Alternative credit scoring: A new landscape of opportunities and risks
In many parts of the world, access to centralised, accurate credit histories is taken for granted. However this is not the case everywhere, for example, in India nearly 80 per cent of the adult population does not have a credit record, and as a result they have difficulty accessing formal credit. Clearly many of these people are creditworthy but don’t have credit histories as they work in the unorganised sector, in temporary jobs, or do not have steady sources of income.
The high costs of serving these consumers via conventional models, the relatively small transaction values involved, and the difficulty of knowing and assessing many of these consumers through traditional means (such as the credit bureaus) dramatically limit the products and services that are made available to this huge but underserved market.
The rapid proliferation of mobile phones, internet and social media is proving to be a game changer. India already has over 462 million active Internet users and 153 million actively using social media platforms— 130 million out of which access them through their mobile phones. When this is combined with increasing smartphone penetration , the growing acceptability of online payments and digital transactions and favourable demographics, the result is the creation of a digitally visible segment which in turn provides increased access to lenders.
This explosion of e-commerce, Internet and social media usage, increased computing power coupled with the perceived inability of traditional credit scores to adequately capture “thin file” borrowers has prompted the emergence of alternative, big-data tools that promise lenders a way to extract additional results out of their underwriting processes.
Every time the digital consumers browse the internet, make a phone call, send a message and participate in social media networks they leave behind a digital footprint. Data from mobile phone records, mobile bill payments, mobile browsing, app download history or prepaid top-ups can be used to assess consumer risk and determine the creditworthiness of underserved customers. Lenders can use the output of their credit scoring to offer unsecured, small ticket, short-term credit at a much lower cost than traditional loans. “All data is credit data’ – this mantra is increasingly being followed by lenders to use non-traditional sources of data—many of them not directly related to money—to augment their traditional underwriting mechanism.
Numerous lending startups or fintechs are seeking to capitalize on this opportunity, with many using proprietary “machine-learning” algorithms to sift and sort through thousands of data points available for each consumer. Banks and NBFCs are also assessing how they can gain advantage by partnering with the startups. So, is alternative credit scoring the perfect solution to the problem of financial inclusion? Is it going to be the key driver for banks and NBFCs to grow their lending portfolios while managing risk?
The early indications are quite promising, however a few challenges and concerns could throw a spanner in the works. Major concerns exist around privacy and transparency – are thousands of data points analysed by credit-scoring tools being collected with consumers’ knowledge? Not likely. Even if consumers’ have agreed to the terms and conditions do they actually understand what they have agreed to? The accuracy of data is also questionable. If consumers are denied access to financial products will they be able to determine which decisions are unfair? Will they understand what actions they should take to improve their credit profiles? Then there are concerns about discrimination towards certain groups – Big data tools may risk creating a system of “creditworthiness by association” in which consumers’ family, religious, social, and other affiliations would determine their eligibility for an affordable loan. Such discriminatory scoring may not be intentional but a result of sophisticated algorithms that generate insights of patterns. However, the implications are serious in terms of circumventing existing non-discrimination laws and systematically denying credit access to certain groups.
The limited information available about these alternative credit-assessment tools and the utilization of alternative data in scoring models and credit decisions is already attracting the attention of regulators. Across the globe they are attempting to get a better understanding and to ensure that innovators proceed responsibly and have strong legal incentives to ensure that their scoring decisions are transparent, accurate, unbiased, and fair. For example, the Consumer Financial Protection Bureau in the US recently issued a request for information (RFI) regarding the use of alternative data and modeling techniques in the consumer credit marketplace. Some prominent industry bodies globally have expressed concerns about the use of alternative credit scoring. As per recent news, the Indian government is expected to come up with regulations in the fintech space in the near future.
So what should banks and NBFCs do in the current scenario? While these institutions should look at alternative data as a source of opportunity, they need to carefully consider the distinct advantages and disadvantages inherent in each data source. It is important to make the right call on which alternative data to leverage, especially given that there are significant operational and cost considerations of acquiring, maintaining and updating such data. They need to consider whether the compliance risks and costs, coupled with the uncertain predictability of alternative data, justify the potential return and benefits. As of now, it may be worthwhile for them to leverage the alternative data to augment the credit scores gathered from traditional means rather than using it as a complete replacement.
Also, banks and NBFCs need to consider if they are leaving any money on the table when it comes to driving growth and profitability. Can they take advantage of the transaction data that they have accumulated over time to grow their lending business while mitigating risk? Advanced analytics solutions can churn large volumes of data and generate predictive insights on the most appropriate target segments and the most relevant sales channels. Similarly, the data can be used to identify pattern based characteristics for ideal and delinquent customers. These insights can be used only to automate the initial application filtering but also to reduce the chance of customer turning delinquent later.
Leave a comment