We live in times when credit literally 'pings' on our phones. Whether it is a personal loan, a loan to start a new business or a working capital loan, borrowers are now spoilt for choice. Digital lending is the popular kid on the block, and everybody wants to get acquainted.
According to a report titled 'Digital Lending' by Boston Consulting Group, digital lending in India is estimated to exceed $1 trillion over the next five years. This underscores the need for a committed and consistent risk management agenda and makes a compelling argument for both the regulator and fintech companies to work in tandem to create a robust and enduring framework.
From Know Your Customer (KYC) to Know You Risk (KYR)
Digital credit has extended access to instant, automated, remote credit to millions of borrowers. However, lending companies need to stop and ask a few questions: Who is accessing these loans? How are borrowers using the money? What is the risk of late repayment and defaults? Is there lack of transparency in loan terms and requirements? Does the borrower have debt stress or over-indebtedness? Additionally, digital lenders need to set up risk frameworks that adhere to the same enterprise risk standards as the larger, well established lenders and ensure operational strength and enterprise liquidity.
It all starts with Credit Assessment
At present, much of the focus of digital lending occurs at the credit assessment level. An effective underwriting and loan approval process is important for achieving favourable portfolio quality and risk mitigation. Multiple sources of alternate data are analysed to determine the credibility of the borrower. Easy access to credit records, enhanced data analytics software coupled with online PAN and other identification records that facilitate e-KYC has made it relatively simple to determine credit worthiness.
Players in the digital lending space employ rigorous data checks that help waive the risk factor associated with potential defaulters. However, despite various checks in place, lenders are always exposed to the risk of fraud and data sanctity. Consequently, there is the risk of lending to below grade borrowers as well as the risk of credit not being extended to worthy borrowers. Most digital lenders in India do build in some form of cost around frauds within their business models. However, they need to do much more in order to mitigate this risk. Learnings can be around multiple facets, for example, distinguishing fraudulent applications from genuine customers who have missed specific repayments.
Data privacy and security risks
Data is the oil of the 21st century, a valuable commodity. Fintech companies collect large amounts of data about their customers, typically stored, analysed and filtered. Access to such vast reserves of data makes it incumbent for fintech companies to have a robust infrastructure and systems in place to protect this data. Data is also at risk from cyber-attacks, making it imperative for lending firms to adopt the best in class security standards.
Credit screening models
Digital lenders make use of advanced algorithms to determine the credit worthiness of the loan applicant. These models are often designed to accommodate for a greater level of initial defaults so that the algorithms can ‘learn’ from the emerging patterns and adjust. This is primarily done by allowing almost anyone to borrow up to a low threshold limit, and then analysing their repayment behaviour and its corroboration with the alternative data collected.
Another strategy adopted by some Indian players is that of controlled disbursement. The strategy is to first disburse a small amount of loan and then evaluate consumer behaviour in terms of loan repayment. This analysis helps in determining the credit eligibility of the customer. While this can be effective in assessing the credit risk associated with existing customers, its efficacy is suspected in the case of customers who are availing of credit for the first time.
It all ends with a robust recovery mechanism
Absence of collateral makes these loans riskier and highlights the importance of a good recovery mechanism. Unfortunately, most digital lending players are yet to establish a robust ground collection mechanism that effectively reduces the risk associated with loan recovery.
Continuous monitoring and predictive analytics can enable lenders to identify and quickly respond to changes in borrower circumstances and repayment behaviour. Additionally, separate recovery mechanisms need to be developed for different sets of customers. While many solutions are being generated to address risks in customer acquisition and evaluation, we have merely scratched the surface in terms of risks associated with collections.
Fostering new partnerships: Regulators and fintechs
Since digital lending deals with public money, it should be subject to stringent regulatory oversight. The Reserve Bank of India in February 2018, stipulated that fintech companies that are in the business of online lending should fall under the 'full-fledged supervision' category. The regulatory infrastructure in the digital lending space is continuously evolving, making the regulatory risk in the sector high.
Anticipating the growth in the sector, it would be prudent for the regulator to outline a general set of regulatory principles and then work in tandem with the financial institutions and fintech players to implement the most effective regulations without impeding innovation and free-market practice.
Mitigating future risks
In a perennially evolving digital landscape, it is challenging to tether risk. Hence, it is imperative that risk management systems are not only robust but are also agile. Risk management needs to primarily focus on the sanctity of the credit assessment models, the quality and security of the data gleaned, setting up robust recovery mechanisms and adherence to the regulatory environment. Databases need to be updated in real time to avoid potentially outdated information while credit assessment models should regularly undergo stress tests. Even though algorithms are intended to avoid human biases, they should be periodically reviewed for unwanted discriminations.
Risk analytics and mitigation techniques can help firms navigate the potential landmines in the digital lending space. Advances in digital technology, data science and machine learning/ artificial intelligence present new opportunities to manage lending risks and maintain a strong lending book that is resilient to changes in the credit cycle.
The author is President, Risk Advisory at Deloitte India.