Microfinance industry used to be an industry which was hit with adversities once a decade, but since 2010 it seems some event or the other throws the wrench into the engine every few years now.
The lenders have learnt from each difficult situation thrown at them and tuned their business model in various ways to survive these in the future. But I doubt they would ever become immune or anti-fragile to such events.
After the AP crisis the lenders came together to form a code of practices that would help them avoid the political/public consequences/lash-backs of coercive collection practices. Who and how well these are followed is anyone’s guess. The situation on the ground is dynamic, even a pan-India survey would reveal information about the past and present and not how these lenders will behave in the future in times of crisis.
Demonetization would not have had such a severe impact on this industry if RBI had allowed NBFCs to collect old currency notes. One NBFC escaped the carnage by becoming a bank at the right time. Others were not so lucky. A few others did collect the old currency up to 4% of their loan book and 8.4% of the book was ever-greened to control NPAs. What was good about the crisis was that RBI did not allow any moratorium and true vulnerabilities of the industry were allowed to get out to the public. As you will notice in the charts below NPAs spike up after the 90 days delayed NPA recognition facility provided by RBI. The industry was able to provide for all loan losses from one financial year’s of profitability and the lending engine chugged on.
Corona virus has definitely brought about the biggest challenge the industry has had to face till date, whether this is the peak limit of difficulty any event can through at them or something bigger can also occur is an unknown unknown. Till now the nature of adversities has been such that their magnitude of disruption in the livelihoods of the borrowers is varied but majority of the damage is determined by the geographic scale and time duration of the adversity.
DEMON was pan-India event, but the duration impact to livelihoods was short lived. Here also I would define duration impact into two categories one would have been the immediate complete stop of business which is very short lived and another is a general slowdown in economy which is a more medium term impact.
Natural calamities also have a similar impact in disrupting livelihoods though their geographic scale is limited. Such disruption could range from complete wipe-out in certain cases to loss of income for a few weeks. In both situations the borrowers generally fall behind in payment schedules.
The lenders have, since the AP crisis and recent floods, famines focused on diversifying their loan book across the length and breadth of the country. I will touch upon the loss bearing capacity of the business model later on but unless we see mega devastating climate change influenced natural disasters like in the movies, the industry can absorb the losses from such events.
Coming back to the corona virus, the impact is yet to be seen given the moratorium, delay in recognition of NPAs, restructuring of loan schedules. The collection efficiencies have inched up to 70-80% in August-September for most lenders. There is a big discrepancy between institutions as well in the way they calculate the collection efficiency. Generally it should be the collections of the months divided by the total amount that was due to be collected without the moratorium impact, meaning collections due in normal business course. We shall investigate from here on how much hit on collection efficiency the business model can structurally absorb before getting into trouble.
Let’s start with the book yield, the average return the lenders makes on the loan book. These generally range from 15% to 24% for most lenders. While institutions used to have some control over these yields due to lower credit penetration levels across untapped geographies, general regulatory oversight and competition are forcing these down.
Then comes the cost of funds, which is what these lenders pay to their depositors (in case of SFBs/Banks) or to the institutions they borrow from. These costs are a function of market confidence in your institution, both depositors and lenders will give you capital for lower interest rates only if they have confidence in your lending practices and ability to get back the capital that is lent on forward. These are also affected by liquidity in the markets. The cost of funds for SFBs and banks have dropped more than that for NBFC-MFIs because of their access to retail deposits which are generally cheaper capital sources to institutional loans.
What is left after the cost of funds is the interest spread. The financing margin the institution makes.
The chart I have below is the NIMs (net interest margin) which is slightly different in definition from interest spread but serves the same purpose.
After this come the cost of operations, these are the cost of running the branches, employee expenses etc. The have ranged from 4% for the most efficient operators to 8% of the loan book. The management has some control over these costs, but the microfinance model is a high touch one and requires alot of differentiated personnel, systems, and services to chase after smaller quantum of capital. The general practice is to improve branch and employee efficiency and have them disburse and collect as many loans as humanly possible. The group center meeting model required a certain time investment which limited the scalability of these productivity metrics.
With the social distancing practices becoming prevalent after the virus spread many lenders have shown interest in tweaking their model to adopt technological solutions that will enable them to increase their collection and disbursement productivity multi-fold. A few have not expressed any such initiatives.
Time will tell how well such solutions can be integrated into the current business model and how it will affect the credit discipline of borrowers. There are a lot of moving parts here and where the ball will roll is difficult to ascertain but one thing is clear, if the productivity increases dramatically for such business models, the industry is figuratively keeping a larger gold pile in the center of the castle and I am not sure whether their entry barriers derived from higher costs, lower productivity and first mover advantage will protect them from bigger competitors.
This is perhaps the reasons also why bigger banks find it problematic to fit this borrower group into their high loan value business operations. They prefer to do this via an organic or inorganic subsidiary to keep the two separate.
These are the major cost heads of the industry, what comes after this determines the quality of the management in the lending space. Being able to disburse to good borrowers who are willing and motivated to repay back the loans is a skill in itself. The job only gets harder as credit penetration levels increase in a country. The prime borrowers are essentially locked in by the biggest of the banking institutions, what are left are definitely sub prime in conventional sense.
These borrowers have essentially unaccounted income sources which are sometimes irregular and have no particular assets to mortgage for a loan. The management has to find new buffer systems that mitigate some of these risks. The group lending and joint liability model is almost 3 decades old now, and while it has worked beautifully in ensuring a social pressure to pay back the loans taken, a few institutions have removed the joint liability covenant but still continue with group lending model. Perhaps they feel that the social embarrassment of not being able to pay back a loan is deterrent enough to enforce borrower discipline.
These measures work in good times but whenever difficulties arise, the borrowers are still easily hit, as they have little savings to fall back on and cannot payback the loan even if they want to.
So, the credit costs for these institutions range from near 0.5% in good times to 5% when difficult times hit. The GNPAs are also under 2% for most of the time.
What is left after all these costs is the profit which is depicted by the ROA of the institution. As can be seen in the chart, the business model of all institutions have these costs baked in to the book yield to earn decent profits every year. The years in which exceptional disruptions happen the industry is able to pay for the losses from one year of profitability.
It remains to be seen whether the last statement above will remain true in the COVID aftermath. Most institutions have started providing for expected loan losses from the last quarter of the FY20. They have already built up provisions of anywhere between 1-2.5% of the book and they still have 3 more quarters of FY21 to provision from.
That being said, if you look at the averages, with a book yield of 20%, cost of funds of 9% and OPEX costs of 6%, the average institution should have the ability to absorb 5% of credit costs in a year before dipping in to their balance sheet reserves. This is the ideal scenario which says the industry can absorb losses up to a collection efficiency of 95% below that is troublesome.
We still have a long way to go from current 70-80% CE to 95-99% efficiency and it is the last bucket which is the most troublesome to regularize. Once the NPA recognition starts we will get to know the real damage and how long it will take to normalize operations. My guess is that it will take the industry 1-1.5 years to normalize back.
Disclaimer: This post is for research purposes only. Please do not construe this as a buy/sell recommendation.