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.

This thesis surveys the method of extending Cox proportional hazard models (1972) and the general class of semiparametric model (2004) in the upgrades or downgrades of credit ratings by S&P. The two kinds of models can be used to modify the relationship of covariates to a recurrent event data of upgrades or downgrades. The benchmark credit-scoring model with a quintet of financial ratios which is inspired by the Z-Score model is employed. These financial ratios include measures of short-term liquidity, leverage, sales efficiency, historical profitability and productivity. The evidences of empirical results show that the financial ratios of historical profitability, leverage, and sales efficiency are significant factors on the rating transitions of upgrades. For the downgrades data setting, the financial ratios of short-term liquidity, productivity, and leverage are significant factors in the extending Cox models, whereas only the historical profitability is significant in the general class of semiparametric model. The empirical analysis of S&P credit ratings provide evidence supporting that the transitions of credit ratings are related to some determined financial ratios under these new econometrics methods.

信用風險是金融機構最關切的風險之一，信用評等則是具有公信力的評等公司對企業債信良窳的客觀評估。本研究目的即在於探討在效率市場的假設前提下，股票價格所蘊含的信用風險與信用評等間之相互關係。我們以接受信用評等的上市櫃公司為研究對象，利用KMV模型求解出違約距離(Distance-to-Default, DD)，再使用Kalman Filter粹取出符合公司之市場資訊(Adjusted-DD)替代股票價格，本研究分成兩部分討論，第一部份以順序羅吉斯模型(Ordered Logit Model)及羅吉斯模型(Logit Model)探討股價變動是否能領先告知未來公司信用評等的變化情形。第二部分則利用一般化自我相關條件異質變異模型(GARCH Model)觀察信用評等變動是否為市場帶來新的資訊。 第一部分實證結果發現：當長期信評調降時，電子、通訊相關產業及金融業之市場資訊(Adjusted-DD)的變動與長期信用評等為負相關，而長期信用評等調升時，仍得到負向關係，短期信用評等之部分得到當信用評等調降時，電子與通訊相關產業市場資訊變動有負向關係，與長期信用評等得到一致性之關係，但金融業則顯示市場資訊與信用評等調降為正相關，而傳統產業顯示短期信用評等調升與市場資訊呈現負向關係。 第二部分實證結果：與大多學者之研究相符，當信用評等調降時股票價格有負的異常報酬，而本研究更進一步發現當信用評等調升時，股票價格同樣隨著上漲且有顯著結果，兩者具有對稱關係。故信用評等改變時能夠為市場帶來新的資訊，可視為投資的重要參考指標之一。但以股票價格所蘊含的信用風險與信用評等間的關係卻仍無法得到應證。

2008年，美國華爾街危機影響全球金融市場，即使美國擬出許多救市計畫，全球股市依舊暴跌。在此危機衝擊下，各大金融機構不但利潤下滑，且資產減記和信貸損失也愈來愈嚴重。造成此一現象的主因即是次級房貸的影響，次級房貸主要是針對收入低、信用不佳卻需要貸款購屋的民眾，這類客戶通常借貸不易，倘若銀行內部沒有完善的評等機制那放款則需承受較大的違約風險。為因應此趨勢，本研究以台灣未上市中小企業為實例，資料的觀察期間為2003至2005年，透過資料採礦流程，建構企業違約風險模型及其信用評等系統。 本研究分別利用羅吉斯迴歸、類神經網路、和分類迴歸樹三種方法建立模型並加以評估比較其預測能力。發現羅吉斯迴歸模型對於違約戶的預測能力及有效性皆優於其他兩者，並選定為本研究之最終模型，並對選定之模型作評估及驗證，發現模型的預測能力表現尚屬穩定，確實能夠在銀行授信流程實務中加以應用。 / In 2008, the financial crisis on Wall Street had severe impacted the global economy. Although the US government has drawn up regulatory policies in an attempt to save the stock market, the value of global stock market has shrunk drastically. As such, the profits of many financial institutes’ have not only plunged, their value of assets have decreased while loss related to mortgage became more severe. The main cause behind this global phenomenon can be attributed to the effect of subprime mortgages. Subprime mortgages are mainly aimed at consumers who have low income and poor credit history but wish to purchase homes through the means of mortgage. These consumers usually find it difficult to obtain mortgage loans. If banks do not have a well structured evaluation system, they would have to bear more risks in the case of a default. To better understand this trend, this research chooses middle and small private enterprises as its samples. The period of observation is 2003 to 2005. Using the data mining process, this research builds a model that shows the risk associated with contract failure and credit score system. The research builds a model based on logistic regression, Neural Network, and cart to compare and contrast each of the three model’s ability to predict. The result shows that logistic regression is better at predicting defaults and is more effective than the other two models. The research, therefore, concludes logistic regression model as the research’s final model to study and evaluate. In process, the research result demonstrates that the logistic regression model makes more precise prediction and its prediction is fairly stable. Logistic regression model is capable for banks to employ in performing credit check.

為了研究總體因子與產業違約率之間的關聯性, 本文以信用投資組合觀點模型(CPV) 做為開端, 建立在具評等基礎下的違約損失模型, 並以投機等級違約率估計出移轉係數矩陣, 進而模擬各產業條件移轉矩陣, 藉以反應在各種不同總體情境下, 產業內各評等的移轉機率及違約機率。此外, 本文亦建立不分評等的簡化違約損失模型, 並將兩模型做一比較。最後, 我們以台灣537 家上市櫃公司做為投資組合樣本, 分別模擬出兩模型的條件違約損失分配。進一步計算風險指標,以此做為未來規劃資本計提的基礎。最後結果顯示, 投資組合違約情況確實受總體因子影響, 且發現若投資組合中評等越差公司之曝險越小, 將有助於降低組合資產風險。

本研究以權益基礎報酬占總報酬的比率來捕捉薪酬結構，並據以探討給予審計委員會的薪酬結構對於公司權益資金成本及信用評等之影響。利用 2006 至 2010 年間納入美國 S&P1500指數之公司 (排除金融服務與保險業) 為樣本，本研究發現，若權益基礎報酬佔審計委員會薪酬比率越高，其公司之權益資金成本越低，但該公司之信用評等卻越差。顯示權益基礎報酬之比重在二種財報使用者眼中具有不同涵義。投資人認為給予審計委員會較高之權益基礎報酬比重，可使監督更有效，投資人承擔之資訊風險降低，進而願意降低其要求報酬；信用評等機構則認為，給予較高的權益基礎報酬比重將傷害審計委員會獨立性，影響公司治理結構，並降低財務報導之品質，因而給予此類公司較差之信用評等。 / This study examines how investors and credit rating agents react to audit committees’ equity-based compensation. Based on a sample of S&P 1500 firms during 2006-2010, the empirical results show that firms who pay audit committees higher portion of equity-based compensation have lower cost of equity capital and lower credit rating. These results suggest different information users perceive and react to equity-based compensation in different ways. Particularly, investors appear to perceive that higher portion of equity-based compensation can align audit committee members’ interest with the shareholders’, leading to more effective monitoring and smaller information risk. Therefore, investors react by reducing their cost of equity capital. In contrast, credit rating agents appear to perceive that higher portion of equity-based compensation may harm audit committees’ independence, resulting in decreased quality of financial reporting. Therefore, credit rating agents react by downgrading firms’ credit ratings.

本文主要是討論市場上有哪些因素會影響信用違約交換的價格(價差)，並且透過2005年初發生的通用汽車與福特汽車信用評等調降事件，研究信用評等的改變對於股票、債券與信用違約交換市場的影響。 一開始先介紹信用衍生性商品市場的發展。第二部份則介紹評價信用違約交換的模型，並由模型中找出可能影響信用違約交換價格的因素，並且提出公司本身發佈的消息也可能會影響價差的改變，甚至更為明顯，但沒有任何一個評價模型包含這樣一個因素。而透過對通用汽車與福特汽車事件的研究，我們發現兩家公司的股票、公司債或是信用違約交換價格(價差)都在評等調降的消息發布前已經事先反映公司經營不善的狀況，或是所面臨的困境，而在評等調降結果真正公佈的時點時，市場的反應反而沒有預期的明顯。對於公司內部發佈的消息，或是預期之外的事件，價格或價差則會大幅波動。

長久以來，信用風險一直是各銀行經營風險中最主要的來源，而就信用風險的衡量部份，巴塞爾委員會希望國際性銀行最低限度必須採用中等複雜程度的風險計算方法。也就是希望銀行能以新巴塞爾協定中信用風險的內部評等法為基本精神建置一套內部自有的信用風險模型來評估交易對手的信用風險。 同時，由於目前國內對於自有信用風險模型的建置和效力驗證的相關研究付之闕如，故本研究以新巴塞爾協定中信用風險的內部評等基礎法為基本精神，並且應用倒傳遞類神經網路方法，建構一套有效的信用風險模型並加以驗證以期能應用於銀行授信決策系統之中，也擬扮演一拋磚引玉的角色，以期未來有更多資源投入相關研究。 首先，本研究藉由文獻探討的方式，決定模型的輸入變數，接著利用ROE來做為評斷企業總體財務表現的指標，同時使用來對上市/櫃公司進行評分，根據評分的結果，再使用K-Means方法來針對所有ROE值為正的上市/櫃公司進行評等等級的切割，以計算所有上市/櫃公司各年度的評等。 研究結果發現： (1) 利用建模資料帶入模型，分別計算每一筆資料的違約機率，也就是該公司當年度的違約機率，再將每一個等級的所有資料的PD值求平均數，即可得到代表該等級的違約機率，而此估計出的違約機率也的確能隨著評等等級的遞增而增加。 因此，當我們要判斷一間公司的違約等級時，可利用本研究所建構出的信用評等模型，估計出該公司違約機率，以判斷該公司的違約等級，以為決策者提供重要的參考依據。 (2) 信用風險預警模型在預測公司下一年度違約與否的能力上，也有不錯的預測準確率；同時，本研究利用預測結果的型I誤差、型II誤差、模型區別率和模型預測率分析來分析預警模型的效度，經實證結果得知，預警模型在效度驗證方面也能有效滿足要求。 由以上的結果得知，本研究所自行發展的信用風險評等模型與信用預警模型相關建構流程、架構與方法論，可有效應用於銀行授信決策系統之中。

信用評等制度在美國已有百年以上歷史，而我國自民國80幾年開始發展評等制度，截至目前，僅有中華信用評等公司與台灣經濟新報社兩家公司提供評等服務，而台灣經濟新報社更將金融保險業排除於評等對象之外。站在穩定市場競爭、保障消費者權益、配合監理需求，以及輔助專案投標等方面來看，市場上的確需要一套能反映產險業行業特性之評等模式。 本文以美國接受A.M.Best評等之產險公司為研究對象，運用三種統計方法：多元區別分析(Multiple Discriminant Analysis，MDA)、羅吉斯迴歸(Unordered Logistic Regression，ULR)、順序性羅吉斯迴歸(Ordered Logistic Regression，OLR)，來建構產險公司之信用評等模式。樣本選擇方面：估計樣本，選取美國1993年到1996年接受A.M.Best評等之產險公司327家；保留樣本，為1997年78筆資料。 而本文預定達成目標如下： 一、建立等級預測模型：參考Ederington(1985)所作債券等級預測模型，以獲利能力、槓桿、流動性、投資風險、準備金適足性五類指標共38個財務比率，透過三種統計模型，建構等級預測模型。 二、藉由等級預測之建立，尋找能有效區別產險公司評等等級之財務指標，並分析其影響程度。 三、力求模型公信力：無論變數選擇或權數決定，皆由統計軟體按照樣本特性選取產生，減少人為主觀判斷。 在決定研究對象之初，因考慮到國內產險公司接受評等之家數不多，且年數又太短，資料數量無法據以建立評等模式，因而決定以美國的產險公司為對象，再以台灣樣本作為保留樣本，預測之等級結果僅供參考之用。 / Three possible models of the P-L Insurers rating process are estimated and compared：1. Muitiple Discriminant Model, 2. Unordered Logistic Model, 3. Ordered Logistic Model. Each model is estimated for a sample of 327 American P-L insurance companies using the same 38 independent variables. The three estimated models are then employed to predict ratings for a holdout sample of 78 companies. The study analyzes 1993 through 1997 data for a sample of P-L insurers that acquired A.M.Best Financial strength ratings between December 31,1993, and December 31, 1997. Empirical evidence suggests that even when models with the same basic structure were compared, differences in estimation procedures resulted in quite different coefficient estimates and classifications. The muitiple discriminant model clearly outperformed the regression model, while the unordered logistic model was clearly superior to the ordered logistic model.

【第一篇論文中文摘要】 本文檢測公司負債合約中的利差是否可被最終的違約後償還率所解釋。透過1962年至2007年間在美國金融市場上發行但最後卻違約的負債合約資料來進行實證，發現違約後償還率的確有反映在發行時的利差上，且此關聯性會隨著美國開放商業銀行進行證券承銷業務後隨之更加顯著。我們並且進一步發現此償還率的資訊能更加有效反映原因與發行公司的資訊不對稱程度降低有關。此外，我們同時又發現此負債合約中的利差與違約後償還率的關聯性對於公司治理較差、以及非投資等級的發行公司會更為顯著。最後，我們的實證結果在考量內生問題、潛在可能遺漏解釋變數、以及其他模型設定後，仍同樣具有堅實性。 【第二篇論文中文摘要】 本文使用臺指選擇權的日內資料來探討選擇權提前交易期間是否具有資訊內涵與價格發現的功能。就作者所知，我們是第一篇透過選擇權資料探討提前交易期間資訊內涵的研究。首先，我們分別透過價、量、與高階動差三類資訊變數指標來衡量提前交易期間的資訊內涵。實證結果顯示：選擇權提前交易期間不只能有效反映隔夜資訊 (公開資訊)，且具有預測當日現貨指數開盤後5分鐘內股價指數移動的能力 (反應私有資訊)，說明提前交易期間的確具有資訊內涵與價格發現的功能。此外，我們進一步發現價平選擇權包含最強的資訊內涵，此應與投資人尋求交易流動性最高的價平選擇權來迅速實現其利潤以反映其資訊有關。最後，本研究亦發現前一日海外市場 (美國) 投資人情緒傳染效果的強度會影響提前交易期間選擇權的資訊內涵，而前一日是否交易 (週末效果與假日效果)則不會影響此資訊內涵。 / 【第一篇論文英文摘要】 We investigate whether the spread of corporate debt contacts can be explained by their ultimate recovery rates. Using the actual realized recovery rates of defaulted debt instruments issued in the U.S. from 1962 to 2007, we find that recovery rate is reflected in the spread at issuance, and that this relationship has become more significant since commercial banks were allowed to underwrite corporate securities. Our further investigation indicates that the enhanced informativeness of recovery rate can be attributed to the lowering of information asymmetry of individual firms. Besides, the relation between the spread at issuance and the recovery rate is stronger for weak corporate governance and non-investment grade issuers. Our conclusions are found to be robust to endogeneity issues, potentially omitted variables and alternative model specifications. 【第二篇論文英文摘要】 This study uses tick-by-tick data to examine the information content and price discovery of TAIEX option trading during the pre-opening period. To the best of our knowledge, this is the first study that focuses on the options market. We construct three groups of information variables to measure the information content of the pre-opening period, including the price, volume, and high moment information variables. We find that option trading during the pre-opening period not only can reflect the overnight information (public information) but also predict the 5-minute intraday returns after the opening of spot market (private information), showing the information content and price discovery of option trading during the pre-opening period. We also find that at-the-money options contain the strongest richness of information content, which may result from its highest liquidity. Finally, we also find that the empirical results would be stronger depending on the intensity of investor sentiment from overseas (U.S. market) of last day but not the length of hours without trading (weekend and holiday effect).