Tracking error is the standard deviation of the differences between an index fund's returns and the returns of the index it copies, measured over a series of days, months or years. A low figure means the fund follows its benchmark index steadily; it does not say how far behind the fund runs. That shortfall has its own name, tracking difference, and the two are easy to mix up. The International Organization of Securities Commissions (IOSCO) and the EU's European Securities and Markets Authority (ESMA) both define tracking error as the volatility of the return difference. Index funds explained covers choosing a fund, and the stocks hub gathers the related guides.
How is tracking error calculated?
Subtract the index's return from the fund's for each period, then take the standard deviation of those differences.
Both returns should be total returns, with dividends reinvested, and the fund's should be its return at net asset value, which reflects the portfolio rather than the trading of its shares. The sample standard deviation, dividing by the number of periods minus one, suits short histories; the standard deviation entry shows the arithmetic. Tracking error is often computed from daily or monthly differences and then annualized. A monthly figure is multiplied by the square root of 12, so 0.05 points a month becomes about 0.17 points a year, and a daily figure by the square root of the number of trading days, about 252.
How does tracking difference differ from tracking error?
Tracking difference tells you how far a fund lagged or beat its index, while tracking error tells you how much that gap moved from one period to the next.
Each hypothetical fund below follows the same index. The figures are each year's fund return minus the index return, in percentage points.
| Year | Fund A | Fund B | Fund C |
|---|---|---|---|
| 1 | −0.05 | −1.00 | +0.40 |
| 2 | −0.04 | −1.01 | −1.10 |
| 3 | −0.06 | −0.99 | −0.20 |
| 4 | −0.05 | −1.00 | +0.30 |
| 5 | −0.05 | −1.00 | −0.90 |
| Average tracking difference | −0.05 | −1.00 | −0.30 |
| Tracking error (sample standard deviation) | 0.01 | 0.01 | 0.68 |
Funds A and B are equally steady, so their tracking errors match, but B trails by 0.95 points a year more, about what a 1% fee would cost next to a 0.05% one. Fund C lags less than B on average yet is the hardest to predict: in a given year it might beat the index or fall more than a point behind. A low tracking error says nothing about cost, and a modest average gap can hide wide swings, so read the two together.
How closely did IVV track the S&P 500 from 2021 to 2025?
Very closely: the ETF's return at net asset value trailed the index by between 0.02 and 0.05 points in each of the five calendar years.
Worked example: Tracking difference and tracking error for IVV, 2021 to 2025
BlackRock publishes IVV's calendar-year total returns at NAV beside the S&P 500's. The last column is the fund minus the index.
| Calendar year | IVV at NAV | S&P 500 | Tracking difference (points) |
|---|---|---|---|
| 2021 | 28.66% | 28.71% | −0.05 |
| 2022 | −18.13% | −18.11% | −0.02 |
| 2023 | 26.27% | 26.29% | −0.02 |
| 2024 | 24.98% | 25.02% | −0.04 |
| 2025 | 17.85% | 17.88% | −0.03 |
- Average tracking difference: (−0.05 − 0.02 − 0.02 − 0.04 − 0.03) ÷ 5 = −0.032, or about −0.03 points a year, the same as IVV's 0.03% expense ratio
- Each year's distance from that average: −0.018, +0.012, +0.012, −0.008 and +0.002
- Squared and added: 0.00068; divided by 4, one fewer than the number of years: 0.00017
- Square root: a tracking error of about 0.013 points
- In dollars: $10,000 grew to about $19,590 at NAV over the five years, against $19,617 for the index, a $27 shortfall
On market prices the picture is noisier. IVV's market-price returns beat the index by 0.04 points in 2021 and trailed it by 0.08 in 2024, giving a tracking error of about 0.05 points, because a return measured on closing prices also picks up any change in the premium or discount between the start and end of the year. Five annual figures make only a rough estimate, since monthly or daily data catch gaps that cancel out within a year. The SPY, VOO and IVV comparison sets three S&P 500 funds' records side by side, and SPY's price page shows how the oldest of them trades.
What causes tracking error?
Costs and the practical limits on copying an index cause most of it, while a few factors, such as securities lending income, can push a fund ahead.
| Cause | Effect on tracking difference | Effect on tracking error |
|---|---|---|
| Expense ratio | Lowers it by about the fee, every year | Little, since the drag is steady |
| Trading when the index changes or rebalances | Lowers it, more for indexes with frequent changes or illiquid holdings | Raises it in busy years |
| Cash from dividends and inflows waiting to be invested | Lowers it when markets rise, helps when they fall | Raises it |
| Sampling, or holding a representative subset of the index | Either direction | Raises it |
| Tax diversification limits on concentrated indexes | Either direction | Raises it while weights differ |
| Foreign dividend withholding tax below the index's assumption | Raises it | Small |
| Securities lending income | Raises it | Small |
ESMA expects a UCITS prospectus to describe these factors and to say whether the fund replicates the index fully, samples it or uses swaps. To limit cash drag, IVV's fund page says it may use exchange-traded index futures to offset cash and receivables. IOSCO notes that securities lending income can partly offset an ETF's management fee.
Three causes are less obvious. MSCI's net total return indexes reinvest dividends after withholding tax at the maximum rate for non-resident institutional investors who don't benefit from tax treaties, so a fund that pays a lower treaty rate can beat its net index. A US fund keeps its tax status only if, at each quarter-end, no single issuer exceeds 25% of its assets and the issuers above 5% add up to no more than half, so a fund following a very concentrated index may have to hold less of its largest names. And MSCI calculates its indexes from each exchange's official closing price, so a fund that values foreign stocks at fair value as of the New York close can show day-to-day gaps that have nothing to do with costs.
Where do you find a fund's tracking error?
In Europe, an index fund's prospectus and reports must state it, while US investors usually work out tracking difference from published returns.
ESMA's guidelines, which national regulators such as the Central Bank of Ireland apply, require an index-tracking UCITS to give its anticipated tracking error in normal market conditions in the prospectus and the tracking error at the end of the period in its annual and half-yearly reports. The annual report must also explain any divergence from the anticipated level and disclose the annual tracking difference, so it is the first place to look.
Form N-1A, the SEC form that US mutual funds and most ETFs file, has no tracking error item. The prospectus summary does have a table of 1-, 5- and 10-year average annual returns that must include an appropriate broad-based market index, and a fund may add the narrower one it tracks; subtract the index row from the fund row for the tracking difference. Many issuer websites also post calendar-year returns at NAV, at market price and for the benchmark, the source of the IVV figures above.
How is tracking error different from similar measures?
Several fund statistics compare a fund with an index or with its own holdings, and each answers a different question.
| Measure | What it compares | What it tells you |
|---|---|---|
| Tracking difference | Fund total return minus index total return over a period | What following the index cost or earned |
| Tracking error | The standard deviation of those differences | How steady the gap is |
| Expense ratio | Yearly operating costs divided by average net assets | The predictable part of the gap |
| Premium or discount | An ETF's market price against its NAV at one moment | Whether buyers paid more or less than the holdings were worth |
| Alpha | Return beyond what the fund's market exposure predicts | What is left after adjusting for market risk |
| Information ratio | Average return above the benchmark divided by tracking error | Excess return per unit of active risk, for active funds |
| R-squared | The share of a fund's movement the index explains | Whether the index is the right yardstick |
The expense ratio and tracking difference are easy to conflate. The fee forecasts the drag; the tracking difference reports it after trading costs, lending income and taxes. A fund that regularly trails its index by much more than its fee is losing money in ways the ratio does not show.
What mistakes do people make with tracking error?
The most common is treating tracking error as the amount a fund lags its index, when it measures only how much that lag varies.
- Taking the phrase at face value. Many investor guides use tracking error for any gap between a fund and its index; FINRA's guide to exchange-traded products, for one, says it occurs when a product's returns deviate from its benchmark's. Check whether a quoted figure is a standard deviation or a return difference before comparing funds.
- Using the wrong version of the index. A fund measured against a price index, which ignores dividends, appears to beat it by roughly the dividend yield. Against a gross total return index, which reinvests dividends before withholding tax, a foreign stock fund looks worse than against the net version, so check which one the fund's documents name.
- Mixing currencies or share classes. A fund priced in euros or pounds compared with an index in dollars, or a currency-hedged share class compared with an unhedged index, shows currency moves, not tracking. MSCI publishes its indexes in US dollars and in local currency, so match the two.
How ETF share prices are kept close to the value of their holdings, the other half of what a buyer earns, is covered in what is an ETF.