This master’s thesis explains what the term risk means, how the project of risk management is running and financial risks existing in the company ABC, s.r.o. are described. These risks are currency risk, credit risk and liquidity risk. The methods of their analysis and measurement and also instruments are stated. These procedures and project of the risk management are explained.

Cette thèse porte sur l’étude des mécanismes internes de gouvernance bancaire et leurs effets sur le risque de crédit et la performance bancaire. A partir d’un échantillon composé de 13 banques françaises, 13 allemandes et 20 japonaises cotées durant la période 2005 – 2012, nous essayons de détecter l’impact des caractéristiques de gouvernance bancaire, à travers les conseils d’administration et leurs différents comités, en premier lieu sur le risque du crédit, et en deuxième lieu sur la performance bancaire. Nos résultats montrent que les mécanismes internes de gouvernance affectent certes, le niveau des crédits non performants et la performance financière des banques, mais avec des effets plutôt mitigés. Ils révèlent également des divergences entre les banques des pays étudiés. / The study of the internal mechanisms of governance in particular the board of directors and its relationship with the credit risk as well as the performance is the main subject of this thesis. From a sample of 13 French banks, 13 German and 20 Japanese rated during the period 2005 - 2012, we are trying to detect the impact of the characteristics of the Board of Directors and its committees on the credit risk, and on the banking performance. Our results show that the internal mechanisms of governance affect certainly, the level of appropriations non-performing assets and the financial performance of banks, but with mixed effects; they reflect this divergence between countries.

We assume pairs of random variables (X_i, Y_i), whereby the real variable X_i measures the creditworthiness of individual i with i = 1, . . . , n. The Bernoulli variable Y_i represents the default indicator of individual i. Our main target is a division of the creditworthiness into a given number of groups with a homogeneous default risk, i.e. to estimate rating classes. The framework of change point analysis provides a nonparametric method to estimate the breakpoints between the rating classes under quite weak assumptions. Up to now, the theory of breakpoint estimation is developed under the assumption of exactly one breakpoint. The contribution at hand, basically implements this theory, but extends it into a multi-stage heuristic. That means, we sequentially apply the theory for only one breakpoint as a multi-stage procedure. With this article we transfer the interesting theoretical issue of breakpoint estimation into an applicable form. Thereby, all the results are checked and obtained by simulation. The main results are as follows. Applying a sequential breakpoint estimation basically works and leads to outcomes of practical purpose. Thereby, the multistage heuristic reveals some weakness esp. in the case of quite huge differences between default probabilities that can be resolved by some interventions.

El presente trabajo de investigación tiene como objetivo desarrollar un modelo de rating que permita mejorar la gestión del riesgo de crédito de las operaciones crediticias de los deudores de medianas empresas (según el tipo de crédito definido por la Superintendencia de Banca, Seguros y AFP), y complementar la evaluación cuantitativa y cualitativa que realiza el funcionario de créditos. Además, el modelo de rating permite determinar la probabilidad de incumplimiento de pago durante el año posterior al momento de evaluar a los deudores de medianas empresa en base a la información de sus características, comportamiento de pago y estados financieros, otorgándole una visión prospectiva, y finalmente segmentarlos según su nivel de riesgo. Al respecto, se elaboró un modelo de rating para deudores de medianas empresas, que se basa en un modelo de regresión logística con 8 variables, que permite discriminar a los deudores que cumplen sus pagos respecto a los que incumplen sus pagos. Dicho modelo cuenta con niveles adecuados de predicción y discriminación tanto en la muestra de construcción como en la de validación. Asimismo, en función de los resultados del modelo, se clasificó a los deudores de medianas empresas en grupos según su nivel de riesgo, y en base a esta segmentación se pueden tomar mejores decisiones en la gestión de riesgo de crédito. / The objective of this research paper is to develop a rating model that allows to improve the management of the credit risk of the credit operations of the debtors of medium-sized companies (according to the type of credit defined by the Superintendency of Banking, Insurance and AFP), and complement the quantitative and qualitative evaluation carried out by the loan officer. In addition, the rating model makes it possible to determine the probability of payment default during the year after the time of evaluating the debtors of medium-sized companies based on information on their characteristics, payment behavior and financial statements, giving it a prospective vision, and finally segment them according to their risk level. In this regard, a rating model was developed for debtors of medium-sized companies, which is based on a logistic regression model with 8 variables, which allows us to discriminate between debtors who make their payments with respect to those who default on their payments. This model has adequate levels of prediction and discrimination in the development sample and in the validation sample. Likewise, based on the results of the model, debtors of medium-sized companies were classified into groups according to their level of risk, and with this segmentation, better decisions can be made in credit risk management. / Trabajo de investigación

Le but de cette thèse était d'explorer la thématique du scoring dans le cadre de son utilisation dans le monde bancaire, et plus particulièrement pour contrôler le risque de crédit. En effet, la diversification et la globalisation des activités bancaires dans la deuxième moitié du XXe siècle ont conduit à l'instauration d'un certain nombre de régulations, afin de pouvoir s'assurer que les établissements bancaires disposent de capitaux nécessaires à couvrir le risque qu'ils prennent. Cette régulation impose ainsi la modélisation de certains indicateurs de risque, dont la probabilité de défaut, qui est pour un prêt en particulier la probabilité que le client se retrouve dans l'impossibilité de rembourser la somme qu'il doit. La modélisation de cet indicateur passe par la définition d'une variable d'intérêt appelée critère de risque, dénotant les "bons payeurs" et les "mauvais payeurs". Retranscrit dans un cadre statistique plus formel, cela signifie que nous cherchons à modéliser une variable à valeurs dans {0,1} par un ensemble de variables explicatives. Cette problématique est en pratique traitée comme une question de scoring. Le scoring consiste en la définition de fonction, appelées fonctions de score, qui retransmettent l'information contenue dans l'ensemble des variables explicatives dans une note de score réelle. L'objectif d'une telle fonction sera de donner sur les individus le même ordonnancement que la probabilité a posteriori du modèle, de manière à ce que les individus ayant une forte probabilité d'être "bons" aient une note élevée, et inversement que les individus ayant une forte probabilité d'être "mauvais" (et donc un risque fort pour la banque) aient une note faible. Des critères de performance tels que la courbe ROC et l'AUC ont été définis, permettant de quantifier à quel point l'ordonnancement produit par la fonction de score est pertinent. La méthode de référence pour obtenir des fonctions de score est la régression logistique, que nous présentons ici. Une problématique majeure dans le scoring pour le risque de crédit est celle de la sélection de variables. En effet, les banques disposent de larges bases de données recensant toutes les informations dont elles disposent sur leurs clients, aussi bien sociodémographiques que comportementales, et toutes ne permettent pas d'expliquer le critère de risque. Afin d'aborder ce sujet, nous avons choisi de considérer la technique du Lasso, reposant sur l'application d'une contrainte sur les coefficients, de manière à fixer les valeurs des coefficients les moins significatifs à zéro. Nous avons envisagé cette méthode dans le cadre des régressions linéaires et logistiques, ainsi qu'une extension appelée Group Lasso, permettant de considérer les variables explicatives par groupes. Nous avons ensuite considéré le cas où la variable réponse n'est plus binaire, mais polytomique, c'est-à-dire avec plusieurs niveaux de réponse possibles. La première étape a été de présenter une définition du scoring équivalente à celle présentée précédemment dans le cas binaire. Nous avons ensuite présenté différentes méthodes de régression adaptées à ce nouveau cas d'étude : une généralisation de la régression logistique binaire, des méthodes semi-paramétriques, ainsi qu'une application à la régression logistique polytomique du principe du Lasso. Enfin, le dernier chapitre est consacré à l'application de certaines des méthodes évoquées dans le manuscrit sur des jeux de données réelles, permettant de les confronter aux besoins réels de l'entreprise. / The objective of this thesis was to explore the subject of scoring in the banking world, and more precisely to study how to control credit risk. The diversification and globalization of the banking business in the second half of the twentieth century led to introduce regulations, which require banks to make reserves to cover the risk they take. These regulations also dictate that they should model different risk indicators, among which the probability of default. This indicator represents the probability for a client to find himself in the incapacity to pay back his debt. In order to predict this probability, one should define a risk criterion, that allows to distinguish the "bad clients" from the "good clients". In a more formal statistical approach, that means we want to model a binary variable by an ensemble of explanatory variables. This problem is usually treated as a scoring problem. It consists in the definition of functions, called scoring functions, which interpret the information contained in the explanatory variables and transform it into a real-value score note. The goal of such a function is to induce the same order on the observations than the a posteriori probability, so that the observations that have a high probability to be "good" have a high score, and those that have a high probability to be "bad" (and thus a high risk for the bank) have a low score. Performance criteria such as the ROC curve and the AUC allow us to quantify the quality of the order given by the scoring function. The reference method to obtain such scoring functions is the logistic regression, which we present here. A major subject in credit scoring is the variable selection. The banks have access to large databases, which gather information on the profile of their clients and their past behavior. However, those variables may not all be discriminating regarding the risk criterion. In order to select the variables, we proposed to use the Lasso method, based on the restriction of the coefficients of the model, so that the less significative coefficients will be fixed to zero. We applied the Lasso method on linear regression and logistic regression. We also considered an extension of the Lasso method called Group Lasso on logistic regression, which allows us to select groups of variables rather than individual variables. Then, we considered the case in which the response variable is not binary, but polytomous, that is to say with more than two response levels. The first step in this new context was to extend the scoring problem as we knew in the binary case to the polytomous case. We then presented some models adapted to this case: an extension of the binary logistic regression, semi-parametric methods, and an application of the Lasso method on the polytomous logistic regression. Finally, the last chapter deals with some application studies, in which the methods presented in this manuscript are applied to real data from the bank, to see how they meet the needs of the real world.

Bakgrund: Den 1 november 2010 avskaffades revisionsplikten för små företag i Sverige, vilket innebar att små företag får välja om de vill ha en revisor eller inte. Ett av motiven bakom lagändringen var att småföretagens tillväxt skulle öka. För att öka tillväxten krävs finansiering. Småföretagens främsta finansieringskällor är eget kapital och löpande intäkter. Detta räcker emellertid inte utan företagen behöver ofta komplettera med externfinansiering, i form av banklån. Trots vikten av externt kapital har små företag svårare än stora företag att finansiera tillväxten med banklån. En av anledningarna till detta är att den finansiella informationen som bankerna samlar in är bristfällig. För att informationens trovärdighet ska öka önskar därför bankerna att den finansiella informationen är granskad av en revisor. Syfte: Syftet med studien är att förklara hur bankerna upplever att kreditbedömningen av små företag har påverkats av frivillig revision. Således syftar studien till att öka förståelsen för hur kreditbedömningen går till samt vilken roll revisorn har i denna. Metod: Studien har baserats på den kvalitativa metoden där sex stycken intervjuer har genomförts. Fem av dessa har varit telefonintervjuer medan en har varit en personlig intervju på informantens arbetsplats. Alla informanter arbetar inom banksektorn och var tillgängliga inom Skaraborgsområdet, därmed har både ett strategiskt- och bekvämlighetsurval gjorts. Resultat och slutsats: Studien visar på att revisionen har en viss roll vid kreditbedömningen genom sin granskningsfunktion. Däremot kan inte studiens resultat påvisa att bankerna i Skaraborg har upplevt att den frivilliga revisionen har haft en avgörande påverkan på kreditbedömning. / Background: On 1 November 2010 the statutory audit for small firms in Sweden was abolished, which meant that small firms were allowed to choose whether or not to havean auditor. One of the motives behind the change in the law was that the growth of small firms would increase. In order to increase growth, financing is required. The main source of financing for small firms are equity and current income. However, this is not enough,and small firms often need to supplement with external financing, in the form of bankloans. Despite the importance of external capital, small firms find it more difficult than large firms to finance economic growth with bank loans. One of the reasons for this is that the financial information collected by banks is inadequate. Therefore, in order to increase the credibility of the information, the banks wish that the financial information is audited by an auditor. Purpose: The purpose of the study is to explain how the banks feel that the credit assessment of small firms has been affected by the voluntary auditing. Thus, the aim of the study is to increase the understanding of how the credit assessment is carried out and what role the auditor has in it. Methods: The study has been based on the qualitative method where six interviews have been conducted. Five of these have been telephone interviews while one has been a personal interview at the informant’s workplace. All the informants work in the banking sector and were available in the Skaraborg area, thus both a strategic and convenience selection has been. Results and conclusion: The study shows that the audit has a certain role in the credit assessment through its function. However, the results of the study cannot show that the banks in Skaraborg have experienced that the voluntary audit has had a decisive impact on credit assessment.

The deregulation of the financial markets that started over two decades ago in the developed countries has led to increased house prices and loan to value ratios. Home owners in western countries have over the last two decades steadily decreased their savings and at the same time increased the size of their mortgages and the amount of leverage used to purchase their homes. This development has increased the financial risk for homeowners which recently became clear in the United States when prices on homes started to fall rapidly in 2007. Due to this development Finansinspektionen in Sweden has enforced new regulation on mortgage lending making it more expensive for home owners to use high leverage ratios. Finansinspektionen is responsible for consumer protection in terms of financial products and the new regulation aims to protect mortgage borrowers. Finansinspektionen suggests that an insurance that protects the borrower from loss could be used as an alternative to the regulation restricting the amount of leverage. Finansinspektionen also mentions the Canadian mortgage market as an example where compulsory mortgage insurances are enforced today. In Canada the borrower must take out a mortgage insurance when the mortgage exceeds 80 percent of the house value. However, we find that the Canadian mortgage insurance system would not fulfil the aim of Finansinspektionen’s regulation. The Canadian mortgage insurances are constructed to protect the lender against default and there purpose was initially to increase lending. When examining the basic structure of mortgage and home value insurance products we find that such products and systems are complicated to construct to match the Finansinspektionen requirements and purpose due to issues such as moral hazard, adverse selection, price, willingness to pay and systemic risk.

The use of machine learning methods in credit risk modelling has been proven to yield good results in terms of increasing the accuracy of the risk score as- signed to customers. In this thesis, the aim is to examine the performance of the machine learning boosting algorithms XGBoost and CatBoost, with logis- tic regression as a benchmark model, in terms of assessing credit risk. These methods were applied to two different data sets where grid search was used for hyperparameter optimization of XGBoost and CatBoost. The evaluation metrics used to examine the classification accuracy of the methods were model accuracy, ROC curves, AUC and cross validation. According to our results, the machine learning boosting methods outperformed logistic regression on the test data for both data sets and CatBoost yield the highest results in terms of both accuracy and AUC.

No description available.

No description available.