Tracking error: how far a fund lags behind its index
· 6 min · beginner
Automated material · TradeAlmanac editorial deskDraft prepared by a language model from our stored data; not reviewed by an editor.
Tracking error is a measure of how far a fund's return diverges from the return of the index it has promised to replicate. The divergence is almost never zero, and that is not a defect but the price of how the structure is built: an index is a calculated value with no costs, while a fund is a real portfolio that pays a fee, keeps a cash balance in its account, buys and sells securities at market prices and lives inside a tax regime. The question to ask of a fund is not "does it lag or not" but "by how much and for what reasons" — because some sources of the lag are predictable and constant, while others mean that the manager is not coping with the mechanics of replication.
Lag and dispersion are different questions
In everyday use "tracking error" means one thing, and in fund reporting it means another, and mixing these meanings up spoils every conclusion.
The first is the accumulated lag: the difference between the return of the fund unit and the return of the index over a period. This answers the question "how much have I lost compared with the benchmark". It is a directional figure: it can carry a minus sign, and sometimes a plus.
The second is the dispersion of the differences: how evenly the fund follows the index day after day. It is calculated as the standard deviation of the difference between the daily returns of the fund and of the index. Here the sign does not matter; what matters is stability. A fund may have a small accumulated lag and still jerk around the index — which means the replication is held together with stopgaps and will break in a crisis at exactly the moment when that is most expensive.
A good index fund is the kind whose lag is close to the sum of its explicit costs and whose dispersion is small and not growing. The site keeps the terminology separately: the glossary definition — tracking error.
Where the lag comes from
The management fee. The most predictable part. It is deducted from the fund's assets daily and reduces the unit value regardless of whether the market is rising or falling. This part of the lag can be estimated in advance from the fund's trust management rules.
Trading costs. The index is reviewed on a schedule: the constituents change and the weights are recalculated. The fund has to repeat this with real trades — with a spread, with slippage, sometimes against funds of the same kind that need the same securities on that day. The index does not bear these costs.
The cash balance in the portfolio. The fund needs free cash for payouts and to prepare for rebalancing. In a rising market the uninvested balance drags the return down; in a falling market, on the contrary, it helps a little.
Incomplete replication. If the index holds many securities, or some of them are illiquid, the manager buys not the whole list of constituents but a sample chosen so as to reproduce the behaviour of the index. A sample diverges from the benchmark more than a full copy does, and it diverges unpredictably.
Taxes inside the fund. Dividends and coupons paid into the fund's assets fall under a tax regime, while a total return index is calculated without it. The specific regime depends on the type of fund and on the current version of the Tax Code — this is the point where it is worth reading the fund's rules rather than a journal article.
Currency conversion. If the index is calculated in one currency and the unit in another, a further source of divergence is added, and it has nothing to do with the quality of replication.
Why a fund sometimes "beats" the index
The most common reason for outperformance is an error in the comparison, not the manager's prowess. An index has a price version and a total return version: the first takes into account only the change in prices, the second also the reinvested dividends. The fund receives dividends and reinvests them, so when it is compared with the price version it is bound to look better than the benchmark. That is arithmetic, not skill. The site keeps the payout calendar that creates this difference separately: upcoming dividends.
| Security | Dividend | Record date |
|---|---|---|
| CNRUIPJSC Cian | ₽33.00 | 30 Nov 2026 |
Genuine outperformance happens too: income from a tax refund, from taking part in offerings, from selling rights — all of this settles in the fund and is not reflected in the index. But for an index fund, a persistent lead over the benchmark is a reason to check which index it is actually measured against.
How to calculate it yourself
Step 1 — take the right benchmark: the index named in the fund's rules, in its total return version.
Step 2 — take the fund's return based on the calculated unit value, not on the exchange price. The exchange price includes a premium or a discount to net asset value, and that is a separate effect that has nothing to do with replication.
Step 3 — align the dates. A mismatch of one trading session in the start and end dates can distort the result more than a year's fee.
Step 4 — compare the lag you obtain with the fund's stated costs. If the lag is noticeably larger, the remainder is what the manager is losing on the mechanics.
What tracking error does not measure
It does not measure the quality of the index itself. A fund may replicate perfectly a benchmark that describes the market badly — and the reader will get precise tracking of a poor idea. To discuss how the instrument itself reacts to the market, a different measure is needed — beta.
It does not measure the liquidity of the unit on the exchange. A narrow spread and the presence of a market maker are a separate characteristic, and it is exactly this that determines what getting in and getting out will cost you.
It does not measure the risk of the portfolio. The lag is counted from the index, not from zero: a fund with minimal tracking error falls together with the market by exactly as much as its benchmark falls.
And it does not protect against a biased selection of funds: comparison tables usually keep the survivors, while closed funds disappear from the statistics — a separate trap, examined in the article on survivorship bias.
Where to look for the data
This article does not give specific lag figures for individual funds: they change and are calculated against different benchmarks, so their place is in the instrument profiles, not in the text. It is worth starting with the list of funds, and looking at the composition and behaviour of the underlying assets on the stocks and bonds pages.
Draft prepared by a language model from our stored data; not reviewed by an editor.
Model: claude-opus-5
How we use language modelsSimilar articles
- Crypto ETFs: what the fund actually holds and what you pay to the coin holderA crypto ETF is an exchange-traded fund whose units trade like an ordinary share, backed either by the coin itself held in a custodian's accounts or by derivatives on it.
- How to choose a fund: five checks in orderA fund is chosen not by its name or by a chart of a past period but by a sequence of checks: what is inside, how much the management company charges for its work, how closely the fund tracks its benchmark, how it trades on the exchange and what legal wrapper all of this comes in.
- Gold funds: how they differ from the metalWhen you buy a gold fund, you are buying not the metal but a unit — a book entry recording your share in the fund's assets, which are controlled by the management company.
- Active and Passive Management: What the Fee Pays ForThe fee in active management is a payment for the attempt to beat a benchmark: for the work of analysts and for decisions on which security to buy and when to sell it.