Credit Risk Models Research Articles

Reinforcement of the Bank Loan Model using the Feature Selection Method of Machine Learning

Does feature selection and machine learning (ML) guarantee the effectiveness of the bank credit system model? This article aims to analyze this problem. In fact, in finance, expert-based credit risk models still dominate. In this study, we establish a new benchmark using consumer data and present machine learning methods. A risk prediction that is as accurate as possible is an important requirements for credit scoring models. In addition, regulators expect that the models should to be auditable and transparent. As a result, the superior predictive power of contemporary machine learning algorithms cannot be fully utilized in credit scoring because very simple predictive models, such as several ML classifiers, are still widely used. As a result, significant potential is missed, increasing reserves or the number of credit defaults. A framework for comparing scores before and after feature selection machine learning models that are transparent, auditable, and explainable is presented in this article, as well as the various dimensions that need to be taken into consideration in order to make credit scoring models understandable. In accordance with this framework, we give an overview of the models which demonstrate how it can be used in credit scoring, and compare the results to scorecards' interpretability. The model presented demonstrates that machine learning techniques can maintain their ability to enhance predictive power while still maintaining a comparable level of interpretability.

Boosting credit risk models

In this article, we give various recommendations to boost the performance of credit risk models. It is based upon more than two decades of research and consulting on the topic. Building credit risk models typically entails four steps: gathering and preprocessing data, modelling of probability of default (PD), Loss Given Default (LGD) and Exposure at Default (EAD), evaluating the credit risk models built and then the deployment step to put them into production. We give recommendations to boost credit risk models during each of these steps. Furthermore, we also define and review model risk as an all-encompassing challenge one needs to be properly aware of during each step of the process. We conclude by presenting a research agenda of topics we believe are in high need for further investigation and study.

Modern Approaches in Credit Risk Modeling: A Literature Review

Abstract In the financial industry, models are pervasive, and their quantity and complexity continue to increase. Constant advancements are made in econometric and statistical theory, but a fast-developing body of rules and regulations governing their use needs modeling specialists to remain vigilant and adaptable. The tendency of these regulations to be ambiguous necessitates that industry professionals and institutions interpret them independently and jointly. This leads in what is referred to as a “industry standard,” or a set of procedures that are recognized among modeling professionals but not necessarily to those outside of the industry. Non-practitioners in the industry may view the modeling department as a “black box” for these reasons. The accurate evaluation of financial credit risk and the forecasting of bankruptcy are crucial to both the economy and society. In recent years, more and more approaches and algorithms have been advanced for this purpose. At this point, it is of the highest concern to investigate the current credit risk assessment methods. In this paper, we review the traditional statistical models and cutting-edge intelligent methods for forecasting financial distress, with a focus on the greatest advances in the academic literature, as the promising trend in this field. Lastly, the paper will conclude with an overview of the evolution of methodologies and conceptual frameworks in credit risk management research, as well as possible future research directions.

State guarantees to counteract the financial effects of the COVID-19 pandemic on industrial supply chains

As the recent COVID-19 pandemic has clearly demonstrated, appropriate state support policies are crucial for supporting industrial supply chains during crises to prevent viable businesses from defaulting. In this context, this study proposes a hybrid credit risk model to appropriately size public interventions for viable (worthy) businesses through systematic risk assessment during a period of turmoil. This study discusses the effects of the credit crunch-based economic downturn and proposes a methodology to assist policymakers in managing limited public resources to effectively support industrial supply chains. The proposed approach initially focuses on the dynamics of credit risk during economic recession periods, identifying the conditions that may justify a public intervention strategy based on public guarantees. Subsequently, a hybrid credit risk model is developed to appropriately size public interventions by quantifying systematic risk. Finally, a numerical application is presented to demonstrate the effectiveness of the proposed approach.

Managing the risk of embedded options in non-traded credit using portfolio modeling

A framework for measuring and managing the risk of embedded options in non-traded credit is developed. For typical bank clients there is no market information related to their ability to pay (bond or CDS spreads) available. The absence of market information is a key assumption of this paper. In this case, a bank has to rely solely on statistical data to judge the credit quality of a borrower. To value a loan with embedded options like prepayment rights, a model is proposed that combines an interest rate derivatives pricing model with statistical information on default and recovery rates. Using this for evaluating the risk of embedded options in loans, it is shown how the concepts of credit risk management can be applied after defining a suitable concept of risk. It turns out that this modeling framework combines the theories of derivatives pricing and credit risk modeling in the sense that derivatives pricing theory measures the costs for hedging optional components in loans while credit risk modeling measures the risk that these hedging costs turn out to be inadequate. This risk depends not only on the single loan’s risk characteristics but also on the dependence structure and the granularity of the total loan portfolio.

METRIC BINARY CLASSIFIER WITH SELECTION OF FEATURE WEIGHTS

The family of metric algorithms based on determining the distance from one observation to another has a number of advantages, such as their suitability for many types of problems and results have a clear interpretation. Therefore, metric algorithms are widely used in credit risk modeling, non-destructive quality control of products, medical diagnostics, geology, and many other practical areas. The most common metric algorithm in practice is the k-nearest neighbors method. At the same time, one of the key problems of metric algorithms is the problem of dimension, since the decision is made on the basis of all observations of the training sample. In addition, usually all variables have the same weight when calculating the distance, which leads to a drop in the quality of the algorithm with an increase in the number of features. The article discusses a new machine learning method for solving classification problems – a metric classifier with the selection of feature weights, which allows to solve these problems to a large extent. Nine algorithms were used to optimize the function. Classification quality based on them is checked on 3 problems from the UCI repository. As a result of the comparison, the truncated Newton method was chosen to build a new metric classifier. The quality of the new classifier was tested on 8 datasets from the same repository and compared with the quality of the classical nearest neighbor method. This classifier has a higher quality for problems with a large number of features in comparison to the classical approach. Data set characteristics and calculation results are presented in the corresponding tables.

A bottom-up, reduced form credit risk model approach for the determination of collateralised loan obligation capital

This paper uses a bottom-up, reduced form credit risk model with hazard rate estimated default probabilities to compute various collateralised loan obligation (CLO) tranches' loss probabilities and capital factors. It is shown that with respect to the loss probabilities, credit rated CLO tranches are less risky than comparably rated corporate bonds. In addition, a similar argument can be made that corporate debt loss rates will be on average larger than equally rated CLO tranche loss rates. And, it is shown that the National Association of Insurance Commissioners (NAIC) capital factors are typically larger than value-at-risk based capital factors. The policy implication is that NAIC capital factors distort investment incentives by requiring too much capital for CLOs relative to equally rated corporate debt.

SPECIFICATION ANALYSIS OF STRUCTURAL CREDIT RISK MODEL WITH STOCHASTIC VOLATILITY

In this research, we conduct specification analysis of structural credit risk models with stochastic volatility using term structure of Credit Default Swap (CDS) spreads and equity volatility form high-frequency return data. We also test five representatives of structural credit risk models using samples of 93 single name CDS contracts from January 2010 -2022. The model we consider are; the standard Merton(1974) model, the Black & Cox (1976) model with flat barrier, the Longstaff and Schwartz(1995)model with stochastic interest rates, the Collin- Dufresne and Goldstein (2001) model with stationary leverage, and the double exponential jump diffusion model used in Huang & Huang(2003). Our study provides consistent econometric estimation of the pricing model parameters and specification tests based on the joint behavior of time-series asset dynamics and cross-sectional pricing errors. Our empirical tests reject strongly the standard Merton (1974) model, the Black and Cox (1976) with flat barrier model, the Longstaff, Schwartz (1995) model with stochastic interest rates. The double exponential jump-diffusion barrier model used in (Huang and Huang, 2003) improves significantly over the five models. The best model is the stationary leverage model of Collin-Dufresne and Goldstein (2001), which we cannot reject at 0.5 level of significance in our sample firm. However, our empirical results document the inability of the existing structural models to capture the dynamic behavior of CDS spreads and equity volatility, especially for high investment grade names derivatives. These points to a potential role of time-varying asset volatility, a feature that is missing in the standard structural models.

EVALUATING CREDIT RISK MODELS

Empirical testing of competing theories lies at the heart of social science and applied Science research. We demonstrate that a well-known class of statistical models, called finite mixture models, provides an effective way of rival theory testing. In the proposed framework, each observation is assumed to be generated either from statistical models implied by one of the competing theories or more generally from a weighted combination of multiple statistical models under consideration. Researchers can then estimate the probability that a specific observation is consistent with each rival theory. By modeling this probability with covariates, one can also explore the condition under which a particular theory applies. We discuss a principles way to identify a list of observations that are statistically consistent with each theory and propose measure of the overall performance of each competing theory. We illustrate the relatives’ advantages of our method over existing methods through empirical and stimulation studies.

Bankruptcy Prediction for Micro and Small Enterprises Using Financial, Non-Financial, Business Sector and Macroeconomic Variables: The Case of the Lithuanian Construction Sector

Credit-risk models that are designed for general application across sectors may not be suitable for the construction industry, which has unique characteristics and financial risks that require specialised modelling approaches. Moreover, advanced bankruptcy-prediction models are often used to achieve the highest accuracy in large modern datasets. Therefore, the aim of this research is the creation of enterprise-bankruptcy prediction (EBP) models for Lithuanian micro and small enterprises (MiSEs) in the construction sector. This issue is analysed based on classification models and the specific types of variable used. Firstly, four types of variable are proposed. In EBP models, financial variables substantially explain an enterprise’s financial statements and performance from different perspectives. Including enterprises’ non-financial, construction-sector and macroeconomic variables improves the characteristics of EBP models. The inclusion of macroeconomic variables in the model has a particularly significant impact. These findings can be of great significance to investors, creditors, policymakers and practitioners in assessing financial risks and making informed decisions. The second question is related to the classification models used. To develop the EBP models, logistic regression (LR), artificial neural networks (ANNs) and multivariate adaptive regression splines (MARS) were used. In addition, this study developed two-stage hybrid models, i.e., the LR is combined with ANNs. The findings show that two-stage hybrid models do not improve bankruptcy prediction. It cannot be argued that ANN models are more accurate in predicting bankruptcy. The MARS model demonstrates the best bankruptcy prediction, i.e., this model could be a valuable tool for stakeholders to evaluate enterprises’ financial risk.

Efficacies of artificial neural networks ushering improvement in the prediction of extant credit risk models

Efficacies of artificial neural networks ushering improvement in the prediction of extant credit risk models

Sea-Level Rise Exposure and Municipal Bond Yields

Abstract Municipal bond markets began pricing sea-level rise (SLR) exposure risk in 2013, coinciding with upward revisions to worst-case SLR projections and accompanying uncertainty around these projections. The effect is larger for long-maturity bonds and not solely driven by near-term flood risk. We use a structural model of credit risk to quantify the implied economic impact and distinguish between the effects of underlying asset values and of uncertainty. The SLR exposure premium exhibits a trend different from house prices and is unaffected by house price controls. Together, our results highlight the importance of climate uncertainty in driving municipal bond prices. Authors have furnished an Internet Appendix, which is available on the Oxford University Press Web site next to the link to the final published paper online

Structural credit risk model driven by Lévy process under knight uncertainty.

As per the projections of conventional credit risk structured model, the risky asset values tend to adhere to the geometric Brownian motion. On the contrary, the risky asset values remain a non-continuous and dynamic ones and jump based on the conditions. Is not possible to measure the real Knight Uncertainty risks in financial markets with the help of a single probability measure. In this background, the current research work analyzes a structural credit risk model that belongs to Levy market under Knight Uncertainty. With the help of Lévy-Laplace exponent, the authors developed a dynamic pricing model in this study and acquired the price intervals for default probability, stock value and the bond value of enterprise. To be specific, the study intended to establish explicit solutions for three value processes, discussed earlier, with an assumption that the jump process follows a log-normal distribution. At the end, the study also conducted numerical analysis to understand the crucial role played by Knight Uncertainty upon the pricing of default probability and the stock value of the enterprise.

A Framework for Trustworthy AI in Credit Risk Management: Perspectives and Practices

We present a reference architecture for trustworthy artificial intelligence based on three key tenets (usable, reliable, and transparent) and analyze its applicability in managing the challenges involved in the credit risk model from a practitioner’s perspective. We propose a holistic approach covering relevant concerns in practical implementations.

Default Prediction with Industry-Specific Default Heterogeneity Indicators Based on the Forward Intensity Model

When predicting the defaults of a large number of samples in a region, this will be affected by industry default heterogeneity. To build a credit risk model that is more suitable for Chinese-listed firms, which have highly industry-specific default heterogeneity, we extend the forward intensity model to predict the defaults of Chinese-listed firms with information about the default heterogeneity of industries. Compared with the original model, we combine the Bayes approach with the forward intensity model to generate time-varying industry-specific default heterogeneity indicators. Our model can capture co-movements of different industries that cannot be observed based on the original forward intensity model so that the model can flexibly adjust the firm’s PD according to the industry. In addition, we also consider the impact of default heterogeneity in other industries by studying the influence of the level and trends of other industries’ default heterogeneity on a firm’s credit risk. Finally, we compute PDs for 4476 firms from January 2001 to December 2019 for 36 prediction horizons. The extended model improves the prediction accuracy ratios both for the in-sample and out-of-sample firm’s PDs for all 36 horizons. Almost all the accuracy ratios of the prediction horizons’ PDs are increased by more than 6%. In addition, our model also reduces the gap between the aggregated PDs and the realized number of defaults. Our industry-specific default heterogeneity indicator is helpful to improve the model’s performance, especially for predicting defaults in a large portfolio, which is of significance for credit risk management in China and other regions.

Default Feature Selection in Credit Risk Modeling: Evidence From Chinese Small Enterprises

This paper aims to design a novel AFCM-SMOTENC-APRIORI model to mine the default feature attributes of small enterprises. It can overcome the problem that the data characteristics of “small defaulting small enterprises and large non-defaulting small enterprises” make it difficult to mine the defaulting feature attributes of existing small enterprises. We used 1,231 small enterprise credit data from a city commercial bank in China to make an empirical analysis. We found that 23 feature attributes are strongly associated with default and 87% of the association rules are the same between the extended data and the original data mining. It shows that the data mining results with SMOTE-NC are highly consistent with the results of the original data mining, and the model is robust and reliable. It can be used as a reference for the credit risk identification of small enterprises in commercial banks.

The Saga of Ruchi Soya Industries Limited : Could Credit Risk Models Predict Bankruptcy?

Purpose : The primary objective of this study was to examine the efficacy of credit risk models in predicting bankruptcy and evaluating the firm's post-acquisition performance. Design/Methodology : Ruchi Soya Industries Limited, an Indian listed firm that went into bankruptcy in 2018, was identified for back-testing and evaluating the predictive ability of four models: three accounting-based (Altman Z-score, Altman’s emerging market score, and Zmijewski) and the market-based KMV model. In the second stage of analysis, the operating and financial performance of the company was evaluated for pre-bankruptcy and post-acquisition by Patanjali Foods using t-tests on financial parameters of solvency, profitability, and efficiency. Findings : The findings of this study revealed that while all the models were accurate in predicting default accurately for up to 1 year before bankruptcy, they failed to do so accurately beyond that. The predictive ability of the models was highest for the KMV model, followed by Zmijewski and Altman's Z-score. The performance of the firm improved significantly post-acquisition on profitability and solvency parameters. Practical Implications : According to the findings of this research, credit risk models are accurate at predicting financial trouble and bankruptcy up to 1 year in advance. These findings can be used for credit appraisal by lenders to assess any financial trouble and enable effective risk management. These models can also be used to avoid business failure to develop proactive and preventive financial and managerial decisions. The impact on performance parameters post-acquisition can help consultants and advisors evaluate the restructuring process to see whether there has been value creation post-restructuring.

A Practical Guide to the Valuation of Coupon-Bearing Fixed Income Securities

The purpose of this article is to illustrate the use of reduced-form credit risk models for valuing defaultable coupon-bearing securities, such as risky sovereign bonds, corporate bonds, and retail loans like auto loans and mortgage loans. We focus on issues in the application of these models and not the mathematical derivations of the formulas employed.

Использование марковских моделей со множеством состояний для прогнозирования вероятности дефолта заемщиков

Introduction. After the crises, lenders realized the importance of assessing the risk of default on loan portfolios in various economic conditions. Modeling of credit risk assessment occurs mainly using internal ratings of banks based on probabilistic models of defaults of borrowers over a certain period of time. Theoretical models. Three models are considered. The first is a naive Markov model with R states. The transition matrix is given. The second is a Markov model with multiple states with covariates. Macroeconomic indicators are proposed as covariates. The third model is multinomial logit regression. Approbation of Markov models and multinomial regression on simulated and real data of borrowers’ defaults. We investigate the possibility of using multi-state Markov models to predict borrower defaults in financial institutions over time. Three approaches are considered for credit risk modeling. The first approach assumes that the transition probability matrix is constant over time, and the residuals of the Markov model and logistic regression are taken into account further when forecasting over the time horizon. The second one is supplemented by the Markov model, which takes into account the impact of default risks on migration, both individual factors of borrowers and the economic situation in the country. Using covariates, the models made it possible to simultaneously estimate the transition rate and the probability of erroneous classification of states. A multinomial logistic regression model is considered to compare the results obtained using multi-state Markov models. The proposed models are tested both on real and simulated data. Conclusion. The presented models show good predictive results with high accuracy of default estimates. The models reproduce the structure of the generated data quite well. The peculiarity of the multinomial regression model in predicting defaults is its adjustability, and Markov models estimate the probabilities of defaults. To implement the model, software R was used.

A de-singularized meshfree approach to default probability estimation under a regime-switching synchronous-jump tempered stable Lévy model

The default probability of a publicly-traded company is modeled by a synchronous-jump regime-switching model in the paper (Hainaut and Colwell, 2016). In this investigation, we first generalize the proposed Lévy model to a more general setting of tempered stable processes recently introduced into the finance literature. However, the resulting integro-partial differential operator suffers from a singularity, thus a general framework based on strictly positive-definite functions is proposed as a remedy to de-singularize the operator. We then analyze an efficient meshfree collocation method based on radial basis functions to approximate the solution of the corresponding system of partial integro-differential equations arising from the structural credit risk model. We prove that considering some regularity assumptions, the proposed method automatically de-singularizes the problem in the tempered stable case. Finally, we perform some numerical experiments using the proposed method on some standard examples from the literature which confirm the validity of our theoretical results and the stability of the numerical algorithm.