Custom Baskets

What a custom basket does

The standard basket (Chapter 69) is one-size-fits-all. A custom basket is negotiated for the transaction in front of you, and it earns its keep in four situations:

1. The AP cannot source the standard basket. A market is closed, a bond has not traded in weeks, a constituent is capacity-constrained. Rather than defaulting to cash, which forces the fund to trade and bear the cost, the parties agree a substitute list the AP can actually deliver.

2. The fund wants specific securities in or out. A creation is an opportunity to acquire what the portfolio is short of; a redemption is an opportunity to send out what it holds too much of. Custom baskets let the manager use flows to keep the portfolio aligned with the index, replacing trades the fund would otherwise pay to execute. This is real, unglamorous, and worth basis points of tracking difference.

3. Index changes are approaching. Around reconstitution (Chapter 41), custom baskets let a fund transition holdings through flows rather than through market trades, reducing the crowd-trade cost every tracker pays.

4. Embedded gains need managing. In the US, redemption baskets loaded with low-cost-basis lots remove unrealized gains in kind (Chapter 54). This is the tax-efficiency mechanism, operating at transaction level.

The regulatory history, and why it mattered

Before 2019, the ability to use custom baskets in the US depended on the terms of a fund's individual exemptive order. Older, larger complexes had generous relief; some newer entrants did not. The result was a structural advantage for incumbents, unrelated to skill: a fund that could use custom baskets had better tax efficiency and lower transition costs than an identical fund that could not.

Rule 6c-11 (2019) replaced the patchwork with a standard: any ETF within its scope may use custom baskets, provided the fund adopts written policies and procedures specifying the parameters under which they will be used, sets conditions ensuring baskets are in the fund's best interest, and subjects the arrangements to board oversight and record-keeping. The regulator's judgment was that custom baskets are operationally beneficial and that the risk (favouring one AP over another, or dumping unwanted securities on the fund) is better managed through governance than prohibition.

The honest assessment of the outcome: it levelled a real competitive asymmetry, improved bond ETF functioning materially, and simultaneously made the aggressive tax-management techniques of Chapter 54 available to everyone rather than to incumbents alone. Both consequences are real.

Where the criticism bites

Custom baskets are the tool behind heartbeat trades: a short-lived inflow arranged with a partner, followed shortly by a large custom redemption carrying out low-basis positions, scrubbing embedded gains beyond what organic flows would achieve. Journalistic investigations documented the pattern, academics have measured it, and it has drawn Congressional attention without legislative change so far.

The balanced position this book takes:

• The underlying mechanism is legislated and beneficial. In-kind redemption not triggering fund-level gain recognition is decades-old law, and it protects ordinary index-fund holders from other investors' exits.
• Custom baskets make the mechanism sharper, and sharpness has an edge. Using a partner's temporary capital purely to manufacture redemption capacity is closer to tax engineering than to portfolio management, and the fact that it is legal does not make it the same activity as ordinary basket flexibility.
• The governance answer is plausible but untested at the margin. "Written policies plus board oversight" is exactly the kind of standard that works well in normal operation and is difficult to police at the aggressive end.
• Policy risk is real and prospective. If the treatment changes, it would most plausibly apply to future gains, which limits the practical exposure of a current holder.

An investor's practical takeaway is modest: the funds most exposed to a rule change are those whose tax efficiency depends most on aggressive techniques, which tends to mean high-turnover strategies rather than broad index funds.

The bond market case, where it is uncontroversial

Fixed income is where custom baskets are least debated and most necessary. A bond ETF cannot demand delivery of a pro-rata slice of 5,000 bonds; the AP would have to source instruments that simply are not available. Custom baskets let the parties agree what the AP can deliver, matched to the fund's duration, sector and credit needs.

Without them, bond ETF creations would rely far more heavily on cash, the fund would bear the trading costs of assembling bonds itself, and both tracking and spreads would deteriorate. A meaningful share of the improvement in bond ETF functioning over the past decade traces back to more flexible basket practice; and it is a large part of why the wrapper held up in March 2020 (Chapter 84).

Governance: what good looks like

For anyone assessing a fund's practices, the questions are:

• Are custom basket policies disclosed in the prospectus or SAI, with the parameters described rather than merely asserted?
• Is there evidence of even-handedness across APs, rather than one favoured counterparty?
• Does the fund's realized capital-gains history look like an index fund's, or does it show the signature of heavy engineering (unusually perfect zero distributions in a high-turnover strategy)?
• Does the annual report reconcile basket practice with actual portfolio turnover?

None of these is a red flag on its own. Together they distinguish operational flexibility from tax engineering: which is the distinction the rule intends and does not always achieve.

A worked example

A €3bn EUR corporate bond ETF receives a €150m redemption request.

• Standard basket: a pro-rata slice of 1,800 bonds. The AP cannot source many of them; the fund would have to sell bonds itself and deliver cash, paying spreads in an illiquid market and realizing gains.
• Custom basket, negotiated: the fund proposes 240 specific bonds: positions it is overweight after recent index changes, weighted to match the portfolio's duration and credit profile. The AP confirms it can distribute them into client demand it already has.
• Result: the fund's overweights are removed at no trading cost, its duration and credit exposure remain matched to the index, continuing holders pay nothing, and the AP receives exactly what it can place.

Nobody was disadvantaged and the fund's tracking difference improved. That is what custom baskets do on an ordinary day; and it is worth holding that picture alongside the heartbeat critique, because both are accurate descriptions of the same tool.