Index Fund

Full replication buys every constituent at its index weight and is feasible when the index is liquid and narrow enough: large-cap equity benchmarks are the standard case. As the constituent list lengthens or liquidity thins, transaction costs of holding thousands of small positions overwhelm the precision gained. Stratified sampling then divides the index into cells by sector, country, size or, in fixed income, by duration and credit bucket, and holds representative securities matching each cell's weight. Optimisation goes further, using a risk model to select a smaller portfolio whose factor exposures match the index while minimising expected tracking error.

The trade-off is consistent: replication minimises tracking error but maximises trading and custody costs; sampling and optimisation cut costs but accept residual mismatch. Broad fixed-income indices, containing bonds that rarely trade, are almost never fully replicated.

Two distinct measures matter. Tracking difference is the realised gap over a period: fund return minus index return. It is normally negative, because the index is a costless arithmetic construct while the fund pays management fees, trading commissions, bid-ask spreads on rebalances, and irrecoverable withholding tax on foreign dividends. Cash held for redemptions creates drag in rising markets. Securities lending revenue and favourable tax treatment can push the difference the other way.

Tracking error is the standard deviation of the periodic return differences, annualised. It measures consistency rather than level. A fund can have a large negative tracking difference and near-zero tracking error if its shortfall equals the expense ratio every period. Both should be assessed together: expected shortfall approximates total costs; volatility around it reveals how faithfully the replication method works.

The index chosen embeds active choices, so the fund inherits them. Float-adjusted market-capitalisation weighting is self-rebalancing (weights drift with prices, requiring trades only at reconstitution, corporate actions and cash flows) which is why it dominates low-cost products. Equal weighting and fundamental weighting require systematic rebalancing and therefore carry higher turnover and cost.

Reconstitution creates a visible market effect. When additions and deletions are pre-announced, index-tracking demand concentrates on the effective date, and front-running by other participants raises the implementation cost borne by the fund. Concentration is another inherited feature: a cap-weighted equity index can carry substantial weight in a handful of names, so the fund's exposure is neither equally diversified nor immune to sector bubbles.

Separate the two risk concepts. Total risk, measured by the standard deviation of fund returns, is close to that of the index and reflects full market exposure. Active risk, measured by tracking error, is deliberately near zero. A passive fund removes benchmark-relative risk, not market risk, and it delivers the full drawdown of its index in a bear market.

Be able to justify replication choice from index characteristics, decompose expected tracking difference into fees, trading costs, cash drag and tax, and explain why compounding a lower expense ratio produces a materially higher terminal value over long horizons. Also be able to evaluate benchmark appropriateness: a fund tracking an unrepresentative index has low tracking error but may be an unsuitable vehicle for the investor's objective.