A Study On Stock Returns Based On A Random Effects Poisson Modeling Approach

In this paper, we study the portfolio selection problem in security market. Our approach is to employ a random effect Poisson regression model to decide what factors may contribute to gaining more days of positive return. The method we use for analyzing is so called Orthodox best linear unbiased predictor(BLUP) approach, which was developed by Renjun Ma etc. in recent years.And the stock data is from January 4,2013 to December 31, 2014, and the data is divided into three periods: the period of bull market, the period if bear market and the period with some ups and downs(PUD). The industry characteristics can be used as random effects so we can eliminate the impact of industry characteristics to stocks’ returns. The Orthodox BLUP approach to this model depends only on the first- and second-moment assumptions of unobserved random effects. And this is a great approach for processing a large number of high-dimensional data in study. This paper illustrates that change, turnover and PB have negative effects on the number of days of stock positive returns in bull market. Moreover, in bear market, change has positive effects on the number of days of stock positive returns, while PS.TTM negatively affects that.