# The real economics of fintech lending
A fintech advertises a personal loan at 15% APR. Sounds like a fat margin, right? Cost of funds at, say, 6%, so the lender pockets 9 points. Except that is not what happens. By the time you subtract losses, servicing, and acquisition, that "9 points" can shrink to almost nothing, or go negative.
APR (Annual Percentage Rate) is the top line. It is not the profit. Let us build the actual P&L, one loan at a time.
Think of every loan as its own tiny business. It earns interest and it incurs costs. The difference is your net interest margin, and then some.
Here is a simplified worked example. All figures are illustrative, not market quotes.
Say you originate a $10,000 unsecured consumer loan at 15% APR over a 3-year term.
Revenue (gross yield): 15%
Now subtract the costs, each expressed as a percentage of the loan balance per year.
This is what the fintech pays to get the money it lends out. Unless a lender is a chartered bank taking deposits, it borrows from somewhere: a warehouse line (a revolving credit facility from a large bank), a forward flow agreement (selling loans to an institutional buyer), or securitization (bundling loans into bonds sold to investors).
Assume cost of funds is 6%.
Running total: 15% - 6% = 9% left.
This is the big one, and the one newcomers underestimate. Some borrowers will not pay. Expected loss is usually written as:
Expected Loss = Probability of Default (PD) x Loss Given Default (LGD)
PD is the share of borrowers who default. LGD is how much you lose when they do (for unsecured loans, often most of the balance, since there is no collateral to seize).
Suppose 8% of borrowers default (PD) and you recover only 20%, so LGD is 80%.
Annualized expected loss is roughly 8% x 80% = 6.4%. For simplicity call it 6%.
Running total: 9% - 6% = 3% left.
Someone has to send statements, run the app, staff a call center, chase late payments (collections), and handle compliance. Call this 2%.
Running total: 3% - 2% = 1% left.
Fintechs spend heavily on marketing. If you pay, say, $200 to acquire a borrower on a $10,000 loan, that is 2% of the balance, spread over the loan life it might be roughly 0.7% per year. Call it 1% to be generous about efficiency.
Running total: 1% - 1% = roughly 0% left.
That is the punchline. A headline 15% APR can net out to near break-even before you have covered your fixed overhead, equity cost, or a single dollar of profit.
Notice that expected loss (6%) was the single largest deduction after cost of funds. This is why underwriting is the whole game in lending.
A 1-point swing in loss rate moves the P&L more than most marketing optimizations ever will. If your model is even slightly wrong about who defaults, you do not lose a little. You wipe out the margin.
This is also why fintechs obsess over the vintage curve: tracking how each monthly cohort ("vintage") of loans performs over time. Losses show up late, often 6 to 18 months after origination, so a book that looks pristine today can sour tomorrow. Growth hides bad underwriting until it does not.
The Consumer Financial Protection Bureau publishes accessible research on consumer credit performance if you want real-world context on delinquency patterns.
Now the part that flips a profitable book into a losing one.
Many fintech loans are fixed-rate: the borrower locks 15% for the full 3 years. But the lender's funding is often floating-rate: the warehouse line or securitization coupon resets with market rates.
So watch what happens when rates rise.
Before (rates low):
After (rates rise 3 points):
You just lost 3 full points of spread. Feed that through the same P&L:
15% - 9% funding - 6% loss - 2% servicing - 1% CACCACCustomer Acquisition Cost (CAC) is the total sales and marketing spend divided by the number of new customers gained in a period. It measures how efficiently you grow.View full definition → = -3%.
The book is now losing 3 cents on every dollar, on loans you cannot reprice because they are locked in. This is a classic asset-liability mismatch: long, fixed-rate assets funded by short, floating-rate liabilities. It is the same structural risk that has toppled lenders for centuries.
Rising rates rarely arrive alone. They usually come with economic stress, which pushes defaults up. So at the exact moment your funding gets more expensive, your PD climbs. Both the funding line and the loss line move against you at once. The margin does not just shrink, it collapses.
Higher funding costs force the lender to raise APRs on new loans to protect the spread. But higher APRs attract riskier borrowers (the ones who cannot get cheaper credit elsewhere), a phenomenon called adverse selection. Good borrowers walk away; risky ones stay. Loss rates on the new vintages climb further.
This is the trap. A fintech that grew fast in a low-rate era, funded short and lent long, can find every lever moving the wrong way simultaneously.
Knowledge check
1. Why is APR a poor proxy for a fintech lender's profitability on a loan?
2. A lender that is not a deposit-taking chartered bank must obtain the money it lends from external sources such as warehouse lines, forward flow agreements, or securitization. What does this imply for its economics?
3. Why is Loss Given Default (LGD) typically high for unsecured consumer loans?
4. Select ALL correct answers about how expected loss is derived and interpreted in loan unit economics.
Select all the correct answers.
5. Select ALL correct answers about treating each loan as its own tiny business (unit economics).
Select all the correct answers.
Smart lenders do not just hope rates stay low. They engineer around the risk.
Match funding. Fund fixed-rate loans with fixed-rate liabilities so both sides move together. Securitizations can lock in a fixed coupon for the life of the bonds, neutralizing the mismatch.
Hedge with interest rate swaps. A swap lets a lender exchange floating payments for fixed. If funding costs rise, the hedge pays off, offsetting the pain. It costs money up front, like insurance.
Shorter-duration products. Buy-now-pay-later (BNPL) loans and short installment plans repay in weeks or months, not years. Short assets reprice fast, so a rate shock hurts far less. This is part of why short-tenor products proliferated.
Risk-based pricing. Charge each borrower an APR that reflects their individual default probability, rather than one blended rate. This protects the spread when loss rates shift across segmentssegmentsDividing a market into distinct groups of customers who share similar needs, characteristics or behaviours, so each group can be served with a tailored approach.View full definition →.
Reserves and capital. Set aside expected losses in advance (an allowance), so a bad vintage does not become an existential event.
When you evaluate a lending business, ignore the marketing APR. Ask the P&L questions:
A fintech can post booming origination growth and glowing "revenue" while quietly running a negative unit economic loan book. The numbers that matter are buried below the top line.