Research Of Investment Strategies Based On Idiosyncrastics Volatility

Existing research shows that the relationship between idiosyncratic risk of firms and expected returns of stock may be related to factors such as the method used to estimate the idiosyncratic volatility.It is important to figure out the influence of these factors on the application of idiosyncratic volatility to actual investment.This article uses a variety of methods to estimate the idiosyncratic volatility by data from different frequencies and different time ranges.Based on this,stocks are sorted into different groups to construct investment portfolios,and comparisons of these portfolios are made based on indicators such as annualized yields and Sharpe ratios.Results shows that use 1-month-daily data,and take standard deviation of Fama-French three factor model's residual as idiosyncratic volatility leads to equal-weighted grouped portfolios that show best monotonicity,and the corresponding long-short portfolio performs better than any portfolios obtained using other methods;Further analysis show that the group with higher idiosyncratic volatility,also has lower market value,higher book-to-book ratio,higher previous month's yield,and turnover rate.Using the heterogeneous beliefs variable and the short-selling dummy variable defined in this article,regression analysis of the monthly returns of long-short portfolios and the Fama-Macbeth regression of individual stock-month yields was performed;Both indicated that the heterogeneous beliefs,can explain to some extent,but not fully,the significant negative correlation between the realized variability and future returns.Based on the above research,this paper tries three methods to combine the realized variability with other factors to construct a portfolio.One is to use the reverse rank order of the realized variability as the weight of individual stocks,on the grouped portfolios sorted according to other factors;The second is to cross-sort the realized variability with other grouping variables;Third is to add the realized variability to the multi-factor regression model and multi-factor scoring model,then construct portfolios based on the expected yield or score.The portfolios constructed by these methods all have good performance and provides ideas for actual investment.