Banks are faced with intense competition and ever-increasing pressures to seek options to generate revenue. From an asset generation stance, banks are searching for effective ways to attract creditworthy customers through improved business processes and effective use of technology. During the origination stage of the lending lifecycle, banks use the decisioning process to identify sound customers. However, over the course of time the same set of customers who were once creditworthy can potentially turn untrustworthy because of various factors. A successful underwriter should be enabled with powerful decisioning tools that assist in identifying customers that have a high chance of default.
The market has a plethora of credit decisioning solutions provided by 3rd parties, and they offer services such as derived credit scores and critical KYC checks for a fee. However, due to technology advancements, modeling, and data mining techniques, banks can now be self-empowered with newer ways of engaging with their customers to meet their decisioning objectives.
Dependency on credit bureaus
Banks typically source credit scores from credit bureaus as a paid service. To arrive at a customer credit score, banks supply a plethora of customer data such as customer income, expenditure, demographics, details of previous loans and their credit policies to the bureaus among other parameters. Bureaus then run this data through their proprietary data modeling tools and scorecard development functions to arrive at the customer credit score.
So, any bank that talks to multiple bureaus (for credit score, checking for politically exposed persons (PEPs) and local level checks) will need to either collate responses manually or build multiple integrations. Also, banks now have the challenge to supply underwriting solutions that can not only satisfactorily assess creditworthiness, but also keep the per-unit processing cost low while also reducing turnaround times for customers. Thus, a combination of factors such as business priorities, sharing of bank IP, expensive integrations, and viable technology options have caused banks to look inwards or to their solution providers to build solutions that will support decisioning capabilities. Typically, these decisioning capabilities will cover aspects such as arriving at probability of customer default, calculating optimal lendable value, individual risk ratings among others.
New opportunities for reducing cost and building competency
In a competitive environment, financial institutions must retain their existing customers while their underwriting teams are expected to build a portfolio of low risk customers. Banks must not only identify risky customers, but also identify ways to effectively minimize loss while recouping receivables.
A combination of internally developed risk scorecards and dynamic lending models offers a viable, powerful, and cost effective option that meets business needs. These have been used to predict delinquency, bankruptcy, and receivable collections without banks having to reach out to credit bureaus for a score, and at the same time retaining their IP (credit policies, models, and calculations) on their platform. For banks that have made investments in teams that make use of AI/data analytics capabilities and in-house development, it provides significant returns. As an example, banks can leverage their internal capabilities to better understand the risk behavior of customers and thereby develop efficient strategies to mitigate risk while allowing teams to experiment with a variety of factors such as segmentation and probability of default derivations.
From a technology standpoint and with focus on improved productivity, flexibility and maintainability, banks can consider building their platform to be modular and scalable and may want to extend it to country or region-specific lending models or import any existing or generic lending models to avoid cost in recreating new models. Industry standard Credit lending processes can be built as standalone components and the readymade lending models can support calculations that determine how much to lend, whom to lend along with other factors such as market risk, rate of return and loan volatility.
Banks know their business and customers best and therefore leaving key modeling and sampling decisions to external agencies is a suboptimal and costly route in the effort to create an effective decisioning solution. Experience seems to suggest that in-house development of decisioning tools and frameworks can be done faster, cheaper and with far more flexibility. Furthermore, a flexible and modular solution can empower Banks to implement new strategies with low time to market whilst giving them an opportunity to revisit and enhance policy rules. Automation can not only assess creditworthiness satisfactorily, but also ensure low processing costs, minimal instances of credit denial to worthy customers and ensure that potentially delinquent customers are kept out.