What Is Tracking Error?

If you are comparing ETFs or index funds, you will quickly meet two similar-sounding phrases: tracking error and tracking difference.

They sound technical, but the idea is simple. An index fund is trying to follow an index. Tracking error and tracking difference are two ways of checking how closely it has done that.

That matters because when I buy an index fund, I am not buying the index itself. I am buying a fund that attempts to follow it. A tracking objective does not promise a perfect match.

This is not financial advice or a recommendation to buy any particular fund. It is a glossary-style explainer to help you understand the language before reading a fund page or factsheet for yourself. Investing puts your capital at risk and past performance does not guarantee future results.

Quick answer: what is tracking error?

Tracking error measures how consistently a fund follows its benchmark index.

In plainer English:

• the index produces a return
• the fund produces a return
• subtract the index return from the fund return for each observation
• tracking error measures how much those return gaps vary around their average

A low figure means the return gap has been relatively steady. It does not necessarily mean the gap was small.

The formal measure is the annualised standard deviation of the periodic fund-minus-index returns. Standard deviation measures how spread out those observations are. Vanguard’s explanation of index tracking distinguishes this variability from the overall return gap.

For a beginner, the key point is:

Tracking error is about consistency of following the index, not whether the fund beat or lagged the index overall.

That last bit is where tracking difference comes in.

Tracking error vs tracking difference

These two terms are easy to mix up, so I think of them like this:

Tracking difference is the outcome gap

Tracking difference compares the fund’s return with the index’s return over a chosen period.

A simple version looks like this:

Tracking difference = fund return - index return

In a hypothetical example, if an index returned 8% over a year and the fund returned 7.7%, the tracking difference would be -0.3 percentage points.

If the fund returned 8.1%, the tracking difference would be +0.1 percentage points.

The tracking difference can be negative or positive. It depends on the period, the index, the fund, fees, taxes, dividend treatment and other practical details.

Tracking error measures variability in the return gap

Tracking error is not just the final gap at the end of the year. It looks at how variable the gap was along the way.

Two funds could both end the year 0.3 percentage points behind the index, but behave differently during the year:

• Fund A stayed close to the index every month
• Fund B was sometimes well ahead and sometimes well behind, but ended in the same place

They might have similar tracking differences, but Fund B would likely have higher tracking error.

That is why I find it helpful to separate the ideas:

• tracking difference is about the gap
• tracking error is about the consistency

A simple example

These invented monthly figures illustrate the distinction. They are not actual fund performance or forecasts. The gaps are in percentage points, not percentage changes in the index value.

This fund lagged the index by the same amount each month. With these exact, unrounded observations, the standard deviation of the monthly gaps is zero. The fund still underperformed. Its full three-month tracking difference would use compounded returns, not simply add the monthly gaps.

Now imagine another fund:

This fund’s gap jumps around more. Even if its final return ended up close to the index, its tracking error would be higher.

That is the heart of the concept.

Why an ETF or index fund may not match its index exactly

An index is a calculation. A fund has costs and practical decisions to make. The main sources of a mismatch include:

• Fees: ongoing charges and trading costs reduce the fund’s return. A steady cost can create a persistent shortfall without making the return gap particularly variable.
• Tax: withholding tax on investment income can differ from the assumptions used in the benchmark. The effect depends on the fund and benchmark, so this is not a statement about an individual investor’s tax bill.
• Cash and trading: money waiting to be invested and trades made when an index changes can affect returns. Cash can help relative returns during a fall as well as hold them back during a rise.
• Replication: the fund’s holdings or other arrangements for following the index may not produce exactly the benchmark’s return.

The iShares II prospectus, under “Anticipated Tracking Error” discusses holdings differences, cash management, rebalancing costs and withholding tax. These are possible causes, not a way to predict a particular fund’s future gap.

For the structure itself, I explain physical and synthetic ETFs separately. The current documents for the specific fund remain the place to check its approach.

Why tracking error can be low but tracking difference still matters

A fund can follow an index very consistently and still lag it over time.

The first table shows why: a consistently negative return gap can have zero variability. A smaller tracking error does not tell me whether the overall shortfall was acceptable.

A positive tracking difference is possible too. It describes what happened over the measured period, not a promise that the fund will keep beating its benchmark.

I want both figures in context. Vanguard’s research on looking beyond expense ratios explains why tracking consistency, return differences and costs answer different questions.

Where to find tracking error and tracking difference

When I am looking at fund information, I treat the issuer’s current documents as the starting point.

Useful places can include:

• the fund issuer’s website
• the latest factsheet
• the key information document for costs and risks, where available
• the prospectus
• the annual report or semi-annual report
• the fund’s performance and index comparison charts

Not every document contains a tracking error figure. Also distinguish anticipated tracking error from realised tracking error. The former is an estimate; the latter describes observed tracking over a past period. The iShares prospectus directs readers to annual and semi-annual reports for realised figures.

Different providers present the data differently. Some show tracking error clearly. Some focus more on performance against the benchmark. Some third-party websites calculate their own figures.

I would be careful about comparing numbers from different sources unless I understood the method and time period being used.

What to check before relying on a tracking figure

A tracking error number is only useful if I know what it is measuring.

Before putting too much weight on it, I would want to know:

What benchmark is being used?

The fund should have a named benchmark index. If I am comparing funds, I need to check they are actually tracking the same index, or at least understand how the indices differ.

Two global equity funds can sound similar but follow different benchmarks, include different countries or have different rules.

What time period is being measured?

A one-year tracking figure may tell a different story from a five-year figure.

Short periods can be distorted by unusual markets, index changes or one-off events. A longer history is still a record of the past, not a guarantee about the future.

Is the number annualised?

Tracking error is often annualised. That means the figure has been converted into an annual equivalent.

I would also check whether the calculation uses daily, weekly or monthly observations. For example, the iShares II prospectus defines its measure using monthly return differences. Figures with different methods are not automatically comparable.

Are the benchmark and return figures comparable?

I would check the exact benchmark version, currency, dates and share class. For income, the comparison needs the same treatment of reinvested distributions. A price-only chart and a total-return series answer different questions.

A net total-return benchmark includes withholding-tax assumptions, while a gross version does not deduct that withholding tax. Neither is a statement of my personal tax position. The issuer’s comparison notes should explain which version it uses. MSCI’s current index calculation methodology, section 2.2, sets out its gross and net dividend-reinvestment approaches.

I cover the document checks in how to read an ETF factsheet.

Common beginner mistake: treating one metric as the answer

The biggest trap is turning tracking error into a league table and assuming the lowest number wins.

I do not think that is a sensible way to learn about ETFs or index funds.

I would also look at what the fund owns, its costs, structure and risks. A fund can track its index extremely consistently while that index falls sharply. Low tracking error is not protection against losing money.

If the costs terminology is unfamiliar, what is an expense ratio? covers that part of the comparison.

How I think about tracking error as a beginner investor

For my own learning, I use tracking error as a quality-control question:

How steady has the gap between the fund and benchmark returns been?

Then I use tracking difference as a second question:

After real-world costs and frictions, how has the fund actually compared with the index?

That still does not make the decision for me. It just helps me understand the product better.

What I take from the number

Tracking error measures the variability of the periodic return gap between a fund and its benchmark index.

It is not the same as tracking difference. It is not a guarantee of future behaviour. And it should not be used on its own to judge an ETF or index fund.

For me, it is one useful check in a wider process: understand the fund, read the current issuer documents, recognise that capital is at risk and avoid turning a single metric into a decision machine.

Nothing on this site is financial advice. You can read the wider site boundaries in my disclaimer.