Introduction

Tracking error and tracking difference measure different aspects of an ETF's performance relative to its benchmark.

Tracking error

Tracking error is a statistical measure that quantifies the volatility of the difference in returns between the ETF and its underlying index.

It is often expressed as a standard deviation percentage and can highlight the consistency of an ETF's performance.

A lower tracking error indicates a more consistent replication of the index.

Tracking difference

On the other hand, tracking difference takes a more straightforward approach by calculating the annualised difference in returns between the ETF and its benchmark.

Unlike tracking error, which considers the variability of returns, tracking difference directly measures the performance gap.

A positive tracking difference means the ETF outperformed its index while a negative value indicates underperformance.

Tracking difference is rarely nil as several factors prevent the ETF from perfectly mirroring its index.

For example, indices that are weighted by market capitalisation adjust their holdings based on the changing market value of their securities. This means index-based ETFs must frequently trade a large number of securities to match these shifts.

ETFs aim to mirror these indices by holding securities in the same proportion as their index weightings, which requires trading potentially hundreds or thousands of different securities to align with the ever-changing index.

However, in practice, when investors buy or sell shares of an ETF, the simultaneous trades of all these securities at their current prices is idealised rather than actual.

Large transactions by the ETF can slightly alter the prices of the securities involved. Moreover, the timing of these trades can vary due to differences in exchange speeds and trading volumes for each security.

Physical vs synthetic ETFs

Additionally, the replication method used by the ETF – whether physical (holding the underlying assets) or synthetic (using derivatives to mimic the index performance) – can significantly influence both tracking error and tracking difference.

Synthetic ETFs tend to experience lower tracking errors than their physical peers.

Physical ETFs, especially those tracking broad or complex indices, might not hold every asset in the benchmark. Instead, they hold a representative sample, which can introduce sampling error – discrepancies between the performance of the sample portfolio and the full index.

In contrast, synthetic ETFs use swap agreements to secure the precise return of the index, effectively eliminating sampling error.

In addition, another reason for tracking error can be how closely a fund's holdings mirror those of its benchmark index.

For example, due to UCITS securities weighting limitations, a non-replicating ETF might sometimes be at risk of being unable to hold the full weighting of the constituents in the underlying index, which would risk increasing the tracking error against the performance of its benchmark.

In contrast, fully replicating strategies – buying all of the securities that make up the index – leaves less room for tracking error.

Finally, illiquid securities can amplify tracking error due to significant price deviations from the market price caused by larger bid-ask spreads when buying or selling the securities.