Lesson 7 of 7
What trading costs an edge
Spread, commission and tax come off a measured edge before anything reaches the account. Compute the drag, and what deferring a tax bill is worth.
By Javier Sanz
Founder of ValueScreener. Built Ninety Nine, a securities services provider, and sold it to a tier-one global bank; ran operations and internationalization at Alpaca, the brokerage API.
Who writes thisEditorial processLinkedIn (opens in a new tab)X (opens in a new tab)
In one minute
- Every round trip pays a spread and a commission, so the cost of a rule is its turnover multiplied by the cost of one round trip.
- Tax is not only a rate: paying it later rather than each year is worth a large amount over a long horizon.
- Both drags are certain, and the edge they are subtracted from is not.
An edge is measured before costs. An account experiences it after them. The gap is computable, it is larger than most readers expect, and it is the one part of the return that does not depend on being right.
Cost drag at three levels of turnover
- 25 percent turnover a year0.15%
- 100 percent turnover a year0.60%
- 200 percent turnover a year1.20%
The first subtraction
Turnover of 100 percent means the portfolio is replaced once in the year, which is one round trip per unit of capital. At 0.6 percent a round trip that is 0.6 percentage points of return, gone before anything else happens.
The number is not a constant. A round trip in a large, heavily traded company may cost a tenth of that, and one in a small company with a wide spread may cost several times more. The arithmetic does not change: cost drag is turnover multiplied by the cost of one round trip.
The second subtraction, which is about timing
That is the compounding argument for low turnover, and it is stronger than the commission argument. A tax deferred is capital still working; a tax paid is capital that has left the account and cannot compound.
Putting both against a measured edge
What is left of three points
-
Start with an edge of 3 percentage points a year, which is generous after the publication decay measured elsewhere in this academy.
-
Subtract the cost drag of a rule that turns over once a year: 0.6 points, leaving 2.4.
-
Note that the same rule realises its gains annually, so it is on the 6.4 percent path in the table rather than the 8 percent one.
-
What reaches the account is smaller than the headline by an amount that is certain, while the headline itself is an estimate.
The break-even a reader can compute
There is a useful number hiding in this. A trade has to beat the cost of making it, and the cost is known before the trade.
Selling one holding to buy another costs a round trip, say 0.6 percent, plus whatever tax the sale triggers. If the sale realises a 40 percent gain at a 20 percent rate, that is 8 percent of the position gone immediately. The replacement holding therefore has to be better than the one being sold by roughly 8.6 percent before the switch has paid for itself, which is a very large difference to be confident about.
That calculation is why a taxable account and a tax-free account are genuinely different instruments running the same method, and why a rule designed in a backtest with no tax line is describing an account nobody has.
What this argues for, and what it does not
It argues for low turnover, for caring about the spread in small companies, and for knowing whether an account is taxable before choosing a method. None of that requires any view about whether markets are efficient.
It does not argue that trading is always wrong. A guardrail that breaks is a reason to act, and refusing to act in order to protect a tax position is how people end up holding businesses they no longer believe in. The arithmetic sets the price of a decision. It does not make the decision.
Check yourself
A portfolio turns over twice a year at a round trip cost of 0.6 percent. What is the drag?
- 0.6 percent
- 1.2 percent
- 0.3 percent
Show the answer
Answer: B.
Cost drag is turnover multiplied by the cost of one round trip, so two round trips at 0.6 percent is 1.2 percentage points a year.In the example, what does paying the same tax rate annually rather than at the end cost over twenty years?
- Nothing, the rate is the same
- About 47 percent of the starting amount
- The full 20 percent rate
Show the answer
Answer: B.
The annual path ends at 3.46 times the starting amount and the deferred path at 3.93 times, a difference of 0.47 times the start.Why is the cost drag more certain than the edge it is subtracted from?
- Because costs are set by law
- Because the spread and the tax are paid whatever happens, while the edge is an estimate that may not repeat
- Because commissions never change
Show the answer
Answer: B.
Costs arrive with certainty on every trade. The premium being measured is an estimate from past data that has already shrunk once on publication.Sources
- Computed for this lesson. The 0.6 percent round trip, the 8 percent return, the 20 percent tax rate and the twenty year horizon are stated assumptions, and every figure follows from them by arithmetic.
What this leaves out
- Trading costs vary enormously by company size, order size and market conditions. The 0.6 percent figure is a stated assumption, not a measurement.
- Tax rules differ by country, by account type and by holding period, and nothing here describes any reader's tax position or constitutes tax advice.
- The comparison assumes gains are realised annually in one case and once at the end in the other. Real accounts sit between the two.
- Nothing here is advice about any security or strategy.