傳統BSM (Black and Scholes 1973; Merton 1974) 結構式模型由於過度簡化之假設，可能造成較不精確之企業違約預測效力。有鑑於此，本文應用複合選擇權理論，建構一個較符合實際經濟社會債務結構的違約預警模型。實證結果顯示，CO模型與BSM模型皆具有即時偵測違約發生之能力，然而以複合選擇權理論建立之CO模型，其違約預測效力相較BSM模型而言有進一步提升之效果。另外，透過censored Tobit迴歸模式探討影響兩種結構式模型違約預測效力表現的相關因素，亦可以發現傳統BSM模型在建構過程中僅考慮企業償債性與市場面因素；相較而言，CO模型則多考慮企業獲利性層面因素，亦能有效地即時偵測企業違約之發生。故本文CO模型可作爲除市場基礎之BSM模型外，另一個判斷企業違約風險的良好預警工具。

Due to the improperly simplified assumptions of the traditional BSM (Black-Scholes-Merton) model, the prediction of performance in detective default is not very accurate. The application of compound option theory to establish the credit risk model of detective default in firms is more fit to the debt structure for firms in real world. From the empirical study results, both the CO model and the BSM model have good ability to detect the default of firms, but compared with the traditional BSM structural model using market information, the result of empirical testing shows the performance of detective default in firms of our CO model established by the using of compound option theory has improved effect. Furthermore, using the censored Tobit regression model to observe the characteristics of detective default performance in two structural models, we also find the traditional BSM model that only considers the solvency factors and the market factors in firms, compared with the CO model that considers the profitability factor in firms also, the latter model might detect default in firms more effectively for our reasonable inference. Besides the BSM model with market-based framework, the CO model can be used as another good tool in the detection of firms' default.