Resource Library

Author: Vasilis Aggelis, Dimitris Christodoulakis

"The introduction of data mining methods in the banking area although conducted in a slower way than in other fields, mainly due to the nature and sensitivity of bank data, can already be considered of great assistance to banks as to prediction, forecasting and decision making. One particular method is the investigation for association rules between products and services a bank offers. Results are generally impressive since in many cases strong relations are established, which are not easily observed at a first glance. These rules are used as additional tools aiming at the continuous improvement of bank services and products helping also in approaching new customers. In addition, the development and continuous training of prediction models is a very significant task, especially for bank organizations. The establishment of such models with the capacity of accurate prediction of future facts enhances the decision making and the fulfillment of the bank goals, especially in case these models are applied on specific bank units. E-banking can be considered such a unit receiving influence from a number of different sides. Scope of this paper is the demonstration of the application of data mining methods to e-banking. In other words association rules concerning e-banking are discovered using different techniques and a prediction model is established, depending on e-banking parameters like the transactions volume conducted through this alternative channel in relation with other crucial parameters like the number of active users."

Keywords: bank data, data mining, algorithms, e-banking

Author: Richard A. Derrig, Karthik Balakrishnan

Research by IBM indicates only 1% of data collected by organizations is used for analysis. Predictive Modeling and Data Mining widely embraced by leading  businesses: 2002 Strategic Decision Making survey by Hackett Best Practices determined that world class companies adopted predictive modeling technologies at twice the rate of other companies; Important commercial application is Customer retention:  5% increase in retention  95% increase in profit; It costs 5 to 10 times more to acquire new business. Another study of 24 leading companies found that they needed to go beyond simple data collection.

Keywords: data mining, predictive modeling, analysis, knowledge-discovery in databases, CRISP-DM

Author: Saksoft

The client is one of the leading banks in Asia Pacific region, and provides a wide range of financial services including personal financial services, private banking, trust services, commercial and corporate banking, corporate finance, capital market activities, treasury services, futures broking, asset management, venture capital management, general insurance and life assurance. As part of their preparation for Basel II compliance, a data mart for Behavioural Scoring and Retail Segmentation was required. This data mart would store data for multiple countries. Saksoft built the Segmentation and Behavioural Scoring data mart for Singapore, Malaysia and Thailand.

Keywords: basel II, basel II compliance, retail exposures, behavior scores

Author: Dr Paul Russell

This study highlights the news headlines regarding the UK credit market. It also covers indebtedness, industry initiatives, measuring consumer indebtedness, affordability, and consumer value management.

Keywords: indebtedness, consumer credit, credit cards, retail credit market, credit bureau

Author: Luis Rodriguez and Michelle Katics

Working with McKinsey & Company and Sociedad Hipotecaria Federal (SHF) in Mexico, Fair Isaac has integrated mortgage credit scores into an advanced portfolio valuation methodology. The resulting scoringbacked methodology is specifically designed to help lenders boost portfolio profitability by better managing their forecasting and capital reserve functions while meeting compliance requirements.

Keywords: retail portfolio, credit risk modeling, compensation, compliance, borrowers, risk segments

Author: Portfolio Defense

The purpose of this document is to provide you with key information about your Custom Empirical Credit Scoring Model. In it, we review how the models were developed, how they can best be implemented and used, certain regulatory requirements associated with the use of scoring models, as well as recommended actions your organization will likely want to take for portfolio and model tracking.

Keywords: credit scoring, Custom Credit Scoring Model, good/bad, analysis

Author: Sam Buttrey

Regarding scorecards: Interactions exist in real data, so we should include them in the model. They show how the effect of one variable on the response can depend on the value of another. Useful for segmentation, potentially for scorecard construction.

Keywords: variables, scorecards, interaction, Logistic regression, modeling

Author: Daniela Şchiopu

"In this paper we analyze information about the customers of a bank, dividing them into three clusters, using SPSS TwoStep Cluster method. This method is perfect for our case study, because, compared to other classical clustering methods, TwoStep uses mixture data (both continuous and categorical variables) and it also finds the optimal number of clusters. TwoStep creates three customers’ profiles. The largest group contains skilled customers, whose purpose of the loan is education or business. The second group consists in persons with real estate, but mostly unemployed, which asked for a credit for retraining or for household goods. The third profile gathers people with unknown properties, who make a request for a car or a television and then for education. The benefit of the study is reinforcing the company’s profits by managing its clients more effectively"

Keywords: TwoStep Cluster, clustering, pre-clustering, CF tree

Author: Tony Bellotti and Jonathan Crook

Typically models of credit card default are built on static data, often collected at time of application. We consider alternative models that also include behavioural data about credit card holders and macroeconomic conditions across the credit card lifetime, using a discrete survival analysis framework. We find that dynamic models that include these behavioural and macroeconomic variables give statistically significant improvements in model fit which translates into better forecasts of default at both account and portfolio level when applied to an out-of-sample data set. Additionally, by simulating extreme economic conditions, we show how these models can be used to stress test credit card portfolios.

Keywords: discrete survival models, stress testing, loss distributions, choleski decomposition, credit risk.

Author: James Marinopoulos

This PowerPoint presentation covers risk families and retail decision models. It also covers RDM structure and responsibilities, scoring, scorecard modeling, business objectives, world banks, monitoring, and future direction.

Keywords: risk, credit scoring, world banks, retail, FICO, behavioral scoring, application scoring

Author: Edward Huang and Christopher Scott

Credit risk scoring has gone a long way since Fair Isaac introduced the first commercial scorecard to assist banks in making their credit lending decisions over 50 years ago. It now becomes the cornerstone in modern credit risk management thanks to the advancement in computing technologies and availability of affordable computing power. Credit scoring is no longer only applied in assessing lending decisions, but also on-going credit risk management and collection strategies. Better designed, optimally developed and hence more powerful credit risk scorecard is a key for banks and retail finance companies alike to achieve competitive advantage in today’s competitive financial services market under the tough economic environment with severe consumer indebtedness. Several books have been published which serve as a good introduction to credit management and scoring.

Keywords: credit risk, credit risk scoring, scorecard, genetic algorithm, survival analysis modeling

Author: Lungi Mazibuko, Lindiwe Dlukulu, Dimakatso Qocha, Veli Mfetane, Desiree Thloaele, Shokie Bopape

The rapid growth, since 1994, of the black South African middle class termed “black diamonds”, brought many challenges including the rapid rise of indebtedness in these strata of the community. There are several views about the existence or non existence of this group even though some studies have shown that this group does indeed exist. Not only does the group exist but they have heftily contributed to the debt growth of South Africa and they continue to do so. This study attempts to answer the matter raised above, by indicating that this group does exists and to further prove that it is currently largely in debt. Amidst good and steady growth in the past years, South Africa is experiencing a slowdown in the economy indicated by high interest rates, and soaring fuel and food prices. It has been proven that many South Africans, especially, the middle class, are currently struggling to service their debt.

Keywords: South Africa, debt, middle class, reckless spending, consumers, reckless lending, Credit Bureaus, indebtedness, Credit Bureaus, fiscal policy

Author: Hian Chye Koh, Wei Chin Tan, Chwee Peng Goh

Credit scoring can be defined as a technique that helps credit providers decide whether to grant credit to consumers or customers.  Its increasing importance can be seen from the growing popularity and application of credit scoring in consumer credit.  There are advantages not only to construct effective credit scoring models to help improve the bottom-line of credit providers, but also to combine models to yield a better performing combined model.  This paper has two objectives.  First, it illustrates the use of data mining techniques to construct credit scoring models.  Second, it illustrates the combination of credit scoring models to give a superior final model.  The paper also highlights the prerequisites and limitations of the data mining approach.

Keywords: Credit scoring; data mining techniques; construction of models; combination of model; meta-modeling

Author: Linda Allen, Gayle DeLong, Anthony Saunders

We survey the most recent BIS proposals for the credit risk measurement of retail credits in capital regulations. We also describe the recent trend away from relationship lending toward transactional lending in the small business loan arena. These trends create the opportunity to adopt more analytical, data-based approaches to credit risk measurement. We survey proprietary credit scoring models (such as Fair Isaac), as well as options-theoretic structural models (such as KMV and Moody’s RiskCalc), and reduced-form models (such as Credit Risk Plus). These models allow lenders and regulators to develop techniques that rely on portfolio aggregation to measure retail credit risk exposure.

Keywords: Banks; Government policy and regulation

Author: Bryan D. Nelson

"Most direct mail and e-commerce companies have hundreds if not thousands of variables for each customer on their database. When statisticians try to build segmentation or other types of models with a large number of variables, it becomes difficult to figure out the correct relationships between the dependent and independent variables. In fact, when redundant variables are included in some of the model building procedures, the model can degrade the segmentation model by: destabilizing the parameter estimates, increasing computation time, confounding the interpretation, and increasing the amount of time spent building a segmentation model. PROC VARCLUS can help a statistician quickly reduce the number of variables used to build a segmentation model. PROC VARCLUS will cluster variables - it will find groups of variables that are as correlated as possible among themselves and as uncorrelated as possible with variables in other clusters. The algorithm used by PROC VARCLUS is binary and divisive - all variables start in one cluster. If the second eigenvalue is above the current threshold (i.e. there is more than one dominant dimension) then the cluster is split. By default, PROC VARCLUS does a nonhierarchical version where variables can be reassigned to other clusters."

Keywords: variable reduction, modeling, direct mailing, e-commerce, PROC VARCLUS, models

Author: David J. Fogarty

This paper analyzes the importance of using proper techniques for the reject inference to develop consumer credit scoring. The focus is treating reject inference as a missing data problem and using model-based imputation techniques as a way to enhance the information inferred from the rejects over that of traditional approaches when developing credit scorecards. An overview and comparison of the standard missing data approaches to reject inference are provided. Multiple imputation is also discussed as a method of reject inference which can potentially reduce some of the biases which can occur from using some of the traditional missing data techniques. A quantitative analysis is then provided to confirm the hypothesis that model-based multiple imputation is an enhancement over traditional missing data approaches to reject inference.

Keywords: consumer credit, credit scoring, missing data mechanisms, missing data

Author: Pang-Ning Tan, Michael Steinbach, Vipin Kumar

Cluster analysis divides data into groups (clusters) that are meaningful, useful, or both. If meaningful groups are the goal, then the clusters should capture the natural structure of the data. In some cases, however, cluster analysis is only a useful starting point for other purposes, such as data summarization. Whether for understanding or utility, cluster analysis has long played an important role in a wide variety of fields: psychology and other social sciences, biology, statistics, pattern recognition, information retrieval, machine learning, and data mining.

Keywords: algorithms, cluster analysis, data, utility, complete clustering, fuzzy clustering

Author: Manju Puri, Jörg Rocholl, and Sascha Steffen

This paper analyzes the importance of retail consumers’ banking relationships for loan defaults using a unique, comprehensive dataset of over one million loans. We find that loans of retail customers, who have a relationship with their savings bank prior to applying for a loan, default significantly less than customers with no prior relationship. We find relationships matter in different forms (transaction accounts, savings accounts, prior loans), in scope (credit and debit cards, credit lines), and depth (relationship length, utilization of credit line, money invested in savings account). Importantly, though, even the simplest forms of relationships such as transaction accounts (e.g., savings or checking accounts) are economically meaningful in reducing defaults, even after controlling for other borrower characteristics as well as internal and external credit scores. We are able to access data on loan applications to assess how banks screen. We find that relationships are important in screening but even after taking screening into account relationships have a first order impact in reducing borrower default. Our results suggest that relationships of all kinds have inherent private information and are valuable in screening, inmonitoring, and in reducing consumers’ incentives to default.

Keywords: banking, loans, retail customers, borrower characteristics, loan characteristics

Author: KL Wong

"Lessons learned from the recent financial crisis have been driving changes in risk management,with firms recognizing the importance of a comprehensive and consistent approach to the datathat drives their risk analytics.While risk management has traditionally been driven by analytics,data has been increasingly scrutinized,both internally within firms and externally through regulators,due to oversights that were recognized during the crisis.Adding to the pressure of meeting increasingly demanding regulatory requirements,organizationsalso have unmet needs to reduce project costs and implementation risks by centralizing data, andto eliminate operational risks associated with ad-hoc,non-auditable,manual data-related processes.From a risk management perspective, data centralization yields invaluable benefits such as ensuring consistency in analysis, meeting regulatory requirements, reducing implementation timeframes and costs for new infrastructure projects, and eliminating the operational risks of manual data processes that are ad-hoc and non-auditable. This allows users to focus more on understanding and managing risk across the enterprise, rather than on troubleshooting data issues."

Keywords: data, risk management, data management, data modeling, data security, auditing

Author: Sumit Agarwala, Souphala Chomsisengphetb, and Chunlin Liuc

"An individual’s decision to maximize his investment in social capital is determined by his socio-economic characteristics, such as homeownership and mobility (Glaeser, Laibson and Sacerdote, 2002). In this paper, we empirically assess the role of individual social capital formation characteristics on personal bankruptcy and default outcomes in the consumer credit market. After controlling for a borrower’s risk score, debt, income, wealth as well as legal and economic environments, we find that default/bankruptcy risk rises and then falls over the lifecycle, while a borrower who owns a home or is married has a lower risk of default/bankruptcy. Moreover, a borrower who migrates 325 miles from his “state of birth” is 26 percent more likely to default and 28 percent more likely to file for bankruptcy, while a borrower who continues to live in his state of birth is 14 and 10 percent less likely to default and file for bankruptcy, respectively. Finally, a borrower who moves to a rural area is 9 and 7 percent less likely to default and declare bankruptcy, respectively."

Keywords: Social Capital; Consumer Bankruptcy; Default; Credit Risk; Credit Cards; Banking.

Author: Subhash Sharma, Ajith Kumar

"Consider the following situations:  The marketing manager of a large financial institution is interested in developing portfolios of product offerings that would appeal to various market segments. The manager of a large telecommunication company is interested in developing and managing a portfolio of customers that would generate substantial profits over their lifetime. The store manager is interesting in identifying the optimal mix of products demanded by its customers. In each of the above scenarios, the manager’s objective is to group stimuli (e.g., product offerings, customers, and mutual funds) into groups such that stimuli in each group are similar, and stimuli in each group are different from stimuli in other groups. Cluster analysis is one of the techniques in the managers’ toolkit that can be used to achieve this objective. The purpose of this chapter is to present a non-technical discussion of cluster analysis."

Keywords: cluster analysis, factor analysis, marketing

Author: Tony Bellotti and Jonathan Crook

Typically models of credit card default are built on static data, often collected at time of application. We consider alternative models that also include behavioural dat about credit card holders and macroeconomic conditions across the credit car lifetime, using a discrete survival analysis framework. We find that dynamic model that include these behavioural and macroeconomic variables give statistically significant improvements in model fit which translates into better forecasts of default at both account and portfolio level when applied to an out-of-sample data set. Additionally, by simulating extreme economic conditions, we show how these models can be used to stress test credit card portfolios.

Keywords: discrete survival models, stress testing, loss distributions, cholesky decomposition, credit risk

Author: SAS Institute

Industry analysts expect the use of data mining to sustain double-digit growth into the 21st century. One recent study, for example, predicts the worldwide statistical and data mining software market to grow at a compound annual growth rate of 16.1 percent over the next five years, reaching $1.13 billion in the year 2002 (International Data Corporation 1998 #15932). Many large- to mid-sized organizations in the mainstream of business, industry, and the public sector already rely heavily on the use of data mining as a way to search for relationships that would otherwise be “hidden” in their transaction data. However, even with powerful data mining techniques, it is possible for relationships in data to remain hidden due to the presence of one or more conditions.

Keywords: data mining, aggregation, business intelligence cycle, data, SEMMA methodology, database

Author: Jeff Morrison

Understanding credit behavior (or any business) at the aggregate portfolio level is becoming increasingly important in an environment where competition abounds and management is under constant pressure to stress test baseline (business as usual) forecasts under a variety of conditions. The term stress test refers to applying different assumptions to your business portfolio-especially in terms of anticipated economic conditions- to see their impact on demand, revenue, and expenses.

Keywords: credit behavior, stress test, charge-offs, modeling, credit cards, Pillar 2, Pillar II

Author: David J. Hand

This chapter is concerned with the problem of ‘reject inference’ in the credit granting process. That is, it addresses the issue of how to take proper account of people who were previously rejected for a loan, for example when trying to construct a new scorecard.

Keywords: credit granting, rejected loan, scorecard, reject inference, credit scoring, good bad, odds

Author: Charlie Gluckman and Ah Mun Kuan

This analytical supplement accompanies the publication of the National Audit Office value for money report, Helping Over-Indebted Consumer, published on 4 February 2010 and available on our website www.nao.org.uk/publications. The supplement covers the results of an NAO survey of over-indebted consumers into consumers’ debt situation, attitudes, and behaviour. It may give useful data and insights for those who want or need to understand the nature of over-indebtedness, in particular in designing services around user needs.

Keywords: over-indebted consumers, over-indebtedness, data collection, clusters