信用評等差異決定因素之分析

An analysis on determinants of differences in credit rating

蕭明輝(Ming-Wui Hsiao)

指導教授 ： 池祥麟

國立臺北大學/商學院/國際財務金融碩士在職專班/碩士(2009年)

摘要

我國已自2007年起正式實施新巴賽爾資本協定（Basel II），該協定規定，銀行須依信評機構對借款人之信用作一評比，以預估借款人可能倒帳之發生率，據以計算信用風險。並允許銀行在符合最低適用標準下，得以採用內部風險模型以計提信用風險所須的資本額。實施以來，銀行在提升資本適足率、積極建置內部信用風險模型及資料庫、及要求信用風險得以數量化等工作下，均面臨極大挑戰。

國內因信用評等發展歷史並不長久，以往信用評等研究大多以預測信用風險模型之建立，或是信用評等與經營績效關聯性之評估為主，本次研究有別於以往，係將探討銀行給予授信戶之信評等級，與外部專業信評機構如中華信評公司及台灣經濟新報，所給予之評等差異程度，期待透過選取可能的變數，利用模型分析找出造成差異之決定因素，冀望提供銀行未來修改信用風險模型或進行信評人工調整時之參考。

本研究將以Probit Model實證分析包括成立時間、營業規模、產業類別、董監持股比率、董監持股質押比率、關係人銷貨比率及關係人進貨比率等七大類變數。由於不同的信評機構係分成二組不同的分析模型，故實證分析結果顯示，在第一組全球銀行與台灣經濟新報的模型中，成立時間在第一階段是顯著的，後續二階段則均不顯著；至於營業規模則在三階段均是極為顯著，且與評等差距係呈現明顯負相關。

在第二組的全球銀行與中華信評公司模型中，再細分為金融業及一般產業。其中除營業規模在二種產業模型中的第三階段均呈現相當顯著水準外，其他變數均不顯著；而營業規模變數在這二種產業模型雖是相當顯著，然其與評等差距的關係則是呈現正向的，與前一組產出的結果恰好相反。另一個變數則是關係人進貨比率，其在一般產業中的第二、三階段的變數陸續投入後，顯著水準亦持續提升。整體而言，唯有營業規模變數不論對中華信評或台灣經濟新報均屬相當顯著，然與評等差異之關係卻存有正負向之二極結果，至於其他變數則有成立時間及關係人進貨比率在部分模型是顯著的。

關鍵字

信用評等 ； 新巴賽爾資本協定 ； Probit Model

並列摘要

New Basel Capital Accord (Basel II) has implemented from 2007, provide that bank should adopt credit rating from credit rating organisation to estimate probability of default by each borrower. In addition, it allows that bank conform to the lowest standard could adopt internal risk model to calculate capital amount for credit risk. Since Basel II implemented, Bank has faced challenge in increasing BIS, building internal risk model and database, using quantitative method to mange credit risk and so on.

The developing history of credit rating is not for a long time in Taiwan, study of credit rating in the past aims to building credit risk model or estimate relationship between credit rating and operate performance. This dissertation analyses differences degree between credit ratings from bank and external rating organization such as Taiwan Ratings and Taiwan Economic Journal Company to be reference to adjust credit risk model and rating.

This dissertation adopts 7 variables (established time, business scale, industry, Shares Ratio of Directors’and Supervisors’Shareholdings, Pledged Shares Ratio of Directors’and Supervisors’Shareholdings, related party sales ratio, related party purchase ratio) to analysis by Probit Model. Since this dissertation adopts two external rating organizations, there are two parts of result as below：

Firstly, in global bank with Taiwan Rating model, established time is significant in first phase but is not significant in next two phases. Business scale is significant in three phases and has negative relation with rating gap. Secondly, in global bank with Taiwan Economic Journal Company model, it divides two groups: banking and other industries. Only business scale is both significant in these two groups in third phase and business scale and rating gap have positive relationship. This result is opposite to Taiwan Ratings’. Besides, related party sales ratio is significant in second and third phases in other industries.

Therefore, only business scale is significant in two external rating organizations：Taiwan Ratings and Taiwan Economic Journal Company. However, they have opposite result in relationship with rating gap. Established time and related party sales ratio are significant in some models.

並列關鍵字

credit rating ； New Basel Capital Accord ； Probit Model

參考文獻

19.儲蓉，「對發展信用評等應有的態度與做法」，台灣經濟研究院副研究員。

9.李建然、廖益興，「董事會結構特徵與盈餘管理－台灣家族控制企業因素之影響」，台北大學會計系教授及中原大學會計系講師。

18.洪德生（民96），「信用評等對公司治理的重要性」，台灣經濟研究院院長。

1. Donald P. Morgan. “Rating Banks:Risk and Uncertainty in an Opaque Industry”American Economic Review, September,2002.

3. Cantor, Richard and Frank Packer (1994), “The Credit Rating Industry.” Federal Reserve Bank of New York Quarterly Review.