Rs 50,000 is not a round number somebody picked because it looks generous. In Indian credit data, Rs 50,000 is the ceiling that defines a small-ticket personal loan. That bracket is where fintech lenders hold 56.8 per cent of the market, where 70.5 per cent of the book is unsecured, and where delinquency ran 6.4 per cent in March 2026. About half of those loans went to people under 35.
THE VERDICT: THE SYSTEM
The Rs 50,000 in a pre-approved BNPL offer is not a generous round number. Rs 50,000 is the ceiling that defines a small-ticket personal loan in Indian regulatory data, and it is the exact segment where fintechs hold 56.8 per cent of the market and 70.5 per cent of the book is unsecured. The offer is calibrated to the top of the least-secured slice of Indian credit. That is a design decision, not a coincidence. What it is not, on the current numbers, is a solvency crisis.
Every BNPL offer you have ever been shown was sized against that line.
This is not a piece arguing that buy now, pay later is a scam. The numbers do not support that, and we will get to the numbers that argue the other way. It is a piece about where the actual manipulation sits, because it is not where the outrage usually points.
of small-ticket personal loans in India, the ones under Rs 50,000, are held by fintech lenders. 70.5 per cent of their loan books are unsecured.
Source: The Indian Express, 17 August 2026, reporting Reserve Bank of India and credit bureau data on household borrowing.
Why does every offer stop at Rs 50,000?
Because that is the top of the bracket.
Indian lending data splits personal loans by size, and “small-ticket” means under Rs 50,000. It is a reporting category, used by the Reserve Bank of India and the credit bureaus to track a slice of the market separately.
That slice behaves differently from the rest. The Indian Express reported on 17 August 2026 that fintech firms hold 56.8 per cent of small-ticket personal loans, and that 70.5 per cent of their loan books are unsecured. Banks do not dominate here. Fintechs do.
It is also the slice with the worst repayment. Delinquency on small-ticket personal loans stood at 6.4 per cent in March 2026. Roughly half of those loans went to borrowers under 35.
So an offer capped at Rs 50,000 sits precisely at the top edge of the least-secured, worst-performing, youngest-skewing part of Indian consumer credit. Go a rupee above and the loan lands in a different reporting bracket, a different competitive set, and a different level of attention.
Nobody has to have written that down for it to be true. Products get tuned to the segment they are measured in.
How bad is the repayment problem, really?
Smaller than the headlines, and worse than the averages. Both things are true, and the chart is the fastest way to see it.
Where the risk actually sits
Bad loans as a share of the book, March 2026, and the small-ticket comparison
Sources: gross NPA of 0.7 per cent on secured retail and 1.7 per cent on unsecured retail at end-March 2026, from the Reserve Bank of India Financial Stability Report, June 2026, as reported by Business Today, 22 July 2026. The 6.4 per cent small-ticket personal loan delinquency figure at March 2026 is from The Indian Express, 17 August 2026. The three figures are drawn from the same period but are not identical measures: 6.4 per cent is a delinquency rate on one loan category, the other two are gross NPA ratios across retail. They are shown together to compare scale, not as like for like.
Gross bad loans on secured retail credit were 0.7 per cent at the end of March 2026. On unsecured retail, all of it, 1.7 per cent. Those are low numbers. An Indian bank looking at that book is not in trouble.
Then the small-ticket category runs 6.4 per cent.
One caution on that comparison, because it matters. The 6.4 per cent is a delinquency rate on one loan category. The other two are gross NPA ratios across retail lending. They are not the same measure, and stacking them is a scale comparison, not a like-for-like one. The point survives the caveat: the small end is several times worse than the average it is hiding inside.
This is what concentration looks like. A portfolio can be healthy in aggregate while one segment inside it is carrying most of the damage, and that segment can be the one aimed hardest at the youngest borrowers.
So where is the manipulation?
Not in the notification. In the placement.
The bit that is actually the manipulation
It is not the push notification. It is the placement. A credit decision has been moved to the checkout screen, where the question in the buyer’s head is whether they want the product, not whether they want the debt. Every other consumer loan in India makes you go somewhere and apply. This one arrives as a payment option, sitting in a list next to a debit card, formatted like a way to pay rather than a thing to owe.
Consider what BNPL actually did to the sequence of events. Applying for consumer credit in India used to be a separate act. You went to a bank, or opened an app, and the whole exercise announced itself as borrowing. You were, at that moment, a person taking a loan.
BNPL moves the decision into the checkout, and puts it in the list of payment methods. Debit card. UPI. Pay in 3.
The question in your head at that moment is whether you want the shoes. It is not whether you want a debt. The interface has answered a credit question inside a shopping question, and it has done it by formatting the loan to look like a way to pay.
Rs 50,000 is not a marketing number. It is the top of the bracket where the least-secured lending in India lives.
Then the instalment does the second half of the work. A price of Rs 6,000 shown as “Rs 2,000 x 3” is the same money, but it is not the same number, and the number is what people compare against their balance.
The same rupee, framed two ways
How the offer is presented
- Pre-approved, so it reads as something you already earned
- A round number, so it reads as a limit rather than a debt
- Offered at checkout, where the decision is about the product
- Split into instalments, so the price shown is a fraction of the price
- No interest named, because the cost sits in fees and late charges
What the same rupee is in the data
- An unsecured personal loan, counted as household debt
- Sitting at the top of the under-Rs-50,000 bracket
- In the segment where fintechs hold 56.8 per cent share
- In a category running 6.4 per cent delinquency at March 2026
- On a credit file, where a missed instalment behaves like any default
What about the argument that BNPL is good?
It is a real argument, and the data behind it is stronger than most people criticising BNPL have bothered to look at.
THE OTHER SIDE
The moral panic is louder than the numbers justify, and the honest version has to say so. India’s credit-active population has gone from about 11 per cent a decade ago to roughly 28 per cent. A large part of that is people who were never going to be given a credit card. Gross NPA on unsecured retail was 1.7 per cent at March 2026, which is low by any standard. And the share of over-leveraged consumers actually fell, from 18 per cent in FY24 to 15 per cent in FY26. If BNPL were manufacturing a generation of defaulters, that number would be climbing. It is not. The system-level story is credit expansion that is mostly performing. The problem is concentration, not collapse.
Take that seriously before dismissing it. A rise in the credit-active population from roughly 11 per cent to about 28 per cent is millions of people getting a credit file for the first time. In a country where a formal credit history is the gate to almost every later loan, that is not nothing. It is the thing decades of financial-inclusion policy was trying to do.
And the over-leverage number is the one that should stop a lazy argument dead. The share of over-leveraged consumers rose from 5 per cent in FY17 to 18 per cent in FY24, and then fell to 15 per cent in FY26. If instalment credit were manufacturing a generation of defaulters, that line would not be falling.
So the honest verdict is narrower than the outrage. The system is not breaking. A specific segment is expensive, and the people in it are young.
What is actually changing?
The composition of Indian household debt, and quickly.
Non-housing retail loans were 58.4 per cent of household borrowing at March 2026, up from 54.9 per cent a year earlier. Household debt is now 45.5 per cent of GDP.
Read those two together. A rising share of a rising total. Indians are borrowing more, and a growing majority of it is not a house. It is consumption.
A mortgage buys an asset that is still there in year ten. A pay-in-3 on a pair of headphones does not. Both count as household debt in the same table, and only one of them leaves something behind.
That is the shift worth watching, and it is happening whether or not any individual BNPL product is well designed.
What should anyone do about it?
If your checkout offers BNPL, you are not adding a payment method. You are originating unsecured consumer credit at the point of sale, and the framing you choose is doing regulatory work. Naming the total repayable next to the instalment costs you almost nothing in conversion and removes the single most defensible complaint anyone can make about the flow. Also stop calling it interest-free in the headline if the late fee is where the economics live. That is the line that gets pulled first when the rules tighten.
Pre-approved does not mean you passed a test. It means the lender has decided the segment is worth the loss rate. Rs 50,000 is not a compliment. Before you tap accept, find the total repayable and the late fee, and read the instalment as what it is, which is a loan that will sit on your credit file whether or not you think of it as one.
The uncomfortable part
The regulator has the numbers. The bureaus have the numbers. None of it is secret, and none of it is new.
What has not happened is any requirement that the checkout screen carry the same information the regulatory table does. The customer sees an instalment. The system sees an unsecured personal loan in the worst-performing bracket it tracks.
Both descriptions are accurate. Only one of them is on the screen at the moment the decision gets made.
Sources: Fintech share of 56.8 per cent in small-ticket personal loans under Rs 50,000, the 70.5 per cent unsecured share of fintech loan books, the 6.4 per cent delinquency on small-ticket personal loans at March 2026 with roughly half extended to borrowers under 35, the over-leveraged consumer share moving from 5 per cent in FY17 to 18 per cent in FY24 and easing to 15 per cent in FY26, the rise in the credit-active population from about 11 per cent a decade ago to roughly 28 per cent, and average debt per borrower of Rs 4.78 lakh at March 2025 against Rs 3.41 lakh at March 2018: The Indian Express, 17 August 2026, reporting Reserve Bank of India and credit bureau data. Household debt at 45.5 per cent of GDP, non-housing retail loans at 58.4 per cent of household borrowing at March 2026 against 54.9 per cent at March 2025, and gross NPA of 0.7 per cent on secured retail and 1.7 per cent on unsecured retail at end-March 2026: Business Today, 22 July 2026, reporting the Reserve Bank of India Financial Stability Report, June 2026. Nothing in this piece alleges wrongdoing by any named lender, platform or individual. No BNPL provider is accused of breaching any law or regulation. It is analysis and opinion on published regulatory and bureau data about a category. Figures this piece deliberately does not use: a specific ‘pre-approved Rs 50,000’ push notification wording, and any claim about how often such notifications are sent, because neither could be traced to a named, dated source. The Rs 50,000 figure is used here only as the documented regulatory boundary of the small-ticket personal loan category. A limit on the data: the 6.4 per cent delinquency figure covers small-ticket personal loans as a category, not BNPL specifically. BNPL sits inside that segment. It is not the whole of it.
