由報酬離散度預測超額報酬離散度－台灣股票市場實證

Abstract

波動性和超額報酬之間的關係，為投資人最關注的議題之一。波動性可以定義為歷史資料統計而得之歷史波動性、利用選擇權市場價格反推而得之隱含波動性、由ARCH及其衍生模型所預測出之波動性以及橫斷面波動性。 本研究應用Gorman, Sapra and Weigand(2009)所提出之新方法，由報酬率橫斷面波動性(離散度)預測市場之超額報酬，以台灣50指數成分股橫斷面離散度及台指VIX依序對未來一個月、兩個月、三個月、半年及一年後之台灣股票市場超額報酬的離散度以確定預測的效果。 實證結果發現，報酬率的離散度及VIX對於台灣的資料，能有效的預測一個月及兩個月後之超額報酬離散度，但對於更長時期之預測，則效果不佳。本預測模式對於對於尋求絕對報酬之投資者或是經理人，提供了何時超額報酬會有大幅變動的一個新方法。

The relationship between volatility and alpha is one of the most important issues for investors. Main types of volatility model include the historical volatility, implied volatility, ARCH family of models, and cross-sectional volatility(dispersion). This paper applied a new method proposed by Gorman, Sapra and Weigand(2009) that uses cross-sectional standard deviation of stock return to forecast alpha dispersion to Taiwan stock market. We use the component (constituent) stocks of TSEC Taiwan 50 index and VIX to forecasted the alpha dispersion of one month later, two month later, three month later, half an year later and even one year later. The results showe that the return dispersion and VIX do have the ability to forecast the alpha dispersion in 1 month and 2 month, but losing forecast power in longer periods. This research provides the absolute return investors a new method to measure the timing of significant changes of alpha.