Funds · Indexes and index funds
Tracking difference in index funds: why a fund trails its index
How to measure the gap between an index fund and its index, what causes it besides the fee, and the 1, 5 and 10-year gaps of 13 index ETFs from their 2026 fact sheets.
On this page
An index fund promises the return of its index. It never delivers exactly that. Over ten years to September 30, 2026, Vanguard's S&P 500 ETF returned 15.29% a year and the S&P 500 returned 15.33%. The 0.04 of a percentage point between them is the fund's tracking difference: the return of the fund minus the return of the index over the same period.
It is the most useful single number for judging how well an index fund does its job, because it is the total of everything that separates the fund from its index: the fee, trading costs, cash, taxes withheld abroad, and income the fund earns that the index does not. It is different from tracking error, which measures how much the gap wobbles from period to period and is covered in Tracking error and information ratio: a fund against its benchmark. A fund can trail its index by a steady 0.5% a year with almost no tracking error.
The measure
Three details make the comparison fair.
Use net asset value, not market price. An exchange-traded fund has two returns: one from its net asset value (NAV) and one from its closing market price. Over a year the market-price return also contains any change in the premium or discount (Premiums and discounts in ETFs: why price and NAV disagree). The NAV return is the one that reflects what the portfolio did.
Use the index's total return. The fund collects dividends, so the comparison is with the index including dividends (How a stock index is built: eligibility, float and the divisor).
Use annualized figures over several years. A single year is noisy for reasons given below. Fund fact sheets and prospectuses print the fund and its benchmark side by side for 1, 5 and 10 years, which is all that is needed.
Worked example
From Vanguard's September 30, 2026 fact sheet for VOO:
10-year average annual return, fund at NAV: 15.29%
10-year average annual return, S&P 500: 15.33%
tracking difference = 15.29 − 15.33 = −0.04 points a year
The fund's expense ratio is 0.03%. Add it back and the fund trailed by 0.01 beyond its fee. In money, $10,000 growing at the index's 15.33% for ten years becomes $41,632, and at the fund's 15.29% it becomes $41,487. Copying the index cost $145 over a decade.
What causes the difference
The expense ratio
The fee is deducted from the fund's assets a little every day, so a fund that copied its index perfectly and did nothing else would trail by exactly its expense ratio (Expense ratios and fund costs: what you pay and how it compounds). The index pays no fee.
Trading costs
When the index changes, the fund trades. It pays bid-ask spreads and commissions and may move prices. The index is calculated as if the change happened at the closing price for free. These costs are small for a fund on a low-turnover index of large companies and larger for small-company and emerging-market funds. They are not in the expense ratio.
Sampling
A fund can hold every member of its index in the index's proportions (full replication) or a selection designed to behave like the whole (sampling). Vanguard's fact sheet describes VOO as full replication and VTI, whose index has more than 3,400 stocks, as index sampling. Sampling saves the cost of trading hundreds of tiny positions and introduces small random differences that can go either way.
Cash
A fund always holds a little cash: dividends received and not yet paid out, and money waiting to be invested. In a rising market cash earns less than stocks and drags the fund below its index. In a falling market it helps. Funds reduce the effect with index futures.
Securities lending
A fund can lend its shares to borrowers, usually short sellers, and earn a fee. The index earns nothing from lending, so the income pushes the fund's return toward or above the index. It matters most in funds holding small or hard-to-borrow stocks. The details and the risks are in Securities lending in funds: who borrows, who earns, what goes wrong.
Foreign withholding tax
Foreign countries tax dividends paid to outside investors. Index providers publish "net" versions of international indexes that deduct this tax at an assumed rate. iShares' fact sheets explain that the net indexes assume the maximum withholding rate applicable to non-residents, which is often higher than the rate the fund itself is subject to. When the fund pays less tax than its index assumes, it gains on the index for a reason that has nothing to do with skill.
Pricing dates
A US-listed fund holding foreign stocks sets its NAV at 4 p.m. New York time, hours after most of its markets have closed, and may adjust stale closing prices to a fair value. The index uses the local closing prices. On any single day the two can differ by a percent or more, and that difference lands in whichever period ends that day (International index funds: country classification, taxes and pricing).
Thirteen funds against their indexes
The table gives the same funds over three periods. Each row is taken from the fund's own document, and the dates differ, so compare a fund with its own index and not with the fund in the next row.
| Fund | Index | As of | Expense ratio | 1 year | 5 years | 10 years |
|---|---|---|---|---|---|---|
| VOO | S&P 500 | Sep. 30, 2026 | 0.03% | −0.04 | −0.04 | −0.04 |
| IVV | S&P 500 | Jun. 30, 2026 | 0.03% | −0.03 | −0.03 | −0.03 |
| SPY | S&P 500 | Oct. 31, 2026 | 0.0945% | −0.13 | −0.13 | −0.15 |
| VTI | US total market | Sep. 30, 2026 | 0.03% | 0.00 | 0.00 | 0.00 |
| ITOT | S&P Total Market | Jun. 30, 2026 | 0.03% | −0.02 | 0.00 | 0.00 |
| IWB | Russell 1000 | Jun. 30, 2026 | 0.15% | −0.19 | −0.15 | −0.15 |
| IWM | Russell 2000 | Sep. 30, 2026 | 0.19% | −0.14 | −0.11 | −0.08 |
| EFA | MSCI EAFE (net) | Sep. 30, 2026 | 0.32% | −0.98 | −0.02 | −0.07 |
| IEFA | MSCI EAFE IMI (net) | Sep. 30, 2026 | 0.07% | −0.65 | +0.25 | +0.18 |
| VEA | FTSE Developed ex US | Sep. 30, 2026 | 0.03% | −0.84 | +0.03 | −0.05 |
| EEM | MSCI Emerging Markets | Sep. 30, 2026 | 0.72% | −1.15 | −0.63 | −0.68 |
| VWO | FTSE Emerging | Sep. 30, 2026 | 0.06% | +0.22 | −0.03 | −0.15 |
| VXUS | FTSE Global ex US | Sep. 30, 2026 | 0.05% | −0.53 | +0.01 | −0.10 |
Tracking difference in percentage points a year: fund NAV return minus benchmark return.
What the numbers show
The S&P 500 funds trail by about their fee. VOO and IVV, each charging 0.03%, were 0.04 and 0.03 behind in every period shown. SPY, charging 0.0945%, was 0.13 to 0.15 behind. SPY is a unit investment trust, an older legal form with less flexibility than the open-end structure of the other two, and its shortfall has been a few hundredths more than its fee. Over ten years at the fact-sheet rates, $10,000 in the index would have become $42,030 and in SPY $41,487, a gap of $543.
The total-market funds gave the fee back. VTI and ITOT matched their indexes to two decimal places over five and ten years while charging 0.03%. Both hold thousands of small companies, where lending income is highest, and that income was enough to cover the fee and trading costs.
The Russell 2000 fund recovered more than half its fee. IWM charges 0.19% and trailed by 0.08 over ten years.
The iShares international funds beat their fee. EFA charges 0.32% and trailed by 0.07 over ten years; IEFA charges 0.07% and was 0.18 ahead. Both are measured against net indexes that assume a higher tax rate than the funds pay. EEM, at 0.72%, trailed by 0.68 over ten years, just inside its fee, which at those rates is $1,446 on $10,000. Vanguard's VEA, VWO and VXUS, measured against FTSE benchmarks that Vanguard adjusts for withholding taxes, trailed by 0.02 to 0.09 points a year more than their fees.
One-year figures for international funds are noise. EFA, IEFA, VEA and VXUS were all 0.5 to 1.0 points behind over the year to September 30, 2026, while their five-year figures were within 0.03 of zero or positive. EEM was 1.15 points behind over the year to September 30, 2026, yet its June 30 fact sheet had it 1.55 points ahead over the year to that date. A gap that large and that inconsistent is the pricing-date effect at the start and end of the year, not a cost.
The same index, three funds
A second way to see tracking difference is to compare funds on the same index with each other. This needs no index data at all.
From Finabase's monthly prices and distributions, IVV returned more than SPY in each of the six calendar years from 2020 to 2025, by 0.03 to 0.13 points, and VOO did in five of the six. The average lead was 0.07 points a year for each, against a difference in expense ratios of 0.0645. Over the six years SPY compounded at 14.98% a year and IVV and VOO at 15.05%.
The yearly figures bounce around more than the fee difference alone would explain. These are market-price returns, not NAV returns, so each includes a few hundredths of a point of closing premium or discount, and distributions are assumed reinvested at month end. The comparison starts in 2020 because the distribution records behind it contain errors in earlier years.
Where tracking difference misleads
A change of benchmark. Funds sometimes switch index. The "benchmark" line in a fact sheet is then a splice of the old and new indexes. VTI's fact sheet splices three indexes, with changes in 2005 and 2013. The spliced comparison is fair, but it is not a comparison with any one index.
Net and gross indexes. An international fund measured against a net index looks better than the same fund against a gross index. Check which the fact sheet uses before comparing two funds from different issuers.
Rounding. Fact sheets print two decimals. A difference of 0.01 between two funds is within rounding.
Different end dates. Tracking difference to June 30 and to September 30 cover different markets. Only compare funds over identical periods.
It measures copying, not choosing. A fund that tracks a poor index perfectly has a tracking difference of zero. Over the same ten years in which VTI matched its index, that index returned 14.70% a year against 15.33% for the S&P 500. That gap of 0.63 points a year between two indexes is about fifteen times VOO's tracking difference (S&P 500 vs total market funds: what the extra 3,000 stocks change).
How Finabase does it
Finabase does not publish index returns, so it does not calculate a tracking difference. The figures on this page come from the funds' own fact sheets, which is the right place to look for any fund.
What Finabase offers is the fund-against-fund comparison. Every fund has a symbol page
with quote, chart, news and a Dividends & splits tab. On the comparison tab, opening SPY
with IVV and VOO rebases the three to percent change from the first bar in view and
tabulates price returns from one week to two years. Those returns exclude dividends, so
small differences between funds on the same index partly reflect the timing of their
payouts as well as their costs. The Holders tab can show a fund card with the issuer,
expense ratio, assets, net asset value, holdings and sector weights when the feed has
them, and /portfolio tracks holdings and the dividends they pay.
Frequently asked questions
Is tracking difference the same as tracking error?
No. Tracking difference is the average gap between fund and index: how far behind or ahead the fund ended up. Tracking error is the variability of that gap from period to period. An index fund should have a small, steady, slightly negative tracking difference and a tracking error close to zero.
Can an index fund beat its index?
Yes, by small amounts. Securities lending income, and for international funds a lower foreign tax rate than the net index assumes, can outweigh the fee. IEFA was 0.18 points a year ahead of its net index over ten years to September 30, 2026.
Where do I find a fund's tracking difference?
In its fact sheet or the performance table of its prospectus, which show the fund's average annual returns next to the benchmark's for 1, 5 and 10 years. Subtract one from the other. The annual report gives the same comparison for the fiscal year.
Why is my return different from the fund's published return?
Published returns assume a single investment held for the whole period with distributions reinvested. Your own return depends on when you bought, the market price you paid, and whether you reinvested. Tracking difference describes the fund, not your account.
Does a lower expense ratio always mean a smaller tracking difference?
Usually, but not always. EFA charges more than four times as much as IEFA and trailed its own index by only 0.07 over ten years. Expense ratio predicts tracking difference well among funds on the same index and poorly across different indexes.
Summary
Tracking difference is the fund's return minus the index's, and it sums up every cost and every offsetting gain of running an index fund. For large US stock funds it is close to the expense ratio: 0.03 to 0.04 points a year for the cheapest S&P 500 funds and zero for two total-market funds over the ten years to 2026. For international funds it is distorted by tax assumptions in the index and, over short periods, by pricing dates. It is found by subtracting two numbers printed in every fact sheet, and it is worth more attention than the expense ratio alone. It does not say whether the index itself was a good choice.
Check your understanding
5 questions on this lesson. Each answer comes with the reason, and nothing is scored.
A fund’s fact sheet shows a 10-year average annual return of 9.75% at NAV and 9.83% for its index. What is the tracking difference?
- −0.08 points a year
- +0.08 points a year
- It cannot be found without the expense ratio
Show the answer
A. −0.08 points a year Tracking difference is the fund’s return minus the index’s: 9.75 − 9.83 = −0.08. The expense ratio is one of its causes but is not needed to calculate it.
That fund charges 0.19% a year. What does a tracking difference of −0.08 tell you?
- The fund has hidden costs of 0.27%
- The fund recovered more than half of its fee from other sources, such as securities lending
- The index was calculated wrongly
Show the answer
B. The fund recovered more than half of its fee from other sources, such as securities lending A fund that only paid its fee would trail by 0.19. Trailing by 0.08 means income or advantages worth about 0.11 a year offset part of the fee and the trading costs.
An international index fund beat its net index by 0.18 points a year over ten years. What is the most likely reason?
- The manager picked better stocks than the index
- The net index assumes a higher foreign withholding tax rate than the fund paid
- The fund holds more cash than the index
Show the answer
B. The net index assumes a higher foreign withholding tax rate than the fund paid Net indexes deduct foreign dividend tax at an assumed maximum rate. A fund that is taxed at a lower rate keeps the difference, and lending income can add to it.
A developed-markets index fund is 0.98 points behind its index over one year and 0.02 behind over five years. How should the one-year figure be read?
- The fund’s costs rose sharply last year
- Mostly as a pricing-date effect: the fund’s NAV and the index were struck at different times on the first and last days
- The fund is failing to track its index
Show the answer
B. Mostly as a pricing-date effect: the fund’s NAV and the index were struck at different times on the first and last days A fund holding foreign stocks sets its NAV after those markets close and may use fair-value prices, while the index uses local closes. The mismatch on the end dates swamps a single year and washes out over longer periods.
Which return should be used to measure an ETF’s tracking difference?
- The market-price return
- The NAV return, against the index’s total return
- The NAV return, against the index’s price level
Show the answer
B. The NAV return, against the index’s total return The NAV return reflects what the portfolio earned, without changes in premium or discount, and the fund collects dividends, so the index’s total return is the fair comparison.
Sources and further reading
- Vanguard — fact sheets of VOO, VTI, VEA, VWO and VXUS, September 30, 2026: average annual returns at NAV and of the benchmark, expense ratios, replication method
- iShares — fact sheets of IVV, ITOT, IWB and EEM (June 30, 2026) and of IWM, EFA and IEFA (September 30, 2026), and EEM’s product page (quarter-end September 30, 2026): average annual returns at NAV and of the benchmark; note on net index withholding rates
- State Street Investment Management — SPDR S&P 500 ETF Trust (SPY) product page, performance at month end October 31, 2026; gross expense ratio 0.0945%; unit investment trust
- U.S. Securities and Exchange Commission, Investor.gov — Index funds
- Finabase chart data — monthly bars and dividends of SPY, IVV and VOO, 2020 through 2025; statistics computed by Finabase
