臺灣博碩士論文加值系統

本研究的目的係運用統計方法上，具有顯著偵測能力的ＣＵＳＵＭ模
型，架構一動態化財務危機預測模式，以提升財務危機之預警能力。　　
本論文係以全額交割股公司為失敗公司，配對相同產業、規模相近之公司
為健全公司，利用固定資產／總資產比率、淨營運資金／總資產比率、每
股盈餘／每股市價比率、營業利益／總資產比率、存貨／營業收入比率等
五種比率作為自變數，導入ＣＵＳＵＭ模型以偵測財務危機之發生。實證
結果獲致以下重要結論：一、失敗公司的固定資產／總資產比率、淨營運
資金／總資產比率、每股盈餘／每股市價 比率、營業利益／總資產比
率均低於健全公司，而失敗公司的存貨／營業收入比率均　　較健全公司
為高。表示失敗公司因銷貨遲滯、存貨積壓、營業收入衰退，造成營運資
　　金週轉不靈，最終陷於財務危機的處境。二、多變量ＣＵＳＵＭ模式
的偵測能力，在失敗公司財務危機發生（被列為全額交割股）　　的前六
季，ＣＵＳＵＭ模式即可偵測出徵兆，精確度比區別分析高，顯現其具有
更佳　　的預測能力。

The purpose of this thesis is to provide a dynamic financial
predictingmodel that is constructed from a statistical "CUSUM"
model, by which it issignificantly able to predict the financial
distress in advance. The stock required full delivery firms
are selected as failed firms in comparison with healthy firms of
the same industry. We adapt five financialratios as independent
variables and put them into CUSUM model to test the financial
performance. The major results are as follows:1.For the ratios
such as FA/TA, NWC/TA, EPS/P, OI/TA if the failed firms are
lower than those of healthy firms except INV/SALE ratio, this
means that in the long run the failed firms will encounter
financial distress due to the bad inventory management.2.The
CUSUM model is well to predict the financial distress before six
sea- sons that is more accurate than a better prediction and
results in the Dis- criminate analysis.