景氣循環的擴張、收縮期的認定，以及轉折點的預測，向來是經濟學家、政府機構、企業及一般大眾所關注的焦點。越能掌握經濟發展的狀況，而適切的提早做出反應，越能充分利用經濟情勢的不同，以達成獲利的目標，或是減少企業的虧損。 本文即嘗試以票券、債券、股票、外匯及金融機構等金融變數，透過Kaminsky, Lizondo and Reinhart (1998)用以建構銀行危機預警綜合指標之訊號法（Signal Extraction Approach），以及可以解釋logit模型係數之發生比率（odds ratio），分別結合二元 logit 模型，建構預測景氣循環轉折點之金融綜合指標。 研究結果顯示，於1990年1月至2009年6月期間，金融指標對景氣循環轉折點的確有預測能力，且不論是預測景氣谷底或是高峰，表現最優良之綜合指標正確率皆高達九成以上，且不論是預測景氣之高峰或谷底，以發生比率建構之二元logit模型都有最好的預測能力。 / To confirm the expansion and contraction of business cycle, and to forecast the turning point, has always been economists, government institutions, enterprises and the public cared about. The more control on the state of economic development and take advantage of the different economic situation, the more they could response appropriately, and furthermore this could maximize their profit or reduce business losses. This article attempts to use financial variables, such as the indicators about bills, bonds, equities, foreign exchange, financial institutions, and use signal approach which was created by Kaminsky, Lizondo and Reinhart in 1998 to construct a composite index warning the banking crisis early. I also try to use odds ratio combine with binary logit model in order to predict business cycle turning point. The results show that in January 1990 to June 2009, the financial indicators of business cycle turning points does have predictive power, and the accurate prediction of the index over ninety percent. The result also show that the best performance of the model is combination of odds ratio and binary logit method.