Please use this identifier to cite or link to this item: https://ah.lib.nccu.edu.tw/handle/140.119/114696
題名: 財務困難公司下市櫃之離散時間涉險預測模式
Using Discrete-Time Hazard Models to Forecast Financially Distressed Firms Delisted from the Taiwan Stock Exchange and the TAISDAQ
作者: 吳清在
謝宛庭
Wu, Tsing-Zai C.
Hsieh, Wan-Ting
關鍵詞: 全額交割股 ; 離散時間涉險模式 ; 基準涉險模式 ; 簡單涉險模式
Margin trading ; Discrete-time hazard model ; Baseline hazard model ; Simple hazard model
日期: Jul-2004
上傳時間: 15-Nov-2017
摘要: 本研究旨在探討台灣股市全額交割股下市櫃之涉險預測模式。本研究根據國內外有關財務困難公司之相關文獻,就可能影響全額交割股下市櫃之決定因素,以一般結構、財務結構、營運結構及股權結構,建構財務困難公司從列為全額交割股至下市櫃期間之預測模式。異於傳統企業財務困難之研究,本研究採用結合存活分析及「logit模式」之「離散時間涉險模式」。此種涉險模式考慮上市櫃公司之股票從因發生財務困難而列為全額交割股後至下市櫃之存活時間內每一期之風險,故所採用解釋變數之數值亦隨時間而變動。因此,「離散時間涉險模式」使用研究期間所有可能之資訊來預測每一公司-季之下市櫃機率,此與單期logit模式僅用下市櫃當期之公司資訊有所不同。實證結果發現,在顯著水準10%下,「離散時間涉險模式」之時間虛擬變數及解釋變數如公司規模、負債比率、權益市價對總負債比率、現金流量對總負債比率與董監持股質押比率為顯著且符合預期方向,而「logit模式」中僅有負債比率、資產報酬率、董監持股質押比率與董監經理人持股比率為顯著變數。就模式預測精確度而言,本研究亦發現「離散時間涉險模式」優於「logit模式」,此一動態涉險預測模式較能有效預測財務困難公司下市櫃之風險。因此在建立預警模式時,已經考量企業存活時間每期風險之「離散時間涉險模式」較僅考慮企業在下市櫃當期資訊之「logit模式」更加周延。
This study develops a hazard model to forecast financially distressed firms that face the delisting risk in the Taiwan Stock Exchange and the TAISDAQ. We first identify determinants for firms delisted from these exchanges in terms of a firm’s general structure, financial structure, operational structure, and ownership structure, and then construct the forecast model in the discrete-time hazard method. Unlike prior studies, we employ discrete-time hazard model that combines survival analysis and logit model to forecast the delistings. A hazard model corrects coefficients for periods at risk and allows time-varying covariates. It uses all available information to estimate the delisting probability for all firms at each point in time. The empirical results show that significant independent variables of our discrete-time hazard model include the time dummy variable, the firm size, the debt ratio, the equity to liabilities ratio, the cash flow to liabilities ratio and the extent of shares as collateral by the board of directors (hereafter SCBD). Under the logit model, however, only the debt ratio, the net income to assets ratio, SCBD, directors’ and managers’ shareholdings are significant independent variables. Based on the forecast accuracy of the model, we find that discrete-time hazard model is more preferable to the logit model. In developing a forecast model to predict the delistings, discrete-time hazard model is more comprehensive.
關聯: 會計評論, 39, 55-88
資料類型: article
DOI: http://dx.doi.org/10.6552%2fJOAR.2004.39.3
Appears in Collections:期刊論文

Files in This Item:
File Description SizeFormat
index.html106 BHTML2View/Open
Show full item record

Google ScholarTM

Check

Altmetric

Altmetric


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.