金融市場預警系統之建立-以紡織化纖業為例

A Study of Establishment of Early Warning System of Financial Markets—An Example of the Textile and Chemical

戴惠滿(Huey-Maan Day)

指導教授 ： 楊智超博士 ； 邱國欽博士

朝陽科技大學/管理學院/企業管理系/碩士(2014年)

摘要

摘要

金融海嘯(2008年)造成全球股市波動造成投資人投資恐慌及投資意願低落，近年接至而來的歐債危機及美國財政懸崖等，讓投資人在投資方向更為小心，企業的獲利及財務風險，若能透過預警系統讓投資大眾及金融授信人員參考，以降低成本及面臨危機來時所受害的損失，本研究以學者常用之分析模式及方法建立財務迴歸預警模式，比較模式間判別正確率之優劣，以求較佳之模式及分析方法組合。透過正確率較高的預警模式及顯著性高的變數，信用風險量化及新巴塞爾資本協定(Basel II)，內部評等法之重要內涵，透過明確的風險區分，透過內部信用評等系統的建置，企業依其信用及違約風險程度高低區分等級，可有效評估企業信用風險。本研究以台灣在2009年到2011年間上市之紡織化纖業為例，共55家公司(8家為警示股，47家正常公司)為研究對象，在變數選取方面以「台灣經濟新報」的財務合併報表數據為主。
本文選取13個變數來判別公司是否具有財務危機，利用邏輯斯迴歸與區別分析(完全輸入法及逐步迴歸)判別正確率。本文彙整這2種模式針對年限取前一年、前兩年及前三年，來判別哪一種組合之正確率較高，並透過以圖表輔助說明哪一種模式及年限組合之判別率較高。結果顯示不論年限期間為何，以邏輯斯迴歸使用變數群(13個)之判別正確率為最佳，而對任何模式分析而言，以年限為前一年之整體判別正確率較佳。本研究之結果中對變數選取、不同判別模式及年限謹供日後作相關研究者之參考。

關鍵字：財務危機預警、區別分析、邏輯斯迴歸分析、t檢定、新巴塞爾資本協定

關鍵字

邏輯斯迴歸分析 ； 區別分析 ； 財務危機預警 ； t檢定 ； 新巴塞爾資本協定

並列摘要

Abstract
Financial crisis in 2008 has led to global stock market volatility and causing the of investors decline to invest. Due to the recent European Debt Crisis and United States fiscal cliff, investors carefully consider before investing.The cost as well as the lost of the company while facing crisis could be reduced if the profit and financial crisis of the company provide a reference for investor and financial credit officer through an early warning system.This study applied an analysis model and method that frequently used by researchers to build a financial early warning system to distinguish the advantages and disadvantages of correct rate between distinct models, in order to investigate a better combination of model and analysis method.
According to a higher correct rate of the financial early warning model, significant variables, as well as Basel II, the importance of Internal Ratings Based Approach, differentiated by financial risk.The credit and the level of the default risk of the company could be differentiated and effectively evaluated through the Internal Ratings Based Approach system.This research was conducted at Taiwan during the period of 2009-2011,by using the textile and chemical fiber industry as an example. A sample of 55 companies (8 companies as warning shares, 47 companies as usual companies) was collected as the research target, and the data of consolidated financial statement from Taiwan Economic Journal was used as the variables.
This paper selected 13 variables to determine whether the company has financial crisis by using the Logistic Regression and Discriminant Analysis to differentiate correct rate. Two models were collected and emphasized on last year, previous two and three years to determine which combination has a higher correct rate, and explained which model and years limit combination has a higher discrimination rate by using chart.The result indicated that no matter what period of years, by applying the logistic regression is the best way to distinguish the correct rate of thirteen variables. A year before is the best predication for every analysis models to differentiate the correct rate. The result of this research on the selection of variables, distinguish models and years limit only provided as a reference.

Keywords: Financial Crisis Early Warning System, Discriminant Analysis, Logical Regression Analysis, T-test, Basel II

並列關鍵字

Logical Regression Analysis ； Discriminant Analysis ； Financial Crisis Early Warning System ； T-test ； Basel II

參考文獻

7.葉珮伊(2009)「台灣電子產業預警模型之研究」，朝陽科技大學研究所碩士論文。

1.Beaver, W. H. (1966). Financial ratios as predictors of failure. Journal of Accounting Research, 4, 71-111.

2.Foster, B. P. (1997). A Note on Selecting a Response Measure or Financial Distress, Journal of Business Finance & Accounting, 869.

4.Martin, D. (1977). Early Warning of Banking Failure, Journal of Banking and Finance, 1, 249-276.

5.Ohlson, J. A. (1980). Financial Ratios and the Probabilistic Prediction of Bankruptcy, Journal of Accounting Research, 18(1), 109-131.