Track 4 · The House · lesson 2
How a bank makes money
12 min
Track 4 · The House · lesson 2
12 min
Pays you 1.5 percent. Lends it at 6. The difference is the business, and it has been the business for about eight hundred years.
Everything else a bank does — the app, the branches, the card, the mortgage adviser — is either a way of collecting deposits or a way of placing loans.
A bank stands between two groups who cannot deal with each other directly.
Savers want their money back on demand, in small amounts, with no risk. Borrowers want large sums for twenty-five years and cannot promise anything. Those two requirements are incompatible, and the bank makes them fit by taking the mismatch onto itself.
That is a genuine service, and it is not free of danger. The bank has promised to return money on demand that it has lent out for a quarter of a century. The gap between those two promises is the reason banks exist and the reason they occasionally fail.
Size. Ten thousand deposits of 3,000 become one loan of 30,000,000. No saver could have made that loan; no borrower could have negotiated with ten thousand savers.
Time. Deposits are repayable today. Loans are repayable over decades. The bank absorbs the difference, which works perfectly until a large number of depositors want their money on the same morning.
Risk. Some borrowers do not repay. The saver is not exposed to any individual borrower; the bank is exposed to all of them, and prices that in.
For those three services it charges the gap between the two rates. Not a fee, a gap — which is why most people never think of themselves as paying a bank anything.
Predict
Below is a bank with a million of deposits. Everything that makes it profitable or not is on screen. The controls start at plausible settings; only the write-off rate moves.
Net interest margin
Hypothetical model, not a forecast
A bank's core economics: pay one rate on deposits, charge another on loans, add fees, subtract losses.
It assumes
It ignores
Everything else held still, find the write-off rate at which this bank stops making money altogether. Note how small it is.
| Year | Net interest income | Operating costs | Written off | Profit before tax |
|---|---|---|---|---|
| 0 | 0 | 0 | 0 | 0 |
| 1 | 36k | 17k | 4.3k | 19.9k |
| 2 | 37.4k | 17.7k | 4.4k | 20.6k |
| 3 | 38.9k | 18.4k | 4.6k | 21.5k |
| 4 | 40.5k | 19.1k | 4.8k | 22.3k |
| 5 | 42.1k | 19.9k | 5k | 23.2k |
| 6 | 43.8k | 20.7k | 5.2k | 24.2k |
| 7 | 45.6k | 21.5k | 5.4k | 25.1k |
| 8 | 47.4k | 22.4k | 5.6k | 26.1k |
| 9 | 49.3k | 23.3k | 5.8k | 27.2k |
| 10 | 51.2k | 24.2k | 6k | 28.3k |
Not there yet — keep moving the controls.
Under three percent. A bank lending at 6 against deposits costing 1.5 is wiped out by a write-off rate of about 2.8 percent of its book.
That is the single most important thing to understand about banking, and it explains almost every regulation attached to it. The margin is thin, the leverage is enormous, and a modest deterioration in the quality of the loan book is not a bad year — it is the whole business.
This is why a bank's most important decision is not the rate it charges. It is who it says no to.
An extra point of interest on a loan is worth one point. A borrower who does not repay costs the entire principal, which is worth twenty or thirty points of margin on somebody else's loan. Underwriting beats pricing, and it is not close.
The model has a fee income line at 0.6 percent of the book, which looks like a detail. In a real retail bank it is nothing of the kind — it is often a third or more of the total, and it behaves quite differently from the spread. That is the subject of the next lesson.
There is also a structural point hidden in the controls. Notice what happens when the deposit rate and the loan rate move by the same amount: nothing much. Notice what happens when they move by different amounts, which is what actually occurs — deposit rates tend to follow the market up slowly and come down quickly. Most of the variation in a retail bank's fortunes is in that lag.
Check
Who pays. The borrower pays the visible price, and the depositor pays an invisible one: the difference between what the money could have earned and what it was paid. Both halves of the spread come out of a customer.
Who benefits. The bank, on the gap. And genuinely, both customers — the borrower gets a mortgage no individual saver would have written, the saver gets instant access to money that has been lent for twenty-five years. Neither could have arranged that alone.
Who carries the risk. Mostly the bank, which is why underwriting is the whole job. But not entirely: the depositor carries the risk that the bank fails, which is why deposit protection schemes exist in most countries — and the state, and therefore the taxpayer, carries the tail of it.