資訊管理學報
黃金生 ; 施東河; 劉建利;
頁: 63-80
日期: 1996/06
摘要: 本研究嘗試以類神經網路及GARCH模型來預測台灣人壽保險業股票之風險溢酬。基於Ross(1976)的套利定價理論，本研究的預測模型擴充原Chen, Roll and Ross(1986)及Mei and Saunders(1994)之 財務預測模型，並涵蓋台灣保險業市場特徵及政治環境變數。本研究經由類神經網路模型之訓練及測試以及GARCH的統計檢定，來確認影響台灣保險業風險溢酬之重要經濟變數。本研究結果指出保險產業特徵及政治變數對風險溢酬有顯著貢獻，同時台灣壽險產業與不動產業間亦存在外溢效果。實證結果顯示本類神經網路模型將提供保險業風險溢酬之財務預測領域另一有潛力之方向。
關鍵字: 類神經網路; 股票風險溢酬; 套利定價理論; 一般化自迴歸條件異質變異數;;
Abstract: In this study, we employ an artificial neural network model and GARCH model as finan-cial vehicles for forecasting risk premiums on Taiwan life insurance industry. Specifically, in the spirit of Ross's(l976) arbitrage pricing theory, our models of this research extend the origin-nal Chen, Roll, and Ross(1986), and Mei and Saunders(l994) to include the characteristics of Taiwan insurance market and variables of political environment. Our neural network model and GARCH model are able to identify the important economic variables of risk premium on insurer stocks. Moreover, the insurance industry characteristics and political variable are shown significantly contributing to the insurance risk premium. Meanwhile, there is significant spill-over effect between Taiwan life insurer and real estate industry. The encouraging results of this study suggest this neural network model is promising in the field of risk premium forecast-ing on insurer stocks.
Keywords: Artifical neural network; Stock risk premiums; Arbitrage pricing theory; GARCH;;
瀏覽次數: 6818 下載次數: 49
引用 導入Endnote
頁: 63-80
日期: 1996/06
摘要: 本研究嘗試以類神經網路及GARCH模型來預測台灣人壽保險業股票之風險溢酬。基於Ross(1976)的套利定價理論，本研究的預測模型擴充原Chen, Roll and Ross(1986)及Mei and Saunders(1994)之 財務預測模型，並涵蓋台灣保險業市場特徵及政治環境變數。本研究經由類神經網路模型之訓練及測試以及GARCH的統計檢定，來確認影響台灣保險業風險溢酬之重要經濟變數。本研究結果指出保險產業特徵及政治變數對風險溢酬有顯著貢獻，同時台灣壽險產業與不動產業間亦存在外溢效果。實證結果顯示本類神經網路模型將提供保險業風險溢酬之財務預測領域另一有潛力之方向。
關鍵字: 類神經網路; 股票風險溢酬; 套利定價理論; 一般化自迴歸條件異質變異數;;
Neural Network Prediction of Risk Premiums on Taiwan Life Insurer Stocks
Abstract: In this study, we employ an artificial neural network model and GARCH model as finan-cial vehicles for forecasting risk premiums on Taiwan life insurance industry. Specifically, in the spirit of Ross's(l976) arbitrage pricing theory, our models of this research extend the origin-nal Chen, Roll, and Ross(1986), and Mei and Saunders(l994) to include the characteristics of Taiwan insurance market and variables of political environment. Our neural network model and GARCH model are able to identify the important economic variables of risk premium on insurer stocks. Moreover, the insurance industry characteristics and political variable are shown significantly contributing to the insurance risk premium. Meanwhile, there is significant spill-over effect between Taiwan life insurer and real estate industry. The encouraging results of this study suggest this neural network model is promising in the field of risk premium forecast-ing on insurer stocks.
Keywords: Artifical neural network; Stock risk premiums; Arbitrage pricing theory; GARCH;;
瀏覽次數: 6818 下載次數: 49
引用 導入Endnote