DISRUPTING FINANCIAL LENDING USING ML/AI
Lenders are experiencing disruption due to the rise of online credit intermediaries and fintechs such as Chime, Upstart, SoFi1. Their online presence powered by ML/AI technology is democratizing the credit lending industry, which in turn, has enabled the financial services startups to cut costs and acquire significant market share faster across lending products. According to FDIC 2021 survey, ~ 96%2 of U.S. households are banked but 22% 3 of Americans don’t have any credit history (thin credit file), possibly due to being new to credit or because of fresh immigration. According to a BankRate survey, 57%4 of adult Americans can’t afford even $1000 emergency expense. This is a matter of momentous concern when there are insufficient funds in the bank to even make it to the next paycheck, let alone for coverage of unexpected expenses. In such a time, people aim for small dollar loans as a credit option. However, it becomes more challenging for these thin file borrowers to get credit as they don’t have a credit history. Therefore, innovation of small dollar lending loans using alternative bank data powered by sophisticated machine learning models could be a game changer which can revolutionize the lending industry.
Challenges with existing financial ecosystem
Small-dollar lending has long been a critical tool for borrowers in need of quick cash to cover unexpected expenses. However, traditional (credit bureau) small-dollar lenders often charge high fees and interest rates, making it challenging for low-income and underbanked or unbanked borrowers to access affordable credit. As a result, many of these borrowers have turned to alternative lending options, such as payday loans and title loans, which are often predatory and can trap them in a cycle of debt.
These service providers claim the higher interest rates and fees of payday loans is also due to the high cost of loan processing using traditional methods. The predatory lending practices have been tightly securitized by the regulators and many lenders are facing several fines5. In order to be compliant and ensure they can capitalize on the huge market opportunity with low risk, lenders need to find cost-effective and intelligent lending decisioning.
Today, Lenders can rely on sophisticated risk decisioning machine learning models powered by a strong technology infrastructure that considers various data points without the credit bureau reports. As these borrowers don’t have sufficient credit history, lenders can rely on alternative data sources such as employment history, social media, and bank data, etc.
The evolution of technology has disrupted access and usability of consumer data. The innovation fueled by alternative data can offer tremendous value to the consumer as well as lenders. It has unlocked the potential to serve credit to creditworthy borrowers without a credit score while helping them build their credit history. To that end, borrowers with “thin file” credit histories are an untapped source of opportunities for various lenders. Alternative data sources can help assess the creditworthiness of these individuals with the proper risk management.
One of the most important sources of alternative data for risk decision-making is the bank data of the borrower. This data contains information about the borrower’s income, expenses, assets, liabilities, and other financial metrics that can be used to assess their creditworthiness. In this white paper, we will explore the various ways in which bank data can be used for risk decisioning in small lending loans, especially for these targeted thin file borrowers.
By enabling financial data access and delivering intelligent insights, we can empower borrowers to unlock the potential of their own financial data to improve their financial health. The digital data spine can help build those neural connections across lenders which can help the borrowers to meet financial stability and have a faster way to access credit at the time of need.
This is also an opportunity to ensure that alternative data is used responsibly and transparently to ensure practices are compliant with relevant regulation guidelines. The borrowers need to be informed timely and have control over the data that is collected and how it is utilized.
Benefits for lenders
Lenders have several benefits in using bank data for small lending loans, such as
Accessibility of financial data: Lenders already have access to a wealth of data on their customers' financial history, including intimate knowledge of their income and expenses. This data can help them assess a borrower's ability to repay a loan and make more informed lending risk assessment without depending on the 3rd party bureau data.
Affordable loans: Using bank data, fintechs or lenders can save significant costs on 3rd party data fees. This gives them an opportunity to offer affordable small dollar loans by understanding the market dynamics.
Improved customer experience: By using bank data, fintechs or lenders can offer a more streamlined application process and faster funding as there are less players and dependency in the value chain. This also mitigates technological hurdles for various lenders.
Benefits for borrowers
Over the past several years, the online lending market has exploded with rapid growth both in proliferation of new companies and product offerings. Borrowers, like online shopping, don’t need to go the bank branch for banking. Similarly, to be a lender, one doesn’t need to be a bank to offer loans, ensuring it is compliant with the FCRA guidelines. While the opportunities for lenders are immense, there are several benefits for borrowers who need a quick, convenient way to bridge their liquidity needs, which include:
Affordable loans: With the advancement of technology, borrowers have access to cost effective small dollar loans with affordable interest rates.
Protect personally identifiable information (PII): As many times these small dollar loans are with existing bank relationships, no PIIs need to be provided or transmitted.
Build credit history and opportunities: Borrowers can build credit history by ensuring timely repayments. Even banks can leverage machine learning models to track consumer financial behavior and notify them proactively for any miss events. This can help them to build their credit history faster and create opportunities for immediate small dollar funds availability when needed.
Key considerations for lenders
Innovation through a customer centric approach is the key to the success of offering small dollar lending using alternative data. However, it is essential for all industry players such as regulators, policy makers, lenders, and borrowers to collaborate in driving informed lending decisions. Working collaboratively in an ecosystem-based approach can benefit all stakeholders and drive economic empowerment and financial resiliency across the board. There are a few key considerations for lenders, such as
T.R.A.C.E. Framework: In the new lending ecosystem, it is necessary to improve consumer awareness and ways to protect unfair lending practices. These can be achieved through transparency, regulatory, accessibility, collaboration, and ease of information available by the lenders to the borrowers.
Financial education: Lenders should provide the borrowers with personalized financial education and guidance. One way is to identify areas where borrowers need support, such as budgeting or debt management, and provide them with targeted resources and advice.
Regulatory Insights and modernization: Any innovation needs to be protected and practiced with proper governance. Lenders need to work under the guidance of regulatory policies to ensure fair and transparent lending practices are being followed with their borrowers per the guidelines from the regulatory institutions such as CFPB, FCRA and other regulatory institutions.
Importance of sophisticated decisioning technology
A Small dollar loan is typically used for short-term or unexpected financing needs, such as medical bills, utility bills. However, lenders face a high level of risk when offering these loans, as the borrowers often have limited credit histories and may not have significant collateral or assets.
To minimize these risks, lenders should build sophisticated risk decisioning machine learning models that consider various data points of consumer bank data, evaluate customer financial behavior and segment them into the risk profile. These models should evaluate the creditworthiness of the borrowers and assess the likelihood of default. By using these models, lenders can make informed lending decisions and minimize their risk exposure.
Here are a few options where alternate bank data and technology can innovate the risk decisioning process:
Machine learning and artificial intelligence: Using machine learning algorithms, lenders can analyze large amounts of data to identify patterns and predict the likelihood of a borrower defaulting and fraud on a loan.
Digital identity verification: Leveraging technology, digital identity verification solutions can help lenders verify the identity of borrowers and prevent fraud. This can also speed up the loan approval process, making it a seamless customer experience for the borrowers.
Collaboration and data sharing: With the mindset of open banking, lenders can collaborate and share data with other fintech’s to build a more comprehensive picture of a consumer's creditworthiness. This can help lenders make better-informed decisions and reduce risk and fraud losses.
The new way of risk decision-making using bank data can help to improve the efficiency and accuracy of lending practices and enable affordable credit to all creditworthy borrowers.
Type of bank data required in credit risk decisioning
Bank data can play a crucial role in providing low-cost, small-dollar loans to borrowers who need them. It is one of the rich sources of data for risk decisioning in small dollar loans. This data contains information about the consumer’s income, expenses, assets, liabilities, and other financial metrics that can be used to assess their creditworthiness. By analyzing this data, lending institutions can gain insights into the borrowers financial behavior and evaluate their ability to repay the loan.
Here are some of the types of bank data that can be extracted for risk decisioning:
Income and expenses: By analyzing the borrower’s income and expenses, lending institutions can evaluate their ability to repay the loan. This data can also be used to assess the borrower’s financial stability and predict their future income and expenses.
Assets and liabilities: Bank data can also provide information about the borrower’s assets and liabilities. This data can be used to assess the borrower’s net worth and evaluate their ability to provide collateral.
Payment history: Bank data can also provide information about the borrower’s payment history. This data can be used to assess the borrower’s creditworthiness and predict their likelihood of default.
Account balances: Bank data can also provide information about the borrower’s account balances. This data can be used to assess the borrower’s financial stability and predict their ability to repay the loan.
This helps the lenders to be equipped with essential information to assess consumer’s creditworthiness, verify their income, prevent fraud, manage risk, and underwrite loans more efficiently at a low cost and with trustworthy data.
Credit Risk Decisioning life cycle using bank data
To use alternative data such as bank data for small lending loans, there are various ways to make informed lending decisions to evaluate creditworthiness of their borrowers. Lenders can leverage modern technology architecture, machine learning, and offer complete real-time small dollar loan processing and traceability instead of traditional semi-automated ways. In the new world the loan origination and servicing can be outlined below but not limited to:
Identify and collect bank data you need: After consumer’s permission, the modern system can identify and extract the relevant bank data automatically to evaluate loan applications, such as bank statements, and transaction history.
Analyze the data: Once the bank data, machine learning, and credit rules can determine the creditworthiness of the consumer, the lender can perform the risk profiling of each applicant.
Make lending decisions: Based on the analysis of the bank data, modern systems can make lending decisions as approve or decline.
Monitor loan performance: After you have made lending decisions, machine learning models can be leveraged to monitor the performance of the loans. Keep track of the repayment history of each loan to determine whether they are being paid on time or if there are any defaults or frauds.
Challenges of Using Bank Data for Small-Dollar Lending
While using bank data for small-dollar lending has many potential benefits, there are also several challenges to consider, including:
Data privacy concerns: Banks must comply with strict data privacy laws, such as the Gramm-Leach-Bliley Act and the Fair Credit Reporting Act. Lenders must ensure that they are using bank data in a compliant manner and protect borrowers' personal information.
Limited data access: Not all borrowers have bank accounts, and though most do their financial lives can be fragmented among multiple institutions. Lenders must ensure that they are not excluding potential borrowers based on limited data access.
Bias and discrimination: The use of bank data for lending decisions can raise concerns about bias and discrimination. Lenders must ensure that their algorithms and decision-making processes are fair and unbiased.
Using bank data for small-dollar lending has the potential to provide low-cost, accessible credit to borrowers in need. However, it is essential to address the challenges and concerns associated with using this data to ensure that borrowers are protected and that lending decisions are fair and unbiased. As technology and data analytics continue to advance, we expect to see more lenders using bank data for small-dollar lending, leading to a more inclusive and accessible financial system.
Bank data has enormous potential to empower borrowers to borrow and enable lenders to drive innovation in lending. By making bank data more accessible, using it to improve credit risk assessments and offer personalized lending products, providing financial education, preventing fraud, and better risk assessment in lending products, we can create a more diverse and inclusive financial ecosystem.
This innovation journey in alternate data is not limited to small dollar loan and thin file borrowers only. One can even expand to several other lending products such as credit card and expand the consumer base to small businesses. This is only the tip of the iceberg in revolutionizing financial access and inclusion in the world as we know it.