一般化風險指標於廣泛資產類別之應用

Application of Generalized Measure of Riskiness to Various Asset Classes

黃楷瀚(Kai-Han Huang)

指導教授 ： 曾郁仁

國立臺灣大學/管理學院/財務金融學研究所/碩士(2017年)

https://doi.org/10.6342/NTU201701517

摘要

本篇論文旨在測試由Bali, Cakici, and Chabi-Yo (2011)發展的一般化風險指標(Generalized Measure of Riskiness, GMR)在股票大盤指數、投資級別固定收益商品、外匯、及黃金等不同資產類別的適用性，並發掘GMR在跨資產類別與時間序列趨勢。本文採用一般化動差法(Generalized Method of Moments, GMM)估計GMR、並利用GMM估計量逼近常態分配的特性，就跨資產與時間段的估計值差異進行統計檢定。
結果發現，以GMM採滾動方式估計的GMR能在一定程度上反映各類資產風險性的差異、其風險報酬受總體經濟指標的影響，並作為標的資產當期夏普指數與索提諾指標的反指標；在時間序列趨勢方面，採滾動方式估計下，GMR在樣本期間內呈現穩定上揚趨勢，反映同期間內標的資產風險調整後報酬下滑的現象。採六個月及十二個月為觀察窗口作非滾動估計的GMR，除了在時間序列上能顯著反映同期各類資產基於負面標準差、標準差、超額報酬、索提諾指標、及夏普指標衡量之風險報酬狀況外，其趨勢在指定不同風險趨避水準下仍有穩固性，顯示GMR指標對風險偏好不同的決策者而言，皆能做為良好的中長期標的資產風險報酬衡量指標。

關鍵字

風險衡量指標 ； 報酬指標 ； 風險趨避程度 ； 一般化動差法 ； 時間序列 ； 超額報酬 ； 負面標準差

並列摘要

This paper is aimed at expanding the application of the Generalized Measure of Riskiness(GMR), introduced by Bali, Cakici, and Chabi-Yo (2011), to various kinds of assets, and testing for its feasibility for properly reflecting the pattern of riskiness and return of the underlying assets.
The paper conducts the estimation of riskiness measures based on historical data between December 31, 1988 to January 31, 2009 using Generalized Method of Moments (GMM). The paper examined the time series trend, the correlation with concurrent and predictability of future Sharpe ratio and Sortino ratio, along with the relationship with macroeconomic variables of GMR calculated with GMM based on rolling and non-rolling estimation alike. The empirical results indicate that for the rolling-estimated GMR, the riskiness of stock and investment grade bond index had been gradually increasing during the sample period, which is in harmony with the results proposed by M Taboga (2009).
For the non-rolling estimated GMR, it can properly reflect the concurrent risk and return of the underlying assets, and the results are robust to different degree of risk aversion. As for the predictability of future Sharpe and Sortino ratio, the indicator shows no statistically significant predicting power, which may be due to the backward-looking nature of the GMM method. Finally, The coefficients with regard to macroeconomic variables are in accordance with general economic rationale.
Based on our empirical findings, we conclude that GMR is an appropriate measure for middle- to long-term historical performance feasible for various asset classes.

並列關鍵字

riskiness measure ； return index ； excess return ； risk aversion ； time series ； negative standard deviation ； generalized measure of riskiness

參考文獻

1. Acerbi, Carlo, and Dirk Tasche. "Expected shortfall: a natural coherent alternative to value at risk." Economic notes 31.2 (2002): 379-388.

2. Aumann, Robert J., and Roberto Serrano. "An economic index of riskiness." Journal of Political Economy 116.5 (2008): 810-836.

3. Bakshi, Gurdip, and Dilip Madan. "Spanning and derivative-security valuation." Journal of Financial Economics 55.2 (2000): 205-238.

4. Bakshi, Gurdip, Charles Cao, and Zhiwu Chen. "Empirical performance of alternative option pricing models." The Journal of finance 52.5 (1997): 2003-2049.

5. Bali, Turan G. "The intertemporal relation between expected returns and risk." Journal of Financial Economics 87.1 (2008): 101-131.