Portfolio Turnover Explained: What Drives It, and What It Costs

Portfolio turnover measures how much of a portfolio was actually traded over a period, usually a year, compared with the size of the portfolio. A turnover of 100 percent means the fund or strategy did buying and selling equivalent to replacing its entire holdings once during the year. It matters because turnover is the meter on which trading costs and the timing of taxable gains are billed.

It is one of the few portfolio statistics that is almost purely about friction. It says nothing about whether the trading was any good. It only tells you how much of it there was, and therefore how large a hurdle the trading has to clear before it has added anything.

What the ratio measures

The idea in words: add up the value of what was bought and the value of what was sold during the period, compare that trading activity to the average value of the portfolio, and express it as a percentage.

The most common convention takes the lesser of total purchases or total sales over the period and divides it by the average net assets of the portfolio. Taking the lesser of the two is deliberate. It stops money flowing into or out of the fund from being counted as trading activity the manager chose to do. Indian mutual funds disclose a portfolio turnover ratio in their published factsheets on a broadly similar basis.

Other conventions exist. Some reports add purchases and sales together and divide by two. Some sum both and divide by average assets without halving, which roughly doubles the printed figure for the same underlying activity. Because of this, a turnover number from one platform is not automatically comparable to one from another. Always check whether the figure is one-sided or two-sided before comparing.

A useful mental translation: the reciprocal of turnover approximates the average holding period. Turnover of 100 percent implies an average holding of about a year. Turnover of 400 percent implies about three months. Turnover of 25 percent implies about four years. The approximation is rough, because it assumes trading is spread evenly rather than concentrated in a few positions, but it converts an abstract percentage into something intuitive.

What drives turnover

Turnover is mostly a consequence of design choices, not a dial someone sets directly.

• Signal speed. Rules built on fast-moving price data reshuffle their rankings often, because the underlying inputs change every day. Rules built on reported financials change only when new results arrive, which is a few times a year at most.
• Rebalancing frequency. Reviewing monthly gives the portfolio twelve chances a year to trade. Reviewing annually gives it one. Frequency multiplies whatever churn the rules would otherwise produce.
• Position count and weighting scheme. An equal-weighted portfolio has to trim winners and top up laggards at every review just to stay equal-weighted, which creates turnover even when the holdings do not change. A market-cap-weighted portfolio largely reweights itself as prices move.
• Rank instability at the boundary. Names sitting near the cut-off of a screen cross in and out repeatedly, generating trades without any real change in view. Buffer rules, which require a name to fall well below the threshold before it is dropped, exist to dampen exactly this.
• Flows. For a pooled vehicle, subscriptions and redemptions force buying and selling regardless of what the manager wants. This is precisely why the standard formula uses the lesser of purchases and sales.
• Events outside the portfolio. Index reconstitutions, mergers, delistings and corporate actions all force trades.

How to read it

Read it against the style, not against a universal standard. There is no single acceptable level. A slow fundamental approach and a fast price-based one are supposed to sit at very different turnover levels. The question is whether the figure matches the stated process. Turnover that is far higher than the described approach implies is a signal worth investigating, because it suggests the portfolio is doing something the label does not describe.

Convert it into a cost estimate. This is the practical use. Suppose, purely as an illustration, that a full round trip of buying and selling costs about 0.4 percent of the traded value once brokerage, statutory charges and the gap between the modelled and traded price are included. At 100 percent annual turnover that arithmetic implies roughly 0.4 percent a year of drag. At 300 percent turnover it implies roughly 1.2 percent a year. The numbers here are hypothetical and every desk’s real cost differs, but the structure of the calculation is the point: turnover multiplied by round-trip cost is the recurring hurdle the trading must clear.

Watch the trend, and separate chosen turnover from forced turnover. A figure that climbs steadily over several years without any stated change in process usually means either that ranks have become less stable or that the process has quietly changed. A one-year spike often means something duller: heavy redemptions for a fund, or an index review for a rules-based portfolio.

Put it next to the tax picture. Turnover determines how quickly unrealised gains become realised ones. In India, equity gains are taxed differently depending on how long the holding was held, with a shorter-term rate and a longer-term rate. Higher turnover means more gains crystallised at the shorter-term treatment and less of the compounding that comes from leaving a gain unrealised. This is a structural feature of the arithmetic and applies whatever the prevailing rates happen to be.

What it does not tell you

It does not tell you whether the trading made money. Turnover is a volume measure. A portfolio that traded heavily and profitably and one that traded heavily and destructively can print the same number. Only a comparison of returns net of costs, against a baseline, speaks to that.

It does not tell you what the trading actually cost. Cost per unit of turnover varies enormously with the liquidity of what is traded and the size of the position. Trading large, heavily traded companies in modest size is cheap. Trading small, thinly traded ones in size can cost several multiples of that. Two portfolios with identical turnover can face very different bills.

It does not show where the churn sat. A single annual number can describe a portfolio that trimmed every position slightly or one that completely replaced a quarter of its holdings while leaving the rest untouched. Those are different portfolios with the same statistic.

It does not distinguish deliberate from forced activity, unless the report breaks it out. The standard formula reduces the distortion from flows but does not remove the effect of index events or corporate actions.

It does not measure risk. A concentrated portfolio held for years without a single trade has very low turnover and can still carry substantial concentration and drawdown risk.

It is not comparable across sources without checking the definition. The one-sided and two-sided conventions differ by roughly a factor of two, and reporting periods vary, so an apparent difference can be pure arithmetic.

Read this way, turnover is best treated as a cost and tax meter rather than a verdict. It sets the size of the hurdle. Whether the trading cleared that hurdle is a separate question, and it needs a return comparison to answer.

Related reading

• Portfolio metrics explained: the hub that maps every risk, return and cost metric to the question it answers.
• Transaction costs in backtests: the brokerage, statutory charges and duties that turn gross results into net ones.
• Slippage and impact cost: the gap between the price you model and the price you actually get.
• Rebalancing frequency and backtest results: how the review schedule drives turnover and cost.
• Tax on portfolio rebalancing in India: what happens to gains when a position is sold rather than held.

This article is educational. Altys Labs is not a registered research analyst or investment adviser, and nothing here is investment advice or a recommendation to buy, sell, or hold any security.