Default Probability Prediction Research Articles

Depicting Risk Profile over Time: A Novel Multiperiod Loan Default Prediction Approach

With the rapid development of fintech, the need for dynamic credit risk evaluation is becoming increasingly important. While previous studies on credit scoring have mostly focused on single-period loan default prediction, we call for a new avenue—multiperiod default prediction (MPDP)—to depict risk profiles over time. To address the challenges raised by MPDP, such as monotonic default probability prediction and complex relationship accommodation, we propose a novel approach, hybrid and collective scoring (HACS). We design a hybrid modeling strategy to predict whether and when a borrower will default separately through a default discrimination model and a default time estimation model, respectively, and synthesize them through a probabilistic framework. To accommodate various possible patterns of default time and measure the distribution of default probability over successive time intervals, we propose a joint default modeling method to train the default time estimation model. Empirical evaluations at the model (time-to-default prediction performance and discrimination performance) and mechanism (identifiability and discriminability) levels, as well as impact analyses at the application (granting performance and profitability performance) level, show that HACS outperforms the benchmarked survival analysis and multilabel learning methods on all fronts. It can more accurately predict time-to-default and provide financial institutions and investors better decision-support in granting loans and selecting loan portfolios.

An Overview on the Landscape of R Packages for Open Source Scorecard Modelling

The credit scoring industry has a long tradition of using statistical models for loan default probability prediction. Since this time methodology has strongly evolved, and most of the current research is dedicated to modern machine learning algorithms which contrasts with common practice in the finance industry where traditional regression models still denote the gold standard. In addition, strong emphasis is put on a preliminary binning of variables. Reasons for this may be not only the regulatory requirement of model comprehensiveness but also the possibility to integrate analysts’ expert knowledge in the modelling process. Although several commercial software companies offer specific solutions for modelling credit scorecards, open-source frameworks for this purpose have been missing for a long time. In recent years, this has changed, and today several R packages for credit scorecard modelling are available. This brings the potential to bridge the gap between academic research and industrial practice. The aim of this paper is to give a structured overview of these packages. It may guide users to select the appropriate functions for the desired purpose. Furthermore, this paper will hopefully contribute to future development activities.

Linear discriminant analysis and logistic regression for default probability prediction: the case of an Italian local bank

Linear discriminant analysis and logistic regression for default probability prediction: the case of an Italian local bank

Linear discriminant analysis and logistic regression for default probability prediction: The case of an Italian local bank

Linear discriminant analysis and logistic regression for default probability prediction: The case of an Italian local bank

Comparison of econometric and deep learning approaches for credit default classification

AbstractThe deep‐learning model outperforms the conventional structured models developed by using econometric techniques. Instead, econometric techniques provide an important insight into specific factors and their contribution to default probability. Using data from an Armenian universal credit organization that contains financial and nonfinancial variables of more than 9,000 borrowers of agriculture loans from 2012 to 2017 years, we compare deep neural networks' performance against conventional and widely used econometric techniques. Delays on past loans, together with loan size and currency, are major factors contributing to loan default probability prediction accuracy. The set of statistically significant or important variables differs between econometric and deep learning models proving that the latter can capture nonlinear relationships.

Impawn rate optimisation in inventory financing: A canonical vine copula-based approach

In the inventory financing business, an optimal impawn rate (loan-to-value ratio) can help the inventory financing providers (IFPs, she) maintain competitiveness in the inventory financing market. However, the literature has been silent on how IFPs can optimise the business through the optimisation of the impawn rate. This study examines the role of the optimal impawn rate in the inventory financing business. The key to setting the optimal impawn rate is first evaluating default probability and then incorporating this into the profit function. We use a data-driven approach to explore the copula model in setting the optimal impawn rate. Through numerical analysis, we find that the Clayton canonical vine copula has a better performance for the prediction of default probability than the multivariate normal distribution (MVN) and can thus be used to evaluate default probability. In addition, we uncover that setting multiple impawn rates for different collaterals allows inventory financing to yield a higher profit. Further, although the interest rate, industrial impawn rate, and optimal impawn rate have strong effects on inventory financing profit, interestingly, the relationship between them is marginally diminishing.

Prediction of Default Probability of Credit-Card Bills

The credit-card industry has existed for decades and is a product both of changing consumer habits and improved national incomes. Both the number of card issuers and issuing banks, and transaction volumes themselves, have increased significantly. Nonetheless, with the increase of credit-card transactions, overdue amounts and delinquency rates of credit-card loans have also become problems that cannot be ignored. The successful resolution of this issue is central to the successful future development of the industry. In this work, we have presented a credit-score model to reflect the attributes and credit ratings of clients. We merge 23 variables from the original dataset and 25 additional financial features that are mined from the original financial variables, then apply to the XGBoost model. The model itself provides the 13 most significant variables by listing them according to the calculated scores. It then predicts the probability of individuals’ willingness to pay back a credit-card loan. At last, the default ratio will be converted to a credit-score system to understand the credit ratings of clients more intuitively. This model can make contributions to the resolution of default probability and is very helpful to the credit-card industry.

Prediction of Default Probability by using Statistical Models for rare Events

Summary Prediction models in credit scoring usually involve the use of data sets with highly imbalanced distributions of the event of interest (default). Logistic regression, which is widely used to estimate the probability of default, PD, often suffers from the problem of separation when the event of interest is rare and consequently poor predictive performance of the minority class in small samples. A common solution is to discard majority class examples, to duplicate minority class examples or to use a combination of both to balance the data. These methods may overfit data. It is unclear how penalized regression models such as Firth’s estimator, which reduces bias and mean-square error relative to classical logistic regression, performs in modelling PD. We review some methods for class imbalanced data and compare them in a simulation study using the Taiwan credit card data. We emphasize the effect of events per variable for developing an accurate model—an often neglected concept in PD-modelling. The data balancing techniques that are considered are the random oversampling examples and synthetic minority oversampling technique methods. The results indicate that the synthetic minority oversampling technique improved predictive accuracy of PD regardless of sample size. Among the penalized regression models that are analysed, the log-F prior and ridge regression methods are preferred.

DMDP: A Dynamic Multi-source Default Probability Prediction Framework

In this paper, we propose a dynamic forecasting framework, named DMDP (dynamic multi-source default probability prediction), to predict the default probability of a company. The default probability is a very important factor to assess the credit risk of listed companies on a stock market. Aiming at aiding financial institutions in decision making, our DMDP framework not only analyzes financial data to capture the historical performance of a company, but also utilizes long short-term memory model to dynamically incorporate daily news from social media to take the perceptions of market participants and public opinions into consideration. The study of this paper makes two key contributions. First, we make use of unstructured news crawled from social media to alleviate the impact of financial fraud issue made on default probability prediction. Second, we propose a neural network method to integrate both structured financial factors and unstructured social media data with appropriate time alignment for default probability prediction. Extensive experimental results demonstrate the effectiveness of DMDP in predicting default probability for the listed companies in mainland China, compared with various baselines.

Default Probability Prediction with Static Merton-D-Vine Copula Model

We apply standard Merton and enhanced Merton-D-Vine copula model for the measurement of credit risk on the basis of accounting and stock market data for 4 companies from Prague Stock Exchange, in the midterm horizon of 4 years. Basic Merton structural credit model is based on assumption that firm equity is European option on company assets. Consequently enhanced Merton model take in account market data, dependence between daily returns and its volatility and helps to evaluate and project the credit quality of selected companies, i.e. correlation between assets trajectories through copulas. From our and previous results it is obvious that basic Merton model significantly underestimates actual level, i.e. offers low probabilities of default. Enhanced model support us with higher simulated probability rates which mean that capturing of market risk and transferring it to credit risk estimates is probably a good way or basic step in enhancing Merton methodology.

Impact of borrower's attributes on mortgage default: evidence from Nigerian lending market

The need for proper identification of mortgage default factors has become a major focus of mortgage researches given the debilitating effect of default on mortgage market and real estate finance in particular. This paper therefore analyses the socio-economic attributes of borrowers as default triggers in residential mortgages of Primary Mortgage Institutions (PMIs) in Nigeria. Relevant data were collected on profiles of 305 borrowers randomly drawn from the credit databank of 36 PMIs. Using logistic regression (LR), payment-to-income ratio, type and sex of borrowers are found as significant mortgage default factors. With 68.2% overall prediction accuracy, LR is found appropriate for mortgage default prediction. However, the findings of this study also signal the complexity that is inherent employing socio-economic factors for default probability prediction.

Default Probability Prediction with Static Merton-D-Vine Copula Model

Default Probability Prediction with Static Merton-D-Vine Copula Model

Prediction of default probability in banking industry using CAMELS index: A case study of Iranian banks

Prediction of default probability in banking industry using CAMELS index: A case study of Iranian banks

A Class of Discrete Transformation Survival Models With Application to Default Probability Prediction

Corporate bankruptcy prediction plays a central role in academic finance research, business practice, and government regulation. Consequently, accurate default probability prediction is extremely important. We propose to apply a discrete transformation family of survival models to corporate default risk predictions. A class of Box-Cox transformations and logarithmic transformations is naturally adopted. The proposed transformation model family is shown to include the popular Shumway model and the grouped relative risk model. We show that a transformation parameter different from those two models is needed for default prediction using a bankruptcy dataset. In addition, we show using out-of-sample validation statistics that our model improves performance. We use the estimated default probability to examine a popular asset pricing question and determine whether default risk has carried a premium. Due to some distinct features of the bankruptcy application, the proposed class of discrete transformation survival models with time-varying covariates is different from the continuous survival models in the survival analysis literature. Their similarities and differences are discussed.

Is Jump Risk in iTraxx Sector Indices Diversifiable?

Previous research on credit risk mainly focuses on the prediction of default probability and most of these are based upon bond market analyses. The rapid development of the credit derivatives market make research on credit risk using information from this market more important and attractive. Also, over the last few years, analysis of dynamic behavior of credit risk has become important among academics, practitioners, and regulators. This article focuses on the investigation of nondiversifiable jump risk in iTraxx sector indices based on a multivariate framework that explicitly admits discrete common jumps for an index and its components. Our empirical research shows that both the sector iTraxx indices and the iTraxx Non-Financials, which acts as an index for the sector, experience jumps during sample period. This means that the jump risks in the iTraxx sector index are not diversifiable. Such a result warrants further investigation on credit risk, for example, in pricing and prediction. <b>TOPICS:</b>Credit risk management, credit default swaps, statistical methods, factor-based models

Assessing credit quality from the equity market: can a structural approach forecast credit ratings?

AbstractWe investigate the empirical performance of default probability prediction based on Merton's (1974) structural credit risk model. More specifically, we study if distance‐to‐default is a sufficient statistic for the equity market information concerning the credit quality of the debt‐issuing firm. We show that a simple reduced form model outperforms the Merton (1974) model for both in‐sample fitting and out‐of‐sample predictability for credit ratings, and that both can be greatly improved by including the firm's equity value as an additional variable. Moreover, the empirical performance of this hybrid model is very similar to that of the simple reduced form model. As a result, we conclude that distant‐to‐default alone does not adequately capture the firm's credit quality information from the equity market. Copyright © 2007 ASAC. Published by John Wiley &amp; Sons, Ltd.