Thesis (MSc)--Stellenbosch University, 2015 / ENGLISH ABSTRACT : In recent years, the use of Lévy processes as a modelling tool has come to be viewed more favourably than the use of the classical Brownian motion setup. The reason for this is that these processes provide more flexibility and also capture more of the ’real world’ dynamics of the model. Hence the use of Lévy processes for financial modelling is a motivating factor behind this research presentation. As a starting point a framework for the LIBOR market model with dynamics driven by a Lévy process instead of the classical Brownian motion setup is presented. When modelling LIBOR rates the use of a more realistic driving process is important since these rates are the most realistic interest rates used in the market of financial trading on a daily basis. Since the financial crisis there has been an increasing demand and need for efficient modelling and management of risk within the market. This has further led to the motivation of the use of Lévy based models for the modelling of credit risky financial instruments. The motivation stems from the basic properties of stationary and independent increments of Lévy processes. With these properties, the model is able to better account for any unexpected behaviour within the market, usually referred to as "jumps". Taking both of these factors into account, there is much motivation for the construction of a model driven by Lévy processes which is able to model credit risk and credit risky instruments. The model for LIBOR rates driven by these processes was first introduced by Eberlein and Özkan (2005) and is known as the Lévy-LIBOR model. In order to account for the credit risk in the market, the Lévy-LIBOR model with default risk was constructed. This was initially done by Kluge (2005) and then formally introduced in the paper by Eberlein et al. (2006). This thesis aims to present the theoretical construction of the model as done in the above mentioned references. The construction includes the consideration of recovery rates associated to the default event as well as a pricing formula for some popular credit derivatives. / AFRIKAANSE OPSOMMING : In onlangse jare, is die gebruik van Lévy-prosesse as ’n modellerings instrument baie meer gunstig gevind as die gebruik van die klassieke Brownse bewegingsproses opstel. Die rede hiervoor is dat hierdie prosesse meer buigsaamheid verskaf en die dinamiek van die model wat die praktyk beskryf, beter hierin vervat word. Dus is die gebruik van Lévy-prosesse vir finansiële modellering ’n motiverende faktor vir hierdie navorsingsaanbieding. As beginput word ’n raamwerk vir die LIBOR mark model met dinamika, gedryf deur ’n Lévy-proses in plaas van die klassieke Brownse bewegings opstel, aangebied. Wanneer LIBOR-koerse gemodelleer word is die gebruik van ’n meer realistiese proses belangriker aangesien hierdie koerse die mees realistiese koerse is wat in die finansiële mark op ’n daaglikse basis gebruik word. Sedert die finansiële krisis was daar ’n toenemende aanvraag en behoefte aan doeltreffende modellering en die bestaan van risiko binne die mark. Dit het verder gelei tot die motivering van Lévy-gebaseerde modelle vir die modellering van finansiële instrumente wat in die besonder aan kridietrisiko onderhewig is. Die motivering spruit uit die basiese eienskappe van stasionêre en onafhanklike inkremente van Lévy-prosesse. Met hierdie eienskappe is die model in staat om enige onverwagte gedrag (bekend as spronge) vas te vang. Deur hierdie faktore in ag te neem, is daar genoeg motivering vir die bou van ’n model gedryf deur Lévy-prosesse wat in staat is om kredietrisiko en instrumente onderhewig hieraan te modelleer. Die model vir LIBOR-koerse gedryf deur hierdie prosesse was oorspronklik bekendgestel deur Eberlein and Özkan (2005) en staan beken as die Lévy-LIBOR model. Om die kredietrisiko in die mark te akkommodeer word die Lévy-LIBOR model met "default risk" gekonstrueer. Dit was aanvanklik deur Kluge (2005) gedoen en formeel in die artikel bekendgestel deur Eberlein et al. (2006). Die doel van hierdie tesis is om die teoretiese konstruksie van die model aan te bied soos gedoen in die bogenoemde verwysings. Die konstruksie sluit ondermeer in die terugkrygingskoers wat met die wanbetaling geassosieer word, sowel as ’n prysingsformule vir ’n paar bekende krediet afgeleide instrumente.

Thesis (MBA)--Stellenbosch University, 2004 / ENGLISH ABSTRACT: This paper investigates the vulnerability of four key emerging markets to crises originating in Asia in 1997; Russia in 1998; Brazil in 1999 and Argentina in 2001. The emerging markets examined, Argentina, Malaysia, Poland and South Africa have been chosen to represent different geographic continents. Stock market data is used to measure for changes in unconditional correlation coefficients during and after the crisis periods. This is to establish whether the volatility shocks generated by the crises are what would reasonably be expected. Results suggest that there is evidence of contagion during the Asian crisis but there is little support of significant cross-market correlations transmitted during the Russian, Brazilian or Argentinean crises. Granger Causality tests are calculated to identify the existence of a relationship between stock market returns of countries in crisis and each of the four emerging markets. There is no evidence of causality emanating from the Thai stock market during the Asian crisis or from the Argentinean index during the Argentinean.crisis. Findings show that there is Granger causality from the Russian index during the Russian crisis to the Argentinean stock market but there was no impact on the markets in Malaysia, Poland or South Africa. Interestingly, there is no evidence that the Polish stock market returns were affected by the Russian crisis, the Argentinean returns by the Brazilian crisis or the Malaysian market by the Asian crisis. The paper further examines whether there is a relationship between stock market returns and country credit ratings and if credit risk can explain stock market returns. Significantly for active investment management, past values of country credit ratings can help predict stock market returns in Argentina, Malaysia and South Africa. Therefore, country credit risk contains information about expected stock market returns and potential investors would benefit by devising an asset allocation strategy that incorporates the explanatory powers of credit risk. / AFRIKAANSE OPSOMMING: Hierdie verslag ondersoek die kwesbaarheid van vier sleutelontwikkelende markte ten opsigte van krisisse wat onstaan het in Asië in 1997; Rusland in 1998; Brasilië in 1999 en in Argentinië in 2001. Die Argentynse, Maleisiese, Poolse en Suid Afrikaanse markte is gekies om verskillende geografiese kontinente te verteenwoordig. Effektebeurs data is gebruik om die verandering in onkondisionele korrelasie koeffisiente gedurende en na die krisis tydperk te meet. Dit is gedoen om vas te stel of die wisselvalligheid-skokke wat veroorsaak is deur die krisis ooreenstem met wat wesenlik verwag sal word. Resultate dui daarop dat daar getuienis is van besmetting ("contagion") gedurende die Asiatiese krisis, maar dat daar min ondersteuning gebied word vir die oordraging van beduidende kruis-mark korrelasie gedurende die Russiese, Brasiliaanse of Argentynse krisisse. Granger "causality" toetse is uitgevoer om die bestaan van 'n verwantskap tussen die effektemark opbrengste van die lande in krisis en elkeen van die vier opkomende markte te identifiseer. Daar is geen bewyse van enige veroorsakende verband voortgebring vanuit die Thai effektebeurs gedurende die Asiatiese krisis, of van die Argentynse indeks gedurende die Argentynse krisis nie. Die bevindinge toon dat daar Granger veroorsaking is vanaf die Russiese indeks na die Argentynse effektebeurs gedurende die Russiese krisis, maar dat daar geen impak was op die markte in Maleisië, Pole of Suid Afrika nie. Dit is interessant dat daar geen bewyse is dat die Poolse effektebeurs opbrengste beïnvloed is deur die Russiese krisis, die Argentynse opbrengste deur die Braziliaanse krisis, of die Maleisiese mark deur die Asiatiese krisis nie. Die verslag ondersoek verder of daar 'n verwantskap bestaan tussen effektebeurs opbrengste en die land se kredietgraderings asook of krediet-risiko effektebeurs opbrengste kan verduidelik. Betekenisvol vir aktiewe beleggingsbestuur is dat die historiese kredietgraderings kan help met die vooruitskatting van effektebeurs opbrengste in Argentinië, Maleisië en Suid Afrika. Dus bevat land kredietgraderings informasie rakende verwagte effektebeurs opbrengste. Potensiële beleggers sal dus baat vind in die ontwikkeling van 'n bate-allokasie strategie wat die verduidelikende kragte van krediet risiko inkorporeer.

銀行或金融機構面對的風險眾多，包括信用風險、市場風險、作業風險、法律風險、系統風險與國家主權風險等。2001年公佈的巴塞爾新資本協定(Basel II)，對於信用風險有全新的規範，使得金融機構為了提升競爭力，而致力改革。國內銀行也需加強信用風險管理，減少壞帳，增加獲利，以面對全球化的考驗。 判斷公司違約情形是風險管理者的主要責任，管理者利用各種模式，加以評估公司狀況，而面對模式的不同，會產生不同的結果，使管理者缺乏單一指標，可供遵循。本研究以此為出發點，以統計模式構建組合預測，期能簡化管理者的決策過程。 首先介紹多種信用風險模式，及國內外相關實證。然後針對兩種信用風險模式，Z-score與Merton選擇權模式，研究在台灣的適用情形。Z-score代表公司財務資訊，為一個落後指標；Merton選擇權模式代表市場資訊，是一個領先指標；本研究將兩項指標應用Logit迴歸予以組合，期能得到一個較佳的指標。組合結果顯示，組合預測能夠整合Z-score與Merton選擇權兩種模式，得到一個較好的預測正確率，判斷公司是否可能發生危機。在危機發生前一年時，判斷正確率高達90%，但是離危機發生時點越遠，組合預測的結果會近似於鑑別分析結果，其主要原因為Merton選擇權模式的不適用。 本研究提出了若各種預測指標對同一家公司的信用風險評估不同，則可以經由組合預測，得到一個綜合結果，以減少指標間互相衝突所造成的差異，與風險作業人員誤判的可能，並可提供風險管理者作為執行業務的參考。

This thesis consists of four independent chapters on bank capital regulation and the issue of unsolicited ratings.<p> The first chapter is introductory and reviews the motivation for regulating banks and credit rating agencies while providing a detailed overview of the thesis.<p> The second chapter uses a simultaneous equations model to analyze how banks from six G10 countries adjusted their capital to assets ratios and risk-weighted assets to assets ratio between 1988 and 1995, i.e., just after passage of the 1988 Basel Accord. The results suggest that regulatory pressure brought about by the 1988 capital standards had little effect on both ratios for weakly capitalized banks, except in the US. In addition, the relation between the capital to assets ratios and the risk-weighted assets to assets ratio appears to depend not only on the level of capitalization of banks, but also on the countries or groups of countries considered.<p> The third chapter provides Monte Carlo estimates of the amount of regulatory capital that EMU banks must hold for their corporate, bank, and sovereign exposures both under Basel I and the standardized approach to credit risk in Basel II. In the latter case, Monte Carlo estimates are presented for different combinations of external credit assessment institutions (ECAIs) that banks may choose to risk weight their exposures. Three main results emerge from the analysis. First, although the use of different ECAIs leads to significant differences in minimum capital requirements, these differences never exceed, on average, 10% of EMU banks’ capital requirements for corporate, bank, and sovereign exposures. Second, the standardized approach to credit risk provides a small regulatory capital incentive for banks to use several ECAIs to risk weight their exposures. Third, the minimum capital requirements for the corporate, bank, and sovereign exposures of EMU banks will be higher in Basel II than in Basel I. I also show that the incentive for banks to engage in regulatory arbitrage in the standardized approach to credit risk is limited.<p> The fourth and final chapter analyses the effect of soliciting a rating on the rating outcome of banks. Using a sample of Asian banks rated by Fitch Ratings, I find evidence that unsolicited ratings tend to be lower than solicited ones, after accounting for differences in observed bank characteristics. This downward bias does not seem to be explained by the fact that better-quality banks self-select into the solicited group. Rather, unsolicited ratings appear to be lower because they are based on public information. As a result, they tend to be more conservative than solicited ratings, which incorporate both public and non-public information.

Cette thèse a pour objet l'évaluation du risque de crédit par une approche comportementale dans un contexte d'information asymétrique et de rationalité limitée. Nous cherchons à travers cette analyse, à concilier les « experts métiers » et les statisticiens, en intégrant le comportement humain dans la conception des outils quantitatifs d'évaluation du risque de crédit, en vue d'optimisation de la gestion de ce risque.L'évaluation du risque de crédit est basée sur des modèles et techniques statistiques de plus en plus avancées. Nous citons à titre d'exemple les modèles du Crédit Metrics et JP Morgan, le modèle KMV et le modèle Crédit Portfolio de Mekinsey ou encore les modèles de scoring introduits pour évaluer la qualité du risque des emprunteurs. Plusieurs travaux soulignent l'intérêt de ces modèles quantitatifs. En effet, Scot FRAME et al, ont montré que le recours aux modèles de scoring contribue effectivement à réduire le coût d'information dans les grandes banques américaines. Les résultats de ces modèles dépendent de la réalisation des facteurs de risque spécifiques à chaque emprunteur et de facteurs de risque systémique. Cependant, la crise financière actuelle a mis en lumière la défaillance de ces modèles, aussi bien les modèles théoriques de notation que les modèles opérationnels utilisés par les praticiens, dans l'évaluation du risque de crédit. Toutes ces constructions n'ont pas su intégrer parfaitement l'ensemble de l'information et traiter la complexité d'interactions entre les variables déterminant le risque car elles sont basées sur des techniques purement statistiques qui ne savent représenter que des relations linéaires entre le risque de défaut et les variables qui en sont à l'origine sans prendre en compte le comportement du gestionnaire du risque crédit, dans l'optimisation de sa gestion. Etant données les limites de l'approche quantitative, nous avons convergé vers une approche comportementale qui concilie les techniques statistiques et le comportement humain basée sur la prise en compte et la validation collective des règles de décision émergeant des discussions et confrontations. Cette approche comportementale qui prend en compte la rationalité des décideurs à travers un modèle expert nous permettra d'une part de construire un cadre d'analyse normatif permettant d'identifier et d'évaluer le risque de crédit et d'autre part intégrer ces règles dans les systèmes de décisions opérationnels.Notre recherche a un intérêt multiple. Elle apporte un éclairage théorique sur l'optimisation de la décision des banques, dans un contexte d'incertitude, à travers un modèle portant à la fois sur le caractère quantitatif des modèles d'évaluation du risque de crédit et le comportement humain. L'évaluation du risque de crédit à travers notre approche permettra également de déterminer le montant de capital économique nécessaire à la couverture du risque de crédit. Ainsi, elle permettra aux banques de mettre en place une allocation optimale des fonds propres et une tarification adéquate des crédits basée sur une évaluation précise du risque de crédit. Ce qui porte un grand intérêt aux banques et aux clients aussi. / This thesis focuses on the assessment of credit risk with a behavioral approach in a context of asymmetric information and bounded rationality. We seek through this analysis, to reconcile the "business experts" and statisticians, incorporating human behavior into the design of tools for quantitative assessment of credit risk in order to optimize the management of this risk.The assessment of credit risk is based on models and statistical techniques more advanced. We cite as examples of models JP Morgan Credit Metrics, model and KMV's Portfolio Credit Model Mekinsey or scoring models introduced to assess the quality of the risk of borrowers. Several studies underline the importance of these quantitative models. Indeed, Scot FRAME and AL showed that the use of scoring models is effective in reducing the cost of information in large U.S. banks. The results of these models depend on the realization of the risk factors specific to each borrower and systemic risk factors. However, the current financial crisis has highlighted the failure of these models, both theoretical models that rating business models used by practitioners in assessing credit risk. All these constructions have not been able to integrate fully all the information and treat the complexity of interactions between variables determining the risk because they are based on purely statisti cal techniques who can represent the linear relationships between risk default and the variables that are at the origin without taking into account the behavior of credit risk manager, in optimizing its management.Given the limitations of the quantitative approach, we have converged on a behavioural approach that combines statistical techniques and human behaviour based on consideration and validation of collective decision rules emerging from the discussions and confrontations. This behavioural approach that takes into account the rationality of decision makers through an expert model we will firstly build a normative framework for analysis to identify and assess credit risk and also integrate these rules systems in operational decisions.Our research has a substantial multiple. It sheds light on the theoretical optimization of the decision of the banks in a context of uncertainty through a model bearing both on the quantitative assessment models of credit risk and human behavior. The assessment of credit risk through our approach will also determine the amount of capital necessary to cover credit risk. Thus, it will allow banks to establish an optimal allocation of capital and adequate pricing of loans based on an accurate assessment of credit risk. This brings great interest to banks and customers too.

Our sample (Xit; Yit) consists of pairs of variables. The real variable Xit measures the creditworthiness of individual i in period t. The Bernoulli variable Yit is the default indicator of individual i in period t. The objective is to estimate a credit rating system, i.e. to particularly divide the range of the creditworthiness into several rating classes, each with a homogeneous default risk. The field of change point analysis provides a way to estimate the breakpoints between the rating classes. As yet, the literature only considers models without dependencies or with dependence only in cross section. This contribution proposes multi-period models including dependencies in cross section as well as in longitudinal section. Furthermore, estimators for the model parameters are suggested. The estimators are applied to a data set of a German credit bureau.

L’étude des déterminants internes du risque de crédit des banques tunisiennes est l’objetprincipal de cette thèse. Il s’agit en particulier de la gouvernance bancaire et de la relationbancaire. À partir d’un échantillon de 10 banques tunisiennes cotées durant la période 1998-2009, nous essayons de détecter l’impact de la concentration de la propriété et descaractéristiques du conseil d’administration sur le risque de crédit. Nos résultats montrent queles mécanismes internes de gouvernance n’ont pas assuré jusqu’à présent leur rôle de contrôleet de garant de la bonne gestion des banques tunisiennes et qu’ils ont contribué à une gestionimprudente du risque de crédit. À partir de l’étude de dossiers de crédits accordés par unebanque tunisienne à une clientèle d’entreprises, nous essayons de détecter l’impact del’information hard et soft sur le risque de crédit. Nos résultats plaident en faveur del’hypothèse du laxisme des banques envers certains clients importants et surtout de longuedate. Ce laxisme pourrait être la cause d’une augmentation du risque de non remboursementdes crédits accordés en Tunisie. / This research aims to analyse the internal determinants of Tunisian bank risk. We particularlyfocus on the role of bank governance and relationship banking. Based on a sample of 10Tunisian listed banks during 1998-2009, we examine the impact of ownership concentrationand board characteristics on credit risk. Our results demonstrate that the importance ofTunisian banks’credit risk is rooted in bank governance deficiency. This latter wascontributed to implement a reckless credit policy. From the analysis of credit files granted bya Tunisian bank to corporate customers, we attempt to detect the impact of soft and hardinformation on credit risk. Our results suggest a lax attitude of the bank towards someimportant clients, particularly long-standing ones. This attitude may be the root of the creditrisk increase in Tunisia.

This diploma thesis "Methods of risk aggregation on financial markets" introduces all kinds of risk that are present on the financial markets. In the first part there are explained the ways and methods of measurement of these risks. Next there are shown the methods of aggregation of credit, market and operational risks. One of these methods are copula functions which are constructed in practical part of this thesis.

There has been a long discussion about macroeconomic variables influencing the level of aggregate credit risk in the economy. While literature provides both empirical evidence and theoretical explana- tion of the influence of the business cycle on credit risk, the effect of other macroeconomic variables has not been explored sufficiently. In addition, recent literature suggests the existence of a latent risk factor behind aggregate credit risk, which is regularly interpreted as the latent default cycle. This thesis provides in its first part a discussion of potential aggregate credit risk drivers, which have been previously suggested in literature. We verify using a linear regression model whether the effect of these macroeconomic variables is also apparent in the Czech Republic. Results seem to be stable for both different model specifications and different clients segments and are in line with previous studies. The second part of this thesis explicitly models the latent factor that is assumed behind aggregate credit risk by adding an unobserved component to the already existing model constructed earlier in this thesis. The unobserved component can be estimated by applying Kalman filter. We subsequently discuss the sources of the latent component and whether it can be interpreted as the default cycle. The...

This thesis focuses on application of artificial intelligence techniques in credit risk management. Moreover, these modern tools are compared with the current industry standard - Logistic Regression. We introduce the theory underlying Neural Networks, Support Vector Machines, Random Forests and Logistic Regression. In addition, we present methodology for statistical and business evaluation and comparison of the aforementioned models. We find that models based on Neural Networks approach (specifically Multi-Layer Perceptron and Radial Basis Function Network) are outperforming the Logistic Regression in the standard statistical metrics and in the business metrics as well. The performance of the Random Forest and Support Vector Machines is not satisfactory and these models do not prove to be superior to Logistic Regression in our application.