本研究選取至95年底止,股本金額小於新台幣300億元之我國中小型銀行共24家,其87年至95年共九個年度之財務報表資料,在參考國內外相關機構及相關文獻之後,選出X1∼X28等28個變數,另加上筆者個人意見所增加之29(股東權益規模)及X30(分行家數)二個變數共30個變數,並分別歸類為F1∼F8等8個主要構面。以SAS 9.1版軟體作為 分析之工具,使用集群分析法、因素分析法、一般線性模型、Duncan多重比較檢定法加以比較排序。本研究之目的,希望能建立一套及早發現經營績效不佳銀行之方法,除能督促其早日自行改善,健全自身體質外,亦期能建立一套金融預警系統,供作主管監理機關及社會大眾之參考,避免讓問題銀行再造成浪費巨大之社會成本。
This study selects 24 the common capital stock smaller than NT$ 300 hundred million of domestic mid-banks to 2006 years end, and the finance reports from 1998 to 2006 altogether 9 years. After referring to the domestic and foreign relative organizations and literatures, this study selects X1 to X28 totally 28 variables, and according to my personal opinions, add separately X29 (shareholder equity scale) and X30 (branch numbers) two variables altogether 30 variables, and the next classification is F1 to F8 totally 8 main surfaces. We process the research by the SAS 9.1 edition software, and we use some analytic methods about the cluster analysis , the factor analysis, the discriminated method, the general linear model, the Duncan multiple range test. The purpose of this study is to warn the banks which have worse performance improve themselves automatically and earlier. And also establish a banking crisis early warning system for the reference of the government and the whole civics, to avoid waste huge society costs again.