{{item.videoDuration}}
{{item.title}}
{{item.text}}
{{item.videoDuration}}
{{item.title}}
{{item.text}}
Socially responsible banking: A digital path to financial inclusion
The power of digital and AI to help provide affordable credit without sacrificing profitability
Banks are serious about their commitments to narrowing racial inequality gaps. The sector’s proving that by earmarking over $100 billion to support equality and diversity initiatives, far more than any other industry.1
Even beyond financial services, many organizations agree the costs from systematic racism run deep. Low or lost wages, limited access to equitable lending and disparities in mortgage credit have restricted economic growth. Closing those gaps could add $5 trillion in gross domestic product over the next five years, according to economists at Citi.2
But addressing these problems often comes back to banks and credit, the cornerstone of economic growth, prompting many bankers to ask: What can we do more of and how can we do a better job with it?
In short, it requires banks to lend outside of their current customer base.
Community banking or banking a community?
Banks have long been encouraged to build community into what they do. In fact, helping meet the credit needs of the local community is one of the responsibilities of holding a bank charter. But this is usually achieved in an obligatory way—a means to satisfy a requirement.
In our work with banks over the past year, we’ve come to recognize that the strategic changes brought about in 2020 can help address barriers to financial inclusion. Many institutions are rethinking how consumers access their networks, how technology can help meet customer needs and how to manage risk. But what we’re also learning is that these are the very ways in which banks can shift inclusion from a requirement to part of their purpose.
Banks can make a difference through innovation in lending
Many lenders overly favor borrowers at the upper end of the credit spectrum. The aversion to customers without pristine credit is partly driven by regulatory pressures but also because banks haven’t figured out how to make money from lending to lower-credit-score borrowers. This leaves behind a large population of lower-income, lower-credit-score individuals who would benefit greatly from access to more affordable credit.
In the US, we believe that the need for accessible credit expands beyond the 12 million unbanked individuals highlighted by the FDIC,3 may include up to 42 million people who are considered financially vulnerable and a growing portion of the 125 million individuals who are considered to be financially coping by the Financial Health Network.4 With middle class wages effectively stagnating while healthcare, housing and education costs soar, this population will likely continue to grow.
Addressing financial inclusion in a profitable manner typically runs into several barriers:
- Common customer acquisition methods aren’t necessarily suited for left behind borrowers. Physical or community-based marketing can be more effective.
- More resource-intensive customer service and higher intensity collections amount to greater operating costs for segments of the addressable population.
- Credit losses are generally higher for this population due to a higher volatility in ability to repay.
- Traditional signals of creditworthiness such as payment histories are not as reliable for this population.
But properly designed digital products can make a huge difference in the financial outcomes of individuals while creating new business opportunities for banks. Consider M-Pesa in Kenya. The company was founded on small dollar money transfers—and it’s profitable. Its mobile-based financial services are available to those who aren’t part of the banking system, and its costs are held in check by eliminating cash processing and physical branches. Since its launch in 2007, M-Pesa has contributed to a tripling of financial inclusion in Kenya from 15% to 44%, proving that a successful business model and financial inclusion can go hand in hand.5
In the US market, regional banks in particular have looked to leverage digital to expand access beyond their core geographic footprint. We estimate regional banks entering a new geographic market can now rely on about 80% fewer branches than they would have considered just five years ago.
These same digital tools can be used to connect with otherwise neglected communities to increase access to financial services. Digital financial services have shown great promise at lowering costs to increase access and might be the only way banks can deepen inclusion and make the economics work.
Alternative data and AI for credit risk management
The industry hasn’t always been a success story with financial inclusion, as many banks have been hesitant to lend outside of their comfort zone. But those that are incorporating alternative data and AI models are showing new promise in closing the gap.
Take, for example, the 42 million individuals considered to be financially vulnerable. Some of these individuals might have missed a credit card payment, lowering their credit score below a threshold a bank might be willing to lend to. But by incorporating additional data sets, such as deposit transactions, utilities or rent payment information, banks can likely identify the candidates that are ready to move back up the credit spectrum.
Many credit models in use today rely on traditional sources of data, including the borrower application information as well as the credit bureau information. But when adding alternative data, an upgrade to the modeling techniques might also be needed. A core advantage in using AI models is they’re better suited to handle larger amounts of data, even if it's poorly structured. Some models might consider over 500 variables, working to detect hidden interactions across these data elements to provide the insights and predictions for decision-making.
The predictive patterns identified by AI models can help lenders widen credit access. For example, by:
identifying previously overlooked but credit healthy customers
approving a greater number of traditional borrowers, though the techniques may differ.
Advanced credit modeling techniques are being driven by software companies and larger banks, and while adoption is just starting, our analysis suggests lenders already are seeing 15% to 30% increases in credit approvals with no change in loss rates. On the financial inclusion side, these approvals often include once overlooked borrowers who can now access credit from the formal banking system.
Of course, when incorporating new data in the decisioning process it's important to identify new potential risks. Some data might be more readily available for more digitally active or higher-income individuals, creating potential for discrimination. And as models “learn” they need to be monitored to ensure they’re operating as intended and not inadvertently discriminating.
Besides the benefit to consumers from expanding credit eligibility, use of alternative data and advanced modeling techniques can help better measure and reduce credit risk to the bank. Lenders can further reduce credit risk by positively influencing the likelihood to repay or by developing innovative ways to collateralize or guarantee the loan. Incorporating credit and spending products with payroll or other upstream functions can reduce repayment risk. The same goes for offering alternatives to short-term, high-interest loans through different types of secured lending and collateral.
Reasons why banks have been slow to adopt AI in loan and credit decisioning:
- Regulatory and internal process hurdles to move models to production
- Varying levels of data quality from internal and external systems
- Core bank systems that weren't designed to operate with new types of data
- Implementing processes to explain reasons for a particular decision
- Monitoring to ensure model operates as intended; without adverse action or disparate impact.
Starting small with enhanced pricing and targeting
Many institutions that are developing AI models have faced challenges putting them into practice. We’ve worked with some banks that start smaller, using AI for prescriptive targeting or within their collections team. This approach can help institutions:
- Learn about less discriminatory model alternatives with enhanced pricing methodologies
- Get acclimated with new types of data and technology without changing more sensitive parts of the credit function
We worked with a top 20 US bank using AI to apply customer insights in order to understand individual financial needs and preferences. In analyzing traits such as wallet size, product usage and customer circumstances and mindset, the bank was able to better determine the right product to offer and the right time to offer it. These types of tailored, needs-based offerings can be combined with a relationship-based proposition to help maximize customer stickiness. The net results included more accurate customer targeting, which improved the overall efficiency of the product campaign.
Navigating the next steps
The time is right to tackle the long-standing issue of financial inclusion. But banks need a paradigm shift—by designing product variants and innovating their credit approach specifically for populations that need economic inclusion. This involves marketing initiatives that enable access, products that suit specific needs, digital adoption that reduces costs and smarter models that do a better job at identifying creditworthy customers.
1. Company reports
2. Closing the racial inequality gaps, Citi GPS, September 2020
3. PwC analysis, US Census Bureau, FDIC, How America Banks: Household Use of Banking and Financial Services, October 19, 2020
4. US Financial Health Pulse 2020 Trends, Financial Health Network
5. 2018-2019 Annual Report, Communications Authority of Kenya
{{filterContent.facetedTitle}}
{{contentList.loadingText}}
{{contentList.loadingText}}
Follow us
