Investing
Tracking Error Is The Gap Between Fund And Index
An index fund cannot replicate its benchmark exactly, and the sources of divergence are structural, arising from costs, timing, sampling and the treatment of income.

An index is a calculation. A fund attempting to follow it is a real portfolio with costs, cashflows and settlement, which is why the two never move identically.
The index has no costs and the fund does
Index calculations assume frictionless trading and no expenses. A fund pays management charges, transaction costs, custody and, in some cases, taxes on income or transactions.
Those costs come out of the fund's assets, so a perfectly replicating portfolio would still fall behind its benchmark by approximately the amount of the charges.
This is why the expected long-run divergence has a floor, and why comparing funds on the same index often reduces to comparing what they cost to run.
Index changes force trading at a known moment
Indices add and remove constituents on published schedules, and funds must trade to match. Everyone tracking that index needs the same trades at the same time.
Concentrated demand moves prices against the funds doing the buying and in favour of those selling to them, which imposes a cost that the index calculation does not experience.
Some providers manage this by trading around the event rather than at it, which reduces the cost but increases short-term divergence from the benchmark.
Sampling trades precision for practicality
Broad indices can contain thousands of holdings, many small and thinly traded. Holding every one in exact proportion would be expensive and in some cases impossible.
Funds therefore often hold a representative subset chosen to behave like the whole, which works closely but not perfectly, particularly in unusual market conditions.
The narrower the sample, the lower the trading cost and the wider the potential divergence, which is a genuine trade rather than a defect.
Cash and income create timing differences
Money arriving from investors, or dividends received from holdings, sits as cash until it is invested, and cash does not behave like the index during that interval.
Index calculations typically assume income is reinvested immediately, whereas a fund receives it on payment dates and invests it when practical.
Over short periods these timing effects can push a fund either side of its benchmark, which is why divergence is not always in the direction of the charges.
Synthetic replication substitutes a different exposure
Some funds obtain index returns through agreements with counterparties rather than by holding the underlying securities, which can track very closely indeed.
What that introduces is exposure to the counterparty and to the collateral arrangements supporting the agreement, which is a different kind of risk rather than an absence of one.
Regulatory requirements on collateral and disclosure vary by jurisdiction and change over time, so the protections around such structures are specific to where the fund is domiciled.
Questions readers ask
Does this mean I should sell before a fall?
Only if you can identify falls in advance, which the evidence suggests almost nobody does consistently. The practical response is choosing an allocation whose falls you can sit through.
Why do average returns overstate what I got?
Because averaging percentages ignores that each return applies to a different balance. The compound return accounts for that and is always lower when returns vary.





