以風險組成因子估計'預測及分析實際股票報酬波動之研究

張仲平(Chung-Ping Chang)

指導教授 ： 林修葳

國立臺灣大學/管理學院/國際企業學研究所/碩士(2004年)

https://doi.org/10.6342/NTU.2004.01036

摘要

摘 要

股票報酬為一混合的型態，由各來源成分組成。報酬的組成成分各自波動程度不一，也因此對報酬波動的影響不一。本研究首先由不同角度建立股票期望報酬模型，而後參考VaRC法轉為由風險因子加總組成的報酬波動拆解模型，並以S&P 500股票指數報酬波動作為實証研究對象，以風險來源的加總來估計、預測及分析實際報酬波動的情況。

實証結果顯示，相同的風險因子放在不同的報酬波動模型中，仍會對整體波動有相同的比重影響，因此模型估計效力的差異在於各個風險因子的選擇加總後所產生的誤差大小。本研究並以ARIMA及GARCH模型作預測，預測結果並顯示，以短期高頻的資料預測要比以長期低頻的資料作預測來得好，顯示報酬波動並沒有太長久的記憶性。另一方面，以不同期間來看，各風險因子佔整體報酬波動的比重會有變化，在各期間對報酬波動的影響力不一，而各因子佔報酬的成分比重與佔波動的比重迥異，因此我們無法僅以某因子佔報酬的比重來斷言其對波動的影響，兩者並無絕對關係。

相同的方法亦可應用於其它的金融商品或市場，以分析投資標的的風險結構。瞭解報酬波動的風險來源與結構，可提高投資時避險的效率，也可提供主管機關或企業在決策及風險控管時的有效參考。

關鍵字

GARCH模型 ； 風險因子 ； VaRC方法 ； ARIMA模型

並列摘要

Stock return is garbled and composed of its many origins. Factors of stock return have their own degrees of volatility and make different contributions to total return volatility. We first construct expected stock return models under different circumstances and then transfer them to volatility models by referring to VaRC method. This study focuses on the US S&P 500 stock total return index and sums risk origin factors up to estimate, forecast, and analyze real stock return volatility.

The result shows that the same risk factors will have the same weights of total volatility no matter in other different models. The estimate effectiveness is due to the errors caused by summing other risk factors up. We use ARIMA and GARCH (1,1) models to forecast volatility and find that it would be better using short-term high frequency time series data to forecast than using long-term low frequency data. The forecast result shows that return volatility dose not have long-term memory. On the other hand, each factor has different weights of total volatility during various periods. The weights of each factor to return differ from them to volatility. There is no absolute relationship between weights to return and weights to volatility.

Analyzing the componential contributions of each stock risk factor can let us know the risk structure and help us to make fine decisions when hedging. Also, this method can be utilized to analyze other financial instruments and offer some suggestions to practitioners and regulators.

並列關鍵字

Risk Factor ； VaRC Method ； ARIMA Model ； GARCH

參考文獻

1.Akgiray, V.(1989) “Conditional Heteroscedasticity in Time Series of Stock Returns : Evidence and Forecasts.” Journal of Business, 62, 55-80.

2.Bollerslev, T.(1986)“Generalized Autoregressive Conditional Heteroscedasticity.” Journal of Econometrics, 31, 307-327.

3.Campbell, J. and Shiller, R.(1988) “Stock Prices, Earnings, and Expected Dividends.” Journal of Finance, 43, 661-676.

4.Campbell, J. and Shiller, R.(2001) “Valuation Ratios and the Long-Run Stock Market Outlook: An Update.” NBER Working Paper No.8221.

5.Dickey, A. and Fuller, A.(1979) “Distribution of the Estimators for Autoregressive Time Series With a Unit Root.” Journal of American Statistical Association, 74, 427-431.