An index tracker fund promises to follow a benchmark like the FTSE 100 or the MSCI World Index as closely as possible — but "as closely as possible" is rarely "perfectly". The small, usually modest gap between a fund's actual return and its benchmark's return is called tracking error, and understanding what causes it helps explain why two funds tracking the very same index can still produce slightly different results over time.
What tracking error actually measures
Tracking error refers to the divergence between a fund's performance and the performance of the index it aims to replicate. It is usually expressed as a statistical measure of how much that divergence varies over time, though in everyday use investors often refer more loosely to the simple difference between the fund's return and the index's return over a given period, sometimes called "tracking difference". A well-run index tracker typically has a small tracking error, generally within a fraction of a percentage point per year, though this varies by index and fund structure.
Why index funds don't perfectly replicate their benchmark
Fund charges
The most predictable and persistent source of tracking difference is simply the fund's own OCF. An index itself has no costs — it is a theoretical calculation — while a fund tracking it has to pay for fund management, administration, and dealing. A fund with a 0.15% OCF will, all else being equal, tend to underperform its benchmark by roughly that amount over time, purely because of this cost drag.
Sampling versus full replication
Some indices contain thousands of constituent securities, some in small or illiquid markets. Rather than buying every single constituent in its exact weighting (full replication), some funds use a sampling approach, holding a representative subset of securities chosen to closely mimic the index's overall behaviour. Sampling can introduce small deviations from the index's actual return, in either direction.
Cash drag
Funds typically hold a small amount of cash to manage day-to-day inflows, outflows, and dividend receipts before reinvestment. Because cash doesn't grow at the same rate as the invested index, holding even a small cash buffer can create a slight, usually very small, drag on performance relative to the index during rising markets (and the reverse during falling markets).
Dividend reinvestment timing and withholding tax
Index calculations typically assume dividends are reinvested instantly and, for a "net total return" index, after a standard withholding tax deduction. In practice, funds receive and reinvest dividends with some delay, and the actual withholding tax a fund suffers on foreign dividends can differ from the index calculation's assumption, depending on the fund's domicile and the specific tax treaties in place. Both factors can create small, ongoing tracking differences.
Securities lending
Some funds lend out a portion of their holdings to other market participants (commonly for short-selling purposes) in exchange for a fee, a portion of which is often passed back to the fund, potentially offsetting some of the cost drag described above. This can, in some cases, cause a fund to track slightly ahead of its benchmark before costs, though it also introduces a small additional layer of counterparty risk.
A worked hypothetical example
Suppose a benchmark index returns exactly 8.00% over a year. Three hypothetical funds tracking that index might report the following:
| Fund | OCF | Other factors (sampling, cash drag, securities lending, illustrative) | Reported fund return | Tracking difference vs index |
|---|---|---|---|---|
| Fund A (full replication, low cash drag) | 0.07% | +0.02% (securities lending income) | 7.95% | -0.05% |
| Fund B (full replication) | 0.20% | -0.03% | 7.77% | -0.23% |
| Fund C (sampling approach) | 0.12% | -0.15% (sampling deviation) | 7.73% | -0.27% |
In this hypothetical illustration, all three funds track the same index reasonably closely, but Fund A's lower OCF and modest securities lending income leave it closest to the benchmark, while Fund C's sampling approach introduces a somewhat larger deviation despite a lower OCF than Fund B. These figures are illustrative only and not based on any specific real fund.
Why some indices are harder to track than others
Broad, liquid, developed-market indices
An index like the FTSE 100 or the S&P 500 comprises a relatively small number of large, highly liquid, easily traded constituents, all listed on well-established exchanges with low dealing costs. Funds tracking these indices generally achieve very close tracking, since full replication is straightforward and inexpensive to maintain.
Broad global or emerging market indices
An index like the MSCI World or a broad emerging markets index can include thousands of constituents across dozens of countries, some in markets with higher dealing costs, foreign ownership restrictions, or less liquid trading. Funds tracking these broader or less liquid indices often show somewhat larger tracking error, both because sampling is more commonly used and because the underlying transaction costs of maintaining full replication would be higher.
Bond indices
Bond index tracking presents its own challenges, since bond indices often include an extremely large number of individual bond issues, many of which trade infrequently and in large minimum sizes unsuitable for a fund of modest size to hold directly. Bond tracker funds very commonly use sampling for this reason, and tracking error can be somewhat more variable than for a broad equity index tracker.
Synthetic replication and tracking error
Some index funds and ETFs use a "synthetic" replication method, gaining exposure to an index through a derivative (typically a swap) with a counterparty, rather than buying the index's underlying constituents directly. This is discussed in more detail in the companion article on physical versus synthetic ETFs, but it's worth noting here that synthetic replication can, in some cases, produce very close tracking to the stated index, since the swap counterparty contractually agrees to deliver the index's return (minus a fee) rather than the fund needing to replicate the index through direct share purchases. This comes with a different kind of risk — counterparty risk — rather than eliminating risk altogether, and is a separate consideration from tracking error itself.
How to compare tracking error between similar funds
- Look at the fund's factsheet, which often states a tracking error figure or shows historical performance against the benchmark directly.
- Compare the fund's OCF as a first, easily available proxy — lower-cost funds tend, on average, to track more closely, though this isn't guaranteed.
- Check whether the fund uses full replication or sampling, particularly for broad or less liquid indices where sampling is more common.
- Look at consistency over multiple years rather than a single period, since a single year's tracking difference can be affected by one-off factors like dividend timing.
How much should tracking error matter?
For most long-term fund investors, a small, consistent tracking error is a normal and expected feature of index investing, not a red flag — what matters more is whether it stays broadly stable and roughly in line with the fund's stated OCF over time. A tracking error that is significantly larger than the fund's OCF, or that varies erratically from year to year, may be worth investigating further, as it could point to a less efficient replication process. That said, tracking error is generally a secondary consideration compared with the OCF itself and the platform charges applied to hold the fund, both of which are known in advance and tend to matter more over the long run.
Tracking error versus tracking difference: a technical distinction worth knowing
Investment professionals sometimes distinguish more precisely between "tracking difference" (the simple gap between a fund's return and the index's return over a specific period, which can be directly observed after the fact) and "tracking error" in its stricter statistical sense (a measure of how much that gap varies, or how volatile the deviation has been, over time — technically the standard deviation of the return differences). A fund could have a small average tracking difference but a relatively higher tracking error in this stricter sense if its performance relative to the benchmark fluctuates significantly from month to month, even if it evens out over a full year. For most everyday investor purposes, the simpler concept — how closely, on average, has this fund matched its index over time — is the more practically useful one to focus on, but it's worth knowing the more precise definition exists if researching this topic further in more technical fund literature.
Reading a factsheet's performance table with tracking error in mind
Most fund factsheets present a table comparing the fund's return against its benchmark's return over several standard periods (commonly one, three, five, and ten years, where available). Looking across several of these periods, rather than just the most recent one, gives a better sense of whether a fund's tracking has been consistent, since a single period can be distorted by one-off factors such as unusual dividend timing or a specific rebalancing event in the underlying index. A fund that has tracked closely and consistently across multiple periods provides more confidence in its replication process than one where the tracking difference varies considerably between the periods shown.
Key takeaways
- Tracking error is the gap between an index fund's actual return and its benchmark's return, and some degree of it is normal and expected.
- The OCF is usually the single largest, most predictable contributor to tracking difference over time.
- Sampling approaches, cash drag, dividend reinvestment timing, and securities lending income can all cause smaller additional deviations, in either direction.
- A consistent, small tracking error roughly in line with a fund's OCF is generally normal; a large or erratic one may be worth further investigation.
- For most long-term investors, the fund's OCF and the platform charges applied to hold it remain more significant, more predictable cost factors than tracking error itself.