Ask most people what their investments cost and you'll get a shrug, or a number they half-remember from a sign-up page. Ask them what their phone plan costs and you'll get it to the cent. The asymmetry is not stupidity — it's design. Investment costs are quoted in percentages, deducted before the number you see, and spread across five separate layers that no single statement adds up for you.
A fee autopsy is the exercise of adding them up anyway. It takes about 90 minutes with your statements open, it produces one number, and that number changes how you look at everything else. This is a framework for finding and pricing your costs — not a recommendation about what to own or where to hold it. Those calls depend on your situation, your tax position and your timeline.
The five layers
Costs stack. Each layer is quoted separately, in its own units, and often by a different party — which is exactly why almost nobody totals them.
- Product cost. The ongoing charge inside each fund or ETF you hold — expense ratio, OCF, whatever the local label is. Quoted annually as a percentage, deducted daily from the fund's value. You will never see it leave.
- Platform cost. What the broker, custodian or pension provider charges to hold the account. Sometimes a percentage, sometimes a flat annual fee, sometimes both with a cap. The flat-vs-percentage distinction matters enormously as balances grow.
- Advice or wrapper cost. An adviser's ongoing percentage, a robo-adviser's management fee, or the extra layer a managed portfolio or insurance wrapper adds on top of the funds inside it.
- Transaction cost. Commissions, bid-ask spread, foreign exchange margin on non-domestic assets, and the internal turnover cost inside actively traded funds. FX margin is the most-missed line by a distance — it can dwarf the commission next to it.
- Tax drag. Not a fee, but it behaves exactly like one: tax on distributions and realised gains in accounts without shelter. It's the only layer you can sometimes reduce without changing a single holding.
Step 1 — Build the ledger
One page. One row per account. Do not estimate — look each number up, because the gap between what people assume and what they pay is usually the whole point of the exercise.
Three rules for filling it in. First, weight by balance: a 1.2% fee on a small account matters less than 0.4% on your biggest one. Second, convert flat fees to percentages by dividing the annual amount by the balance, so everything is comparable. Third, if you can't find a number in ten minutes, that opacity is itself a finding — write "unknown" and treat it as a flag, not a zero.
Step 2 — Turn percentages into money
Basis points don't feel like anything. Money does. Multiply your weighted all-in percentage by your total balance and write that number down in your own currency.
Worked example. A portfolio of $120,000 across three accounts:
- Product cost, weighted: 0.52% (a mix of index funds at 0.15% and two legacy active funds at 1.1%)
- Platform: 0.25%
- Advice/wrapper: 0.10% amortised
- Transaction and FX: 0.07% estimated from the year's trades
- All-in: 0.94% → $1,128 per year
That is the number. Not "under one percent." One thousand, one hundred and twenty-eight dollars, paid annually, whether the portfolio rose or fell.
Step 3 — Run the 20-year drag
The annual figure understates the damage, because every dollar of fee is also a dollar that never compounds. Do this arithmetic once and you won't forget it.
Take a gross return assumption — 7% is a common planning placeholder, and the specific number matters less than the comparison. Then:
- $120,000 at 7.00% gross for 20 years → about $464,000
- Same money at 6.06% (7% minus the 0.94% all-in) → about $389,000
- Total cost of the fee stack: roughly $75,000 — about 16% of the ending balance, from a number that looked like rounding error.
Now the actionable half. Suppose the same portfolio could be held for 0.24% all-in — a plausible floor for a plain index-and-platform setup in most markets:
- $120,000 at 6.76% for 20 years → about $444,000
- Difference versus the 0.94% version: roughly $55,000.
Our take: $55,000 is not a fee argument, it's a life decision that happens to be denominated in basis points. And unlike returns — which you don't control — costs are the one input you can set with a form. That asymmetry is the entire case for doing this exercise once a year.
Step 4 — The swap test
Cheapest is not automatically right. Some fees buy something real. The swap test forces the question honestly: for each layer, name what you get for the money, then name what you'd lose by moving to the cheapest credible alternative.
Run that block for every layer. The pattern that usually emerges: product cost and platform cost almost never survive the test (nobody has ever received a better index return for paying more), while advice sometimes does — if you can name a specific occasion it earned its keep. "It's reassuring" is not a specific occasion. If you can't remember what the fee bought in the last two years, you're paying a subscription to a service you've stopped using, which is the same problem as any other stack you've never audited.
Step 5 — Before you change anything, price the exit
This is the step that separates an autopsy from a costly stampede. Moving has its own costs, and they're front-loaded:
- Realised gains. Selling inside a taxable account can trigger a bill that takes years of fee savings to repay. Calculate it before you move, not after.
- Exit and transfer fees. Per-holding transfer charges still exist. Ask for the schedule in writing.
- Time out of the market. An in-kind transfer keeps you invested; a sell-and-rebuy doesn't. Ask which one is on offer.
- Lost features. Some wrappers carry benefits — protected allowances, employer matching, guarantees — that don't survive a move at any price.
The rule of thumb worth writing down: if the switching cost exceeds three years of savings, the case has to be strong for reasons beyond fees. If it's under one year of savings, the arithmetic is doing the arguing for you.
Failure modes
- Chasing the last two basis points. Going from 0.94% to 0.24% is transformative. Going from 0.24% to 0.22% is a hobby. Effort should follow the size of the gap.
- Confusing cost with value. A 0.6% fund that gives you exposure you genuinely can't get for less is a different thing from a 0.6% fund tracking the same index as a 0.1% one. Compare like for like, or you're comparing nothing.
- Ignoring the flat-fee crossover. Percentage platform charges are cheap when balances are small and punishing when they're large; flat fees are the reverse. There's a crossover balance — find yours, and diarise a re-check when you approach it.
- Auditing costs while ignoring behaviour. The most expensive line in most portfolios isn't a fee at all; it's the trades made during a drawdown. Fee discipline and a written plan for volatile stretches are the same project — the Selloff Playbook covers the other half.
- Doing it once. Fee schedules change quietly. So do your balances, which changes which structure is cheapest.
How you know it's working
Three checkpoints, one page:
- You can state your all-in number from memory — as a percentage and as an annual dollar figure. If you can't, the autopsy didn't land.
- Every "unknown" from Step 1 is resolved. Opacity is a cost you haven't measured yet. Chase each one until it's a number.
- The number moves in the direction you chose. Re-run the ledger annually — pick a fixed month and put it on the calendar next to your other money reviews, like the mid-year reset. A falling all-in percentage with unchanged exposure is the cleanest evidence that the work paid.
Nobody controls returns. Everybody controls the toll booth. Find out what yours charges.