2007年次級房貸波及全球金融市場後,風險控管的議題再度受到重視,金融機構需要一個可以正確衡量風險的標準,據以對所承擔的風險做相當程度的暴露。自30人團體、巴賽爾委員會等國際權威機構,分別規範以風險值(Value at Risk)做為衡量所暴露的市場風險後,風險值已成為市場風險管理的重要工具。由於我國財政部允許金融機構各自設計內部計算風險值的方法,因此本研究運用風險值概念,找出適合台灣股市的風險值估測模型,尤其針對以個別產業為投資標的的創投公司與私募基金等,提供風險管理上的參考依據。 本研究主要評估GARCH模型、TGARCH模型與EGARCH模型在估計風險值之預測能力。本研究以台灣股票市場之八大類股股價指數做為比較風險值估測模型的研究對象,研究期間為2007年1月2日至2010年3月31日。評估績效方法本研究採用回溯檢定、均方根相對偏差指標(RMSRB)、誤差平方根(RMSE)與絕對平均百分比誤差(MAPE)分別做為衡量模型準確性(accuracy)、保守性(conservatism)及效率性(efficiency)的準則。 實證結果指出台灣八大類產業中,水泥類、食品類、紡織類、機電類與營建類皆是以EGARCH(1,1)為最適估測風險值模型;金融類與塑膠類是以GARCH(1,1)為最適估測風險值模型;造紙類則以TGARCH(1,1)為最適估測風險值模型。而台灣大部分的產業均是以EGARCH(1,1)為最適模型,表示在本研究期間,整體而言EGARCH(1,1)較能捕捉台灣股票市場真實的特性,亦即存在槓桿效果(leverage effect)。
Since 2007 the subprime mortgage crisis attacks financial market in the whole world, many financial institutions take risk management’s subject seriously again. Therefore financial institutions have to apply a correct standard to measure market risk. From Group of Thirty, the Basel Committee and so on international authoritative organization making value at risk (VaR) to measure exposure of market risk, the VaR has become one of the most important tools for market risk management. So the goal of this study is to find the model which is better for forecasting the VaR precisely in Taiwan’ stock portfolio. Especially aims at venture capital firm and privately offered fund and so on to provide a reference of risk management. The purpose of this study is to evaluate three models, which are GARCH model, TGARCH model, and EGARCH model, for measuring VaR precision. The data of Taiwan eight major industries stock market indexes was collected to evaluate these three models. The sampling period is from January 2, 2007 to March 31, 2010. Finally, the performances of three models are tested by the evaluating methods of accuracy, conservatism and efficiency according to the rules of backing test, RMSRB, RMSE, and MAPE. Via the empirical results, this research finds the cement industry, the foodstuff industry, the textile industry, the electrical machinery industry and the construction industry take EGARCH(1,1) as the most suitable model. The financial industry and the plastic industry take GARCH(1,1) as the most suitable model. The paper-making industry takes TGARCH(1,1) as the most suitable model. To sum up, using the EGARCH model to estimate the VaR in Taiwan majority industries is better than the other two models.