Análisis del credit scoring

El problema de la morosidad está cobrándose una gran importancia en los países desarrollados. En este trabajo realizamos un análisis de la capacidad predictiva de dos modelos paramétricos y uno no paramétrico abordando, en este último, el problema del sobreaprendizaje mediante la validación cruzada que, muy habitualmente, se obvia en este tipo de estudios. Además proponemos la distinción de tres tipos de solicitudes dependiendo de su probabilidad cumplimiento: conceder, no conceder (de forma automática), y dudoso y, por consiguiente, proceder a su estudio manual por parte del personal bancario.

Credit Scoring: Aplicação da regressão logística vs redes neuronais artificiais na avaliação do risco de crédito no mercado Cabo-verdiano

Dissertação apresentada como requisito parcial de obtenção do grau de Mestre em Estatística e Gestão de Informação

Poly-bagging predictors for classification modelling for credit scoring

Credit scoring modelling comprises one of the leading formal tools for supporting the granting of credit. Its core objective consists of the generation of a score by means of which potential clients can be listed in the order of the probability of default. A critical factor is whether a credit scoring model is accurate enough in order to provide correct classification of the client as a good or bad payer. In this context the concept of bootstraping aggregating (bagging) arises. The basic idea is to generate multiple classifiers by obtaining the predicted values from the fitted models to several replicated datasets and then combining them into a single predictive classification in order to improve the classification accuracy. In this paper we propose a new bagging-type variant procedure, which we call poly-bagging, consisting of combining predictors over a succession of resamplings. The study is derived by credit scoring modelling. The proposed poly-bagging procedure was applied to some different artificial datasets and to a real granting of credit dataset up to three successions of resamplings. We observed better classification accuracy for the two-bagged and the three-bagged models for all considered setups. These results lead to a strong indication that the poly-bagging approach may promote improvement on the modelling performance measures, while keeping a flexible and straightforward bagging-type structure easy to implement. (C) 2011 Elsevier Ltd. All rights reserved.

Práticas em gestão de sistemas de credit scoring e portfólio de crédito em instituições financeiras brasileiras

Vários estudos foram realizados pela academia brasileira sobre desenvolvimento e aplicabilidade de modelos estatísticos de credit scoring e portfólio de crédito. Porém, faltam estudos relacionados sobre como estes modelos são empregados pelas empresas brasileiras. Esta dissertação apresenta uma pesquisa, até então inédita, sobre como as instituições financeiras brasileiras administram seus sistemas de credit scoring e suas carteiras de crédito. Foram coletados dados, por meio de um questionário, dos principais bancos e financeiras do mercado brasileiro. Para a análise dos resultados, as repostas foram divididas em dois grupos: bancos e financeiras. Os resultados mostraram empregos de métodos diferentes entre os grupos devido a suas características operacionais.

On the impact of disproportional samples in credit scoring models: An application to a Brazilian bank data

Statistical methods have been widely employed to assess the capabilities of credit scoring classification models in order to reduce the risk of wrong decisions when granting credit facilities to clients. The predictive quality of a classification model can be evaluated based on measures such as sensitivity, specificity, predictive values, accuracy, correlation coefficients and information theoretical measures, such as relative entropy and mutual information. In this paper we analyze the performance of a naive logistic regression model (Hosmer & Lemeshow, 1989) and a logistic regression with state-dependent sample selection model (Cramer, 2004) applied to simulated data. Also, as a case study, the methodology is illustrated on a data set extracted from a Brazilian bank portfolio. Our simulation results so far revealed that there is no statistically significant difference in terms of predictive capacity between the naive logistic regression models and the logistic regression with state-dependent sample selection models. However, there is strong difference between the distributions of the estimated default probabilities from these two statistical modeling techniques, with the naive logistic regression models always underestimating such probabilities, particularly in the presence of balanced samples. (C) 2012 Elsevier Ltd. All rights reserved.

Regressão Logística Geograficamente Ponderada aplicada a modelos de Credit Scoring

Dissertação (mestrado)—Universidade de Brasília, Faculdade de Economia, Administração e Contabilidade, Programa de Pós-Graduação em Administração, 2016.

ABOUT PSYCHOLOGICAL VARIABLES IN APPLICATION SCORING MODELS

The purpose of this study is to investigate the contribution of psychological variables and scales suggested by Economic Psychology in predicting individuals’ default. Therefore, a sample of 555 individuals completed a self-completion questionnaire, which was composed of psychological variables and scales. By adopting the methodology of the logistic regression, the following psychological and behavioral characteristics were found associated with the group of individuals in default: a) negative dimensions related to money (suffering, inequality and conflict); b) high scores on the self-efficacy scale, probably indicating a greater degree of optimism and over-confidence; c) buyers classified as compulsive; d) individuals who consider it necessary to give gifts to children and friends on special dates, even though many people consider this a luxury; e) problems of self-control identified by individuals who drink an average of more than four glasses of alcoholic beverage a day.

Credit risk of financial institutions

A Work Project, presented as part of the requirements for the Award of a Masters Degree in Finance from the NOVA – School of Business and Economics

Towards a theory of the credit-risk balance sheet

This article designs what it calls a Credit-Risk Balance Sheet (the risk being that of default by customers), a tool which, in principle, can contribute to revealing, controlling and managing the bad debt risk arising from a company¿s commercial credit, whose amount can represent a significant proportion of both its current and total assets.To construct it, we start from the duality observed in any credit transaction of this nature, whose basic identity can be summed up as Credit = Risk. ¿Credit¿ is granted by a company to its customer, and can be ranked by quality (we suggest the credit scoring system) and ¿risk¿ can either be assumed (interiorised) by the company itself or transferred to third parties (exteriorised).What provides the approach that leads to us being able to talk with confidence of a real Credit-Risk Balance Sheet with its methodological robustness is that the dual vision of the credit transaction is not, as we demonstrate, merely a classificatory duality (a double risk-credit classification of reality) but rather a true causal relationship, that is, a risk-credit causal duality.Once said Credit-Risk Balance Sheet (which bears a certain structural similarity with the classic net asset balance sheet) has been built, and its methodological coherence demonstrated, its properties ¿static and dynamic¿ are studied.Analysis of the temporal evolution of the Credit-Risk Balance Sheet and of its applications will be the object of subsequent works.

Towards a theory of the credit-risk balance sheet

This article designs what it calls a Credit-Risk Balance Sheet (the risk being that of default by customers), a tool which, in principle, can contribute to revealing, controlling and managing the bad debt risk arising from a company¿s commercial credit, whose amount can represent a significant proportion of both its current and total assets.To construct it, we start from the duality observed in any credit transaction of this nature, whose basic identity can be summed up as Credit = Risk. ¿Credit¿ is granted by a company to its customer, and can be ranked by quality (we suggest the credit scoring system) and ¿risk¿ can either be assumed (interiorised) by the company itself or transferred to third parties (exteriorised).What provides the approach that leads to us being able to talk with confidence of a real Credit-Risk Balance Sheet with its methodological robustness is that the dual vision of the credit transaction is not, as we demonstrate, merely a classificatory duality (a double risk-credit classification of reality) but rather a true causal relationship, that is, a risk-credit causal duality.Once said Credit-Risk Balance Sheet (which bears a certain structural similarity with the classic net asset balance sheet) has been built, and its methodological coherence demonstrated, its properties ¿static and dynamic¿ are studied.Analysis of the temporal evolution of the Credit-Risk Balance Sheet and of its applications will be the object of subsequent works.

Modelo de Scoring para o Crédito à Habitação: Um estudo de caso em Portugal

Mestrado em Contabilidade e Gestão das Instituições Financeiras

Modelação e estimação do risco de crédito: estudo de uma carteira

Dissertação apresentada na Faculdade de Ciências e Tecnologia da Universidade Nova de Lisboa para obtenção do grau de Mestre em Matemática e Aplicações - Actuariado, Estatística e Investigação Operacional

Modelação do risco de crédito numa carteira de crédito ao consumo

A análise de risco de crédito nas instituições bancárias e a mensuração do risco é de extrema importância para as instituições, uma vez que a concessão de crédito é a sua principal actividade. A capacidade de distinguir “bom” e “mau” cliente é um processo decisivo na constituição do crédito, pelo que são aplicados modelos de Credit Scoring , modelos quantitativos que consistem numa análise estatística à qualidade do crédito. O objectivo desta dissertação é estimar a probabilidade de incumprimento de cada cliente em função das variáveis sócio-económicas e demográficas, tendo por base dados de uma carteira de crédito ao consumo de uma Instituição Bancária de Cabo Verde, através de uma técnica estatística multivariada: a Regressão Logística. Adicionalmente, estima-se a taxa de recuperação do crédito, para clientes incumpridores, recorrendo à Regressão Beta, com base no histórico do crédito de cada cliente. Neste trabalho propõe-se, ainda, ummodelo para a estimação do spread a aplicar a um novo cliente assumido pela instituição bancária, em função da probabilidade de default(incumprimento) e da taxa de recuperação estimada.

Subprime-kriisi ja luottoriskin hallinta

Tutkielma käsittelee subprime-lainausta ja subprime-kriisiin johtaneita syitä. Lisäksi tutkielmassa käsitellään pankin luottoriskin hallintaa eri näkökulmista. Luottoriski on keskeisin luottolaitostoimintaan liittyvistä riskeistä ja siksi pankkien luottoriskin hallintaa tulee jatkuvasti parantaa. Luottotappiot ja niihin johtaneet syyt ovat keskeisesti tarkastelun alaisina, kun syitä subprime-kriisiin haetaan. Tutkielma perustuu kirjallisuuteen ja siinä pyritään tuomaan esiin eri tekijöitä, jotka vaikuttivat vuonna 2007 alkaneeseen subrime-kriisiin. Työ sisältää lyhyen pankkitoiminnan luonteen ja alan rakennemuutoksen, esittelen aluksi lyhyesti pankkitoiminnan perusperiaatteita ja yleistä toimintaa sekä kuvaan lyhyesti rahoitusmarkkinoiden kehitystä viime vuosikymmenten aikana. Seuraavana esittelen laajemmin luottoriskin roolia subprime-kriisissä sekä sen epäonnistunutta hallintaa osasyynä kriisiä. Tarkemmin käsittelen Value at Risk -mallia luottoriskin mittaamiskeinona ja credit scoring -luottopisteytysmenetelmän osana luottoriskin hallintaa. Toinen tutkimuksen pääaihe on subrime-lainaus. Subrime-kriisin syntyyn vaikuttaneena olennaisena tekijänä pidetään Yhdysvaltojen asuntomarkkinoita ja erityisesti subprime-asuntolainoja, jotka olivat luottoluokitukseltaan heikkoja. Näiden heikkojen lainojen yleistyessä koko rahoitusmarkkinoiden pohja heikkeni. Rahoitusjärjestelmän suurien ongelmien taustalla vaikuttivat osaltaan arvopaperistaminen ja sen avulla muodostetut sijoitusinstrumentit, yli-innokas lainaaminen markkinoilla tarjolla olleen runsaan likviditeetin seurauksena, huolimaton sijoittaminen, riskinottoa liian voimakkaasti painottavat kannustinjärjestelmät, laaja-alainen epäonnistuminen luottoriskinhallinnassa ja sääntelyssä. Näitä tekijöistä tutkielmassa tarkastellaan tarkemmin arvopaperistamista ja kasvanutta lainaamista subrime-kriisin syntymisen osatekijöinä.