Payback time: how many years until the money comes back
Stop thinking of a share as a quote that moves and think of it as a machine that produces cash. Then the only question is how long the machine takes to hand back what you paid for it.
What you'll take away
- Accumulate the cash year by year, with growth included, until the total reaches the price. That count is the answer.
- Eight years, because a conventional buyer accepts about sixteen. The threshold is half, like everything else here.
- A span of years is something you have intuition about. A multiple is not.
A machine, not a quote
Suppose someone offers you a machine for 1,000. It needs no attention and prints 100 a year, and the amount it prints rises about 8% a year because the thing it makes is in steady demand. You would not ask what the machine might resell for next quarter. You would ask how long until it has paid for itself.
That question is the entire model. Everything else is arithmetic you can do in a column.
The reason it is worth asking in exactly that form is that it changes what counts as a good outcome. A quote-based question is answered by somebody else’s willingness to pay more than you did. A payback question is answered by the business itself, on a schedule, whether or not anybody ever quotes you a price again.
Counting the years
There is no formula to memorise. You accumulate, and you stop when the running total passes the price.
| Year | Cash that year | Cumulative |
|---|---|---|
| 1 | 100 | 100 |
| 2 | 108 | 208 |
| 3 | 117 | 325 |
| 4 | 126 | 451 |
| 5 | 136 | 587 |
| 6 | 147 | 734 |
| 7 | 159 | 892 |
| 8 | 171 | 1,064 |
The total crosses 1,000 during year eight, so payback time is eight years. Note what growth did: without it, 100 a year against 1,000 would take ten years. Two years of the answer came from compounding, which is why this model cannot be collapsed into a single division.
Run it backwards and it becomes a price. If eight years is your limit, then the most you can pay is whatever eight years of cash accumulates to — 1,064 in this case. That is the entry price the model hands you, and it is the figure the calculator computes.
Why eight years
A buyer paying a market-average multiple is, roughly speaking, accepting that the business takes about sixteen years to return the purchase price. That is the ordinary bargain, and most of the time it is not an unreasonable one.
eight years = half of what the ordinary buyer accepts
So eight is not a discovery about businesses. It is the same halving that appears everywhere else in this method, expressed in years instead of in price.
Three differences from a multiple
Payback time and a price-to-earnings ratio both compare price to profit, and they are not interchangeable.
- Growth is inside the count. A multiple is a snapshot that treats this year’s profit as the whole story. Payback accumulates a growing stream, so two companies on the same multiple can be years apart.
- It runs on cash, not on reported profit. Net income includes non-cash charges and one-off items; the money that comes back to an owner is cash. Payback refuses the figure that can be shaped.
- A span of time has intuitive grip. "Sixteen times earnings" lands on most people as nothing at all. "Six and a half years to get my money back" — you know immediately whether you accept that, and whether you believe the business will still be there.
That last one is not a presentational nicety. Converting a ratio into a span of years recruits a judgement you already have — about how far ahead you can see — and puts it into the valuation, where it belongs.
Where it breaks
Everything in the count rests on one input, and free cash flow is the noisiest of the four growth metrics. Three handling rules follow.
- Use an average, not last year. Three to five years of free cash flow, averaged, keeps one unusual year from setting the whole schedule.
- When the dip is growth capex, substitute owner earnings. A company that spent heavily on new capacity will show a collapsed free cash flow and an absurd payback period. The cash did not disappear; it was deployed. Run the count on owner earnings instead and say that you did.
- Do not use it at all when cash flow is negative or flips sign. There is nothing to accumulate, and forcing it produces a number with no meaning. This is the case where you fall back on the other two models.
And there is one assumption the count never shows you: that the machine is still printing in year eight. Eight years of accumulation silently requires eight years of survival at something like current economics — not a quarter of it, not half. The arithmetic cannot test that, and no choice of input fixes it.
Which is the same conclusion all three models arrive at from different directions. The valuation prices the certainty; the moat is what supplies it. If you cannot say why this business will still be earning in eight years, the eight-year count is not a valuation — it is a schedule for an event nobody has promised.
Worked examples use invented figures to make the arithmetic visible. Moatrics covers US-listed companies.
All articles