Credit Risk Prediction: A comparative study between logistic regression and logistic regression with random effects (original) (raw)
Related papers
International Journal of Economics and Finance, 2016
This paper aims to develop models for foreseeing default risk of small and medium enterprises (SMEs) for one Tunisian commercial bank using two different methodologies (logistic regression and discriminant analysis). We used a database that consists of 195 credit files granted to Tunisian SMEs which are divided into five sectors “industry, agriculture, tourism, trade and services” for a period from 2012 to 2014. The empirical results that we found support the idea that these two scoring techniques have a statistically significant power in predicting default risk of enterprises. Logistic discrimination classifies enterprises correctly in their original groups with a rate of 76.7% against 76.4% in case of linear discrimination giving so a slight superiority to the first method.
International Journal of Applied Operational Research - An Open Access Journal, 2015
One of the most important issues always facing banks and financial institutes is the issue of credit risk or the possibility of failure in the fulfillment of obligations by applicants who are receiving credit facilities. The considerable number of banks’ delayed loan payments all around the world shows the importance of this issue and the necessary consideration of this topic. Accordingly, many efforts have been made for providing an efficient model for more accurate evaluation and classification of applicants receiving credit facilities for valid decision making about granting or not granting these facilities to them. Different statistic methods have been applied for this purpose, such as Discriminant Analysis, Probit Regression, Logistic Regression, Neural Network and so on. Among these methods, Neural Network has been considered mostly because of its high flexibility in recent years. In this research, many efforts have been made to examine the efficiency of Logistic Regression an...
In this comprehensive study, we delve into the utilization of Logistic Regression (LR) and Artificial Neural Networks (ANN) for predicting credit risk in the English banking sector over the period from 2021 to 2023. Through an in-depth analysis of quarterly financial and non-financial data from various banks, this research aims to discern which predictive modeling technique provides more accuracy and reliability. The comparison between LR and ANN models offers significant insights into their capabilities and limitations, potentially guiding fut ure risk management and decision-making processes in banking. This study also addresses the importance of advanced analytical methods in improving the predictiveness of financial risks, thus contributing to the enhancement of banking operations and the promotion of financial stability.
2022
The search for standards that contribute to the prediction of risk is growing in organizations. The use of credit scoring models seeks to assist the credit analyst in making decisions. This work aims to develop methodological procedures, to structure and improve credit scoring models aimed at the analysis of small and medium-sized companies. With the use of the statistical technique of logistic regression, through the improvements developed in the methodological procedures, such as division of the database into classes according to the companies' framework, it was possible to develop 5 credit scoring models, one model for each class of companies and another for the general database. The models were directed to entities that promote and grant credit to small and medium-sized companies. The accuracy of the models showed significant percentages for the database with non-accounting and nonauditable variables, reaching satisfactory percentages.
Credit risk management of commercial banks in Iran; Using logistic model
African Journal of Business Management, 2013
The observed crisis and profitability decreases in banking system are mainly because of the inefficiency in credit risk control and that's why, utilization of customers' ranking system is the most important tool that is required for managing and controlling the risk. The goal of the study was to present an applied model for credit scoring of real entity customers of banks with reliance on statistical information of credit customers of Parsian Bank in Iran. For this purpose the logistic regression model is used to analyze credit ranking and financial scoring of bank's customers based on their previous and current data record like; job stability, collaterals, income and some other main indicators for estimating non-default probability of facilities offered to each customer. The results of the model estimation showed that non-default probability of facilities have positive relation with variables amount of collaterals received from customer, monthly income amount of the customer, the status of applicant for taking facilities such as place of residence (be owner or tenant of the applicant), the age of applicant for taking facilities, occupational status of the applicant as stability and educational level of the applicant for taking facilities and have negative relation with amount of paid facilities to the customer and payback duration of granted facilities to the applicant.
Expert Systems with Applications, 2012
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 ) and a logistic regression with state-dependent sample selection model 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.
Credit Risk Analysis Applying Logistic Regression, Neural Networks and Genetic Algorithms Models
Most large Brazilian institutions working with credit concession use credit models to evaluate the risk of consumer loans. Any improvement in the techniques that may bring about greater precision of a prediction model will provide financial returns to the institution. The first phase of this study introduces concepts of credit and risk. Subsequently, with a sample set of applicants from a large Brazilian financial institution, three credit scoring models are built applying these distinct techniques: Logistic Regression, Neural Networks and Genetic Algorithms. Finally, the quality and performance of these models are evaluated and compared to identify the best. Results obtained by the logistic regression and neural network models are good and very similar, although the first is slightly better. Results obtained with the genetic algorithm model are also good, but somewhat inferior. This study shows the procedures to be adopted by a financial institution to identify the best credit model to evaluate the risk of consumer loans. Use of the best fitted model will favor the definition of an adequate business strategy thereby increasing profits.
An evaluation of Iranian banking system credit risk: Neural network and logistic regression approach
International Journal of Physical Sciences, 2011
The current study seeks to provide a new solution for evaluation of banking system customers risk by integrating different scientific methodology. Evaluation of banking system customers risk in Iranian banks relies on experts judgment and fingertip rule. This type of evaluation resulted in high rate of postponed claims; therefore, designing new intelligent model for credit risk evaluation will be helpful, thus in this paper, we formulated an intelligent model by neural network and logistic regression that evaluated all individual customers credit risk without prejudice and discrimination. The result revealed that neural network and logistic regression have the same ability in predicting customer credit risk. Their ability in customer credit risk correct evaluation was nearly 79.50%. We suggested that both models could be used by all financial system as consultant model for customer credit risk prediction. The study also involved only one banking system credit customers, which concerns just Tehran city customers and its sample includes only individual customers, thus cannot be for institutional customers. Offering a case study, this paper presents a guide for banking system to predict any customer credit risk and regulate any customer loan in the light of customer risk that was extracted by neural network, and logistic regression employed different scientific methodologies in their service quality development efforts. Intending to offer scientific approaches to risk evaluation as a tool of customer credit risk assessment in banking system loan allocation procedures, this paper tries to bridge the current gap between academicians and practitioners; adds to the relatively limited theoretical literature.
Measuring customer's credit risk in banking industry
Management Science Letters, 2013
This paper presents a comparative study of using a linear probability and Logit models to predict credit risk of the customers in some branches of Bank Mellat in Tehran, Iran. The statistical population of this research includes the applicants of the facilities granted by Bank Mellat in Tehran during the year 2008. Each branches of Bank Mellat of Tehran has been considered as a cluster, where a sample has been taken using simple random method. The sample size consists of 176 companies, 109 legal entities are classified as those ones good at settling their accounts, and 67 as those ones tardy in settling their accounts. The financial ratios of these companies have been calculated based on their audited financial statements and by descriptive and analytical methods of two statistical models. The results show that liquidity ratios are not significant factors for the prediction of credit risks and these two models are not significantly different from each other in this term. Moreover, the accuracy values of credit risk prediction of linear and Logit models are 73.7 percent and 80.3 percent, respectively. Therefore, Logit model is more consistent with reality and more appropriate for such a prediction.