A gold ETF can track consistently and still finish behind its benchmark

Two measurements, two different questions

The term tracking error sounds like a score for everything a fund gets wrong. That intuition is unreliable. Vanguard distinguishes the return gap over an interval from the variability of relative returns through time. iShares makes the same distinction in its comparison guidance. For a gold investor, a vehicle can follow bullion very consistently while delivering a persistently lower return. Consistency and retained return deserve separate entries in the comparison.

This is an analytical method, published on 9 November, rather than a ranking of current funds. No fresh, matched performance panel was obtained for competing gold products. I will use deliberately hypothetical observations to show what the two statistics reveal and what remains outside them.

Source notes: [1] Vanguard Canada — Index tracking · [2] BlackRock iShares — How to Compare Global ETFs

A constant shortfall can have no variability

Suppose a fictional gold benchmark returns zero in each of twelve monthly observations. Product A returns minus 0.10% in every month. Each monthly relative return is therefore minus 0.10 percentage points, and the standard deviation of those observations is zero. Yet compounding twelve returns of minus 0.10% leaves approximately minus 1.19% for the year. Zero variability of the gap did not mean zero cumulative shortfall.

Now imagine Product B alternating between plus 0.20% and minus 0.20% each month against the same flat benchmark. Its relative returns fluctuate, but six pairs compound to roughly minus 0.0024%. These invented sequences are an arithmetic demonstration, not forecasts or realistic product histories. They show why the investor cannot infer the ending wealth difference from the variability statistic alone.

Define the comparison before calculating it

Vanguard describes tracking difference as fund total return minus benchmark total return over a specified period. Its formal tracking-error definition uses an annualized standard deviation of relative returns; it also notes that providers can define the measure differently. I would therefore preserve the published calculation convention, observation frequency, return type and period before comparing two displayed figures. A monthly and a daily estimate are not interchangeable merely because both carry a percentage sign.

For an independent worksheet, calculate each periodic relative return from matched observations, then report its average and variability separately from the full-period compounded return difference. Do not add monthly arithmetic gaps and label the result an exact annual wealth difference. If annualizing variability, state the scaling rule and its assumptions. A shorter or selectively calm sample can change the answer without improving the product.

Source notes: [1] Vanguard Canada — Index tracking

The gold vehicle must have the right benchmark

The SPDR Gold Shares sponsor states an objective of reflecting bullion-price performance less the trust’s expenses and identifies the LBMA Gold Price PM benchmark. That is a specific product objective, not a promise that every exchange-traded instrument with gold in its name behaves identically. My comparison would begin with the legal product, exposure and stated reference before reaching for a performance table.

A miner-equity fund, a leveraged vehicle and a physically backed bullion product should not be ranked as substitute trackers of one simple gold series without accounting for their different mandates. For the same reason, keep a fund’s NAV-based comparison distinct from an investor’s actual entry and exit transactions. A broker account is another measurement layer, even when the fund’s internal tracking is stable.

Source notes: [3] World Gold Trust Services — SPDR Gold Shares

Ask whether the gap is expected or unexplained

My interpretation starts with a reconciliation, not an accusation. Document the product’s stated costs and methodology, then compare the observed gap with what those features could plausibly explain. If the gap changes abruptly, examine the dates, calculation inputs and product notices before concluding that management quality changed. One unexplained observation warrants investigation; it does not identify its own cause.

iShares’ guidance separates ongoing charges from platform fees, trading fees, spreads and currency conversion. This matters because a shareholder’s total experience can differ from a fund-level statistic. A long holding period and frequent short round trips expose different parts of that cost structure. I would retain those account-specific costs in a separate worksheet rather than pretend one tracking measure includes them all.

Source notes: [2] BlackRock iShares — How to Compare Global ETFs

Turn the comparison into a reproducible decision record

A useful record ends with the question the number is meant to answer. For a long-horizon bullion allocation, I would examine the compounded return gap alongside exposure and operational terms. For a short benchmark-sensitive hedge, variability over the relevant horizon may also matter. Neither objective makes an unexplained deviation harmless, and neither permits a universal winner from one statistic.

The Research Lab can apply this measurement discipline without presenting a product comparison as a trading result. The examples above establish no actual fund return, CY66 outcome, fill quality or P&L. The separate Dubai physical-market guides ask a different ownership and transaction-cost question; keeping that alternative visible makes the comparison broader without confusing metal in custody with a share traded through an account.