企業編製財務報表提供給投資者瞭解企業的營運狀況以及財務結構,然而企業陸續爆發作假帳、負責人掏空公司資產、操弄股價,最終引發違約交割,公司股票被迫下市及停止交易,使得投資和金融市場受到很大衝擊。本研究之目的在建立有效的財務危機預警系統,預測公司是否成為危機公司,本文以國內在2011 年到2013 年之間上市櫃食品業為例,選定32家公司(7 家為警示股,25 家正常公司)為研究對象,在變數選取方面以「台灣經濟新報」(TEJ)的財務數據為主,選取27 個變數來預測公司是否有財務危機,並利用區別分析(含逐步區別分析)判別正確率;另外將原始27 個變數運用因素分析萃取之因素分別再作區別分析判別正確率。本文彙整這些模式再針對年限取前一年、前兩年及前三年來判別,並找出那一個模式及年限之組合為最佳預警模式。本文提出三個模式,均以前一年(2013 年)之資料判別正確率為最高,其中又以模式一(區別分析)之判別正確率為最佳。本研究建議以原變數群(27 個)及前一年(2013年)的資料之組合建立區別分析模式為財務預警之最佳模式。
Nowadays, many companies has occurred the issue of accounting fraud. The events of tunneling company assets and market manipulations were manipulated by insiders. It caused the breach of contract, and a huge impact to stock market. This study was desired to forecast the company whether or not it becoming a crisis company, also to establish an effective financial crisis warning system. The participants of this study used the listed food industry company from the years 2011 to 2013 as the examples; 32 selected companies were (7 attention companies, 25 regular companies), the ‘TEJ Smart Wizard ‘(TEJ) was the main financial data in the selected variables; and utilized factor analysis to analyze 27 original variables to extract the factors as discriminant analysis; and year to differentiate the accuracy rates, then figure out which analysis model and year combination was the best warning mode. The results showed that, so as to discriminant analysis, the original variables and the previous years were the best accuracy rates, i.e. year of 2012, the Taiwan domestic food safety scandals might be the reason that disrupted the food industry’s early-financial statements. The recommendation of this study was the best financial warning mode was to use the original variables and combinations of data from the previous year to establish discriminant analysis mode.