Tracking Error and Tracking Difference Are Not the Same Thing

The distinction, stated once

An index fund exists to deliver the index. It never does so exactly, and there are two separate questions about the shortfall.

Tracking difference is how much the fund's return differed from the index over a period. If the index returned a certain amount and the fund returned less, the tracking difference is that gap. It is a level: it accumulates, and it is what actually leaves your account.

Tracking error is how variable that gap was. If the fund lagged the index by a similar amount every month, the tracking error is small even if the shortfall is large. If it beat the index some months and lagged badly in others, the tracking error is large even if the two roughly cancelled by the end.

The names invite exactly the wrong intuition. "Error" sounds like the mistake and "difference" sounds like a technicality, and it is the other way round. Tracking difference is what the fund cost you. Tracking error is how noisily it cost you that.

Why the gap exists at all

A fund cannot be the index because the index is not a portfolio. It is a calculation, and the calculation has no costs.

The fund does. It charges an annual fee, which comes out of returns whatever happens. It must actually buy and sell shares when the index changes its constituents or their weights, and those trades cost money and do not execute at the exact closing price the index used. It holds a small amount of cash to meet redemptions, and cash does not earn the index's return. Dividends arrive on one date and are reinvested on another, and the index assumes they were reinvested instantly.

Those causes have different signatures, which is what makes the two measurements worth separating. An annual fee produces a steady drag: large tracking difference, small tracking error. Clumsy execution around index changes produces occasional lurches: possibly small tracking difference, large tracking error. Two funds can share a tracking difference and have arrived at it by completely different routes.

There is also a direction worth knowing about: tracking difference is not always negative. A fund can beat its index over a period, generally through securities lending income or favourable handling of dividends. That is not skill and should not be extrapolated — a fund that beat the index last year through such an effect will not necessarily do so again, and the effect can reverse.

Which one you should care about

For almost every purpose, tracking difference.

It is the one that compounds. A fund that lags its index steadily every year, quietly and without drama, will over a long holding period leave you meaningfully behind a fund that lags less, and neither will look eventful at any point.

Tracking error matters in two narrower cases. If you have a short and fixed horizon, a large tracking error means the fund's shortfall on your particular exit date is less predictable, which is a genuine risk even if the long-run average shortfall is fine. And a large tracking error is worth investigating for what it reveals: it suggests the fund is having difficulty replicating the index, and that difficulty may eventually show up in the difference as well.

What to check before buying an index fund

The useful comparison is between two funds tracking the same index, over the same period. Comparing across indices tells you about the indices rather than about the funds.

Start with the expense ratio, because it is the largest and most reliable component of the difference, and unlike everything else it is disclosed in advance and applies whatever happens.

Then look at the fund's return against the index's total return over as long a period as exists — total return, meaning with dividends included, because comparing a fund that receives dividends against an index that ignores them manufactures a gap that is not really there. That confusion has its own page: price return against total return.

Look at how large the fund is. A very small index fund carries proportionally higher fixed costs and has more trouble matching index changes cleanly.

And prefer a longer record of a small difference to a short record of none. A fund can have almost no tracking difference over a quiet stretch and considerably more when the index actually changes.

The honest gap

What this page has not done is tell you what tracking difference or tracking error to expect from Indian index funds, or which of them do it well.

That would need the return history of matched pairs of funds tracking the same index, alongside that index's own total-return history, over a period long enough to include some index reconstitutions. This site does not have that data. Our figures come from broad index and interest-rate history, and we hold nothing at the level of an individual fund.

So no figure appears on this page, and none should be added from memory or from a plausible-sounding range. A "typical tracking difference" that nobody measured would be exactly the kind of invented number that makes every other figure on this site worth less. Everything above is a description of what the two measurements are and what causes them, which does not need the data.

What to take away

Ask for the tracking difference, over the longest period available, against the index's total return, for funds tracking the same index. That number is the cost of using this particular fund to own that particular index.

Treat tracking error as a diagnostic rather than a decision: high error is a reason to look at what the fund is doing, not by itself a reason to avoid it. And be suspicious of any comparison that does not hold the index and the period constant, because those two choices can produce whatever answer somebody wants.