Since the dawn of independent India, every incumbent government has tried to loosen the tangled threads of poverty and finance inaccessibility. Financial inclusion has been recognized by the United Nations and World Bank as a crucial goal in reducing global poverty in the 21st century. Even with progress in the field of microfinance, more than 2 billion people remain outside the cloud of the financial system.
However, the catch here is that most accounts lie dormant. World Bank in its Global Findex Database 2018 reported that India has the world’s highest share of inactive accounts despite 80% Indian adults having a bank account thanks to the Jan Dhan Yojana. The absence of an existing credit record haunts the financial dreams of a large portion of under-served yet creditworthy individuals. Without a decent credit score -which is 700 in a range of 300-900, these bank accounts are doomed to stay dormant.
The five Cs based on which the six credit rating agencies registered under Securities and Exchange Board of India (SEBI) namely CRISIL, ICRA, CARE, SMERA, Fitch India and Brickwork Ratings rate the customers to be cherry-picked by banks:
1. CHARACTER: a consumer’s ability to repay his loans assessed via his credit history
2. CAPACITY: comparing the income against recurring debt and assessing his/her debt-to-income ratio
3. CAPITAL: the amount of money a borrower puts towards potential investment usually in cases where mortgages are involved.
4. COLLATERAL: an asset the borrower has to provide to secure a loan
5. CONDITIONS: everything ranging from interest rates to condition of the economy.
There exists another C- Complicated. Doing such an assessment for a population as heterogeneous as that of India across the length and breadth of its 33 lakh sq. km is cumbersome and costly. Numerous efforts have been made to make microfinance available to the rural population but still for a vast majority, formal finance is a faraway dream.
The potential creditworthy who are unable to match certain criteria have doors closed on their faces every bank they visit. Add to this the misery of travelling long distances, standing in long queues, inability in filling up a form that may only accept English or Hindi, to name a few. This is when the informal credit sources pop up and put a noose around their necks. This noose keeps getting tighter and tighter till they end up being Nero’s guests. Around 44 per cent of household surveyed by RFAS 2003 have borrowed informally at least once a year with the interest rates averaging to 48% per annum. It is too expensive to be poor.
The solution to these physical limitations is DIGITAL.
From a population of 1.37 billion, we have 800mn smartphone users and 400 million WhatsApp users. This surge is pushed even further by the Government of India with its aim to build a national optical fibre network that delivers universal internet access at low costs under the Digital India Initiative.
The Jan Dhan Yojana –Aadhaar-Mobile trinity has not only eased the rigmarole of opening a bank account but also makes the process easier with a reduction in authentication costs and reduced incidents of forgery and error. The bank accounts are linked to their Aadhaar number which in turn is linked to the Direct Benefit Transfer scheme.
The launch of Bharat Interface for Money (BHIM) in 2016 provides banking services to both smartphone and non-smartphone users. World Bank reports that from 2015 to 2017, the volume of digital transactions in India grew by compound annual growth rate (CAGR) of 30%.
Advisory bots companies such as Arthyantra and Scripbox provide tailor-made advices or undertake automated investments once a client feeds in his or her goals and risk appetite. Other such efforts like that of Project Financial Literacy and Unified Payments Scheme have taken India a step towards financial inclusion.
Hence we see no or little dearth of digital infrastructure to serve the masses. Once firms are able to assess the risk associated with each customer, the amount of interest rate paid by each can be reduced or increased as per their profile. As more and more people are brought under formal credit, it automatically enrols them for financial literacy, which is much needed for financial inclusion.
As AI is creeping its way into almost every industry, be it agriculture or textiles or teaching, it is now making its way in becoming the backbone of financial infrastructure. It is therefore natural to use this to make sense of different consumer behaviour into makeshift credit profiles in a move to turn away from the traditional methods. Such unconventional ways emerge out of the need to woo the massive untapped, credit invisible population in developing nations.
The alternate measures to assess creditworthiness use mobile phone data, social media networks, psychometric testing, AI, block-chain, etc. Sounds pretty complicated, doesn’t it? But they barely break a sweat in determining those worthy of a loan.
Tracking metrics such as the time one takes to return a call and management of one’s mobile account is a promising sign of strong reliability. Social media data points mainly from Facebook are used to build consumer risk profiles and identify the most responsible borrowers among those who still live precarious financial lives. Blockchains store data over distributed ledgers which allow recovery of contracts chronologically minimizing ambiguities in record keeping.
In pursuit of bringing banking services to the “no-hit” or the “thin file”, certain new companies have come up that use online quizzes to assess a consumer’s financial personality and identify patterns correlated with borrower risk.
As we progress, we turn to machine learning and deep learning approach to assess creditworthiness. Machine learning uses three algorithms- elastic net, random forest and gradient boosting (which outperforms the other two). Deep neural networks have proved to be extremely good at detecting risky customers on complex and highly unstructured data.
The aforementioned algorithms are an advanced version of “unobtrusive measures”. A classic example of this can be how the wear patterns on the floor of a museum show which exhibits are most popular. This analysis uses observed behaviours in order to make predictions, unlike traditional methods that require direct participation by individuals.
Thus we see that data has a better idea and presents a picture of complete win-win.
For traditional lenders such as banks and Non-Banking Financial Company (NBFCs), it opens up the non-served and under-served markets. The Fintech startups, which have increased three-fold in number since 2015, are looking to lend to individuals and SMEs and E-commerce firms looking to verify and create financing options for merchants and online customers. This rise in the number of firms due to easier entry into the financial industry increases competition among firms which will ensure consumer surplus. Above all, data is stored in cloud servers ensures a downward sloping cost curve due to leaner manpower requirement. Over time, this will help anticipate future behaviour which will set a network effect into motion which will lower the costs for consumers too.
This also seems to be a way around the problem of informational asymmetry which is one of the biggest causes of market failure.
The borrowers across India face a spectrum of problems as diverse as the nation itself. For example, it has been observed that the geographical distribution of microfinance in India is highly skewed, with the top ten states accounting for 83% in portfolio outstanding. According to RBI, Kerala, Maharashtra and Karnataka have achieved high financial inclusion with Index on Financial Inclusion (IFI) greater than 0.5. Besides the general lack in demand for credit by the poor, there are very little well established NGOs willing to initiate microfinance in the northeastern and eastern states. Added to this is the misery of not getting a long term loan which is what is required by most of the agriculture borrowers.
India’s MSME sector has a huge unattended need for credit which is as high as $200 billion. The advent of alternate credit assessment techniques by new fintechs has ensured faster and better credit access to SMEs located in tier-3 and tier-4 cities and towns of India.
Now AI-driven algorithms seek to avoid failures due to rigidities such as the case of 1987 Black Monday stock market crash but when it comes to data, a lot of red flags also start to wave. To name a few: cyber-attacks, difficulty in accounting for input bias, inability to rectify errors and the inability of algorithms to differentiate between causation and correlation.
Such bottlenecks are taken care of by Regulatory technology better known as RegTech which is a branch of fintech that is focused on improving compliance systems of financial services companies is yielding sound approaches to managing the risks of algorithmic bias.
“Poverty is deprivation of opportunity,” said Dr Amartya Sen. Without universal access to formal credit, every opportunity will continue to remain a closed-door for our citizens. Hence we need a multidimensional approach that has both the Government and the people on board.
Comments