Thesis (M. Phil.)--University of Hong Kong, 2003. / Also available in print.

Thesis (M. Phil.)--University of Hong Kong, 2003.

Loss given default (LGD) modelling has become increasingly important for banks as they are required to comply with the Basel Accords for their internal computations of economic capital. Banks and financial institutions are encouraged to develop separate models for different types of products. In this thesis we apply and improve several new algorithms including support vector machine (SVM) techniques and mixed effects models to predict LGD for both corporate bonds and retail loans. SVM techniques are known to be powerful for classification problems and have been successfully applied to credit scoring and rating business. We improve the support vector regression models by modifying the SVR model to account for heterogeneity of bond seniorities to increase the predictive accuracy of LGD. We find the proposed improved versions of support vector regression techniques outperform other methods significantly at the aggregated level, and the support vector regression methods demonstrate significantly better predictive abilities compared with the other statistical models at the segmented level. To further investigate the impacts of unobservable firm heterogeneity on modelling recovery rates of corporate bonds a mixed effects model is considered, and we find that an obligor-varying linear factor model presents significant improvements in explaining the variations of recovery rates with a remarkably high intra-class correlation being observed. Our study emphasizes that the inclusion of an obligor-varying random effect term has effectively explained the unobservable firm level information shared by instruments of the same issuer. At last we incorporate the SVM techniques into a two-stage modelling framework to predict recovery rates of credit cards. The two-stage model with a support vector machine classifier is found to be advantageous on an out-of-time sample compared with other methods, suggesting that an SVM model is preferred to a logistic regression at the classification stage. We suggest that the choice of regression models is less influential in prediction of recovery rates than the choice of classification methods in the first step of two-stage models based on the empirical evidence. The risk weighted assets of financial institutions are determined by the estimates of LGD together with PD and EAD. A robust and accurate LGD model impacts banks when making business decisions including setting credit risk strategies and pricing credit products. The regulatory capital determined by the expected and unexpected losses is also important to the financial market stability which should be carefully examined by the regulators. In summary this research highlights the importance of LGD models and provides a new perspective for practitioners and regulators to manage credit risk quantitatively.

Tesis para optar al grado de Magíster en Finanzas / El trabajo a continuación tiene por objetivo encontrar qué variables demográficas, afectan el pago de los créditos hipotecarios de la personas, en qué sentido y en qué medida. En Chile, gran parte de los estudios que afectan la economía del país caen bajo la responsabilidad del Banco Central. El banco central de Chile fue creado bajo el decreto ley N° 486 del 22 de agosto de 1925, con el objetivo de restructurar el sistema monetario y financiero nacional. Este proyecto buscaba estabilizar el valor de la moneda y establecer el patrón oro como base de la unidad monetaria país, además de regular el circulante (Banco Central de Chile n.d.) Posteriormente, luego de la promulgación de la ley orgánica de bancos, el Banco Central pasa a ser una institución autónoma, teniendo por fin “propender al desarrollo ordenado y progresivo de la economía nacional mediante una política monetaria y crediticia que, procurando evitar tendencias inflacionistas o depresivas, permita el mayor aprovechamiento de los recursos productivos del país” (Banco Central de Chile n.d.), por tanto, también pasa a depender del BCCh las acciones pertinentes para priorizar el pleno empleo de los recursos productivos. Luego le siguieron varios cambios en los años 1960 y 1975, afectando el alcance de sus decisiones, la capacidad de endeudamiento, la regulación de los flujos a la Tesorería General de la Republica, las elecciones de sus directores, entre otras características. A cargo del Banco Central quedó la política monetaria, con lo que debe proteger el valor de la moneda nacional, manteniendo la inflación baja y estable, lo cual conlleva un crecimiento sostenido, una situación de pleno empleo y, en general, un bien estar para la población, siendo un incentivo para el ahorro, inversión y las ganancias productivas. La herramienta utilizada para medir la inflación es el Índice de Precios al Consumidor (IPC), el cual debiese estar en torno al 3% según la meta establecida el 2007. También la política financiera quedó bajo la responsabilidad de esta institución, resguardando el funcionamiento del sistema de pago. Por esto mismo, el Banco Central fue denominado como el prestamista de última instancia, dando liquidez a las instituciones en problemas. También adquiere facultades regulatorias en los pagos de intereses sobre cuentas corrientes y razones de crédito. Finalmente, a través de la emisión de instrumentos financieros, ha sido capaz de dar una base para la construcción de una curva de rendimiento cero cupón, información utilizada para la operatoria del mercado de derivados de tasas de interés. Junto con esto, el Banco Central puede establecer un objetivo para la Tasa de Política Monetaria (TPM).

Master of Agribusiness / Department of Agricultural Economics / Brian C. Briggeman / As the recent financial crisis brought to light, the ability of commercial banks to quantify and better manage risk in their loan portfolios is paramount to their continued success and viability. Assessing, managing, and retaining capital is now a larger issue than ever given this event as well as the advent of the Basel III Accord. Pinnacle Bancorp is a community banking organization headquartered in Omaha, Nebraska with roughly $8.6 billion in assets. The company is also one of the largest agricultural lenders in the country and the largest agricultural lender among traditional community banks. Given the ominous outlook heading into 2016 for agricultural producers from lower projected net incomes and increased borrowing costs following Federal Reserve action on the Fed Funds Rate, many banks worry about the increased likelihood of default for agricultural producers. The objective of this thesis is to determine the adequacy of Pinnacle Bank’s equity capital relative to the agricultural loan portfolio. This process begins by employing binary logit regression in an effort to determine the probability of default for the bank’s agricultural loan portfolio. With default likelihood quantified, efforts are then made to determine the bank’s credit value-at-risk at various solvency levels. These figures are then compared to current capital levels in order to determine the adequacy of bank capital as measured by five key regulatory ratios ultimately imposed by Basel III. Finally, recommendations are made to management as to the adequacy of bank capital relative to the agricultural loan portfolio and any future efforts that need to be made in order to determine and ensure the adequacy of bank capital for the entire loan portfolio.

Submitted by Rodrigo Botelho (rodrigobotelho@gmail.com) on 2013-01-15T13:24:36Z No. of bitstreams: 1 Tese Mestrado_RodrigoBotelho.pdf: 783241 bytes, checksum: 6ac3257d172dff384fca0c601e146aa4 (MD5) / Approved for entry into archive by Vitor Souza (vitor.souza@fgv.br) on 2013-01-15T14:50:33Z (GMT) No. of bitstreams: 1 Tese Mestrado_RodrigoBotelho.pdf: 783241 bytes, checksum: 6ac3257d172dff384fca0c601e146aa4 (MD5) / Made available in DSpace on 2013-02-04T16:52:43Z (GMT). No. of bitstreams: 1 Tese Mestrado_RodrigoBotelho.pdf: 783241 bytes, checksum: 6ac3257d172dff384fca0c601e146aa4 (MD5) Previous issue date: 2013-11-27 / This paper explores the sovereign default due to the structure of Credit Default Swap spreads. These spreads show the default probability of a country. The methodology proposed in this paper applied for Argentina, Korea, Ecuador, Indonesia, Mexico, Peru, Turkey, Ukraine, Venezuela and Rússia. We could show that a single factor model following a lognormal process captures the probability of default. We also show that the macro economic variables like inflation, unemployment e growth do not explain the dependent variable of this study. Each country responds differently to the economic crisis that leads to don’t honor their commitments debts. / Este trabalho explora a realização de default soberano em função da estrutura de spreads de CDS (Credit Default Swap). Pode-se dizer que os spreads revelam a probabilidade de default de um país. Aplicamos a metodologia proposta neste trabalho para Argentina, Coreia, Equador, Indonésia, México, Peru, Turquia, Ucrânia, Venezuela e Rússia. Nós mostramos que um modelo de um único fator seguindo um processo lognormal captura a probabilidade de default. Também mostramos que as variáveis macro econômicas inflação, desemprego e crescimento não explicam a variável dependente do estudo (probabilidade de default). Cada país reage de maneira diferente a crise econômica que a leva a não honrar seus compromissos com as dívidas contraídas.

This thesis is based on three essays in corporate credit risk and economic performance analysis. Corporate bankruptcy prediction using past financial information is well established in the literature. Early studies of corporate bankruptcy prediction mainly applied statistical techniques such as discriminant analysis, logit and probit. Although, some of these models such as logit is still widely popular amongst the academics and practitioners due to its simplicity, the shortcomings of such models for bankruptcy prediction have been noticed and criticised in the literature. One of the main shortcomings is that these models as linear classification approach can not explain the possible non-linear relationship between some accounting ratios and the probability of default (PD). This issue has been addressed in the literature by introducing non-linear statistical techniques such as support vector machines (SVM). The first essay of this thesis, presented in Chapter 2, investigates the performance of SVM in corporate bankruptcy prediction and compares its performance with that of logit. This essay analyses bankruptcy risk for firms in the Asian and Pacific region using a list of financial ratios which covers different aspects of a firm's performance. The financial and credit event information for this analysis is provided by the Risk Management Institute of National University of Singapore (RMI NUS). With respect to forecasting accuracy, the findings of this analysis reveal that on average the SVM displays a higher forecasting accuracy and a more robust performance than the logit. Among several financial ratios suggested as predictors of default, leverage ratios and firm size display a higher discriminating power compared to others. Additionally, an analysis of the relationship between PD and financial ratios is provided. The accounting based models in bankruptcy analysis are mostly based on a set of measures which represents current financial position of the firms. However, these models have no indication of the status of the technology competency of a firm amongst its peers, which could be a more significant factor in the survival of a firm. Therefore, a new measure about level of firm's technological knowledge is required for a more precise assessment of firms future financial performance. Considering the rise in the technological competition and patenting activities since the 1990s and also the important role of accurate credit rating modeling in the financial stability, second essay of this thesis examined in Chapter 3 focuses on the relationship between patent applications, as an output of a firm's technological development, and financial survival. Applying a survival analysis model, this relationship is empirically tested on a longitudinal firm-level data set for the listed firms in the US which matches the patent application data from European Patent Offi ce (EPO) against a set of financial variables provided by RMI NUS. The results of this analysis reveal that patent applications are strong identifiers of low default risk companies. In a further analysis, third essay of this thesis presented in Chapter 4 focuses on the impact of patent applications on firm's economic performance. In contrast to the studies which study the overall patent portfolio indicators as proxy for innovation, on a few aspects of firm performance this essay provides a comprehensive analysis of the effect of individual patent applications on several aspects of firm performance including including profitability, leverage, liquidity, size, credit rating quality and stock return. Using the matched data set of patent application data and economic variables for the US listed firms explained earlier, this essay examines whether changing from non-patenting to patenting status when a firm files its first and subsequent applications is associated with significant changes in its firm's performance and stability. The empirical findings of this essay indicates a higher capitalisation, increased liquidity, a lower leverage and an improve in credit quality for the patenting firms.

TESIS PARA OPTAR AL GRADO DE MAGÍSTER EN FINANZAS / Esta investigación se evalúa si los mercados de CDS (Credit Default Swap) de países emergentes son capaces de anticipar cambios en el rating de la deuda soberana. Se utiliza el rating soberano asignado por parte de las tres grandes agencias clasificadoras de riesgo y los Credit Default Swap soberano a 10 años, para una muestra compuesta por 27 países emergentes. Se utilizaron datos de frecuencia mensual para el periodo comprendido entre septiembre de 2008 y enero de 2018, en el cual se incluyen dos crisis financieras internacionales (crisis subprime y la amenaza de contagio de la crisis de deuda soberana de europa). El modelo econométrico consiste en una estimación en dos fases. En la primera, se estima a través de un modelo de regresión lineal de corte transversal el desalineamiento del spread de CDS de un país con respecto a sus pares de igual clasificación. En la segunda, se utiliza esta innovadora variable para estimar a través de un modelo probit con datos de panel la probabilidad de cambio de rating internalizada por el mercado de CDS. Se analiza de manera independiente los eventos de crédito que mejoran el rating (upgrade) y los que lo rebajan (downgrade). Se comprueba que, incluso utilizando diferentes supuestos para la construcción de las variables, los CDS son un instrumento financiero capaz de entregar información relevante para predecir cambios en el rating soberano. Además, mediante un conjunto de pruebas de robustez, se entrega sustento para dos principales conclusiones. Primero, que el mercado de CDS asignaría una mayor probabilidad de cambio de rating (tanto para downgrade como upgrade) a los países de peor clasificación crediticia y, más aún, a aquel grupo de países con grado especulativo. Segundo, los resultados muestran que a medida que se acerca la fecha del evento, el mercado contaría con mayor información para predecir cambios de rating, lo cual se podría esperar intuitivamente. Esta investigación realiza un aporte a la literatura previa tanto en el modelo implementado como su capacidad predictiva de cambios de rating, la cual se mantiene incluso frente a diferentes especificaciones de las variables explicativa relevantes y cambios en los supuestos utilizados.

The objective of this paper is to offer a methodology for sizing credit-sensitive Asset Backed Securities (ABS) used in the prime mortgage lending sector in the U.S. and then to evaluate their relative performance. Using a multi-factor Monte Carlo simulation framework, we perform a four-step analysis. First, we estimate scenario-specific credit losses from a given mortgage pool. We then structure the pool into a 6-pack subordination structure based on statistically-determined stress economic scenarios. Next, we estimate performance indicators of the tranches to compare risk-adjusted returns. Finally, we report our results in terms of tranche-specific risk-adjusted returns. The results indicate that the middle tranches of ABS, e.g., BBB and BB, possess the lowest risk-adjusted returns. We also find and explain a cliff phenomenon in the tranche-level principal cash flows.

The objective of this paper is to offer a methodology for sizing credit-sensitive Asset Backed Securities (ABS) used in the prime mortgage lending sector in the U.S. and then to evaluate their relative performance. Using a multi-factor Monte Carlo simulation framework, we perform a four-step analysis. First, we estimate scenario-specific credit losses from a given mortgage pool. We then structure the pool into a 6-pack subordination structure based on statistically-determined stress economic scenarios. Next, we estimate performance indicators of the tranches to compare risk-adjusted returns. Finally, we report our results in terms of tranche-specific risk-adjusted returns. The results indicate that the middle tranches of ABS, e.g., BBB and BB, possess the lowest risk-adjusted returns. We also find and explain a cliff phenomenon in the tranche-level principal cash flows.