Leveraging Fintech for Risk-based Pricing & Personalisation

By Bindu Ananth

Much of the focus of fintech vis-Ã -vis financial inclusion has been on payments and the ability to transfer money with relative ease using a mobile phone and an app such as the Unified Payment Interface (UPI)/Bharat Interface for Money (BHIM), or using the USSD protocol in case of feature phones. Here, I want to discuss two fintech applications that have received less attention, but which can be transformative for financial inclusion. These include risk-based pricing of microloans and personalised financial advice.

A central feature of finance, especially lending, is information asymmetry. The customer knows more about her creditworthiness than the lender. This is aggravated if the customer has no collateral to offer, which would otherwise serve as a ‘signal’ to the lender. Therefore, there is often a significant risk premium built into the pricing of the loan that buffers the profitability of the lender against credit losses. But we expect that over time, as lenders learn more about the creditworthiness of various customers, the same will be reflected in lower risk premiums.

However, if you look at the microfinance industry as an illustration, you will see in spite of several customer groups having over 10 years of credit track records (this data is available through the credit bureau for at least five years now due to the guidelines of the Reserve Bank of India or RBI), pricing to customers has remained largely the same (the annualised percentage rates of interest lie between 22 and 26 per cent) and there is no distinction between newly-acquired and vintage customers. At the same time, one growing category of fintech companies is digital credit providers – they underwrite loans to customers based on a combination of data points such as credit record, tax data (if available) and bank statement analysis, among others. Some companies also take as input psychometric data such as entrepreneurialism and honesty in dealings to construct a picture of the customer. To a large extent, microlenders and digital lenders currently serve different customer segments – the former tends to serve more unbanked customers and informal sector workers. But these two worlds will soon collide and it is reasonable to expect a lot more risk-based pricing for these customers that will take as input various aspects of a customer’s behaviour and attitude. This will be the great leveller in retail credit: A poor woman who is an agricultural worker with a strong repayment ethic and ambitious goals for the future should be able to borrow at the same risk premium as her urban counterpart who is a salaried worker.

While the rich have private bankers that provide customised financial advice, this service is equally important to low-income households for whom even small financial mistakes can have costly consequences. Yet, most efforts in financial inclusion take a standardised view of customers and have cookie-cutter distribution models.

There have been a few experiments in customised distribution approaches, notably the Kshetriya Gramin Financial Services (KGFS) model that uses a combination of product rules and trained front-line employees in remote rural markets. However, by and large, customisation has been associated with high operating costs and the need for specialised staff at the customer interface. Fintech will mount a significant challenge to this traditionally-held wisdom. One application of fintech, specifically the supervised machine-learning models, is building recommendation engines that can construct customised financial portfolios based on inputs such as age, risk-taking ability and investment horizons. Even if in the near future it does not seem likely or even desirable that a rural customer uses a recommendation engine of this nature in a self-service mode, say, through an app to buy mutual fund and insurance, this can be plugged into systems and processes of existing service providers with a sales force that interacts with this customer segment. Such integrations can significantly enhance the quality and comprehensiveness of the proposition to the customer relative to the mono-product focus (usually loans) prevalent in financial inclusion.

Finally, some cautionary thoughts. Financial inclusion has been notoriously driven by supply-side considerations and poor understanding of customer needs and preferences, resulting in outcomes such as dormant bank accounts. It is not obvious that fintech-based approaches will not fall into the same trap. For example, proponents of mobile banking for financial inclusion have not sufficiently appreciated the challenges regarding a woman’s access to private transactions on the phone even where the household has a phone. Many women seek confidentiality even from other members of the household when it comes to financial transactions, particularly savings activity. Local language interfaces are difficult to support on Chinese-manufactured feature phones that account for a large share of the rural market. These have important design implications and must be taken into account if fintech has to reach its potential. Confidence of the customer is an important factor in widespread adoption of these services. Besides good design of services and affordability, this requires a regulatory framework that enhances customer protection and providers not taking a narrow ‘buyer beware’ approach.

This article first appeared in the latest edition of Business Today magazine under the headline ‘New Tools’.


The Green Stool Innovation

Below is an excerpt from Bindu Ananth’s latest column on the Forbes India Blog.

On Wednesday, we brought together leaders from our Kshetriya Gramin Financial Services (KGFS) companies to talk about best practices from across our operations in Uttarakhand, Orissa and Tamil Nadu. One story really stood out for me – the green stool innovation.

When we were setting up branches in villages which had never been served by formal institutions before, we were very eager to imprint our values in all aspects of our operations. So, our wealth managers hired from the local villages wear uniforms to signal that they are professionals. All visitor cars need to be parked 100 metres away from the branch entrance so that there is no sense of outsider-insider within the branch premises. We insisted that our landlords build bathrooms adjacent to every branch so that we would be able to hire women wealth managers. One of the early things we also did was to have locally made benches painted bright green as the furniture in the branch. The bench seemed to us, to be the metaphoric leveller in a village context often characterised by caste and class divides. Until, one of our employees realised the awkwardness of the bench design.

Click here to read the full post.


The crucial link between financial access and decision making of the poor

A new paper by Anandi Mani et al in the August issue of Science has a stunning finding – that the cognitive impact of being poor may be equivalent to as much as 13 IQ points. The authors study shoppers in a New Jersey mall and sugarcane farmers in Tamil Nadu using an experimental design and are able to show that the poor perform consistently worse on standard non-verbal tests of intelligence when “stressed” than the rich. Very interestingly, in the case of the sugarcane farmers, the comparison is not between rich farmers and poor farmers but the same farmer pre-harvest and post-harvest. Before harvest, the farmer is a poorer version of himself (compared to after harvest) because of the liquidity crunch associated with the time before harvest. Think of it as the equivalent of the last few days of the month for the salaried class.

I think this study has very important implications for thinking about how good financial access will look like for the poor and what it will achieve. Too often, as practitioners, we emphasize the “big factors” such as branches, financial literacy, products, regulation and so on and when we think about the impact we have on our customers, we think about mega metrics like income and empowerment. This study tells us that if done well, perhaps the most important impact we will have is to allow customers to free up their “bandwidth” to focus well on the big decisions in their life – their childrens’ education or choosing where to sell their next crop.

Above excerpt is cross-posted from Bindu Ananth’s latest column on the Forbes India Blog. Click here to read the full post.


Shadow banking as the symptom and not the disease

A friend pointed me to an interesting new book recently, “Inside China’s Shadow Banking: the Next Subprime Crisis”. This is written by an investment banker turned shadow banker, Joe Zhang. His anxious nine year old asked him if he was going to become a “small loan shark” in this transition. The book is a fascinating account of businesses that operate at the periphery of the banking sector in China and the consequences of prolonged negative real rates of interest combined with a banking sector subject to stringent licensing (sound familiar?).

This book also got me thinking how the term “shadow banking” has acquired so much currency in recent years since the US credit crisis. This rather unflattering term is used to refer to an entire spectrum of financial firms including hedge funds on the one end in the US and unlicensed deposit-takers like Saradha Chit Fund in India and micro-credit firms in China on the other. Irrespective of exact identity, being a shadow bank clearly connotes weak to non-existent regulation and risks to customers. A throwback to Shylock, if you will.

Above excerpt is cross-posted from Bindu Ananth’s latest column on the Forbes India Blog. Click here to read the full post.


Housing Rental Markets for the Poor (and not-so-poor)

I read an excellent post by Ajay Shah recently that questioned the policy wisdom of emphasising house ownership over rental housing. His concerns stem from hampering the mobility of labour and worsening risk diversification in the portfolios of households. While these concerns are true for most households, they are much exacerbated in the case of low-income families.

Take a typical worker employed in the construction sector earning, say Rs. 4500 per month. He faces two significant risks to income: accident/health shock resulting in temporary or permanent disability and unemployment resulting in temporary or long-term loss of income. Given this inherent income volatility, the obligation of a fixed Equated Monthly Installment (EMI) over a long period of time to finance the house purchase seems unsuitable. While capital appreciation as a motive might make sense for households with more stable incomes and low exposure to real-estate otherwise, the volatility in this case becomes a real stumbling block. A rental contract provides the much-needed flexibility to reduce housing expenditure when shocks occur and also the ability to migrate when the nature of economic opportunities shift, as they are likely to over a period of time. My colleagues have an interesting paper that simulates household wealth under ownership housing and rental housing that makes this point clearer.

Above excerpt is cross-posted from Bindu Ananth’s latest column on the Forbes India Blog. Click here to read the full post.