[[alternative]]Model Specification for Volatility Forecasting in the Emerging Stock Markets – Evidence from Brics

計畫編號：NSC100-2410-H032-030 研究期間：20110801~20120731 研究經費：299,000[[abstract]]新興國家股票市場的波動性預測模型設定--金磚四國實證 對於如金磚四國(BRICs)等的新興市場國家，其股票市場存在較高波動值，使其產 生高風險與投機行為是新興國家股票市場的主要特徵之ㄧ，因此波動性之估計與預測遂 為一重要又值得探討之議題。在波動性預測上，具厚尾或偏態特性的分配假設為一關鍵 性之設定，波動性不對稱則為另一個波動性估計之重要設定，本文以DM 檢定為基礎， 依其逐次檢定的設計，運用五種GARCH 家族模型(對稱GARCH-N 模型, GJR-N 模型, QGARCH-N 模型, GARCH-t 模型, GARCH-ST 模型)，探索何種關於波動性估計之設定 對於樣本外預測有較佳之改善能力，同時計算逐次檢定之檢定力，本文之結果可作為提 升準確財務預測及風險管理效能之參考。[[abstract]]Model Specification for Volatility Forecasting in the Emerging Stock Markets – Evidence from BRICs The purpose of this research is to employ five competing models (Symmetric GARCH-N model, GJR-N model, QGARCH-N model, GARCH-t model, GARCH-ST model) by a sequential setting of DM-tests to do the volatility model specification, symmetric and asymmetric, for the emerging stock markets of BRICs economies. The finding of the best model specification for volatility forecasting enables us to fully capture the complex characteristics of emerging stock markets. Different loss functions will also be used and the testing power will be calculated to verify the validity of our sequential testing procedure.[[sponsorship]]行政院國家科學委員