A look inside credit risk: theoretical and pratical issues with insights on distributional effects

The Thesis is, in the three articles that compose it, the issue now perhaps more than in the current debate is that scientific practice, on the situation of financial markets today: The Risk of Credit. Much has been said and written about the causes of the financial crisis that is disrupting global markets. Although not unanimous identification of particular sectors of the economy, which triggers outbreaks of great crisis, there seems to be a considerable accession to the theme of the need to improve and enhance the current methodology for calculating risk. Many banks and financial institutions have also devoted significant resources to adjust their standard of measuring risk to the protocol of Basel II. But that apparently did not yield immediate benefits and tangible, as the crisis seems to have taken advantage of obvious weaknesses in the management system of institutions, to infiltrate deeply and dangerously in the European and world economies. It seems urgent a new global system of detection and measurement of financial risks associated with uncovered positions of banks and does not seem able to say more just the mere application of a modeling standard and obsolete, dating back to the early theoretical elaborations fifties. All this has been the driving force that has driven my work for this thesis. The question that I have been from the beginning has been whether and how it can reach a conformation and the identification of the dynamics of risk, and decoupled from the static inefficiency of the classical models. The thesis consists of three scientific papers that deal with various aspects of credit risk. The first feature is developed around a model developed by me personally. After an extensive introduction, we find the presentation of statistical and econometric techniques to determine which fundamental volatility and probability of default. The techniques considered are the functions of copula and GARCH models. The heart of the discussion is centered on a model for determining the performance of risk. The dataset covers a group of American companies with the same characteristics in terms of debt exposure and quality of the debt. The analysis shows some consistency with the assumptions made at the beginning, the correlation effect takes precedence over that of diversification, with the result that a less diversified portfolio, resulting from a restriction on the number of assets, agreed with the performance of first analysis, leads to an improvement in the performance of portfolio risk. The second article deals with a model known in the literature, the multifactorial model of Vasicek for credit spreads. The innovation that is brought in to the model consists of a Kalman filter, having necessarily the same model expressed in state space form. The next step in the analysis of data from a group of companies in the U.S. market, characterized by a common economic and financial situation, centered on an event of distress centered on the dataset. The initial hypothesis is an obvious change in the level of spread as a result of the distress. The estimation of the parameters and the consequent application of the data confirms the hypothesis, indicating a sharp rise in the level of spreads on debt resulting from a disastrous event. Further analysis of default probabilities confirm the results as a direct consequence of the model. The third article is structured as a survey of world literature in the theory of extreme values for the analysis of the tail of the distribution of returns of bonds. The analysis includes the techniques used in the analysis of extreme values and suggests a criticism on classical models of credit risk, whose performance is clearly superseded by more refined techniques. The article concludes with an experiment with data for the comparison of different models in estimating the tail of the distribution. It seems clear that the Value at Risk understates the extreme risk in comparison with more advanced techniques such as the theory of extreme values.