The master thesis deals with the advanced methods for estimating credit risk parameters from market prices: probability of default (PD) and loss given default (LGD). Precise evaluation of these parameters is important not only for banks to calculate their regulatory capital but also for investors to price risky bonds and credit derivatives. We provide forward looking reduced-form analytical method for calculation of PD and LGD of corporate defaultable bonds based on their quoted market prices, prices of equivalent risk-free bonds and quoted credit default swap spreads of the issuer of these bonds. This is reversed to most of the studies on credit risk modeling, as aim is not to price instruments based on estimated credit risk parameters, but to calculate these parameters based on the available market prices. Furthermore, compared to other studies, the LGD parameter is assumed to be endogenous and we provide the method for its simultaneous calculation with the probability of default. Finally, using developed methods, we estimate implied PD and LGD for five European banks assuming that the risk is priced correctly by other investors and the markets are efficient. JEL Classification: C02, C63, G13, G33 Keywords: credit risk, loss given default, probability of default, credit default swap Author's...

none

none

none

The aim of this study is to quantify and investigate the effect of the Covid-19 pandemic on non-financial South African firms listed on the Johannesburg Stock Exchange. The study implemented the Merton (1974) model on the 59 largest non-financial firms and calculated the probability of default for each firm before the pandemic and during the pandemic as at each firm's financial year-end. The default probabilities are calculated predominantly from the value and volatility of firm equity. The results emphasize that the Covid-19 pandemic, on average, had a dramatic impact on the probability of default of publicly traded South African firms. The observed increase in default probability was found to be statistically significant at the 5% significance level.

Submitted by Sara Ribeiro (sara.ribeiro@ucb.br) on 2018-08-08T13:33:03Z No. of bitstreams: 1 RafaelLeiteGregorioDissertacao2018.pdf: 1382550 bytes, checksum: 9c6e4f1d3c561482546aca581262b92b (MD5) / Approved for entry into archive by Sara Ribeiro (sara.ribeiro@ucb.br) on 2018-08-08T13:33:24Z (GMT) No. of bitstreams: 1 RafaelLeiteGregorioDissertacao2018.pdf: 1382550 bytes, checksum: 9c6e4f1d3c561482546aca581262b92b (MD5) / Made available in DSpace on 2018-08-08T13:33:24Z (GMT). No. of bitstreams: 1 RafaelLeiteGregorioDissertacao2018.pdf: 1382550 bytes, checksum: 9c6e4f1d3c561482546aca581262b92b (MD5) Previous issue date: 2018-07-09 / The credit risk assessment has a relevant role for financial institutions because it is associated with possible losses and has a large impact on the balance sheets. Although there are several researches on applications of machine learning and finance models, a study is still lacking that integrates available knowledge about credit risk assessment. This paper aims at specifying the machine learning model of the probability of default of publicly traded companies present in the Bovespa Index (corporations) and, based on the estimations of the model, to obtain risk assessment metrics based on risk letters. We converged methodologies verified in the literature and we estimated models that comprise fundamentalist (balance sheet) and governance data, macroeconomic and even variables resulting from the application of the proprietary model of KMV credit risk assessment. We test the XGboost and LinearSVM algorithms, which have very different characteristics among them, but are potentially useful to the problem. Parameter Grids were performed to identify the most representative variables and to specify the best performing model. The model selected was XGboost, and performance was very similar to the results obtained for the North American stock market in analogous research. The estimated credit ratings suggest that they are more sensitive to the economic and financial situation of the companies than that verified by traditional Rating Agencies. / A avaliação do risco de crédito tem papel relevante para as instituições financeiras por estar associada a possíveis perdas que podem gerar grande impacto nos balanços. Embora existam várias pesquisas sobre aplicações de modelos de aprendizado de máquina e finanças, ainda não há estudo que integre o conhecimento disponível sobre avaliação de risco de crédito. Este trabalho visa especificar modelo de aprendizado de máquina da probabilidade de descumprimento de empresas de capital aberto presentes no Índice Bovespa (corporações) e, fruto das estimações do modelo, obter métrica de avaliação de risco baseada em letras (ratings) de risco. Convergiu-se metodologias verificadas na literatura e estimou-se modelos que compreendem componentes fundamentalistas (de balanço) e de governança corporativa, macroeconômicos e ainda variáveis produto da aplicação do modelo proprietário de avaliação de risco de crédito KMV. Testou-se os algoritmos XGboost e LinearSVM, os quais possuem características bastante distintas entre si, mas são potencialmente úteis ao problema exposto. Foram realizados Grids de parâmetros para identificação das variáveis mais representativas e para a especificação do modelo com melhor desempenho. O modelo selecionado foi o XGboost, tendo sido observado desempenho bastante semelhante aos resultados obtidos para o mercado de ações norte-americano em pesquisa análoga. Os ratings de crédito estimados mostram-se mais sensíveis à situação econômico-financeira das empresas ante o verificado por agências de rating tradicionais.

Universidade Federal do Cearà / Desde o ano de 2000 que as cooperativas de crÃdito brasileiras tÃm experimentado um crescimento contÃnuo no nÃmero de novas unidades. De outro modo os bancos brasileiros tem diminuÃdo em quantidade pelo processo de aquisiÃÃo e concentraÃÃo. Este crescimento das cooperativas pode estar associado com um maior risco para os associados. Este trabalho investiga as causas de falÃncias das cooperativas de crÃdito dos profissionais de saÃde no Brasil. Para tanto busca fornecer um modelo de alerta precoce que informe aos gestores e supervisores do risco de insolvÃncia, fazendo uso de uma anÃlise de regressÃo logÃstica de Ãndices financeiros. Foi estimado um modelo de prediÃÃo de insolvÃncia que fosse parcimonioso e acurado. Este trabalho provà informaÃÃes adicionais a outros estudos brasileiros sobre falÃncia em cooperativas de crÃdito, de trÃs modos: à um estudo de abrangÃncia nacional, trata com cooperativa de crÃdito mÃtuo urbano, usa uma moderna tÃcnica estatÃstica com dados em painel, o que permite capturar as diferenÃas entre as cooperativas. O presente estudo tambÃm fornece uma maneira racional para a escolha do cut-off. Os resultados sugerem que provisÃo para emprÃstimo em atraso para total do ativo, Total de emprÃstimo para Total de ativo, Total de emprÃstimo para Total de depÃsitos e PatrimÃnio LÃquido Passivo total, sÃo os preditores mais significativos da insolvÃncias das cooperativas. De modo contrÃrio as Despesas Operacionais para Receitas Operacionais e Despesas Operacionais para ativo total nÃo indicam ser significativas em prever a insolvÃncia. / Since the year of 2000 Brazilians credit cooperatives has experienced a increasing growth in number of units. On the other hand Brazilians banks decreased their number, by the process of acquisition and concentration. This growth may imply increasing risk for the associates. This paper empirically investigates the causes of failures of credit cooperatives of heath professionals in Brazil. A goal of this paper is provide a early warning model that inform managers and supervisors of a risks of default, by using logistic regression analysis of financial ratios. It was estimate a default prediction model that was parsimonious and accurate. This work provided additional information over other Brazilian studies of credit cooperatives failure by three ways: it is a national wide study, deals with urban mutual credit cooperative, uses modern statistic technique panel data which can capture the differences across cooperatives. It also provided a reasonable for the choosing of cut-off. The results suggest that provision for bad debts over total assets, total loans over total assets, total loans over total deposits are the most significant predictors of credit cooperative failure. Operational expenses over operational incomes and operational expenses over total assets, contrary, do not seem to be significant indicators of failure

新版巴賽爾資本協定的內部評等法中，銀行可自行對借貸戶進行評分，並且根據 評分估算信用風險以提領準備金，因此估算借貸戶評分分數的違約機率(PD)是相當 重要的一環。過去估算違約機率的研究中，大多假定評分分數為離散型式，本文針對 評分分數為連續形式時，提出一種利用曲線函數來配適估計模型。估計模型是使用伽 瑪的截尾分配去配適ROC曲線函數，再利用此ROC曲線函數來估計各評分分數下的 違約機率P(D|S)，在伽瑪分配中的兩參數則是用兩階段的方法求解。本文所提的估 計方法並無假設評分分數的分配，因此在數值方法中使用不同的分配、參數設定、違 約機率等，來驗證此方法的準確度與穩定度，並且與Van der Burgt (2008)、Tasche(2009)的估計方法比較。 / By the internal rating-based approach of Basel II, banks estimate borrowers' default risks to withdraw reserves independently. Hence, estimating default probability (PD) of borrowers is important. Most of previous studies estimating PD assume that evaluation scores are discrete, In this study, we use curve function to t estimation model in the condition that the evaluation scores are continuous . We use truncated gamma distribution to t ROC curve function. And we use the ROC curve function to estimate PD of dierent scores. And use two-step method to nd the value of two parameters in gamma distribution. The estimation method in this study doesn't assume the distribution of estimation scores,so we use dierent distributions, parameters, and default probabilities to test the accuracy and stability of this method. In the end, we also compare our methods with Van der Burgt (2008) and Tasche (2009)' methods.

This thesis has explored the field of internally developed models for measuring the probability of default (PD) in credit risk. As regulators put restrictions on modelling practices and inhibit the advance of risk measurement, the fields of data science and machine learning are advancing. The tradeoff between stricter regulation on internally developed models and the advancement of data analytics was investigated by comparing model performance of the benchmark method Logistic Regression for estimating PD with the machine learning methods Decision Trees, Random Forest, Gradient Boosting and Artificial Neural Networks (ANN). The data was supplied by SEB and contained 45 variables and 24 635 samples. As the machine learning techniques become increasingly complex to favour enhanced performance, it is often at the expense of the interpretability of the model. An exploratory analysis was therefore made with the objective of measuring variable importance in the machine learning techniques. The findings from the exploratory analysis will be compared to the results from benchmark methods that exist for measuring variable importance. The results of this study shows that logistic regression outperformed the machine learning techniques based on the model performance measure AUC with a score of 0.906. The findings from the exploratory analysis did increase the interpretability of the machine learning techniques and were validated by the results from the benchmark methods. / Denna uppsats har undersökt internt utvecklade modeller för att estimera sannolikheten för utebliven betalning (PD) inom kreditrisk. Samtidigt som nya regelverk sätter restriktioner på metoder för modellering av kreditrisk och i viss mån hämmar utvecklingen av riskmätning, utvecklas samtidigt mer avancerade metoder inom maskinlärning för riskmätning. Således har avvägningen mellan strängare regelverk av internt utvecklade modeller och framsteg i dataanalys undersökts genom jämförelse av modellprestanda för referens metoden logistisk regression för uppskattning av PD med maskininlärningsteknikerna beslutsträd, Random Forest, Gradient Boosting och artificiella neurala nätverk (ANN). Dataunderlaget kommer från SEB och består utav 45 variabler och 24 635 observationer. När maskininlärningsteknikerna blir mer komplexa för att gynna förbättrad prestanda är det ofta på bekostnad av modellens tolkbarhet. En undersökande analys gjordes därför med målet att mäta förklarningsvariablers betydelse i maskininlärningsteknikerna. Resultaten från den undersökande analysen kommer att jämföras med resultat från etablerade metoder som mäter variabelsignifikans. Resultatet av studien visar att den logistiska regressionen presterade bättre än maskininlärningsteknikerna baserat på prestandamåttet AUC som mätte 0.906. Resultatet from den undersökande analysen för förklarningsvariablers betydelse ökade tolkbarheten för maskininlärningsteknikerna. Resultatet blev även validerat med utkomsten av de etablerade metoderna för att mäta variabelsignifikans.

Este trabalho faz uma reconstituição histórica da política monetária praticada no Brasil desde a implementação do Plano Real, revisa uma determinada discussão teórica sobre o tema da taxa de juros brasileira e suas possíveis relações perversas com outras variáveis macroeconômicas, e apresenta um modelo para tentar captar esses possíveis efeitos perversos da política monetária, tais como descritos na maior parte dos trabalhos apontados na discussão teórica. No último decênio, a taxa de juros nominal doméstica sempre esteve acima dos 15% ao ano, sendo que em grande parte do período analisado, a taxa de juros real ficou acima deste patamar. Com efeito, essa condução da política monetária trouxe à tona determinados efeitos indesejados, tais como a contaminação do risco-País pela taxa de juros doméstica. Entre os principais resultados obtidos seguindo uma análise com base num modelo VAR em que se avaliam choques nas variáveis por meio de funções impulso-resposta generalizadas (GIR), encontra-se que o risco soberano brasileiro, no período pós-desvalorização cambial, tem como determinantes os fundamentos macroeconômicos, em particular variáveis fiscais, como a dívida líquida do setor público consolidado como proporção do PIB, e a participação da dívida externa como proporção da dívida total. Outro determinante do risco percebido de moratória é a taxa de juros nominal interna. Quanto mais elevada a taxa de juros, mais elevado o risco. Em terceiro lugar, um aumento da taxa de juros pode levar a uma desvalorização cambial, desde que as expectativas dos agentes sejam afetadas pelo aumento dos riscos provocados pela elevação dos juros. / This dissertation revisits the historical background of the monetary policy regime adopted in Brazil in the period after the implementation of the Real stabilization plan, addresses to a determined theoretical framework about the domestic interest rates and its possible undesired relations with other macroeconomic variables, and presents a model to capture these possible relations of monetary policy. In the last decade, domestic nominal interest rate have always been above 15% p.a., and in a significant period of time the real interest rate stood above this level. Therefore, the conduct of monetary policy has brought up some undesired effects, such as the contagion of the Country-Risk to the domestic interest rate. Amongst the main results obtained in this paper, using a VAR model in a Generalized Impulse Response (GIR) framework for the period after the adoption of the floating exchange rate regime, stands out that the sovereign risk of Brazil is determined by macroeconomic fundaments, especially fiscal variables such as the Net Debt of the Public Sector and the share of foreign debt in the total debt. Another significant determinant of the perceived risk of default is the domestic interest rate. The higher the domestic nominal interest rate, the higher the risk. Lastly, a domestic interest rate increase may take to exchange rate depreciation if expectations are affected by the augmented risk derived from the higher domestic interest rate.