The turnover tax on a factor

Pulled a factor study last week and did the boring thing. I read the footnote.

The headline premium was 5.7 percent a year over the market. Nice number. Then a line further down the page: after the cost of actually trading the strategy, closer to 4.2. About a quarter of the edge was gone before it reached a single real account.

Here is the part most pitches skip. That give-back is not a one off. It is turnover, and turnover is a bill you pay whether the factor works or not.

A patient rule, the kind built on value or quality, trades maybe a quarter of itself in a year. It keeps almost all of its edge. A momentum rule turns over one and a half times a year, so more of the premium leaks out to spreads and market impact. Same underlying idea. Very different bill at the end of it.

The research even draws a line. Trade slower than roughly half the book a month and a factor tends to keep a net edge. Trade faster and most of it disappears into the cost of chasing the signal. Short term reversal, rebuilt every single month, looks gorgeous on paper and is basically gone once you pay to trade it.

Honest counter, because it matters. Measured on live trades, real costs run lower than the scariest academic estimates, so the big factors do survive at scale. But survive still means less than the number on the brochure.

So two habits. Read the net premium, not the gross one on the slide. And treat low turnover as a feature, not a sign the manager is asleep. A rule that keeps more of the edge simply because it trades less is not lazy. That is a design choice

What one round trip costs in India

The study in the opening was about spreads and market impact. In India there is a second layer that no one can trade around: the statutory charges on every delivery trade. Here is one round trip, Rs 1,00,000 bought and later sold, at the rates in force in Oct 2026.

Add it up: Rs 200 + Rs 15 + Rs 6.14 + Rs 0.20 = Rs 221.34, or about 0.22 percent of the amount traded. That is before a single paisa of slippage. Slippage is the gap between the price you saw and the price you got, and it grows as you move down the size ladder. For the worked example I will assume 0.20 percent each side, 0.40 percent a round trip. Treat that as an illustration, not a measured figure: in the most liquid large caps it can be smaller, in thin small caps much larger.

So every full round trip costs about 0.22 + 0.40 = 0.62 percent.

Turnover turns that into a yearly bill

Turnover is how much of the portfolio gets sold and replaced in a year. A rule with 25 percent turnover replaces a quarter of its holdings. A rule with 150 percent turnover replaces the whole book one and a half times.

Take Rs 10,00,000 and run the three rules from the opening through the same 0.62 percent:

1. Patient rule, 25 percent a year: Rs 2,50,000 traded both ways, cost about Rs 1,550, or 0.16 percent a year.
2. Momentum rule, 150 percent a year: Rs 15,00,000 traded, cost about Rs 9,300, or 0.93 percent a year.
3. Monthly reversal, the whole book every month: 1,200 percent, Rs 1.2 crore traded, cost about Rs 74,600, or 7.5 percent a year.

Now look back at the footnote. A 1.5 point give-back, 5.7 down to 4.2, is what about 2.4 round trips a year cost at 0.62 percent each. That is a momentum style turnover. And the reversal strategy's 7.5 percent yearly bill is larger than most factor premiums ever were, which is why it "looks gorgeous on paper and is basically gone".

The full trade cost calculator does the same for a single order, line by line: trade cost calculator.

The other turnover tax is the actual tax

High turnover has a second bill, and in India it is often bigger than the first. Sell within 12 months and a listed equity gain is short term. Hold past 12 months and it is long term. Since 23 Jul 2024 the short term rate is 20 percent, and long term gains are taxed at 12.5 percent above Rs 1.25 lakh a year.

Same portfolio, same gain, a Rs 26,875 difference, and it comes from nothing but the holding period. A rule that rebuilds itself every few weeks realises almost every gain as short term. A rule that turns over a quarter of the book a year leaves most gains to age. None of this says a fast rule is wrong. It says the brochure's gross premium has to clear both bills before it means anything to you.

Reading a strategy with turnover in mind

When I look at any rule now, ours included, I walk the premium down to what reaches the account:

1. Gross premiumThe number on the slide, before any trading.
2. Statutory chargesAbout 0.22 percent a round trip, times turnover.
3. SlippageSpread and impact, bigger in thin stocks and fast rules.
4. Tax20 percent on gains realised inside a year, 12.5 percent after.
5. Net premiumThe only figure worth comparing.

A momentum rule can still be worth running after all four steps. A monthly reversal rule usually is not.

The free Learn module on trading costs walks through the full Indian cost stack, NSE's impact cost data and how backtest returns shrink on the way to a live account: The Real Cost of Trading. Two related posts go further: the factor zoo on why most published factors fail after costs, and factor crowding and decay on what happens to an edge once everyone trades it.