Blog article
See all stories »

A BOON but not a Baboon in Credit scoring landscape

The credit scoring and underwriting landscape is undergoing a paradigm shift with the use of enterprise and social data. The banks are planning on utilising the current organizational capabilities including the infrastructure and skills to offset the credit risk associated with borrowers. Scoring models can now use customer consent data from credit bureaus, social media, other lenders, enterprise customer data, publicly available data, location data, mobile data, social footprint, page views, web and behavioral data. These fragments of data together will provide a complete, coherent and a cohesive model that can score credit applications. This has led to the creation of new business models and in turn is leading to economic dependence on the data analysis.

These advancements are now being increasingly used by payday and student loan lenders. Loan applications are being evaluated real time within fraction of seconds, to provide a decision. At the same time, the credit risk posed by the customers to the banks can be mitigated.

The retail adoption of Big Data for efficiency, customized user experience, personalization, marketing and sales is on a high curve when compared to the financial industry. Banks are planning to catch up on this trend to provide a similar experience to the customers. Scoring efficiency coupled with straight through processing and mobility has made it much easier for borrowers to acquire credit products with comfort, similar to an experience in retail transaction.

When it comes to availability of data, there are around 10,000 to 15,000 variables that can be used to provide an input to the scoring models. The banks do have enormous amounts of structured and un-structured data that they are planning to use in underwriting.

Big Data scoring is changing the underwriting landscape as -

  • Customers who do not have a credit score available with the bureaus can avail credit products.
  • Geographies which do not have access to private bureaus can now underwrite loans with reduced credit risk.
  • Efficiency of scoring and underwriting is increasing multifold.
  • With mobility and apps, the customer experience is improved with less TAT.
  • According to a recent study, the bank reduces the risk of default by 10% - 20%.

With every application decision, the model can be fine-tuned to achieve further efficiency. One can find co-relation between thousands of data elements with most of them making sense to analysts while some still remain a mystery. It is amusing to note that a Big Data scoring service provider mentioned that the probability of default in customers with a middle name is much lesser to the customers that do not have one. On a funny note, I don’t think that banks will stop lending to customers without a middle name or the new borns will be provided middle names to assist them in scenarios similar to this in future. Banks usually use Fico score with custom scores to decision an application. It is time for banks to look at the capabilities that a Big Data scoring solution will provide in scoring credit applications.

If banks are looking to foray into emerging markets, they will face a challenge of unavailability of credit bureau data. A recent survey states that 72% of the world’s population is not covered by credit bureaus. While some banks are leveraging the advancements in credit scoring, there are others who are yet to kick start acquiring these capabilities. My suggestion would be to improve the efficiency of their existing models, use feedback mechanism and thereby increase the business benefits arising from such solutions.

Big Data perspective to Credit scoring and Underwriting
4257

Comments: (2)

Ketharaman Swaminathan
Ketharaman Swaminathan - GTM360 Marketing Solutions - Pune 30 September, 2014, 13:43Be the first to give this comment the thumbs up 0 likes

P2P lending portals that are known for using social media and other signals mentioned by you to supplement / replace traditional credit scores still seem to be struggling to attract custom, going by recent reports that they're seeking regulation to even source customers whose loan applications are rejected by banks (https://www.finextra.com/blogs/fullblog.aspx?blogid=9939). Any idea why there's such a big difference between theory and reality?

People with middle names have been believed to be more creditworthy for a long time. The unnamed Big Data scoring service provider referenced by you is recycling very old news.  

The possibility you've surmised on a "funny note" may not be so far-fetched after all: In the early 2000s, a leading American consulting company mandated that all its employees worldwide must have a middle name or at least a middle initial. In countries like UK where middle names are not a common practice, an "X" was inserted as the dummy middle initial! 

Tejasvi Addagada
Tejasvi Addagada - Fortune 500 financial service provider - Mumabi 30 September, 2014, 17:27Be the first to give this comment the thumbs up 0 likes

There is a regulatory need behind the context in my blog, for which I would suggest reference to the mortgage regulatory landscape - The way forward. CFPB in US stands strong, emphasizing banks on changing their credit and lending policies. Banks do have regulatory preparedness in their roadmap, as they are expecting further changes to credit policies depending on the market conditions.

The context in my blog refers to the use of Enterprise and Social data to complement the FICO and the custom scoring models that banks currently use to score applications. A solution/features in these lines would cater to complement the existing model rather than replacing it.

Banks have enormous enterprise data which P2P lenders are in dearth. These data stores serve as a golden source for insight generation. Models that utilize enterprise data along with Social data tend to be far more improvised than traditional models and models that bank only on social data. I was referring to pay day lenders which differ from the P2P lenders in the business model that they operate. But, what remains same are the technology advancements providing enriched, gamified customer experience.

It is a good insight that you brought forth from the blog you have mentioned. Tim brings an impression of banks having a software alliance with reputable P2P lenders that would be advantageous to banks. This would be helpful in entering emerging markets, gaining market share and reducing costs. I don’t consider all payday lenders to be struggling; I suggest reference to some industry leaders which boast a default rate of less than 10% when compared to 30% in the markets they are operating. Further, they are doing well on revenue generation and net profits.

As I write, Credit models of banks are undergoing changes based on their current and regulatory needs. If you have a requirement that would entail a credit scoring model to be improvised, I would be pleased to guide you with industry leading solutions that we offer.

I am still surprised at the fact that a consulting firm mandated its employees to have a middle name.