台灣在2020年1月21日出現首宗新冠肺炎病例，隨後在2020年1月30日台股指數大跌 697 點，並在3月19日來到當年最低點 8,681點；與疫情爆發前的指數水準相比，跌幅約 23%。面對如此震盪的股市，精確地風險控管能為投資人帶來穩定的投資績效。但如何精確地估計風險，則一直是財務及經濟學界重要的議題。

本文採用了Fissler and Ziegel (2016)提出的FZ 損失函數，以半參數方法，不對資產收益分布進行任何假設，來估計兩個財務及經濟學中最常被用到的風險指標：風險值(Value at Risk, VaR)及期望損失(Expected Shortfall, ES)。使用之資料為以下之亞洲新興股票市場：泰國曼谷SET股價指數(SET)、韓國綜合股價指數(KOSPI)、印度Nifty指數(NIFTY500)、中國上海綜合股價指數(SSE)及台灣加權股價指數(TAIEX)。

研究結果顯示，與傳統的計量方法比較，使用FZ損失函數的半參數方法，在某些情況下的確可以有較好的表現。但傳統的計量方法，特別是非對稱GARCH模型(AP-ARCH)，表現總合來說是最佳。

Taiwan's first case of COVID-19 occurred on January 21, 2020, then the Taiwan stock index fell 697 points on January 30, 2020 and reached the lowest point of the year on March 19. Compared with the index level, it is down about 23%. In the face of such a volatile stock market, accurate risk control can bring stable investment performance to investors.

The purpose of this paper is how to accurately estimate risk. This paper uses the FZ loss function proposed by Fissler and Ziegel (2016) to estimate the two most commonly used risk indicators in finance and economics：Value at Risk (Value at Risk, VaR) and Expected Shortfall (ES).

A descriptive survey design was adopted to collect the data. The results of the experiment indicated that compared with the traditional measurement method, the semi-matrix method using the FZ loss function can indeed have better performance in some cases, but traditional econometric methods, especially the asymmetric GARCH model (AP-ARCH), performed the best overall.

1.陳婉淑(2007)，金融市場之風險值模型推論，逢甲大學統計與精算研究所碩士論文。
2.黃博寬(2010)，應用非線性迴歸分量法預測在2008-09年金融危機之風險值，逢甲大學統計與精算所碩士論文。
3.詹雅竹(2007)，金融市場之風險值模型推論，逢甲大學統計與精算所碩士論文。
4.劉衛東(2020)，新冠肺炎疫情對經濟全球化的影響分析，地理研究，39(7)，1439-1449。
5.賴韋任(2011)，風險值與預期尾部損失對原油價格之風險評估，長榮大學經營管理研究所碩士論文。
6.蘇虹朵(2004)，風險值在台灣股市之衡量與驗證，世新大學財務金融所碩士論文。
7.Bassett, G and Koenker, R. (1978). Regression Quantiles, Econometrica, 46(1), 33-50.
8.Bollerslev, T. (1986). Generalized autoregressive conditional heteroskedasticity,Journal of Econometrics, 31(3), 307-327.
9.Chrétien, S and Coggins, F. (2010). Performance and conservatism of monthly FHS VaR: An international investigation. European Journal of Operational Research, 19, 323-333.
10.Engle, R.F, and Manganelli, S. (2004). CAViaR: Conditional Autoregressive Value at Risk by Regression Quantiles. Journal of Business and Economic Statistics, 22, 367-381.
11.Fissler, T. and J. F. Ziegel (2016). Higher order elicitability and Osbands principle, The Annals of Statistics, 44, 1680-1707.
12.Meng, X. and J. W. Taylor (2020). Estimating Value-at-Risk and Expected Shortfall using the intraday low and range data. European Journal of Operational Research, 280, 191 - 202.
13.Patton, A. J., J. F. Ziegel, and R. Chen (2019). Dynamic semiparametric models for expected shortfall (and Value-at-Risk). Journal of Econometrics, 211, 388-413.
14.Chou, R. Y., Yen, T. J. and Yen, Y. M. (2022).Forecasting Expected Shortfall and Value-at-Risk with Realized Variance Measures and the FZ Loss. Taiwan Economic Forecast and Policy, 89-140.
15.Şener, E., Baronyan, S. and Mengütürk, LA. (2012).Ranking the predictive performances of value-at-risk estimation methods. European Journal of Operational Research, 28, 849 - 873.
16.Taylor, J. W. (2019). Forecasting Value at Risk and Expected Shortfall Using a Semiparametric Approach Based on the Asymmetric Laplace Distribution. Journal of Business & Economic Statistics, 37, 121-133.
17.Zhu, D. and Galbraith, J. W. (2011). Modeling and forecasting expected shortfall with the generalized asymmetric Student-T and asymmetric exponential power distributions. Journal of Empirical Finance, 18, 765-778.
18.Zheng, Y., Q. Zhu, G. Li, and Z. Xiao. (2018). Hybrid quantile regression estimation for time series models with conditional heteroscedasticity. Journal of the Royal Statistical Society Series B (Statistical Methodology), 80, 975-993.