| Sharing credit data – the benefits 2 |
 |
 |
 |
| Tuesday, 11 May 2010 02:00 |
|
In this article PIC Solutions continues a discussion on the benefits of financial services organisations sharing credit data. Credit Data Sharing – The Argument Against Historically, several common objections tend to be raised when the concept of credit data sharing is suggested. The more frequent arguments against sharing data include: • Organisations will lose competitive advantage by revealing their customer details or business specific information. In particular, other organizations may ‘steal’ shared customer information and market to these individuals. • Concerns around client confidentiality and data security. If data integrity is compromised, organisations may face liability and public embarrassment. • Concerns around consumer consent issues and consideration of industry or government regulations pertaining to data sharing. • The organisation does not have sufficient IT or business resources, the required data or system changes take a lower priority. • Resistance to change. There is no need to alter the way we do business. Credit Data Sharing – The Argument For There are numerous reasons why industry organisations contemplate, and eventually start sharing credit data and other data sources. The practise of data sharing has now extended across many different industry sectors, and involvement takes many forms. This is not to say these sectors ignored the legitimate concerns or objections raised with regard to data sharing. Indeed, issues such as the security and confidentiality of data must be addressed and accepted as a necessary part of any sharing initiative. What differentiates these industries, however, is that they have seen the opportunities that data sharing presents, rather than just focusing on the perceived negative aspects. The sharing of information generally has become a modern business reality. With information technology and communications infrastructure gradually converging, deliberate as well as informal sharing arrangements become inevitable. This dynamic has also affected the area of risk management with the scope and definition of what risk management entails now including concepts such as enterprise-wide and reputational risk management. Although these tasks are not historically the domain of credit risk professionals, in increasingly complex and challenging business environments, risk managers may be delegated to ensure effective data handling across business areas. As such, the attitude of many modern companies is that risk management is a holistic pursuit, seeking to use all available information resources to hand. What is also likely is that where one organisation fails to use information opportunities, it is a near certainty competitor organisations will be finding ways to do so. In unknown and emerging business landscapes, information creates illumination for the companies that learn to use this resource effectively. Credit Data Sharing – Reasons To Join Where credit bureaux are building new statistical tools or looking to validate existing statistical models, data is often sourced from external, participating companies. In theory, if the organisation has assisted in model validation or the development process, the eventual product or tool should be more predictive and robust in that organisation’s environment since their data was used in initial testing and creation. In many instances, the benefits of credit data and information sharing initiatives are only accessible to contributing parties. Often, this closed user group status becomes a major inducement to join. For example, the full range of products and services offered by the major South African credit bureaux are only accessible to those organisations regularly contributing customer payment profiles to the bureaux. Credit Bureau Scoring Credit bureau scores (CBS) are only available to clients who share credit data with the bureaux. Supply of information from participating companies to the credit bureaux databases is critical since the CBS is, to a significant extent, based on external data sources. It is this diversity in data sources that gives the CBS one of its key strengths - the ability to assess consumers across their entire credit behaviour with different institutions. Not being able to access the CBS means credit grantors cannot use the more sophisticated tools increasingly accepted as standard practice. For example, in the South African retail industry, the use of quarterly batch runs for account and portfolio management is gaining acceptance as ‘best practice’. In this account management scenario, the credit portfolio is scored four times a year with each batch run providing a dynamic re-classification on each account. Based on this, the risk management team then pursues various strategies in credit limit management, authorisations, risk-based pricing or prioritisation of accounts requiring collections actions. As an illustration, if an account shows a consistent decline in the CBS score between batch runs, and has recently moved into an arrears status, the organisation might accelerate collections actions on the assumption that the individual represents an increasingly poor risk. In credit risk management ‘best practice’, the CBS is typically used to further refine credit strategies, impacting the revenues and costs associated with each strategy. This fine-tuning of credit risk decisioning makes practical business sense and adds value, either if the CBS is used as a stand-alone application, or used in combination with an application score or customised scoring system. Ideally, the CBS and the application score are combined, creating what is commonly known as a joint odds matrix. This matrix ensures a better overall assessment of the risk - or odds - associated with each applicant because the CBS takes into account the ‘bigger picture’ for the applicant and looks at how accounts are conducted in the broader environment. Additionally, customized application scoring tends to have inherent deficiencies as missing, exaggerated or incorrect information on the application form impacts adversely on the score derived. Using credit bureau scores ensures segmentation of accounts is done more profitably and that accounts otherwise failing the applications stage are identified and properly processed. Use of the joint odds matrix provides participating organisations with additional ‘lift’ in their risk management processes. For example, applicants who score lower on a customised scorecard may have established, satisfying relationships with other credit granters. Without the benefit of the credit bureau score, an organisation using only customised scoring as a guideline would normally reject this application. Alternatively, based on the information provided to the organisation, applicants may do well on customised scoring models but have a poor credit bureau score. This insight tends to suggest that existing accounts held by this applicant have not been well conducted, warning the credit grantor in question against automatically accepting the account. By splitting applicants according to score range matrix, a more accurate risk profile emerges. Extending this logic, applicants with low application scores, previously referred to an underwriter for a manual decision, are now automatically accepted if they have a high CBS score. Conversely, those with high application scores, previously automatically accepted, would now be referred for an underwriting decision if they obtained a low CBS score. In this matrix, certain groups can now be automatically declined (very low and low CBS and application scores), and others automatically accepted (very high CBS and application scores). This process of ‘swap-setting’ and grading applicants ensures better risk assessment and the more effective use of underwriting resources. As a result, underwriters have lower volumes to assess and are able to focus additional attention on the ‘marginal’ account applications. This refinement should lead to more consistent, objective and efficient credit risk decisions and as well as improving customer service levels. The credit bureaux system is essentially an arrangement through which participating organisations can alert one another to poorer risk prospects, averting what could turn out to be potential losses. At the most basic, this is the very reason why credit bureaux came into existence. Obviously, the organisations that have actually booked poor accounts do not benefit from the hindsight other lenders obtain but, overall, it could be argued that the incremental gains created by improved credit risk decisions offset potential future losses. Payment History and the Payment Hierarchy As part of the consumer profile, credit bureaux usually have a record of how the individual has conducted accounts historically. Participating organizations can refer to this payment history during credit decisions. Particularly for banks, a strong motivation to provide data (and therefore view payment history) is the payment hierarchy. Previous experience has shown consumers have clear payment tendencies when meeting outstanding obligations. Typically, payment is prioritised as follows: • Utilities and household commitments – mortgage repayments, electricity, water • Personal loans, car loans and prime overdraft bank commitments • Bank card or personal expense cards • Private label cards - retail accounts • Mail order accounts Consumers in situations of financial stress, and depending on the nature of their credit obligations, usually service accounts higher in the payment hierarchy before settling those lower down. Trends in payment history would reveal this phenomenon, providing a rudimentary early warning system. For example, where a customer begins regularly defaulting on retail accounts it would be reasonable to assume their prime banking commitments should follow in due course. With this insight, banking institutions could proactively manage the individual, for example, ensuring early settlement on accounts. Shared Credit Information – Additional Uses Organisations not sharing credit data often cite concerns that other participants in the proposed closed user group will ‘steal’ customers. It is important to bear in mind that credit bureaux are normally strictly governed by industry and association bodies in the way they may use contributed data. In South Africa, credit bureaux are expressly forbidden from generating or selling marketing lists drawn up from contributed payment data. This prevents organisations from using competitor data to solicit new business. Perhaps more importantly, the insight gained from sharing information tends to outweigh the information conceded. Applying the same logic behind a joint odds matrix approach; organisations benefit from seeing the ‘bigger picture’ that access to shared credit data creates. In turn, allowing better assessment as to the true risk each account represents. Based on this information, credit grantors may even ‘choose to lose’ certain poor risk accounts that are likely to be less profitable in the longer term. When it comes to pre-approved offers or account solicitation, participating members can make use of credit bureaux information to separate out prospects. By obtaining a CBS prior to undertaking any marketing activity, the business is able to assess risk up front without incurring marketing, mailing or solicitation expenses first. Only those prospects representing an acceptable level of risk are solicited or sent pre-approved offers. Traditional marketing research information tends to rely on census-based or self-reported aggregated data whereas credit information is instead based on current, individual level credit data, which tends offer a far more accurate and realistic assessment of the prospect’s financial circumstances. The same pre-screening principle can also be applied where businesses undertake cross-sell or up sell campaigns with existing customers. In this scenario, a batch run is conducted on the extracted customer base and prospects with acceptable credit bureau scores are selected for marketing campaigns. Alternatively, data contributors can use the credit bureau score to assess portfolio quality, either account portfolios already owned or those under purchase consideration. As a starting point in the assessment process, an analysis of credit bureau score ranges is undertaken to determine the extent of delinquency on the portfolio. CBS can also assist when credit granting institutions enter into a data swap or marketing list agreement. By obtaining a credit bureau scoring prior to the exchange, each organisation can decide which names it would prefer not to purchase, thereby saving the cost of purchase, as well any marketing and mailing costs otherwise incurred. In developing economies, establishing the risk associated with emerging market sectors or consumers who have no previous credit performance is a challenge. For example, credit risk assessment becomes problematic with applicants who have no fixed or formal address. Without the benefit of credit bureaux databases or credit bureaux scoring tools, organisations remain uninformed as to the applicant’s true risk. An Example – The Insurance Industry In recent years the South African insurance industry, previously slow to contribute and share data, has undergone a paradigm shift. The most recent development was the formation of the Insurance Data System (IDS) in May 2001. The IDS is a database of short term insurance claims, partly created to address the crippling levels of fraud and claims duplication affecting the local insurance industry. It has also been proved that credit data, used in conjunction with statistical scoring, is actually highly predictive of insurance risk and adds value in the areas of insurance marketing, underwriting and claims processing. For example, one South African credit bureau score has been shown to be highly predictive of insurance lapse ratios and can assist in forecasting the likelihood a policy will be cancelled or surrendered within a specified time period. This is despite the fact that the particular credit bureaux score was not originally designed or intended for this purpose. When applied to the claims processing environment, the credit bureau score as an indicator of risk also has one immediate advantage: claims that are classified as ‘good’ risk are streamlined for settlement whilst suspicious or ‘poor’ risk claims are routed for further investigation. Summary There are many arguments to be put forward for the sharing of credit data and other information sources. Benefits range from being able to benchmark credit portfolios relative to the industry sector, to confronting the problem of fraud and using credit bureaux products and tools to refine the credit risk decision process. Concerns about the sharing data are well founded but not entirely insurmountable. With regard to objections that there are insufficient IT resources to make the required data or systems changes, it should be remembered that in many cases data submission is done in a standardised and fairly straightforward format. Once the initial programmes and processes have been established, regular running and extraction of data should require only nominal maintenance. Outsourcing of this function to specialist IT companies is also an option. Regarding issues of consent and data security, most data sharing groups have industry representation or regulatory bodies. Not only does this ensure member compliance to wider legislative or legal developments, it also gives members a powerful lobby group to represent industry interests and contribute towards any public debates on proposed legislation or developments, for example, stricter regulation of the industry sector. Ultimately, as the amount of information contributed to independent credit bureaux database increases, so new elements can be incorporated into redeveloped scoring methods and tools. This deepening of the information available enhances predictive capabilities and further refines the process of making good credit risk decisions. About the Contributor PIC Solutions is the leading specialist credit risk solutions company in the EMEA region. With offices in Cape Town, Dubai, Johannesburg, Manama and Nairobi, we deliver integrated analytics, consulting and software solutions to over 150 companies in 30+ countries. We work worldwide with organizations to improve performance, drive strategies and enhance profitability. Analytics | Consulting | Software |
| Last Updated on Tuesday, 11 May 2010 15:23 |