金融市場預警系統之建立-以營建業為例

A Study of Establishment of Early Warning System of Financial Markets—as an Example for the Construction Industry

吳俊德(Chun-Te Wu)

指導教授 ： 王安平 ； 嚴宗銘

朝陽科技大學/管理學院/保險金融管理系/碩士(2011年)

摘要

2008年金融海嘯造成股市動盪造成投資人投資意願低落，所以希望建立一個預警系統讓大眾參考，以降低危機來臨時受害的損失，本研究以學者常用之分析模式及方法建立財務預警模式，比較模式間判別正確率之差異，以求較佳之模式及分析方法組合。本研究以台灣在2004年到2008年間上市櫃之營建業為例，共67家公司(12家為警示股，55家正常公司)為研究對象，在變數選取方面以「台灣經濟新報」的財務數據為主。
本文選取23個變數來判別公司是否具有財務危機，利用邏輯斯迴歸、區別分析及類神經網路判別正確率。另外，將原始23個變數利用因素分析後所萃取之因素再分別作邏輯斯迴歸分析、區別分析和類神經網路分析判別正確率結果。本文彙整這6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，以邏輯斯迴歸使用原變數群(23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。所以本文建議以原變數群(23個)及近期(前一年)資料之組合建立邏輯斯迴歸模式之財務預警為佳。

關鍵字

區別分析 ； 年限 ； 類神經網路分析 ； 營建股 ； 邏輯斯迴歸分析 ； 因素分析

並列摘要

Financial crisis cause low investor willingness to invest and stock market volatility since 2008, therefore, establishing an early warning system for investors to refer to the current research needs issues. the purpose is to reduce the victim''s loss when the crisis comes, In this research, scholars commonly used model and the method of a financial early warning model, comparison of some models of the discriminate accuracy of the advantages and disadvantages, and suggested that a better combination of models and analysis to investors. This research in Taiwan between 2004 and 2008 listed Company and the construction industry for example, total of 67 companies (12 shares for the warning, 55 normal company) as the research object, variable selection in terms of the "Taiwan Economic Journal" financial data based.
This research selection of 23 independent variables to determine whether the company has the financial crisis, use a logistic regression analysis, Discriminate analysis and neural network methods such as 3 way to determine the correct rate. This collection of patterns for the 6 years of taking the previous year, the first two years and the results of the first three years, help through which to chart a pattern of discrimination and the right combination of life is higher. showed that during the period regardless of what from the results, logistic regression model using the original variable group (23) of the independent variables of the discriminate accuracy was the best, For the six mode model, period of the previous year to determine the overall accuracy of better. Therefore, this article suggests the original variable group (23) of the independent variables and recent (previous year) a combination of information logistic regression model to establish the financial early warning is the best.

並列關鍵字

period time ； neural network analysis ； logistic regression analysis ； discriminate analysis ； factor analysis ； construction shares

參考文獻

7. 黃振豊、呂紹強，2000，企業財務危機預警模式之研究－以財務及非財務因素構建，當代會計，第1卷 (第1期)，頁 19-40。

1. Altman, E. I. (1968). Financial ratios, discriminant analysis and the prediction of corporate bankruptcy. Journal of Finance, 23(4), pp. 589-609.

2. Altman, E. I., Haldeman, R., and Narayanan, P. (1977). ZETATM analysis: A new model to identify bankruptcy risk of corporations. Journal of Banking and Finance, 1(1), pp. 29-54.

3. Altman, E. I., Marco, G. V., and Varetto, F. (1994). Corporate distress diagnosis: Comparisons using linear discriminant analysis and neural networks. Journal of Banking and Finance, 18, pp. 505-529.

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