中小企業是台灣經濟貿易發展的命脈，過去以中小企業為主的出口貿易經濟體系，是創造台灣經濟奇蹟的主要動力。隨著2006年底新巴賽爾協定的正式實施，金融機構為符合新協定規範，亦需將中小企業信用評分程序，納入其徵、授信管理系統，以求信用風險評估皆可量化處理。故本研究將資料採礦技術應用於建置中小企業違約風險模型，針對內部評等法中的企業型暴險，根據新協定與金管會的準則，不僅以財務變數為主，也廣泛增加如企業基本特性及總體經濟因子等非財務變數，納入模型作為考慮變數，計算違約機率進而建置一信用評等系統，作為金融機構對於未來新授信戶之風險管理的參考依據。而本研究將以中小企業中製造傳統產業公司為主要的研究對象，建構企業違約風險模型及其信用評等系統，資料的觀察期間為2003至2005年。 本研究分別利用羅吉斯迴歸、類神經網路、和C&R Tree三種方法建立模型並加以評估比較其預測能力。研究結果發現，經評估確立以1:1精細抽樣比例下，使用羅吉斯迴歸技術建模的效果最佳，共選出六個變數作為企業違約機率模型之建模變數。經驗證後，此模型即使應用到不同期間或其他實際資料，仍具有一定的穩定性與預測效力，且符合新巴塞資本協定與金管會的各項規範，表示本研究之信用評等模型，確實能夠在銀行授信流程實務中加以應用。 / To track the development of Taiwan’s economy history, one very important factor that should never be ignored is the role of small enterprise businesses (the SMEs) which has always been played as a main driving force in the growth of Taiwan’s export trade economic system. With the formal implementation of Basel II in the end of 2006, there arises the need in the banking institutions to establish a credit scoring process for the SMEs into their credit evaluation systems in order to conform to the new accords and to quantify the credit risk assessment process. Consequently, in this research we apply data mining techniques to construct the default risk model for the SMEs in accordance to the new accords and the guidelines published by the FSC (the Financial Supervisory Commission). In addition we not only take the financial variables as the core variables but also increase the non- financial variables such as the enterprise basic characteristics and overall economic factors extensively into the default risk model in order to formulate the probability of credit default risk as well as to establish the credit rating system for the enterprise-based at risk for default in the IRB in the second pillars of the Basel II. The data which used in this research is taken from the traditional SMEs industry ranging from the year of 2003 to 2005. We use each of the following three methods, the Logistic Regression, the Neural Network and the C&R Tree, to build the model. Evaluation of the models is carried out using several statistics test results to compare the prediction accuracy of each model. Based on the result of this research under the 1:1 oversampling proportion, we are inclined to adopt the Logistic Regression techniques modeling as our chosen choice of model. There are six variables being selected from the dataset as the final significant variables in the default risk model. After multiple testing of the model, we believe that this model can withstand the testing for its capability of prediction even when applying in a different time frame or on other data set. More importantly this model is in conformity with the Basel II requirements published by the FSC which makes it even more practical in terms of evaluating credit risk default and credit rating system in the banking industry.

全球化潮流方興未艾，基於與國際接軌目標，我國金融業自2006年起實施新巴塞爾資本協定，期於現今日新月異金融環境中以全球一致性的銀行管理方法及制度落實其精神。實施新巴塞爾協定後，首當其衝者便是台灣產業發展主體—中小企業。以信用風險中資本計提為例，中小企業不若大型企業體質健全，且財務透明度亦為人詬病，相對提升金融機構授信風險，進而導致中小企業融資授信審查趨於嚴格與保守，中小企業融資難度與成本皆大幅增加。 有鑑於此，本研究以中小企業中電子業為主要研究對象，採資料採礦流程進行信用評等模型建置。為求配適最佳違約機率模型，分別以不同精細抽樣比例逐一配適羅吉斯迴歸、類神經網路及分類迴歸樹等統計模型，經評估後篩選出羅吉斯迴歸模型建置信用評等系統。再者，為確認模型與信用評等系統建置適當，係遵循新巴塞爾協定相關規範進行各項測試及驗證，結果顯示模型於樣本外資料測試表現良好，信用評等系統亦通過正確性分析、等級區隔同質性檢定及穩定度分析等驗證準則，冀能提供金融機構一套有效且精簡的信用管理機制，建立與中小企業間資訊對稱管道，於兩造雙方取得互利平衡，防範危機於未然。 / Globalization trend is still growing. Because of the objective of connecting to the world, the banking and finance industry in Taiwan has implemented the New Basel Capital Accord since 2006, hoping to make use of globally consistent banking management method and system to implement its spirit in this changing financial environment today. After the implementation of the New Basel Capital Accord, the principal development part in Taiwan industry, medium- and small-sized enterprises, is the first to be affected. For example, with regard to the capital requirements in credit risks, the constitution of medium- and small-sized enterprises is not as sound as large-sized enterprises’, and the financial transparency of medium- and small-sized enterprises is insufficient that the credit risk of financial institution would be lifted comparatively; and then, the finance and credit investigation of medium- and small-sized enterprises would become strict and conservative, thus the finance difficulty and cost of medium- and small-sized enterprises would be increased substantially. In view of this, this study regards the electronics industry from medium- and small-sized enterprises as the main study objects, and data mining procedures are used so as to establish the credit scoring system. To get the best probability model of default, different oversampling ratios are used one by one to match such statistical models and logistic regression, Neural Network Analysis, and C&R Tree; and logistic regression model is selected for the establishment of credit scoring system after assessment. Moreover, relevant the New Basel Capital Accord standards are followed to carry out every test and verification so as to confirm that the establishments of model and credit scoring system are appropriate. The result indicates that the model has good performance in out-sample test, while credit scoring system also passes such verification standards as accuracy analysis, level segment homogeneity test, and stability analysis. Hopefully, this study result can provide a set of effective and simple credit management system for the financial institution to establish information symmetrical channel with the medium- and small-sized enterprises, so that both parties can obtain mutual balance and the crisis can be alerted in advance.

信用評等機構對於企業與市場投資者而言有著重要且無可取代的功能，其所提供之信用評等資訊應當是許多市場投資者所仰賴的重要決策依據，但近年來，卻有許多外界聲浪質疑信評機構評等之準確性，本論文之研究目的即是希望從錨定效應此一行為偏誤之觀點切入，探討國內信評機構在對企業評等時，是否會受到錨定效應影響，導致評等調整不正確或是評等落後其他財務指標等現象發生。 研究結果顯示，國內信評機構對受評企業之過往財務資訊存在錨定現象，但不至於大幅影響整體違約風險之準確性，且信評機構對受評企業財務之惡化較為敏感。另一方面，本研究也發現，信評機構對於非上市櫃公司、金融業以及初次評等等級在「twAA」以上企業之評等，存在較為明顯之錨定現象。 / Credit Rating Agencies (CRAs) play a major role in the financial market. Credit rating purport to provide investors with valuable information they need to make decisions about investing, but the accuracy of the rating itself has been called into question by many investors in recent years. The purpose of this study is to examine the anchoring effect on CRAs while the rating is being given. The results indicate that domestic CRAs tend to be anchored on the past financial information of the issuers. But the impacts are very slightly. Besides this, CRAs seem relatively sensitive to the financial deterioration. Moreover, the anchoring effect are much more significant when the debt issuers are private firms, financial institutions or the companies with greater or equal to twAA initial credit rating.

落實風險管理為公司治理與透明化過程之重要指標－2008年金融風暴之發生再度引起國內外對風險管理之重視，此也顯示風險管理並無真正落實；為達成此目標，企業風險管理之推動為關鍵因素，此部分除依賴監理機關制定規範要求保險業實行外，獨立第三者之信用評等機制亦扮演重要訊息提供之角色，藉由其對企業風險管理發展出之評鑑標準，及其對眾多受評公司進行評等的過程及瞭解，可提供保險業建置與落實企業風險管理之方向及做法。 保險業主要信用評等機構對企業風險管理的評鑑標準不同，本研究找出最佳信用評等機構之評鑑項目－風險管理文化、風險控管程序、新興風險管理、風險及經濟資本模型、策略性風險管理，及相關之評鑑標準；另藉由信用評等機構對保險業企業風險管理之調查分析，瞭解國內保險業者目前所面臨之關鍵問題：(1)大部分仍停留在傳統風險管理階段；(2)風險胃納、風險容忍度及風險限額定義及關聯不明瞭；(3)經濟資本及風險調整後報酬量化能力不足。 歸納本研究結果，(1)藉信評機制對企業風險管理之評鑑項目及標準，可為國內保險業建置企業風險管理之參考依據；(2)從信評機構對企業風險管理之評等，可瞭解國內保險業之執行成果及問題，具有值得參考之價值。本文亦提出風險管理文化、風險及經濟資本模型相關之建議，以落實企業風險管理。 / To truly implement risk management is the important indicator of corporate governance and transparency progresses－the finanacial crisis in 2008 has resulted in the focus of risk management again, of which the implications have presented it hadn’t been done well. The Enterprise Risk Management (ERM) is the key factor to achieve the objective of truly implemented. In addition to the supervisors imposing those regulations so as to request execuation by insurance industy, the independent third party－rating agencies have provided valuable information. By way of the ERM indicators developed by rating agencies, together with the experiences and understandings getting from rating procedures, they can give the ERM directions and practises to insurance industry. The major rating agencies for insurance industry have different views to ERM critirias. The thesis has found out the best items and their relative indicators developed by rating agency－Risk Management Culture, Risk Control Processes, Emerging Risk Management, Risk & Economic Capital Models, and Strategic Risk Management. Moreover, the thesis has explored the key issues faced by domestic insurance industry through the ERM surveys conducted by rating agencies: (1)the majority of insurers have still positioned at the stage of traditional risk management; (2)the understanding to definitions and relatives of risk appetite, risk tolerance and risk limit are inambigious; (3)the quantibility of economic capitals and risk-adjusted returns aren’t appropriate. The thesis has concluded: (1)the ERM items and critirias developed by rating agencies can be the references to establish their own ERM by domestic insurance industy; (2)they are valuable references to know the situations of implementation and issues faced by domestic insurance industry through the ERM rating items and critirias. In the end, the thesis has submited the suggestions within risk management culture and risk & economic capital models in oder to truly implement ERM.

資訊不對稱可能造成市場中買賣雙方無法順利進行交易的阻力之一，而評等制度可以為資本供需雙方架起資訊的橋樑，讓資本市場發揮資金仲介的功能，使得投資者可以降低風險的不確定性、信用風險較低的企業減少付出的籌資成本，進而擴大市場的深度與廣度，增進資本市場的整體效率。為提供金融市場透明化訊息，並有別於過去相關研究方法，本研究多變量統計方法之因素分析(Factor Analysis)及順序羅吉斯迴歸(Ordered Logistic Regression)為基礎，分析影響壽險業信用評等之重要因素，以期在分析過程中獲得一套有效信用評等過程的關鍵方向。 　　本文以1994年至1997年接受A.M. Best評等之591家美國壽險公司為研究對象。美國壽險公司年報資料來源為NAIC DAATABASE PRODUCT光碟資料庫;而等級資料來源則是A.M. Best資料庫。首先取1994年至1996年共三年之財務資料與評等結果為估計樣本建立順序羅吉斯迴歸模型，並以相同公司的1997年的資料為保留樣本，作為等級預測之用，同時以多變量之因素分析建立1997年之分析預測模型，來比較兩模型之差異及解釋影響模型之重要因素，進而驗證壽險公司之財務報表資料與評等結果間是否有直接關係存在。

市場上存在許多傳統資產模型無法解釋的異常現象，本論文將探討台灣證券 市場異常報酬投資策略之獲利與財務危機間的關聯性，重點放在價值型投資策略， 由買進高淨值市價比的公司股票，放空低淨值市價比的股票，建構出價值型投資 組合。此投資策略的主要獲利來源出自投資組合的多頭部位，即高淨值市價比的 公司。且信用風險作為財務危機的代理變數，在解釋異常報酬上扮演重要角色， 價值型投資策略的異常報酬在高信用風險公司是較大的，表示財務危機的影響是 有反映在股票報酬上的。而與美國市場的結果不同，信用評等降評對報酬的影響 在台灣證券市場並不顯著。 / Anomalies exist in the markets that cannot be explained by traditional asset-pricing models. This paper assesses implications of financial distress for the profitability of anomaly-based trading strategies in Taiwan stock market. We focus on the value-based strategy which conditions on the BM ratio. It involves buying highest BM and selling lowest BM stocks. Financial distress, as proxied by rating downgrades, is likely to be a primary ex ante indicator of a company’s future performance. Anomaly returns of value-based strategy are bigger in high credit risk companies. But unlike the evidence in U.S market, rating downgrades only have limited impact on stock returns in Taiwan.

1999年國際清算銀行巴塞爾銀行監督管理委員會發布了新巴塞爾資本協定架構，歷經多次討論修改後，於2004年6月公告新版巴塞爾資本協定(Basel Accord II)，並將於2006年底開始施行。我國金管會於93年9月亦發函銀行業者並公佈因應新制協定相關措施，期能在2006年底與國際同步正式實施,這不僅是國際金融史上重要的大事,也是我國金融史上重要的里程碑,因為它的象徵意義顯示出我國金融業風險管理制度的提升跨向國際同步的企圖心。 本研究從探討新舊版巴塞爾資本協定內涵的差異到我國金管會與銀行業者 因應新制的準備概況,繼而以C銀行為例說明,並聚焦在執行信用風險管理內部 評等模型法(IRB法)之前與之後銀行對服務業的中型與中小型企業授信融資之影響進行實證研究,本研究隨機抽取96家企業授信案件並假設了11項參考變數 分別為: 信評等級、額度、期間、授信產品項目數、擔保比率、風險成本率、台幣放款利率、權限核章人數、申貸流程所需天數、授信主力銀行、授信控管條件項目數等,作為銀行執行新舊不同信用風險評等模型制度對企業授信融資的差異 研究重點。 實證研究結果:(1)各假設變數中僅風險成本率與授信產品項目數兩項在新 舊不同制度中具顯著差異的。(2)較佳的信評等級所具有的融資優勢程度在新制度執行後更顯相對重要。(3)面對中、小型企業,銀行徵提擔保品、增加擔保率是最快速有效降低風險成本率的方法。(4)新制施行初期難敵市場壓力反應成本利率價格。(5)從有無內外帳企業的信評等級結果發現其對融資條件的影響有顯著差異。(6)是否為家族企業，不因新制施行受到差別影響。 本研究並提出建議:(1)政府應繼續並擴大對於中小企業信用基金保證規模 ,尤其應考慮修改服務業產業之中小型企業規模定義。(2)銀行應審慎面對信用評等與融資報價條件之間的兩難。(3)中小企業應加強經營制度透明化,尤其在財務會計制度方面。(4)銀行應加強對企業之產業瞭解深度以利授信金融商品創新與包裝架構。(5)銀行應重新審視績效考核管理機制。(6)銀行實施新制後應每年檢核模型歷史資料,把景氣變動因子列入修正參考。

本研究主要是應用二元分量迴歸BQR（Binary Quantile Regression）模型的方法估計銀行對中小企業授信之信用評等，以期提早偵測出可能會有違約還款的企業，達到授信時的預警效果。信用評等目的為協助金融機構在貸放前更明確的瞭解企業的信用風險，並具以衡量是否核准貸款的重要依據。在過去的研究中最廣為應用的計量方法主要為有母數（parametric）區別迴歸模型，包括Logit Model和Probit Model等區別迴歸模型，這二種模型在正確的條件設定之下，模型的預測結果可以說相當的好，但若是估計資料的分配並未符合所設定的條件，或者是資料具有無法觀察到的異質變異（heteroskedastic），則估計結果會有顯著的偏誤。傳統區別模型的一般設定如下，假設發生違約的機率給定為： ，此處 表示實際上是否真的發生違約逾期還款的情形。 為了在估計時更能控制風險，最近許多有關信用評等的研究方法傾向使用半無母數（semiparametric）單一指數模型以及無母數（nonparametric）的估計方法，如類神經網路與歸納樹（classification trees）分析方法。 而本文主要是將半無母數的分量迴歸區別模型和過去以有母數為主的Probit及Logit區別迴歸模型做比較。Koenker和Bassett（1978）提出分量迴歸估計方法（Quantile Regression Methods），分量迴歸可以更完整的反應出共變異效果對被解釋變數的影響，除此之外，分量迴歸模式提供使用上較多的彈性，在估計時無需對母體的分配做假設，另外，和傳統的最小平方（OLS）估計法不同在於OLS給予估計參數的分量為50%，因此OLS估計出的迴歸線只有一條，因此分析解釋變數對被解釋變數的影響是平均效果；分量迴歸區別模型則給予估計參數不同百分比的分量，從而可在相同樣本下得到不同的分量迴歸線，觀察解釋變數對於被解釋變數影響程度的變化，因此藉由不同分量估計出不同的迴歸係數 ，可以更加瞭解整體分配的全貌。

本研究旨在探討企業盈餘管理行為對其信用評等之影響，以Ordered Probit Model，探討台灣上市櫃(包含興櫃)企業之盈餘管理行為，對台灣經濟新報社(TEJ)的台灣企業信用風險指標(Taiwan Corporate Credit Risk Index, 簡稱TCRI)是否造成影響，並進一步觀察其影響為何。 本研究發現，企業之盈餘管理行會影響其信用評等，而且透過進一步之邊際影響分析，顯示企業盈餘管理對當期盈餘之影響，與信用評等之間存在著反向的關係，意味著，信用評等人員在使用財務報表，進行信用評等程序時，統計上會因為企業之盈餘管理行為而有處理上之差異；本研究另外發現就平均而言，盈餘增加(減少)之裁量性應計數影響TCRI下降機率的增加(減少)幅度，大於TCRI不變以及TCRI上升之機率變動幅度。 / This study investigates the influence of earnings management on credit rating. Using Ordered Probit Model, this study tests whether earnings management of listed companies (emerging stock included) effects Taiwan Corporate Credit Risk Index (TCRI). This study finds that earnings management does affect credit rating. In addition, it suggests that earnings with income-increasing discretionary accruals have more opportunity for getting downgradings than earnings with income-decreasing discretionary accruals through analysis of marginal effects. That is, the credit analysts treat earnings management in a different way when rating companies. Moreover, this study also finds that the marginal effects of income-increasing discretionary accruals of increasing opportunity for getting downgradings are larger than the marginal effects of opportunity for getting unchanged and upgradings. Similarly, the marginal effects of income-decreasing discretionary accruals of decreasing opportunity for getting downgradings are larger than the marginal effects of opportunity for getting unchanged and upgradings.

信用風險或信用違約意旨金融機構提供給客戶服務卻未得償還的機率，故其在銀行信貸決策的領域是常被鑽研的對象，因為其對於金融機構所扮演的角色尤其重要，對商業銀行來說更是常難以解釋或控制，然而拜現今進步的科技所賜，金融機構可以藉由操控較過去低的成本即可進一步發展強健且精煉的系統與模型去做預測還有信用風險的控管，有鑑於對客戶的評分自大數據時代來臨起，即使是學生亦開始有了可以評鑑的痕跡，憑藉前人所實驗或仰賴的基本考量面向如客戶基本資料、財力狀況或是其於該公司今昔的借貸訊息，再輔以藉由開放資料所帶來的資訊，發想可能影響信用違約率的變數如外在規範對該客戶的紀錄，想驗證是否真有尚可開發的方向，若有則其影響可以到多深。 眾所皆知從過去到現在即有很多種方法被開創以及提出以預測信用違約率，當然所使用的方法和金融機構本身的複雜性、規模大小以及信貸類型有關，最常見的有判別分析，但其對於變數有嚴格的假設，而新興的方法神經網路可以克服判別分析的缺陷且預測的效能也不錯，但神經網路只給予預測結果而運算過程是未知的，對於想要了解變數間的關係無濟於事，故還是選擇從可以對二元分類做預測亦可以藉由模型係數看到應變數和自變數間關係的羅吉斯迴歸方法著手，而研究過程即是依著前人對於羅吉斯迴歸在信用風險上的繩索摸索，將資料如何清理、變數如何轉換、模型如何建立以及最後如何篩選做一個完整的陳述，縱然長道漫漫，對於研究假設在結果終得驗證也始見曙光，考慮的新面向確有其影響力，而在模型係數上也看到其影響的大小，為了更彰顯羅吉斯迴歸對於變數間提供的訊息，故在最後將研究結果以較文字易讀的視覺化方式作呈現。 / Credit risk or credit default means the probability of non-repayment that banks or financial institutions get after they provide services to their customers. Credit risk is also studied intensively in the field of bank lending strategy because it’s usually hard to interpret and control. However, thanks to advanced technology nowadays, banks can manipulate reduced cost to develop robust and well-trained system and models so as to predict and mange credit risk. In the light of the score on customers from the beginning of big data era, every single one can be tracked to assess even though he or she is student. Relying on common facets like personal information, financial statement and past relationship of loan in a specific bank, come up with possible variables like regulations which influence credit risk according to information from open data. Try to verify if there is a new aspect of modeling and how far it effects. As everyone knows, there are several created and offered methodologies in order to predict credit default. They differ from complexity of banks and institutions, size and type of loan. One of the most popular method is discriminant analysis, but variables are restricted to its assumption. Neural network can fix the flaws of the assumption and work efficiently. Considering the unknown process of calculation in neural network, choose logistic regression as research method which can see the relationship between variables and predict the binary category. With the posterior research on credit risk, make a complete statement about how to clean data, how to transform variables and how to build or screen models. Although the procedure is complicated, the result of this study still validates original hypothesis that new aspect indeed has an impact on credit risk and the coefficient shows how deep it affects.