2 January 2020 to 3 December 2020. The GM (1,1) has been estimated on a 30-day rolling basis with the out-of-sample one-day predictions of the future index value. In that way, the predictions were used to form a portfolio whose performance was compared to several benchmark ones. The results are promising, in terms of the overall portfolio performance based on the GM results. Thus, such a methodology can be very useful in portfolio selection applications.