Track 7 · Claims · lesson 7
The drag that compounds against you
11 min
Track 7 · Claims · lesson 7
11 min
One percent a year.
It is the most successful number in finance, and the reason is that it is quoted against the wrong thing. One percent sounds like a rounding error against a return of seven. It is not charged on the return. It is charged on everything you have, every year, including the parts that were charged last year and the year before.
A fee is a rate applied to the stock. A return is a rate applied to the same stock. They are the same kind of number and they belong side by side.
Against a 7% return, a 1% fee is not one part in a hundred. It is roughly one part in seven of the growth, taken annually, and it also removes every future year of growth that the money would have produced. Over decades the second effect exceeds the first.
Move the fee and watch the share of the outcome that goes elsewhere. The gross return, the balance and the period are held still; the only thing changing is the annual charge.
Fee drag
Hypothetical model, not a forecast
Compares a balance charged an annual percentage fee against the same balance charged nothing.
It assumes
It ignores
Hold the return at 7% and the period at thirty years. Find the annual fee at which roughly a third of the no-fee outcome is given up — a share between 0.32 and 0.35.
Charged on the whole balance each year, not on the gain.
| Year | No fee | After the fee |
|---|---|---|
| 0 | 100k | 100k |
| 1 | 107k | 106k |
| 2 | 114k | 112k |
| 3 | 123k | 119k |
| 4 | 131k | 126k |
| 5 | 140k | 133k |
| 6 | 150k | 141k |
| 7 | 161k | 150k |
| 8 | 172k | 159k |
| 9 | 184k | 168k |
| 10 | 197k | 178k |
| 11 | 210k | 188k |
| 12 | 225k | 200k |
| 13 | 241k | 211k |
| 14 | 258k | 224k |
| 15 | 276k | 237k |
| 16 | 295k | 251k |
| 17 | 316k | 266k |
| 18 | 338k | 282k |
| 19 | 362k | 299k |
| 20 | 387k | 317k |
| 21 | 414k | 335k |
| 22 | 443k | 355k |
| 23 | 474k | 376k |
| 24 | 507k | 399k |
| 25 | 543k | 422k |
| 26 | 581k | 447k |
| 27 | 621k | 474k |
| 28 | 665k | 502k |
| 29 | 711k | 532k |
| 30 | 761k | 563k |
Not there yet — keep moving the controls.
Somewhere around one and a third percent a year removes about a third of the outcome over thirty years. Not a third of the growth. A third of the whole final amount, against the version with no charge at all.
Now look at the two figures the model reports separately: the fees handed over, and the growth lost on top of them. The second is larger than the first over long periods, and it is the part nobody is billed for. Every unit taken in year three is also not earning for the following twenty-seven years, and that missing growth belongs on the invoice as much as the fee does.
Predict
A tax charged on gains as they are realised behaves almost exactly like a fee. If the rate is a fifth and the return is 7%, then realising every year costs about 1.4% of the balance annually — the same drag as the fee in the question above, arriving from a different direction.
Set the fee in the model to 1.4 and you are looking at a reasonable approximation of it.
That gives the mechanism that actually matters, and it is not the rate. It is when the charge lands.
A gain that is not realised is not taxed yet, and the amount not yet paid continues to compound for you. Over decades that deferral is worth a great deal, and it means two identical portfolios with identical returns can end up substantially apart purely because one of them was traded frequently and the other was left alone. The one that traded paid its tax early, in instalments, losing the compounding on each instalment.
This is the single largest reason turnover is expensive, and it sits on top of the transaction costs that turnover also incurs.
Being fair about this matters, because the arithmetic above can be read as an argument that all charges are theft, and that is not what it says.
A charge buys something: administration, custody, a structure, access to something otherwise unavailable, or somebody's judgement. Some of those are genuinely worth paying for, and a service that is worth 1% a year is worth 1% a year regardless of how large the compounded figure looks.
The useful question is therefore not whether the fee is small, but whether it is small relative to what it buys — and whether what it buys is available more cheaply elsewhere. The arithmetic in this lesson does not answer that. It tells you the size of the thing you are deciding about, which is normally much larger than the way it was quoted to you.
Check