We need to develop solutions to the many shortfalls in the microfinance industry by addressing the true costs of building the market infrastructure.
Fault lines are very significant. They persist for long periods of time, without much damage. But when they make themselves felt, they spread, and then can have disastrous impacts.
Most industries possess such fault lines and it is important for both regulatory and policy interventions to focus on how to reduce the damage they cause.
For microfinance in India, the growth story has been volatile. It was hit by a collapse in the markets by policy action twice, in 2006 and 2010, with almost 7500 crores of loans written off – no small sum.
The replacement supply has been worse. The regional press in Andhra Pradesh and occasionally the national media have carried stories of the new lenders, popularly called telephone or call lenders. These lenders turn up and disburse credit, not only at a high price but also through employing strong-arm debt collection techniques.
Added to all of this capital destruction, there has been a loss of some 40,000 jobs, and deprivation of resources resources to some 6 million households, conservatively.
Clearly, any intervention to buttress fault lines will cost less than all of this.
The fault lines in microfinance
Where therefore are these fault lines?
Recently a series of reports have “murmured” of problems that beset microfinance, the last one of this series being the article by M.S. Sriram. The problems, as identified in these reports, revolve around saturation of the market, indebtedness, and competition. Exacerbating the problems are client and growth rates in excess of 60%, not supported by adequate operating infrastructure. According to Sriram, for microfinance institutions, an average growth of 18% in infrastructure of staff and branches cannot sustain such a large loan portfolio.
However, it can be argued that over the past few years, technology investments in the industry have been huge. These investments include the computerisation of accounting entries, the unique identification of each borrower, the automation of data sheets, and the almost simultaneous transfer of data to head office. Thanks to all this, the work pressure of field staff has been reduced. Newer tools that inform the credit officer of the underwriting risks of the clients, possessed by agencies such as SKS, help with the quality of the portfolio.
A further problem is that efficiency metrics from income and expenditure statements do not alert us to outstanding gains from growth. The operating expense ratio, which measures personal expenses vis-a-vis income, has not had a particularly dramatic fall (it is inversely proportionate to income if windfall gains happen). Between FY 14 – FY 15, for most agencies, the decline in operating expense ratio has been within 100 basis points. If we were to separate the two largest categories just by size, namely those above 100 crore in outstanding portfolio, the difference between FY 14 and FY 15 remains the same. Had there been gains from growth, it would have shown up in a departure from the norm by one of the larger sets of agencies.
The average number of borrower per credit officer has remained the same as well over the last two years, and across the top four sizes of agencies has remained around the 500 mark (all data from ‘Sa-Dhan: The Bharat Microfinance Report 2013-15.’ Pdf here).
Between the changes in income and operational management, it is not possible to discern any significant signs of heating in the loan portfolio. This is borne out partly by the granular analysis of the IFMR team, who examined credit bureau data to conclude that there is limited evidence of overheating.
Portfolio and client growth clear indicators
The one point that remains unanswered is the growth in outstanding portfolio and clients. Is it evenly spread as the aggregate data makes it out to be, or is there more to it than meets the eye?
The microfinance industry has gone from 33 million clients to almost 37 million in 2015. The increase in portfolio has been from approximately 33,000 crores to 48,000 crores.
It is this unwieldy growth to which Sriram calls attention and suggests that the credit bureau data is not sufficient to capture the distress being experienced. Instead he suggests that pipelining is disguising the levels of indebtedness which is manifest in the episodic data, something to which journalists at The Wire are calling attention. Pipelining means that people borrow money on or for other members in the group.
If that were so, today, with nearly 200 million loan records in the bureaus, and supposedly 95% of the portfolio, we would not have people slipping through the net. Clearly, there is a fault line in the bureau data at present, in serving as an indicator of over-indebtedness. A different set of data will have to be looked at to improve predictability.
There are some simple explanations for this fault line, and pipelining is not very significant.
The first is the absence of data.
One reason for this absence is that NGO-based microfinance institutions were not included in the exercise in the early stages. This was the case primarily because they did not possess the technology and MIS that would allow them to be, and secondarily because no one wanted to address the issue of training and building awareness in expanding the strategic utility of the credit bureau. While there has been an increase in not-for-profit institutions reporting to the bureau, few use the information strategically, or even operationally.
There is also absence of data from banks in the microfinance bureau. This is true for both the self-help groups and some of the loans disbursed through business correspondents. It is well recognised that few banks possess information about clients’ identity in a group loan. In addition, some banks do not think it necessary to report some of their loans to the microfinance bureau. Instead they report their loans as part of their larger lending portfolio. Consequently, borrowing is now recognised to be more than what is disclosed.
A systematic estimation of this discrepancy has been undertaken by Parag Jariwala and Vikesh Mehta at Religare Institutional Research (pdf of report). They suggest that the true amount of credit might be larger than 50% of the reported credit to SROs. If we add the informal lenders to this, who are 30% of the rural households lenders and 11% of the urban low-income household, as per RBI data, we are looking at the early stages of a potential problem.
However, the second most important factor in the bureau data fault line has been the submission of data itself. The IFMR blog suggests that four year after the bureaus were started, there is still a “lag between the disbursal of a new loan and its reporting to the credit bureau by the MFI…However, the prevalence of this operational issue is limited.” Given that it has been already four years that we are facing this problem, whether we will solve it easily is unsure. The right approach would be to test the reporting discipline and technology with all the service providers within a particular geographical area and then scale it up.
No more short-sell solutions
One would assume that building appropriate frameworks that constrain Indian markets would be encouraged, but that does not seem to be the case.
Most solutions, even those in relation to the credit bureau, follow cookie cutter approaches, without addressing deep-seated market failures.
We need to develop solutions by addressing the true costs of building the market infrastructure. Our credit market is replete with short-sell solutions. If we are truly desirous of building an inclusive financial market, it has to be one that is holistic, recognises all the elements, and is ready to pay the full price of a robust market. Non-Performing Assets and Write Offs, which we have seen in India, are the true reminder of the costs of building markets. It is always more efficient to pay the price up-front than struggling with cleaning up the mess that follows from not doing so.
Titus Mathew is Managing Partner at Market and EcoSystem Advisory.