以信用循環指標增進財務危機的預測

Abstract

過去的研究在一階段研究模型，都是使用固定的財務危機區分指標來區分財務危機與非財務危機公司，但是，財務危機的區分指標應該是依總體經濟的繁榮與衰退而有所調整，並非為一固定不動的常數。本研究並依二階段的方法針對在台灣上市、櫃公司資料來建立財務危機預警模型。第一，本研究利用各公司個別財務與市場變數，以Shumway(2001)的離散時間涉險模型計算公司個別發生財務危機之機率。第二，本論文依Kim(1999)使用總體經濟變數來建立信用循環指標，利用信用循環指標來調整二階段模型之財務危機區分指標。本研究針對1986年到2004年的財務、市場與總體經濟變數資料建構一階段與二階段財務危機預警模型，並以2005年到2007年的資料檢測模型的精確度。本研究之財務危機公司的選用，以公司遭遇如下法律事件作為判定之依據：歸為全額交割股、下市、重整或破產。本研究所使用的資料為在台灣證券交易所上市、櫃公司。而準確度的衡量係以包含：Vuong檢定、機率分布、ROC曲線與型一型二誤差率。 本研究結果顯示：(1) 在一階段財務危機預警模型當中，結合個別公司財務、市場因素與總體經濟等變數所建構的模型表現並非較結合個別公司財務變數與市場變數所建立的模型為佳。(2) 在一階段與二階段模型的比較中，財務危機預測也以二階段模型較佳，(2) 在一階段與二階段模型的比較中，財務危機預測也以二階段模型較佳，二階段模型中，由於測試期間為經濟不景氣，由總體經濟變數所建構的信用循環指標會提高財務危機區分門檻指標，二階段的財務危機區分門檻指標高於一階段，使得財務危機公司錯誤分類為正常公司的比率降低。特別是在二階段模型中，以財務比率與市場變數所建構的模型能有最低的公司錯誤分類比率。故本研究所探討的二階段模型在財務危機的預測有顯著的用途。

Previous studies used constant cut-off indicator to distinguish distressed firms from non-distressed firms in the one-stage prediction models. However, distressed cut-off indicator must shift according to economic prosperity, rather than remains fixed all the time. This study focuses on Taiwanese listed firms and develops financial distress prediction models based upon the two-stage method. First, this study employs the firm-specific financial ratio and market factors to measure the probability of financial distress based on the Shumway (2001) discrete-time hazard models. Second, this paper further applies macroeconomic factors to form the Kim (1999) credit cycle index and determine the distressed cut-off indicator of the two-stage models based on the credit cycle index. The one-stage and the two-stage prediction models are developed by using the data during 1986 to 2004, and their levels of accuracy are compared by the data during 2005 to 2007. This study focuses on the data of public listed companies which traded on the Taiwan Stock Exchange. The measurements of accuracy include: probability rankings, ROC (receiver operating characteristic) curve, and Type I and Type II errors. The distressed firms used in this paper are defined as the firms which suffer the following legal events: the reclassification of full-deal stock, delisting, reorganization, and bankruptcy. The empirical results reveal that: (1) As for the first stage of the prediction model, discrete-time hazard models are developed with different combinations of financial ratios, market variables, and macroeconomic factors. The models’ overall goodness-of-fits and out of sample prediction accuracy are compared using various criteria. Overall speaking, the model in incorporation with financial ratios and macroeconomic factors does not perform better than the financial-ratio-only model. Also, the model in incorporation with financial ratios, market variables and macroeconomic factors does not perform better than the model with financial ratio and market variables. (2) In regards to the comparison of two-stage models with the one-stage models, the performance of the prediction models can be improved based upon the two-stage models. The result indicates that the two-stage models have significantly better fit than the one-stage ones. The two-stage model incorporated with financial ratios and market variables experiences the lowest sum of Type I and Type II rates. The two-stage models presented in this paper have incremental usefulness in predicting financial distress. Through development of financial distress prediction models for Taiwan public companies, the study aims to understand the usefulness of credit cycle index in financial distress prediction. This study’s contribution can be summarized as follow. First, the two-stage models in this research utilize the importance of macroeconomic factors, in particular, consumer price index change ratio, stock price index change ratio, and interest rate change ratio of commercial paper to form the credit cycle index, which is used to adjust the distressed cut-off points. The two-stage models perform better in financial distress predictions. Second, not only do macroeconomic factors, market variables have incremental information beyond one another. Findings of this study emphasize the importance of taking into account the unique business and economic environment in develop financial distress prediction models.