| 摘要: | 我國自1988年到1999年間發生了許多的地雷股事件，當時的會計舞弊事件不斷發生，讓投資人及國家經濟蒙受了極大的損失。而近一年多來，相關的財務報表操縱問題不斷浮上檯面，地雷股事件又再度重演，投資人及國內股市又再度經歷了信心喪失的危機。足見我國企業管理當局仍會為了自身的利益將財務報表數字操弄於股掌之間，而讓代理問題層出不窮。而當公司對該財務報表進行會計操縱行為，投資人或金融機構卻無法察覺出其中的異樣時，將可能造成未來的投資損失，更甚者會引發地雷股效應，最終除了影響投資人權益外，亦可能會影響到國家社會的總體經濟的投資發展。因此為了如何區分出適當的會計資訊即為本研究的動機。 本研究藉由八大類共三十二項會計操縱變數的檢定，選取出具有代表性的會計操縱衡量變數，除此之外，並藉由複迴歸分析，了解何種型態的公司以及何種型態的籌資方式較具有會計操縱行為。而藉由此一研究能讓投資人與金融機構，在未來面對投資標的以及貸放對象時，能更有效掌握其會計資訊。本研究由證期局與台灣經濟新報社的資料庫中，取得民國九十二年度中有進行公開舉債以及銀行借款公司之樣本，並利用該樣本公司前三年的季財務報表資訊作為會計操縱行為的研究樣本。研究中並且將樣本公司依據MMH模型區分為財務壓力公司與非具財務壓力公司，檢視在不同財務壓力下的公司，其進行不同籌資活動時，是否具有不同程度之會計操縱程度。而在複迴歸分析方面，則將迴歸式分作四大類，希望藉由迴歸式的變數分析，尋找出在不同型態公司從事不同籌資活動下，影響其會計操縱行為的主要因素。 實證結果發現，研究中所列之三十二項會計操縱衡量變數，僅有第一類變數（除了RECPC與RECSLSPC衡量變數外），第三、四、六類變數以及第七類變數中的NILESSCFFO衡量變數與第八類變數中之IVSMAG變數都具有相當的會計操縱衡量能力，為良好之會計操縱衡量指標。而在迴歸分析中，則發現不同財務壓力之公司在進行不同之籌資方式時，有其不同的顯著因子會影響其會計操縱行為。但對於過去實證中所檢視的變數仍為顯著因子之ㄧ，如公司規模變數、負債權益比變數、董監事持股比率變數都為顯著影響會計操縱行為的解釋變數。除此之外，電子產業變數亦對於會計操縱行為有顯著的影響，至於其他類型之變數則會因不同的財務壓力及不同的籌資方式，而有不同的解釋能力。投資人除了要慎選會計操縱指標外，面對不同類型的籌資公司，亦要利用不同的分析因子，了解其會計數字之操縱性。 From 1988 to 1999, there were many incidents about accounting fraud in Taiwan. Those incidents caused the investors and country’s economy to suffer the enormous losses. In the recent year, the incidents of financial fraud appear again in Taiwan, therefore accounting manipulation become the important study in the academic community. When the investors and financial institutions do not detect the inappropriate financial statements, they may suffer the investing losses, extremely cause economic recession. Consequently, to discriminate the appropriate accounting information is the objective of this research. There are 32 accounting manipulation variables of 8 categories in this research. We use t-test, chi-square test, and Wilcoxon Singed–Ranks test to sieve the fitting variables to detect the accounting manipulations which affect the investors to make the investing decision. In following sections, we use the multi-regression analysis to select the remarkable variables which influence the companies to be engaged in the accounting manipulations. The public debts and Bank loans samples of this research are picked from the TEJ and SEC database separately. And we use the season financial data to analyze the accounting manipulation of those sample companies in the four different classifications. In the empirical results, there are only the first (except the RECPC and RECSLSPC variable), the third, forth, and sixth category, and the NILESSCFFO variable of the seventh category and the IVSMAG variable of the eighth category which can detect the accounting manipulation of sample companies remarkably. In the multi-regression analysis of this research, we find that the sample companies of different financial stress would be affected by different explanatory variables. However the common explanatory variables in past empirical researches still play the important roles to explain the level of accounting manipulation, for example: the company size variable, the debt/equity variable, and the holding percentage of the directors. Besides, we also find that the electronic industry variable is the remarkable variable to explain the degree of accounting manipulation in those sample companies. Hence, the investors and financial institutions must select the suitable variables to analyze the accounting manipulations in order to make the proper investment decisions. |
