The global COVID-19 pandemic touched off economic effects that essentially ended the previous credit cycle in most markets. As these markets slowly resume normal activity, a new credit cycle will begin, offering innovative lenders a rare opportunity to expand into credit markets and win market share. The resumption of the cycle also offers a window for new entrants such as utilities, insurance companies, and other nontraditional lenders to join the market.
Although banks provide financing solutions to a significant share of the global population, large segments of consumers are underserved or not served at all. New-to-market lenders can identify the gaps in lending coverage and try to bridge them. Many potential customers would like innovative, tailored solutions that are not always cost-efficient for traditional banks. New entrants can design new offerings quickly and are unencumbered by legacy processes or infrastructure. They can move from concept to fully developed offering in two or three months, compared with one to two years for incumbents.
Unlike incumbents, these new-to-market lenders may not yet have consumer lending operations and may not be serving consumers with credit history. They likely lack the appropriate lending infrastructure, credit-risk models, and reference data. While they develop these capabilities, they will need to take a structured approach to manage the risk of this business.
New-to-market lenders could include traditional banks expanding market share as well as nonbank financial institutions. These lenders will need to actively manage credit-risk decisions and also the enabling technology. By doing the advance work required to establish a credit-decision platform, lenders can move quickly while still taking the right level of credit risk. To that end, new-to-market lenders could follow a four-part framework (Exhibit 1).
1. Utilize data from a wide range of sources
To model credit risk, new-to-market lenders will need to aggregate data from a broad range of sources. They can make up for any lack of credit expertise by capturing diverse data, including data that they own exclusively. Some traditional categories of credit behavior and demographic data are widely available, particularly for established financial institutions. These include loan information from lenders, deposit data with banks, other current-account information, and point-of-sale transaction data. Nonfinancial companies have other internal sources of customer data, such as product usage, interactions with customer-relationship management, call records, email records, customer feedback, and website navigational data.
Respecting all applicable privacy regulations and guidelines, lenders can seek to employ data from further sources. These include external data from sources such as retailers, telecommunications companies, utility providers, other banks, and government agencies. For certain types of lenders, acquiring needed data through partnerships may be an avenue worth exploring. This strategy—a joint venture with companies that have complementary data about consumer segments—may be particularly suited to lenders with a regional presence.
An approach taken by one telecommunications company is instructive. The company launched an unsecured cash-loan product to serve customers lacking access to formal credit. The challenge was that the company had little credit information available to develop the offering. In response, the company turned to its customer-usage data—specifically, data on mobile bill payments. The data enabled the company to devise a proxy target variable that it could use to train its credit model. When back-tested for model development, the target variable performed in the same way as typical credit-related information would perform for banks. From that point, the company was able to extend credit to prepaid customers via a pilot model, which it then refined based on real-world information.
2. Build the decision engine
The second major step is to build the decision engine. In this area, new entrants will have a large advantage over existing lenders with legacy software that they do not want to alter. The new decision engine can largely be built using advanced analytics, machine learning, and other tools that capitalize on speed and agility.
By using machine learning, the new-entrant lenders will be able to automate as much as 95 percent of underwriting processes while also making more accurate credit decisions. Similarly, real-time machine-learning solutions can improve pricing and limit setting and help firms monitor existing customers and credit lines through smarter early-warning systems. Lenders can also use straight-through processing to generate faster transactions and a better customer experience.
The design of the decision engine can be modular for maximum flexibility. That will allow lenders to retain control of strategic processes while potentially outsourcing other parts. The modular format can also facilitate risk assessment. This approach involves a series of steps, completely integrated from the front end to the back end, and is designed for objective and quick decision making (Exhibit 2).
This approach to risk assessment contrasts markedly with the risk engine in place at many large organizations. The traditional setup is often a single, massive system incorporating every aspect of the lending process, from assessing creditworthiness to printing documents. That approach is increasingly outdated, as it constrains incumbent lenders from adapting quickly.
Based on our experience, applying agile development and implementation can reduce the launch time for a credit engine to less than six months—compared with nearly a year for traditional approaches. One European bank, for example, wanted to launch a digital lending unit. The bank was hindered by legacy systems and entrenched processes, which created long development times for new offerings. To manage this challenge, the bank designed a modular credit-decision engine, which blended parts of the existing system and enabled the team to develop new modules where they were needed. The result was a faster time to market for the newly launched digital business.
3. Create scalable infrastructure
In developing the technology infrastructure, new-to-market lenders have a range of options to consider. They can start by identifying their ambition and perceived advantage in the market and the degree to which their current technology and data availability will support the initiative—or hinder progress. From that point, organizations can plot the right path forward.
Companies that aim to compete primarily through strong customer relationships might need only basic risk-assessment processes. These companies can buy turnkey solutions from an established solutions provider. Different standard market solutions are available for acquisition or as an outsourced service. Most have a comprehensive offering that includes credit origination, line management, automated decision making for credit assessment, customer acquisition, renewals, and exposure monitoring. With such end-to-end capabilities, lenders can easily see the performance of the entire portfolio or a single customer; they can also access credit-bureau solutions to enrich their data. The turnkey approach offers lenders the advantage of speed but entails limits in customization. In addition, configuring a turnkey solution with a company’s existing IT architecture can be cumbersome.
At the other end of the spectrum are lenders whose competitive distinctiveness will rely on an integrated, tailored solution. That can mean designing and building infrastructure from the ground up. Such complex tailored solutions demand significant investment in time and money. This approach may also require hiring talent with specialized skills and capabilities.
Between off-the-shelf and fully tailored approaches, lenders can find a middle ground by buying individual solutions and applications that can be fitted together in a modular way. This will serve their competitive edge in the market: lenders will be able to customize infrastructure to better address target customer segments with their own credit-risk models and solutions. Lenders may also choose a variant of hybrid solution, entailing a custom-built front-end infrastructure—such as the workflow manager—and a standard market solution for back-end elements, such as collateral-management or exposure systems.
Another telecommunications company, with a subscriber base that comprised about 80 percent of its country’s population, collaborated with fintech partners to launch a new lending business. The project required designing technology to support the company’s existing data platforms. The company had to train current employees and hire new talent to run the lending business. The long-term goal is to expand the offerings with new products, build the scale of the infrastructure to support the broader portfolio, and collaborate with more financial institutions in the region (by selling credit-scoring services, for example).
Buy-versus-build choices always involve trade-offs between flexibility and cost. The level of spending on development, installation, and maintenance is a determinant of solution flexibility.
4. Monitor and maintain the models over time
Finally, new-to-market lenders need to track key metrics to monitor the performance of the models over time. The development of each model is a one-time effort, but maintaining and monitoring models are ongoing responsibilities. By using established metrics to track changes in the incoming customer population and model performance over time, a lender can spot problems early on.
Metrics include, for example, the population-stability index, which measures a lender’s current customer base against the population for which a risk model was originally established. Similarly, the credit-default rate will determine the financial health of the current portfolio. And metrics based on Gini coefficients will determine whether the risk model is making accurate predictions.
Amid the COVID-19 pandemic, lenders have become acutely aware that their solutions must account for significant disruptions, whether in the form of financial crises or environmental shocks.
One incumbent lender decided to move into and serve a new customer segment exclusively through digital channels. The lender developed a standard set of model-monitoring metrics and frameworks. These aggregate information and feed it to a dashboard where all aspects of model performance are compared against industry benchmarks. All anomalies are flagged for review. This approach is helping the bank assess models in real time and anticipate any necessary maintenance or correction.
Amid the COVID-19 pandemic, lenders have become acutely aware that their solutions must account for significant disruptions, whether these come in the form of financial crises or environmental shocks. Certainly during the pandemic, data anomalies and disjunctures led to model failures. Developers must consequently design mechanisms within models to anticipate future large disruptions. The goal is to build models that can be proactive rather than reactive, even under rapidly changing conditions. That way, credit solutions will keep pace with the lending environment.
The coming resumption of the credit cycle offers a rare opportunity for innovative lenders to gain access to new markets and customer segments. New entrants can be incumbent financial institutions expanding into new segments and markets or nontraditional lenders seeking to establish credit operations. By choosing to follow the steps discussed here, either kind of organization can set up operations to manage credit risk. With a distinctive strategy and the requisite expertise, innovative lenders will be able to overcome obstacles and capitalize on an emerging opportunity.