Portfolio Turnover and Its Cost

Turnover compounds spread, market impact and financing costs, so a strategy's realistic net Sharpe often differs sharply from its paper one.

Portfolio turnover measures how much of a book gets replaced over a given period, and its cost compounds through bid-ask spread, market impact and financing so that a strategy's paper Sharpe ratio can look very different once realistic trading costs are applied. A strategy that appears attractive on a backtest computed from mid prices can turn unprofitable once every trade is charged the spread it would actually have paid, and the effect grows directly with how often the book turns its positions over.

Annual turnover is typically calculated as the sum of buys and sells over a year divided by average assets under management, and its inverse gives an approximate average holding period. A strategy with turnover of 20 times a year holds a position for roughly 1 divided by 20 of a year, close to two and a half weeks, before replacing it, while a strategy with turnover of 2 times a year holds positions for roughly six months on average.

Every round trip trade pays the bid-ask spread, plus market impact that grows with how large the order is relative to available liquidity, plus any commission and, for short positions, a borrow cost charged for the life of the position. Slippage, execution costs and market impact covers how the market impact component scales with participation rate and order size, but the simple version is that impact rises faster than linearly as an order consumes a larger share of available volume, so cutting order size in half more than halves the impact cost per share.

A strategy with 20 times annual turnover and a round trip cost of 5 basis points, a reasonable illustrative figure for a liquid large cap name traded carefully, pays roughly 100 basis points of annual return in transaction costs alone, before financing costs on any short positions are added. A strategy whose raw signal generates 150 basis points of annual excess return before costs is left with only a third of that after this kind of turnover drag, which is why cost modelling has to happen before a strategy is sized, not after.

Financing cost adds a further layer for any strategy running short positions or using leverage, since a short position pays a borrow fee for as long as it is held and a levered long position pays financing on the borrowed capital, both of which scale with holding period rather than with the number of trades. A strategy with a low trade count but long holding periods on borrowed capital can therefore still carry a meaningful cost drag even though its turnover, measured strictly as trading frequency, looks low.