本文研究目的將針對農會信用都經營惡化之事實，以事後分析角度，並依據農業金融法中有關農會信用部退出市場時機條件，模擬信用部在不同時期下，所承受失敗風險程度大小。對於分析工具，將不同於以往文獻所普遍採用之靜態的logistic及probit模型，而是採用同時考慮到時間及截斷資料特性之存活分析(survival analysis)。本文觀察期間設定，起始年為民國84年，而終止年為民國93年，在期間內符合失敗事件定義之信用部，計有184家，而在觀察期間結束後，仍未有失敗事件發生之截斷(censored)樣本，計有90家，全體總樣本合計274家。本文研究所得結果為：在觀察期間中以民國90年事件發生之風險最高，表示在此期間農會信用都面臨高度的逾放比率風險危機。對照民國90年時，正是政府積極從事基層金融整頓，接管36家農漁會信用部，制定農業金融相關法令，金融監理轉趨嚴格時期，也因此民國90年後整體信用部之危機率才漸恢復穩定，高逾放比率的信用都不似先前的眾多。此外，就分組樣本觀察，在不考慮、區位情形下，農會信用部若事前加入存保，發生失敗 (即農會信用部逾放比例高於15%以上) 之風險較低;同樣地，若農會信用部位於北部亦較南部地區有較低之失敗風險。

The purpose of this paper was to simulate financial crisis of Taiwan’s credit department of farmers' associations. Unlike normally, we will not use the traditional approach (e.g., logistic or probit model). In particular, we will focus on the survival probability and hazard rate using the survival analysis method. Furthermore, our paper is also followed the agricultural financial law making the credit department of farmers' associations left out financial industry. The sample used for this paper consists of 274 credit departments in Taiwan for ten consecutive years, 1995-2004. The 90 samples was defined as censored data because they were over 2004.0therwise, the rest of 184 samples belonged to complete data. Our results have clearly demonstrated that the credit department has highly risk in 2001. This implies that Taiwan's credit department has more non-performing loan ratios crisis in this year. This year is just the government to push up Taiwan’s agricultural financial reform .In addition, if we don’t consider location effect, the credit department has joined Taiwan’s deposit insurance system having lower failure events ( namely, non-performing loan ratios beyond 15%) than others. Likewise, the north credit department has less failure events (or risk) than the south parts .