Tracking Difference vs Tracking Error

Tracking difference: the report card

TD over a year = fund total return − index total return. A fund returning 11.62% against an index at 11.80% has a TD of −0.18%: that, not the TER printed on the factsheet, is what tracking the index actually cost you.

Why TD ≠ TER, item by item (each with its home chapter):

• The TER itself: the baseline drag (Chapter 8).
• Withholding taxes on dividends leaking inside the fund: for a world fund, easily as large as the TER; domicile-dependent (Chapters 21, 51).
• Securities-lending income: basis points earned back, sometimes several (Chapter 36).
• Sampling and rebalance execution: the craft term, positive or negative (Chapters 5, 37, 41).
• Cash drag and dividend-timing effects: small, real.
• For synthetics: swap spreads netted against the gross-dividend edge (Chapter 11).

The sum is the number, and the number frequently reorders rankings: Chapter 24's worked example, where the cheaper-TER fund lost on realized TD, is the normal case, not the curiosity. Across the flagship world- and S&P 500-fund peer groups, realized TDs routinely span a range wider than the TER differences between the funds: which is the entire argument for measuring rather than assuming.

The three rules of measuring TD honestly

1. The right benchmark variant, always. Indices come in price, gross-return and net-return flavors (Chapter 24 introduced this). Net-return indices deduct a standardized withholding tax from dividends: the fair yardstick for UCITS funds, and the reason well-run Irish physical funds can show positive TD against MSCI's net indices (the standard assumes ~30% withholding; the fund achieves 15%, and lending adds more). A fund quietly benchmarked against a price index is flattering itself by the entire dividend yield. First check, every time: which variant?

2. Multi-year or nothing. Single-year TD is noisy: dividend timing, rebalance luck, one-off tax items. Three to five years of annualized TD is the signal; consistency across years is the craft indicator. The data sources: the fund's own annual reports and factsheets (Chapter 24's gap-between-rows), issuer TD disclosures, and independent aggregators that tabulate multi-year TD across UCITS funds: this site maintains a dated table for the flagship peer groups rather than baking numbers into prose.

3. Compare within a peer group, not across missions. TD ranks funds on the same index (or near-identical ones). A −0.35% TD on an EM small-cap index may be brilliant craft; −0.10% on the S&P 500 may be mediocre. The index sets the difficulty; TD grades the execution.

Tracking error: the other metric, correctly filed

TE = the standard deviation of the fund-minus-index return differences (usually daily or monthly, annualized). It measures wobble around the index, not distance traveled: a fund could lag by exactly 0.20% every single year (TD −0.20%, TE ≈ 0) or match the index on average while swinging ±1% around it (TD ≈ 0, high TE).

Who legitimately cares about TE: institutions with mandates (a pension replacing index exposure needs the replacement to behave identically week to week), traders and hedgers (basis risk), and regulators (UCITS rules require funds to disclose expected TE and explain breaches). Who mostly shouldn't: the long-term holder, for whom the destination (TD) dominates the smoothness of the ride, and for whom a slightly higher TE from smart sampling that improves TD is a trade worth taking every time.

The terminology trap to disarm: much retail commentary, and some fund marketing, says "tracking error" while meaning TD. When you meet the phrase, ask which one is meant: a fund boasting "tracking error of only 0.05%" may still be losing 0.4% a year to the index with perfect consistency: consistently mediocre, precisely measured.

The complete selection method, assembled

This chapter completes the fund-selection machine the book has been building:

1. Index (Chapters 5, 20, 39): the decision that sets what you own.
2. Structure and class (Chapters 7, 9, 11): domicile, replication, Acc/Dist, listing.
3. Shortlist on TER and size (Chapters 8, 67): eliminate the expensive and the endangered.
4. Decide on 3–5 year TD vs the correct benchmark variant, checking consistency: this chapter.
5. Sanity-check your own layer: spreads, venue, FX, wrapper (Chapters 12, 14, 56–57): the costs TD can't see because they're yours, not the fund's.

Applied to near-identical flagship funds, the method typically finds differences of 0.05–0.25%/yr: small, real, compounding, and, unlike performance chasing, structural: yesterday's withholding efficiency and lending program persist in a way yesterday's returns don't.

A worked example

Lena shortlists three Irish MSCI World UCITS funds; net-return benchmark confirmed on all three factsheets:

Fund C, the "expensive" one, has beaten its net benchmark: aggressive lending program (she reads Chapter 36's disclosures and accepts the collateralized risk) plus withholding efficiency. Fund A nearly matches it with a smaller lending book. Fund B, cheapest on the label, costs the most in reality, every single year: consistency in the wrong direction. She takes A, files C as the equally rational choice for someone comfortable with more lending, and notes what just happened: the TER ranking (B, A, C) and the truth ranking (C, A, B) were exact opposites. That inversion, invisible to everyone who stops at Chapter 8, is Chapter 35's entire reason to exist.