影響股票報酬率波動因素之研究-以24國為例

Abstract

股票市場波動性(volatility)在資產評價、投資組合的管理及動態避險等方面，均扮演極重要的角色。

過去探討股市波動性的文獻中，大部分皆探討單一因子對單一國家股市波動性的影響，較少探討這些因子對不同國家影響性。且大部分文獻皆探討單一因子在不同的模型中所產生的影響，較少學者探討單一模型中放入不同因子所產生的影響。因此本文以GARCH (1,1)模型為主，分別加入三個不同因子，例如：交易量（trading volume）、日內價格波動性（high-low range volatility）與日內波動（intra-day volatility, IDV），以探討這些因子對各國主要股價指數報酬率波動的影響。本研究更進一步將樣本期間區分為樣本內資料與樣本外資料，比較此模型加入不同變數後的預測效率 (forecast efficiency)。本研究以全球共24個國家為主要研究對象，並將國家分為開發中國家及已開發國家進行探討。

本研究主要有兩項發現：首先，交易量、日內價格波動性及日內波動對於開發中國家主要股價指數報酬率波動性的影響程度大於已開發國家。並發現各個因子的預測效率對各個國家而言並沒有明顯的差異。

While the stock volatility has been extensively investigated, the transnational research is relatively unexplored. This paper studies influences on stock return volatility after entering trading volume, intraday high-low range volatility and intra-day volatility (IDV) in GARCH (1,1) model. Moreover, the study compares the forecasting efficiency across the four models by using out-of-sample data. Using stock indexes from the 24 countries as an example, the period is from January 1, 2007 to December 31, 2011. Furthermore, the data of this study are categorized into the two groups, the developed and developing countries, in order to ascertain differences between them. The first finding suggests that the three variables make a greater impact on the stock return volatility in the developing countries in comparison with the developed countries. The second finding suggests that the forecasting efficiency for GARCH (1,1) model entering trading volume, intraday high-low range volatility, intra-day volatility (IDV) and without variables are similar across the 22 countries. The forecasting efficiency of GARCH (1,1) model with trading volume is slightly better than the other models among the 22 countries. However, there are no significant differences among the 4 models.