本研究係以1980 年1 月至2004 年10 月的台灣加權指數之週報酬率資料為研究對象，主要採用最大概似函數值檢定法估出平均數方程式及變異數方程式同時發生結構性轉變之發生時點為1987 年3 月，由於此方法須對整樣本期間每一個時間點作測試以找出最大概似函數值故相當耗時，所以本研究先用Bai and Perron(2003)的內生結構性轉變點最小殘差平方法估平均數方程式發生結構性轉變之發生時點為1988 年12 月的第4週，再以此時間點之90%信賴區間來測試以縮短測試時間。最後用概似比檢定確定結構性轉變發生於變異數方程式或同時發生於平均方程式及變異數方程式，而發生此結構性轉變原因可能為這段期間政府在外匯制度進行多項開放措施所導致國內大量資金浄流出。

In this paper we apply the technique of maximum likelihood ratio test and the GARCH model with dummy variables in the mean and variance equation, to analyze whether if the volatility of Taiwan weighting index has changed significantly over the period 1980-2004 or not. We find that the sup LR statistic in March 1987 is greater than that of the sample period and so is a break point. But the technique consumes much time as it must test the points in the whole sample period each time. So we apply the technique of minimum residual sum of square and the GARCH model with dummy variables in the mean equation only, to detect the structural break point in the mean equation of the GARCH model. We find that the last week of December, 1988 is a structural break point in the model, then use the maximum likelihood ratio test to detect the points in the 90% confidence interval of the structural break point in order to reduce computing time. Finally, from the likelihood ratio test, we find that the structural break point takes place in the variance equation only, or in the mean and variance equation at the same time. In this break point period, our government opens the exchange market which lets the capital to flow out.