Digital Premium Credit Reviews
Laila Berri
These barriers have caused more than one bank to delay or sidetrack digitization efforts. Programs launched with great executive attention and focus lose momentum as the initial excitement of chief risk and lending officers evaporate. Investments needed to sustain programs are partly or wholly withheld. Incremental changes are sometimes substituted for planned end-to-end transformations.
However, numerous banks successfully digitized the credit journey. In the following pages, we offer the practical lessons that have emerged from these experiences, with special emphasis on SME lending, the area that is currently getting the most attention and investment.
Many banks have found that an end-to-end view of the entire customer journey, including a target state set according to the customer experience, was crucial to success. For example, a Benelux bank redesigned its business-lending process from end to end, allowing it to eliminate numerous handovers. The result was about 30 percent greater efficiency. Without an end-to-end orientation, on the other hand, banks have seen disappointing results. Attempts to improve the credit process piece by piece tend to become incremental, lose customer focus, and miss the big-picture opportunity to deliver a fundamental step change in performance and approach. One Northern European bank found such an opportunity by shifting its focus for SME customers from selling products to fulfilling customer needs. As a result it radically rationalized its lending-product range down to just three simple products, massively reducing complexity. This would not have happened with a piecemeal approach.
While taking an end-to-end view, however, successful banks have learned that it pays to limit the scope of the first wave of the transformation and focus on a minimum viable product (MVP). The MVP is scoped to be substantial enough to drive real value, momentous enough to create excitement within the organization, and simple enough to be designed and implemented rapidly. Improvements can then be made progressively in waves of rapid subsequent releases.
At one Scandinavian bank, as many as half of all credit decisions concerned SME customers with existing loans seeking additional credit. The bank decided to focus on improving their experience, since the cost to serve them was significant, but the decisions involved were less complex, as most of the necessary data were already available in the systems. Over an intense 20-week period, the bank designed a new end-to-end digital journey, including an online application process, a framework for making new credit decisions, a revised credit process with automated decision making and fast-track handling for simple cases, as well as radically simplified credit-paper and collateral-review processes. Certain features of the new journey were not included in the MVP but scheduled for later releases. This kind of approach avoids too much early-stage complexity so that a transformative solution can be implemented more quickly, establishing momentum for future change.
With good reason, risk managers can be wary of a fully automated approval process for business loans. Long-standing policies and decision processes often depend on manual reviews and cross-checks. Years of root-cause analysis of defaults and assessments of soft factors have proved reliable but would be missed in an automated approach.
At one bank in central Europe, the long-standing business-lending process features a decision checklist incorporating thousands of criteria and covenants for contracting and disbursement. While time consuming and costly, the process does achieve the desired risk outcome. In fact, risk functions at many banks successfully use experience-based subjective assessments to achieve low default rates. While the accuracy of data-driven model-based decision making continues to improve, risk managers are correct in taking a cautious approach to automation.
In the most sophisticated examples, about 70 to 80 percent of SME-lending decisions are fully automated, with the remainder referred for credit review, allowing valuable expert time to be focused on complex or marginal cases.
Ultimately, RMs were able to provide loan approval in five to ten minutes about three-quarters of the time; more complex cases are decided in an average of 90 minutes (and not more than 24 hours) following a manual review.
To develop models, many banks have expressed interest in using external data (when legally permissible), including novel sources such as social media. While creative use has been made of unusual data sets, it is usually best to begin with readily available data. Transactional data have proved especially powerful. A number of banks and fintechs have developed tools to process transactions from primary operating accounts line by line, classifying them into detailed revenue and expense items. Advanced analytics can use these rich risk data to generate simplified financial statements, affordability ratios, customer- and supplier-concentration analyses, and so on, in real time. These transactional data offer substantially richer and more up-to-date insights about company performance than out-of-date annual accounts. With the second Payment Services Directive (PSD2) and other open-banking initiatives now coming into force, similar analyses can now also be performed on new customers.
Ambitious data-aggregation plans or multiyear data-lake projects are rarely good bases for digital-lending transformations. Such plans are frequently abandoned before completion. Successful transformations generally rely on existing data sources, sometimes using imperfect, robotics-based data integration (such as screen scraping) to get started. Recently, a major bank in Southern Europe successfully completed the early stages of its transformation using readily available demographic and behavioral data. That experience shows how pragmatic data solutions can create real impact quickly, building momentum for subsequent, gradual data-management improvements.
By incorporating regulatory models in their new credit-decision engines, banks can satisfy regulatory requirements in less time and start reaping the benefits of digitization more quickly. A Northern European bank did just this, after applying the existing internal ratings-based system for business lending and building new automated analyses for affordability and cash flow.
Progress in digital-lending transformation occurs when departments and functions with separate priorities are on board. Resistance to change sometimes arises from a general lack of clarity on how digitization will affect the organization and its customers. Senior-management alignment on the goals of the transformation can help counteract emerging cultural issues. A defined end state does more than guide implementation; it can often help overcome opposition to the program. Other elements essential for success include the following:
While the challenges in digital-lending transformations are formidable and the path to ultimate success can be bumpy, experience proves that the efforts expended are more than fully repaid in competitiveness and profitability. Success means much faster credit decisions, with customers getting cash up to 80 percent sooner; lower costs, with 30 to 50 percent less time spent on decision making; and better-quality risk decisions, which translate into greater profitability down the road.
Inadequate access to finance is an important impediment to development in low- and middle-income countries (LMICs). The rapid proliferation of digital financial services in emerging markets has important implications for global financial inclusion. Unbanked individuals in low- and middle-income countries are increasingly able to bypass brick-and-mortar financial institutions by saving, borrowing, and making payments directly through their mobile phones. While digital credit has the potential to improve access, usage, and resilience for underserved populations, there is still much to learn about potential pitfalls including over-indebtedness and fraud, especially in under-regulated markets. We fund researchers who use randomized evaluations, machine learning, and other rigorous methods to answer questions of critical importance to this sector.
While the rise in the use of mobile phones and other digitally-enabled services has generated optimism about transforming financial services for the underserved, the rapid increase in the volume of personal data being collected and used in a wide range of commercial and policy settings introduces privacy-related risks to consumers that are not well understood.
However, numerous questions about the promises and pitfalls of these products remain. For example, it appears that for certain products, some borrowers benefit, while others consistently incur late fees, default, engage in loan stacking (growing their indebtedness), or deplete savings. The DCO is expanding this portfolio of research through 2024, to investigate how the impacts of digital credit vary according to product design features or borrower characteristics, and how products can be improved to better protect and benefit under-served populations.
The DCO will also build a portfolio of research to better understand concerns regarding data privacy, possible enhancements to data privacy protections in LMICs, and how various privacy-preserving practices can unlock responsible innovation.
As the name suggests, a credit builder loan can help you build credit and improve your financial standing. Digital Federal Credit Union (known as DCU) offers one such loan with a low interest rate, nationwide availability, and no credit check requirement, earning it a place among our best credit builder loans. However, DCU imposes several eligibility requirements that may complicate the enrollment process. Read on to see whether DCU's credit builder loan is a good fit for you.
Chartered in 1979, DCU manages over $9.9 billion in assets, making it the largest credit union in New England in that respect. Based in Marlborough, Massachusetts, DCU operates 23 full service branches in Massachusetts and New Hampshire, and boasts over 1 million members nationwide.
c80f0f1006