Default Prediction Models Research Articles

Is the Financial Report Quality Important in the Default Prediction? SME Portuguese Construction Sector Evidence

This work analyses whether financial information quality is relevant to explaining firms’ probability of default. A financial default prediction model for SMEs (Small and Medium Enterprises) is presented, which includes not only traditional measures but also financial reporting quality (FRQ) measures. FRQ influences the decision-making due to its impact on financial information, which has repercussions on the accounting ratios’ informativeness. A panel data of 1560 Portuguese SMEs in the construction sector, from 2012 to 2018, is analysed. First, firms are classified as default or compliant using an ex-ante criterion which allows us to identify signs of financial constraints in advance. Then, the stepwise method is employed to identify which variables are more relevant to explain the default probability. Results show that FRQ measures, namely accruals quality and timeliness, impact firms’ defaulting, supporting their relevance in predicting financial difficulties. Finally, using a logit approach, the accuracy of the model increased when FRQ variables were included. Results are confirmed using “new age” classifiers, namely the random forest methodology. This work is not only relevant to the extant financial distress literature but has also relevant implications for practice since stakeholders can understand the impact of financial reporting quality to prevent additional risks.

A Survival Model for Wilful Default Prediction – Bayesian Approach

This study develops an insolvency model to predict the possible wilful non-payment of debt obligations that turn into bad assets. This paper reveals that financially weak firms have been in deep financial distress somewhere between two to three years prior to their declaration as a wilful defaulter by the initial credit institution and its reporting on the same to the credit information companies. The Cox proportional hazards model (PHM) has been employed, which is a well-known and profusely applied approach not just in medical science but also in forecasting firm bankruptcy. This widely recognized model has been utilized to estimate the effects of different covariates influencing the times-to-event data. The application of Bayesian methods has the benefit in dealing with censored data in small sample over frequentists’ approach. Herein, Bayesian survival framework is applied incorporating normal priors that generally performs better than the traditional likelihood estimation to forecast wilful default. Subsequently, Markov Chain Monte Carlo (MCMC) sampling enables to provide the Bayesian estimator. In the Bayesian structure, the Survival model is used with the help of hazard function. The gamma distribution is selected as the prior for the standard hazard equation in PHM. In order to solve for posterior distribution, the Metropolis Hastings scheme is followed that avoids solving complicated equations with OpenBugs platform.

The role of financial constraint factors in predicting SME default

Research background: SMEs face financial constraints in their development, which limits their access to external funds, tightens their investment possibilities, and limits their growth. Much research effort has been devoted to understanding the nature and sources of this phenomenon. In sharp contrast to this, very little has been said about the role of these factors in explaining the default probability of these types of enterprises. Understanding such interrelationships could help to adopt policies to alleviate the situation of constrained SMEs and lower their default rates.
 Purpose of the article: This study analyses the role of financial constraint factors in SME defaults. This is done by utilising the financial constraint factors in a newly derived default prediction model. A comparison of the derived model and other SME default prediction models is carried out to assess the potential of financial constraints in the discrimination power of the model.
 Methods: In this study, we use the Cox semiparametric model, while leaving the baseline hazard rate unspecified and employing macroeconomic variables as explanatory variables. The discrimination power was addressed in terms of the area under the curve (AUC), resulting in out-of-sample testing. The DeLong test was used to compare the AUC of the created and analysed models. The model was estimated on a set of over 213,731 SMEs from 28 counties, covering the period 2014?2019.
 Findings & value added: It was found that adopting the financial constraint measures can explain the default of small and medium enterprises with high accuracy; however, they do not explain the default of micro enterprises.

Reliability and Accuracy of Alternative Default Prediction Models: Evidence from Slovakia

The aim of this paper is to assess the reliability of alternative default prediction models in local conditions, with subsequent comparison with other generally known and globally disseminated default prediction models, such as Altman’s Z-score, Quick Test, Creditworthiness Index, and Taffler’s Model. The comparison was carried out on a sample of 90 companies operating in the Slovak Republic over a period of 3 years (2016, 2017, and 2018) with a narrower focus on three sectors: construction, retail, and tourism, using alternative default prediction models, such as CH-index, G-index, Binkert’s Model, HGN2 Model, M-model, Gulka’s Model, Hurtošová’s Model, Model of Delina and Packová, and Binkert’s Model. To verify the reliability of these models, tests of the significance of statistical hypotheses were used, such as type I and type II error. According to research results, the highest reliability and accuracy was achieved by an alternative local Model of Delina and Packová. The least reliable results within the list of models were reported by the most globally disseminated model, Altman’s Z-score. Significant differences between sectors were identified.

A novel two-stage hybrid default prediction model with k-means clustering and support vector domain description

Default prediction identifies the probability of a firm to default by establishing a prediction model. It reveals the functional relation between the features’ data at time t-m and default status at t. If the prediction of a defaulting company is wrong, it will mislead banks into making loans to a “defaulter,” causing huge losses; if the prediction of a non-defaulting company is wrong, it will result in a potential churn in high-quality customers. To support the lending decisions of banks and non-banking financial institutions, this study proposes a two-stage default prediction model that integrates k-means clustering for partitioning the sample and support vector domain description (SVDD) for predicting default (credit scoring). It also uses attributes’ data at time t-m (m = 1, 2, 3, 4, 5) and the default status at t to train the proposed model so that it can warn of default m years ahead. The results show that the predictive accuracy of the proposed two-stage default prediction model is better than that of single-stage models using only k-means clustering or support vector domain description, and the proposed model could achieve a five-year default prediction ability (AUC > 0.85). Further, the study implies that “retained earnings/total assets”, “financial expenses/gross revenue”, and “type of audit opinion” are three key features in default forecasting for Chinese listed enterprises. This study contributes to the field of multi-stage credit scoring research by demonstrating that a combination of different methods is worth considering to improve the performance of default prediction models.

Empirics of Korean Shipping Companies’ Default Predictions

In this paper, we use a logit model to predict the probability of default for Korean shipping companies. We explore numerous financial ratios to find predictors of a shipping firm’s failure and construct four default prediction models. The results suggest that a model with industry specific indicators outperforms other models in predictive ability. This finding indicates that utilizing information about unique financial characteristics of the shipping industry may enhance the performance of default prediction models. Given the importance of the shipping industry in the Korean economy, this study can benefit both policymakers and market participants.

Corporate Default Prediction Model: Evidence from the Indian Industrial Sector

The unprecedented pandemic COVID-19 has impacted businesses across the globe. A significant jump in the credit default risk is expected. Credit default is an indicator of financial distress experienced by the business. Credit default often leads to bankruptcy filing against the defaulting company. In India, the Insolvency and Bankruptcy Code (IBC) is the law that governs insolvency and bankruptcy. As reported by the Insolvency and Bankruptcy Board of India (IBBI), the number of companies filing for bankruptcy under IBC is on a rise, and the industrial sector has witnessed the maximum number of bankruptcy filings. The present article attempts to develop a credit default prediction model for the Indian industrial sector based on a sample of 164 companies comprising an equal number of defaulting and nondefaulting companies. A total of 120 companies are used as training samples and 44 companies as the testing samples. Binary logistic regression analysis is employed to develop the model. The diagnostic ability of the model is tested using receiver operating characteristic curve, area under the curve and annual accuracy. According to the study, return on assets, current ratio, debt to total assets ratio, sales to working capital ratio and cash flow to total assets ratio is statistically significant in predicting default. The findings of the study have significant implications in lending and investment decisions.

Emerging Approach for Detection of Financial Frauds Using Machine Learning

The growth of regularly generated data from many financial activities has significant implications for every corner of financial modelling. This study has investigated the utilization of these continuous growing data by a means of an automated process. The automated process can be developed by using Machine learning based techniques that analyze the data and gain experience from the underlying data. Different important domains of financial fields such as Credit card fraud detection, bankruptcy detection, loan default prediction, investment prediction, marketing and many more can be modelled by implementing machine learning methods. Among several machine learning based techniques, the use of parametric and non-parametric based methods are approached by this research. Two parametric models namely Logistic Regression, Gaussian Naive Bayes models and two non-parametric methods such as Random Forest, Decision Tree are implemented in this paper. All the mentioned models are developed and implemented in the field of Credit card fraud detection, bankruptcy detection, loan default prediction. In each of the aforementioned cases, the comparative study among the classification techniques is drawn and the best model is identified. The performance of each classifier on each considered domain is evaluated by various performance metrics such as accuracy, F1-score and mean squared error. In the credit card fraud detection model the decision tree classifier performs the best with an accuracy of 99.1% and, in the loan default prediction and bankruptcy detection model, the random forest classifier gives the best accuracy of 97% and 96.84% respectively.

Financial crises, bank efficiency and survival: Theory, literature and emerging market evidence

The finance literature on efficiency and crisis at the macro-level and efficiency and default at the micro-level has hitherto grown surely but distinctly. In this comprehensive paper, we globally review and theoretically unify these two strands of research in studying the record level of bank failures and the deepest financial crisis of an emerging market, Turkey, with sixteen distinct measures of efficiency, stemming from two alternative methods, stochastic frontier analysis (SFA) and data envelopment analysis (DEA). The results show that efficiency scores tend to deteriorate gradually before crisis, hit bottom during crisis, and rebound after crisis. Inelastic inputs and elastic outputs seem to produce this pattern. The efficient banks have the highest survival rates. Managers of survivor banks are evidently better at controlling costs and scales, utilizing and allocating resources, generating assets, revenues and profits. Demotion to a lower efficiency class is a rare event in normal times but widespread during crises. The least efficient banks are the least likely to be acquired by private bidders. Default prediction models notably improve with DEA scores, off-balance sheet items, definition of failure with “factual insolvency”, deciles of efficiency, changes in some key variables, homogenous dataset, and efficiency scores based on quantities of inputs and outputs rather than their noisy prices.

Application of XGBoost in P2P Default Prediction

P2P network lending is a “ peer-to-peer ” loan through a third-party Internet platform built by P2P companies. The application of machine learning algorithms to the field of P2P loan default prediction will improve the operating capabilities of the platform, and also effectively regulate the lending market. In this paper, we use Lending Club’s loan data and feature engineering technology to apply the XGBoost algorithm to construct a P2P loan default prediction model, and we choose five performances: accuracy, AUC value, error rate, model robustness, and model run time to compare it with Logistic Regression and Decision Trees. The result shows that the prediction accuracy rate of the XGBoost algorithm is 97.705%, which fits the actual results better, and can effectively control the loss cost caused by model errors. In addition, we also select 10 features that have the greatest impact on loan default rates based on the XGBoost algorithm, and provide a reference for P2P lending platforms.

A novel hybrid ensemble model based on tree-based method and deep learning method for default prediction

Default prediction plays an important role in emerging financial market, so it has attracted extensive attention from financial industry and academic community. A slight improvement in default prediction performance can avoid huge economic losses. Many existing studies have used feature selection to improve the performance of default prediction models but paid limited attention to feature generation. Additionally, deep learning methods have been gradually explored for classification problems. In this study, a novel hybrid ensemble model is proposed to improve the performance of default prediction. First, a tree-based method (i.e., LightGBM) is used to learn new feature interactions and enhance the representation of original features. Second, a deep learning method (i.e., Convolutional Neural Network) is used as feature generation method to generate deeper feature interactions. Moreover, the structure of Inner Product-based Neural Network (IPNN) is used as deep learning classifier to learn feature interactions and reach a good trade-off between predictive accuracy and complexity. Third, ensemble learning method is used to combine the deep learning classifier with tree-based classifiers to obtain superior predictive results. Finally, two default datasets and four evaluation metrics are used to measure the predictive performance. The experimental results show that each component of the proposed model has significant improvement on overall performance.

A Novel Financial Risk Identification Algorithm for Online Fintech Platform

The issues we study are risk identification, credit granting, pre-loan and post-loan lending risks in consumer lending and trust loans. How to effectively evaluate and identify the potential default risk of borrowers and calculate the probability of default of borrowers before granting loans is a fundamental part of credit risk management in modern financial institutions. This paper focuses on the statistical analysis of historical loan data from banks and other financial institutions with the help of the idea of non-equilibrium data classification, and the use of the random forest algorithm to establish a loan default prediction model. Then a fixed threshold screening in the form of a black and white list is used. The limitation of the current method is that it cannot meet the lending needs of large-scale transaction classes and cannot be judged quickly. When the number of decision trees in a random forest is large, the space and time required for training are large. In addition, it is insensitive to data and has low accuracy. The experiments found that the machine learning-based financial lending risk prediction method can solve the problem better, and the experimental results phenotyped that the neural network and random forest algorithm outperformed the decision tree and logistic regression classification algorithm in prediction performance. In addition, by using the random forest algorithm to rank the importance of features, the features that have a greater impact on whether the final default can be obtained, so that the lending risk judgment in the financial sector can be more effective.

Financial Credit Default Forecast Based on Big Data Analysis

How to effectively evaluate and identify the potential default risk of borrowers and calculate the default probability of borrowers before issuing loans is the basis and important link of the credit risk management of modern financial institutions. This paper mainly studies the statistical analysis of historical loan data of banks and other financial institutions with the help of the idea of non-balanced data classification, and uses machine learning algorithms (not statistical algorithms) such as random forest, logical regression and decision tree to establish loan default prediction model. The experimental results show that neural network and random forest algorithm outperform decision tree and logistic regression classification algorithm in prediction performance. In addition, by using the random forest algorithm to rank the importance of features, the features that have a greater impact on the final default can be obtained, so as to make a more effective judgment on the loan risk in the financial field.

A deep metric learning approach for weakly supervised loan default prediction1

To deal with the systematic risk of financial institutions and the rapid increasing of loan applications, it is becoming extremely important to automatically predict the default probability of a loan. However, this task is non-trivial due to the insufficient default samples, hard decision boundaries and numerous heterogeneous features. To the best of our knowledge, existing related researches fail in handling these three difficulties simultaneously. In this paper, we propose a weakly supervised loan default prediction model WEAKLOAN that systematically solves all these challenges based on deep metric learning. WEAKLOAN is composed of three key modules which are used for encoding loan features, learning evaluation metrics and calculating default risk scores. By doing so, WEAKLOAN can not only extract the features of a loan itself, but also model the hidden relationships in loan pairs. Extensive experiments on real-life datasets show that WEAKLOAN significantly outperforms all compared baselines even though the default loans for training are limited.

Graph convolutional network-based credit default prediction utilizing three types of virtual distances among borrowers

Machine learning models have been actively utilized to quantitatively predict the default probability based on the personal information obtained from loan applicants. Although the relationship between loan applicants is receiving attention as important soft information, only the simple individual network features and their relation have been reflected in the prediction model. In this study, we propose a graph convolutional network (GCN)-based credit default prediction model, which can reflect nonlinear relationships between borrower’s attributes and default risk as well as high-order relationships between the borrowers. Three types of information pertaining to the borrowers are separately employed for their relations, namely loan information, credit history information, and soft information. We compare our GCN model to baseline models using the data from an online peer-to-peer lending platform. The results show that our approach outperforms existing classification models and identifies the relative contribution of input attributes.

Machine learning predictivity applied to consumer creditworthiness

Credit risk evaluation has a relevant role to financial institutions, since lending may result in real and immediate losses. In particular, default prediction is one of the most challenging activities for managing credit risk. This study analyzes the adequacy of borrower’s classification models using a Brazilian bank’s loan database, and exploring machine learning techniques. We develop Support Vector Machine, Decision Trees, Bagging, AdaBoost and Random Forest models, and compare their predictive accuracy with a benchmark based on a Logistic Regression model. Comparisons are analyzed based on usual classification performance metrics. Our results show that Random Forest and Adaboost perform better when compared to other models. Moreover, Support Vector Machine models show poor performance using both linear and nonlinear kernels. Our findings suggest that there are value creating opportunities for banks to improve default prediction models by exploring machine learning techniques.

Personal Credit Default Prediction Model Based on Convolution Neural Network

It has great significance for the healthy development of credit industry to control the credit default risk by using the information technology. For some traditional research about the credit default prediction model, more attention is paid to the model accuracy, while the business characteristics of the credit risk prevention are easy to be ignored. Meanwhile, to reduce the complicity of the model, the data features need be extracted manually, which will decrease the high-dimensional correlation among the analyzing data and then result in the low prediction performance of the model. So, in the paper, the CNN (convolutional neural network) is used to establish a personal credit default prediction model, and both ACC (accuracy) and AUC (the area under the ROC curve) are taken as the performance evaluation index of the model. Experimental results show the model ACC (accuracy) is above 95% and AUC (the area under the ROC curve) is above 99%, and the model performance is much better than the classical algorithm including the SVM (support vector machine), Bayes, and RF (random forest).

SME Default Prediction Framework with the Effective Use of External Public Credit Data

Traditional default prediction models mainly rely on financial data. However, financial data on small and medium-sized enterprises (SMEs) are difficult to obtain, and even when they are available, their opaqueness may hinder analysis. Therefore, traditional prediction models encounter serious problems when being utilized to predict the defaulting of SMEs. In this paper, a novel prediction framework utilizing only external public credit data is proposed. The external public credit data used include SMEs’ basic information (BI), credit information from the government (CIG), and court verdict information (CVI), which can be collected from publicly accessible websites. Records on 15,605 sample companies were collected from approximately 300,000 companies. Among them, 8183 have defaulted. The empirical data were applied to construct prediction models using logistic regression, the classification and regression tree (CART) model, and LightGBM. The best results achieved 0.87 accuracy and 0.92 area under receiver operating characteristic (AUC). The results show that the model only uses the external credit data proven to have significant predict ability, and CIG variables offer the best prediction capacities.

Corporate Default Predictions Using Machine Learning: Literature Review

Corporate default predictions play an essential role in each sector of the economy, as highlighted by the global financial crisis and the increase in credit risk. This study reviews the corporate default prediction literature from the perspectives of financial engineering and machine learning. We define three generations of statistical models: discriminant analyses, binary response models, and hazard models. In addition, we introduce three representative machine learning methodologies: support vector machines, decision trees, and artificial neural network algorithms. For both the statistical models and machine learning methodologies, we identify the key studies used in corporate default prediction. By comparing these methods with findings from the interdisciplinary literature, our review suggests some new tasks in the field of machine learning for predicting corporate defaults. First, a corporate default prediction model should be a multi-period model in which future outcomes are affected by past decisions. Second, the stock price and the corporate value determined by the stock market are important factors to use in default predictions. Finally, a corporate default prediction model should be able to suggest the cause of default.

Default Prediction Model: The Significant Role of Data Engineering in the Quality of Outcomes

For financial institutions and the banking industry, it is very crucial to have predictive models for their core financial activities, and especially those activities which play major roles in risk management. Predicting loan default is one of the critical issues that banks and financial institutions focus on, as huge revenue loss could be prevented by predicting customer’s ability not only to pay back, but also to be able to do that on time. Customer loan default prediction is a task of proactively identifying customers who are most probably to stop paying back their loans. This is usually done by dynamically analyzing customers’ relevant information and behaviors. This is significant so as the bank or the financial institution can estimate the borrowers’ risk. Many different machine learning classification models and algorithms have been used to predict customers’ ability to pay back loans. In this paper, three different classification methods (Naïve Bayes, Decision Tree, and Random Forest) are used for prediction, comprehensive different pre-processing techniques are being applied on the dataset in order to gain better data through fixing some of the main data issues like missing values and imbalanced data, and three different feature extractions algorithms are used to enhance the accuracy and the performance. Results of the competing models were varied after applying data preprocessing techniques and features selections. The results were compared using F1 accuracy measure. The best model achieved an improvement of about 40%, whilst the least performing model achieved an improvement of 3% only. This implies the significance and importance of data engineering (e.g., data preprocessing techniques and features selections) course of action in machine learning exercises