What Are Trading Fees?
Trading fees are the total cost of transacting and holding a position, including commission, the spread, financing, slippage and any annual percentage charged on the balance. They are taken from winners and losers alike, and the annual ones compound against you.
Costs are the one input in trading you can know exactly before you start. They are also the one most people never total up, largely because they arrive in four different forms and only one of them appears on a statement.
How it works
There are four costs and they behave differently. Commission is charged per trade and is visible. The spread is taken from the fill price and is not. Financing is charged per night on borrowed size. An annual percentage — a fund’s expense ratio, a platform fee — is charged on the balance regardless of activity.
Only the first one gets compared. Brokers advertise on commission because it is the number people check, and a zero-commission account with a wider spread can easily be the more expensive of two. Spread covers that cost on its own.
And frequency multiplies all of the per-trade ones. The per-trade figure is designed to look negligible. The annual figure is the one that decides whether a strategy is viable.
A worked example
Start with the per-trade cost on this site’s shared series. A round trip costs 0.0098 against a median bar range of 0.493 — about 2% of a typical bar.
Scale it by frequency. One trade a day over 250 days costs 2.45 in price terms. Twenty trades a day costs 49.0 — roughly 99 times the median bar range, surrendered over a year before any judgement about the market is tested.
Now the annual drag, measured over thirty years on this site’s fee study:
| Annual charge | Cost of the final balance |
|---|---|
| 5 basis points | 1.5% |
| 20 basis points | 5.8% |
| 75 basis points | 20.2% |
| 150 basis points | 36.5% |
The relationship is not linear, and that is the finding. Going from 5 to 20 basis points — a change of fifteen hundredths of a percent a year — nearly quadruples the lifetime cost. Going from 75 to 150 takes it from a fifth of the balance to well over a third.
A tenth of a percent a year does not feel like a decision. Across thirty years it is a larger number than most people’s best trade.
The fifth cost
Tax is charged on the outcome, and it changes which strategies survive. Short holding periods are usually taxed at a higher rate than long ones, dividends at a different rate again, and every realised gain is an event whether or not the money was withdrawn.
That makes frequency expensive twice over. A strategy trading often pays the per-trade costs more times and converts unrealised gains into taxable ones earlier, which removes money that would otherwise have kept compounding. Two strategies with identical gross returns can differ substantially after tax purely on how often they transact.
What it costs to hold
Borrowed positions pay interest on the full size, not on the deposit, so a leveraged position held for weeks accumulates a cost several times larger than the capital committed would suggest.
Slippage is the fourth cost and the only one nobody can quote. A stop triggers and fills at whatever is available, which is reliably a little worse than the level chosen — so the realised average loss is consistently larger than the planned one.
And every one of these rises in thin conditions. The same instrument at the open, at the close and in the quiet hours is three different prices for identical activity.
The original data
Per trade: a round trip on this site’s shared series costs 0.0098, about 2% of the median bar range of 0.493 and 1.6% of the median ATR14 of 0.5994.
Per year at frequency: 2.45 at one trade a day, 49.0 at twenty.
Per lifetime as an annual percentage: 1.5% of the final balance at 5 basis points, 5.8% at 20, 20.2% at 75, 36.5% at 150.
Three different scales of the same problem. A day trader is destroyed by the middle figure and barely touched by the last. A long-term investor is the reverse. Almost all published advice about costs is written for one of those two people without saying which.
When it fails
The characteristic failure is optimising the visible cost and paying the invisible one. Commission is published, so it gets compared and driven to zero; the spread is embedded in the fill, so it does not get compared at all. An account moves to a zero-commission broker, the per-trade cost rises, and the change registers as an improvement because the only number anyone looked at went down. For an active strategy the hidden cost is usually the larger of the two, which makes the comparison exactly backwards.
A second failure is backtesting without costs. A strategy with a small measured edge and no cost assumption is not a strategy with a small edge — it is very often a losing one, and the backtest cannot tell you which.
A third is ignoring the annual percentage because it is small. It is small per year and it compounds, which is the entire content of the table above.
A fourth is assuming the stop caps the loss, when slippage and gaps make the realised loss reliably larger than the planned one.
And a fifth is treating costs as a detail rather than a lever. They are the only variable in trading that is knowable in advance and reducible by decision. Expectancy does the arithmetic: below a certain target size the edge is gone to costs entirely, however well the trading is executed.
Related
Spread covers the largest hidden per-trade cost. Expectancy covers where costs enter the arithmetic and what they rule out. And expense ratio covers the annual percentage over long horizons.
Costs are the only part of trading that is completely certain in advance, which makes them the only part you can improve with certainty. Everything else is a probability. I spent years trying to raise a win rate by a point or two and the same effort spent on frequency and spread would have done more, faster, with no uncertainty attached.
— Michael Whitman
This page is educational, not financial advice. Test every idea on your own charts before risking money.