藉由台灣電視節目收視率、台股成交量與台股波動性的資料，本文利用Hausman （1978）檢定指出成交量為自變數來解釋股市波動的內生問題。我們可以用兩階段的 工具變數法來克服這個問題並得到具一致性的估計結果。在工具變數的選取上，我們 首度嘗試以「財經資訊分析」和「新聞播報節目」的收視率為候選變數，依據Stock and Watson（2003, p. 371）和Stock and Yogo（2005）所建議的準則，確認「財經資 訊分析」的收視率為強工具變數；「新聞播報節目」的收視率為弱工具變數。接著， 再以Sargan（1958）的外生性檢定確認「財經資訊分析」的收視率的外生性。本研究 的工具變數法的估計結果顯示，成交量所帶動的股市波動遠大於以最小平方法所估得的結果。根據這樣的結果，本研究進一步模擬投資人在觀察到「財經資訊分析」的收視率以後，對股市的波動形成一致性的預期，並且透過台指選擇權的操作，形成一種可能的獲利空間。

This paper investigates the impacts of trading volume on the volatility of Taiwanese stock by adopting TV rating as instrument variable. According to the Hausman (1978) test, using trading volume as regressor to explain the volatility of stock will yield the endogeneity problem. One can get the consistent empirical estimations by adopting the two-step instrumental estimation. We carry the two-stage instrument estimation by choosing the TV rating of “analysis of financial information” and “news reporting” program as candidate instrumental variables. We find that the TV rating of “analysis of financial information” is strong instrumental variable and the TV rating of “news reporting” is weak instrumental variable based the criterion of Stock and Watson (2003, p. 371) and Stock and Yogo (2005). Besides, this paper applies the Sargan (1958) test to test the exogeneity of strong instrumental variable and show that the TV rating of “analysis of financial information” is the exogenous variable. We find the stock volatility results by 2SLS are larger than the empirical results by OLS. We further show the scenario for investors to form consistent expectation on stock market volatility after observing the TV rating and based on it, they may earn some possible gains by the operation on the options market.