Microfinance: joint-liability lending, collection efficiency and cyclical stress

A microfinance lender reports a flawless run - profit compounding, bad loans near zero, borrowers numbering in the millions - and then, in a single quarter, its collections falter, its bad loans explode, its credit costs multiply, and its profit is nearly wiped out. To a reader watching the profit line the reversal looks sudden and inexplicable. It is neither. The warning was there a quarter or two earlier, in a number the profit statement does not feature: collection efficiency, the share of instalments actually collected. When that number dips, the crash is already coming; the P&L is simply the last to know. illustrative

A microfinance institution lends tiny, unsecured loans to low-income borrowers, traditionally through a joint-liability group - a small group, usually of women, who guarantee each other's loans, so peer pressure substitutes for collateral. In normal times this works remarkably well and collections run at 98-99%. But the borrowers are economically fragile and geographically clustered, so a local shock - a flood, a crop failure, a political disturbance, a wave of over-indebtedness - can make a whole area's borrowers stop paying at once. Collection efficiency falls first, the loans age into non-performing, the lender must provide against them, and only then does the profit collapse. So this module reads a microfinance lender through its leading indicator, not its lagging profit - collection efficiency as the early warning, the joint-liability structure and geographic concentration as the shape of the risk, and the quality of growth as the thing that decides how bad the next shock will be.

Each period a microfinance lender has a schedule of instalments due, and collection efficiency is the share it actually collects. At 98-99% the book is calm and the small wobbles are noise; a fall - to 95%, 92%, lower - means borrowers have started missing payments. Because the loans are weekly or fortnightly and the borrowers fragile, that dip shows up before the loans have aged enough to become non-performing, so it leads the NPA, the credit cost, and the profit. The reader watching collections sees the trouble a quarter or two before the reader watching the profit does.

From there the sequence runs mechanically. Missed instalments age into portfolio-at-risk (PAR) - loans overdue by 30, 60, 90 days. As they age, the lender must provide against them, so the credit cost spikes; and because a microfinance lender's profit is a thin margin on a large book, that spike can crash the profit outright. Then, as the shock passes and collections recover, the sequence reverses and profit rebounds. The profit is the last and most amplified point in a chain that begins with the collection number.

Two structural features decide how violent that chain is. The joint-liability model makes collections resilient in good times, because peer pressure keeps members paying - but it can amplify a shock, since a local event stresses a whole area at once and the peer guarantee fails precisely when everyone needs it. And the model clusters borrowers geographically, so a book concentrated in one region can see a large share crack from a single event, where a book spread across many districts and many small groups barely registers the same shock.

Growth quality does the rest. Microfinance profits are high in the good years, which tempts lenders to grow fast - and the dangerous way to grow is to lend larger tickets to borrowers who already carry loans, over-indebting them. Over-indebted borrowers keep paying while their incomes hold, so near-term profit and collections look fine; but they have no buffer, so a shock turns them into mass default. This is why a shock that hits a concentrated, over-lent book does damage that a diversified, prudently grown one recovers from and a concentrated one does not. Growth from reaching more borrowers at sustainable ticket sizes is healthy; growth built on rising tickets to the existing base is fragility being manufactured, waiting for an event to detonate it.

A microfinance lender reads unlike its cousins in lending, and unlike them the numbers that reassure most are the ones that move last.

The inversion is temporal, not cross-sector. A stable business's profit tells you where it is; a microfinance lender's profit tells you where it was a quarter or two ago, before the collection dip already underway feeds through. Its most reassuring numbers - a healthy profit and a low NPA - are the last to move, so they lull a reader into safety exactly when the leading indicator is flashing. And the joint-liability model that makes collections so reliable in good times is the same structure that turns a local shock into a correlated wave of defaults. So the reader inverts the instinct: distrust the profit and the current asset quality, and read instead the collection number that leads them, the concentration that shapes the severity, and the growth quality that determines the fragility. (A business outside lending altogether - an FMCG maker, say - has no collection cliff at all, and its reported profit is the fair, timely guide a microfinance lender's never is.)

Read the composite microfinance lender across its five years, and watch the third - the stress year - where the cascade runs in full. In the calm years, collection efficiency held at 99%, PAR-30 sat at 1.5%, credit cost was a normal 2%, and profit compounded from ₹400 crore to ₹520 crore. Everything looked pristine. Then a shock hit, and collection efficiency dipped to 92%. That single number was the warning, and it moved first. illustrative

The cascade followed, with a lag. The missed instalments aged into portfolio-at-risk, which spiked from 1.5% to 6%. The lender had to provide against them, so credit cost jumped from 2% to 7% of the book. And that spike, hitting a thin-margin lender, crashed profit from ₹520 crore to ₹180 crore - a two-thirds fall in a single year. A reader watching only the profit would have seen a sudden, shocking collapse; a reader watching collection efficiency would have seen the 92% dip a quarter or two earlier and known the credit-cost spike and the profit crash were on their way.

Then the recovery, which the same indicator foreshadowed. As the shock passed, collection efficiency climbed back to 98% and then 99%, PAR normalised, credit cost fell to 3% and then 2.2%, and profit rebounded to ₹650 crore and ₹900 crore - above the pre-stress level, as the lender resumed growth on a cleaned-up book. The collection number led both the fall and the recovery. And the book's shape decided the depth: a diversified book lent at sustainable ticket sizes weathered the shock and bounced, where a concentrated, over-indebted one would have fallen deeper and recovered slower.

Drag the collection efficiency down and watch the chain react - PAR climbing, credit cost jumping, projected profit collapsing - while the reported P&L would show none of it for a quarter or two. The tool makes the module's point physical: the collection number is the early warning, and by the time the profit reflects the stress the damage is already done. The habit to build is to read collection efficiency first, every quarter, as the leading indicator; read the geographic diversification and joint-liability structure for how correlated and severe a shock would be; read the growth quality for how much fragility is being built; and treat the reported profit and current NPA as the lagging numbers they are.

Collection efficiency leads the profit, but it cannot forecast the shock that will move it. The events that crater microfinance collections - a flood, a crop failure, a political disturbance, a loan-waiver announcement, a wave of over-indebtedness cresting - are largely exogenous, and the number only dips once a shock has begun to bite. So it gives you the earliest warning available in the accounts, not early warning of the world: a book that looks pristine can be one event away from stress.

Nor can the reported number always be trusted at face value, because it can be managed. A lender under stress can flatter its collection figure by restructuring loans, refinancing struggling borrowers with fresh ones (evergreening), or reporting on a basis that folds in prepayments and advance instalments. A suspiciously stable collection number through a visibly stressed period is itself a warning, and the honest read looks at the definition, the restructured book, and whether growth is being used to paper over non-collection.

And the sector carries political and regulatory risk no operating metric captures. Microfinance lends to a politically salient constituency, so loan waivers, interest-rate caps, tighter lending norms, or local interference with collections can reshape the economics or trigger mass defaults overnight - policy decisions, not credit events the numbers foresee. A well-run lender with pristine collections can be hit by a state announcing a waiver. The metrics read the credit health of the book; the political fragility of lending to the poor is a separate risk they gesture at but cannot measure.

When a microfinance lender reports strong numbers, a sharp analyst goes straight to collections and concentration: "Profit and NPA look healthy, but what was collection efficiency this quarter and last, is any of it flattered by restructuring, how concentrated is the book by state, and how much of your growth is bigger tickets to existing borrowers?" The question reads the leading indicator, its honesty, and the fragility being built.

A candid answer sounds like this:

"Fair to focus there. Collection efficiency was 99.2%, steady, and that's pure collections - no restructuring or advance instalments flattering it; our restructured book is negligible. Our largest state is 14% of the book and no district is over 3%, so we're diversified against local shocks. On growth, about 80% is from new borrowers at steady ticket sizes; we've deliberately capped exposure per borrower to avoid over-indebtedness."

It gives the collection number and confirms it is unmanaged, quantifies the diversification, and shows the growth is broad-based with per-borrower caps - enough to judge both the leading indicator and the fragility.

An evasive one sounds like this:

"We're delighted with our strong profitability and pristine asset quality, reflecting our robust underwriting and deep customer relationships. Collection efficiency remains healthy and we continue to grow rapidly across our markets. We're confident in our risk management and well-positioned for the future."

It cites "pristine asset quality" (a lagging number) and "healthy" collections without the figure or its basis, and says nothing about concentration or whether the rapid growth is over-indebting the base. A lender flattering collections through restructuring, or growing by over-lending, would answer in exactly these words.

The follow-up nobody asks is the one that forces the truth open: "What is collection efficiency excluding restructured and refinanced accounts, and what is your exposure to your single largest state and to borrowers with loans from three or more lenders?" If "pristine asset quality, healthy collections" is allowed to stand unchallenged, an investor is trusting lagging numbers and a possibly-flattered collection figure while the leading risk builds. The silence is the tell.