Tracking Error: Variation, Not Shortfall

Michael Whitman · Trading since 2017, full-time since 2020 · Reviewed 2026-09-07

Tracking error measures the variability of a fund's return difference from its benchmark — how much that gap moves about, not how large it is. The average size of the gap is tracking difference, and for an index fund that second figure is the one that costs you money.

How it works

A candlestick chart of the site's shared price history. The headline on the chart reads: How far a fund wanders from its index. — How far a fund wanders from its index. Illustrative chart - not real market data.

Take the fund’s return and the index’s return over the same periods, and subtract. That series of differences is what the measure describes — not any single one of them.

A gently rising stretch of the long price series with an account equity curve beneath it. The headline on the chart reads: It is the variation of the difference, not the difference. — It is the variation of the difference, not the difference. Illustrative chart - not real market data.

It is the standard deviation of that series. A fund that trails its index by the same amount every month has a small tracking error, because the gap barely varies — even though the gap itself may be substantial.

A calmly advancing stretch of the long price series with a slowly rising equity curve beneath it. The headline on the chart reads: Which is not the same as the shortfall itself. — Which is not the same as the shortfall itself. Illustrative chart - not real market data.

The average of the differences is tracking difference, and it is the number that costs you. The two are routinely confused, including in fund marketing. A low tracking error with a large tracking difference is a fund that is reliably behind — which is worse, not better.

A flat, quiet stretch of the long price series with a gradually rising equity curve beneath it. The headline on the chart reads: Fees, sampling, cash and taxes all create it. — Fees, sampling, cash and taxes all create it. Illustrative chart - not real market data.

Four things produce the gap. The ongoing charge, deducted continuously. Sampling, where the fund holds a representative subset rather than every constituent. Cash held for redemptions, which is not invested. And withholding tax on dividends, which the index calculation may ignore.

What a good figure looks like depends on the fund

A strongly rising stretch of the long price series with an account curve breaching its limit. The headline on the chart reads: A low figure is the whole job of an index fund. — A low figure is the whole job of an index fund. Illustrative chart - not real market data.

For a tracker, matching the index is the entire product. A large tracking error means the fund is not doing the one thing it was bought to do, whichever direction the deviation runs.

A choppy, directionless stretch of the long price series. The headline on the chart reads: And a high one is the point of an active fund. — And a high one is the point of an active fund. Illustrative chart - not real market data.

For an active fund the reverse holds. A manager charging active fees while producing a very small tracking error is delivering an index at a premium — the situation sometimes called closet indexing, and it is visible in this statistic before it is visible anywhere else.

A declining stretch of the long price series. The headline on the chart reads: An index fund with a large one is not doing its job. — An index fund with a large one is not doing its job. Illustrative chart - not real market data.

So the same number is good or bad depending on what was promised. It is a measure of difference, and whether difference is a defect is a question about the mandate.

In practice

A 72-bar candlestick section of the shared price history with an account curve shown with and without fees. The headline on the chart reads: The fee is the most predictable part of the gap. — The fee is the most predictable part of the gap. Illustrative chart - not real market data.

The charge is the component you can know in advance. On this site’s arithmetic, compounding the fee alone over thirty years, 5 basis points removes 1.5% of the pot and 75 basis points removes 20.2% — and that is before any question of how well the index was tracked.

A candlestick chart with a volume histogram beneath it, with the volume histogram emphasised. The headline on the chart reads: Sampling exists because some holdings barely trade. — Sampling exists because some holdings barely trade. Illustrative chart - not real market data.

Sampling is a liquidity decision. A broad index contains constituents too small or too thinly traded to buy efficiently, so the fund approximates them — and the approximation shows up here.

A long-horizon candlestick view of the same price series. The headline on the chart reads: It is measured over years to mean anything. — It is measured over years to mean anything. Illustrative chart - not real market data.

A short measurement period tells you very little. The statistic needs enough observations to be stable, and comparing funds over different windows compares different things.

A candlestick series containing several opening gaps, with the largest opening gap marked. The headline on the chart reads: And an index reconstitution is where it widens. — And an index reconstitution is where it widens. Illustrative chart - not real market data.

Index changes are the stress test. When constituents are added or removed on a known date, every tracker must trade the same names at the same time, and the deviation shows up in that period.

A declining stretch of the long price series, with the entry price and the level at which a stop would trigger drawn as horizontal lines. The headline on the chart reads: There is no stop on a fund holding. — There is no stop on a fund holding. Illustrative chart - not real market data.

It says nothing about risk of loss. A fund can track a falling index perfectly, and a stop is not part of this conversation at all.

A candlestick chart of the site's shared price history, annotated with the round-trip cost. The headline on the chart reads: Every switch between funds costs 2% of a bar. — Every switch between funds costs 2% of a bar. Illustrative chart - not real market data.

Switching funds has a price too. A round trip on this site’s shared history is 2% of a median bar’s range, plus any tax on the disposal — so a small improvement in tracking rarely pays for the move.

How to compare two trackers honestly

Start with the tracking difference over five years, not the tracking error. How far behind the index did the fund actually finish, on average, per year. That figure is what you paid, and it should be close to the ongoing charge.

If it is much larger than the fee, ask why. Sampling, tax treatment or cash drag will explain it, and a fund that cannot explain it is one to avoid. If it is smaller than the fee, ask that too — securities lending income can legitimately close the gap, and it carries a risk the fund should disclose.

Only then look at the variability. It matters if you are matching a liability on a date, and it matters much less if you are holding for twenty years and only the endpoint counts.

What tracking error is not

It is not the shortfall. That is tracking difference.

It is not a risk measure. The index can fall too.

It is not always bad. Active funds should have it.

And a low figure is not a cheap fund. Check the difference.

When it fails

A sideways, range-bound candlestick series. The headline on the chart reads: In a flat market it is almost entirely the fee. — In a flat market it is almost entirely the fee. Illustrative chart - not real market data.

In a flat market the whole gap is the charge. Nothing else is moving, so the fund’s shortfall is its cost — which makes a sideways period the clearest read on what a fund is really taking.

The second failure is choosing on tracking error alone. The steadiest laggard wins that contest.

A third is comparing figures over different periods. The statistic is not portable between windows.

A fourth is ignoring what the index itself measures. Two funds tracking similar-sounding indices are not comparable at all.

A fifth is expecting perfect tracking from a sampled fund. It was never going to hold everything.

And a sixth is switching for a marginal improvement. The trading cost and any tax usually exceed it.

The original data

Of the 24,971 videos in research/search-study-corpus.jsonl, 2 have “tracking error” in the title, at a median of 20,003 views across 2 channels, with a maximum of 39,809. “Index fund” appears in 30 at a median of 74,230. The counts are in research/corpus-coverage.json, produced by site/measure_corpus.py.

A strongly rising stretch of the long price series, cut short at the decision bar. The headline on the chart reads: The fund lagged its index again. Switch? — The fund lagged its index again. Switch? Illustrative chart - not real market data.

Two videos, and the median audience for them is 20,003. Index funds are among the most-covered products in this corpus and the measurement of whether one is doing its job appears twice. That pattern recurs across this whole site: the product is taught, the diagnostic is not.

The answer to that final question is usually no, and the fee arithmetic says why. A tracker lagging its index by roughly its charge is working exactly as designed. Compare the five-year tracking difference against the ongoing charge before considering a switch — if they are close, the fund is fine and the trading cost of moving is a certain loss against an uncertain gain.

Index funds is the product this measures and where the fee comparison belongs. Exchange-traded fund investing covers the exchange-listed version and its own tracking quirks. And beta is the neighbouring statistic people most often confuse this one with.

What I actually do

I chose a fund on its tracking error once and felt clever about it. What I had actually chosen was the one whose shortfall was most consistent, which is not the same as the one that cost me least. The steadier fund was reliably a bit further behind. Consistency is a nice property and it is not the property I was paying for.

— Michael Whitman

This page is educational, not financial advice. Test every idea on your own charts before risking money.