[This article previously appeared on our sister site, IndexUniverse.eu.]
Exchange-traded funds work by engaging specialised intermediaries (so-called “authorised participants” or “APs”) to exchange the shares or bonds in an ETF’s underlying index for the ETF shares themselves. Such “creations” of ETF shares take place in standardised units, often worth a few million dollars each. When the AP returns ETF shares to the issuer in the reverse process, a redemption, he receives the index shares or bonds, leaving him where he started.
For a liquid equity ETF, the constituents and constituent weightings of the creation “basket” supplied by the AP to the ETF issuer are pretty much identical to those of the index itself. Look into the creation basket of an S&P 500 or Euro STOXX 50 ETF, and you’d see the index in microcosm, in other words.
But for ETFs tracking less liquid indices or asset classes, the creation and redemption process has never been that simple.
“When we wanted to create a unit of iShares’ MSCI World ETF in the early 2000s, we had consistent problems delivering the less liquid ‘tail’ of index stocks, and this regularly caused the whole creation process to fail,” one former ETF trader told IndexUniverse.eu on condition of anonymity.
The MSCI World index contains around 6,000 stocks from 24 countries.
“The penalty for a failed settlement on a large stock basket was US$30 per stock, and that’s before counting the funding costs for delayed settlement and potential buy-in penalties levied by clearing systems. As a result it was extremely expensive to fail an ETF creation. So we worked with iShares to develop a new methodology: cash creation,” said the trader.
In a cash creation, the AP delivers cash, rather than the index securities, to the ETF issuer. The task of buying the right index constituents then becomes the responsibility of the issuer, which charges APs a variable spread (usually expressed as a percentage of the fund’s net asset value) to cover the costs of acquiring the necessary stocks or bonds.
As well as being used in the traditional (physically replicated) ETF model for less-easy-to-track underlying benchmarks, cash creations (and redemptions) then became the norm for the synthetic ETFs that boomed in Europe after the mid-2000s.
As a synthetic ETF doesn’t actually own the stocks or shares in the index being tracked (instead, it owns a “substitute basket” of unrelated securities, or holds collateral under a pledge agreement with a derivatives counterparty), there’s no obvious need for an AP to supply the index constituents to create a fund unit, and cash creation is the usual practice for such funds.
But the old model of an AP delivering the index shares or bonds to the ETF issuer has evolved as well. In so-called “custom” or “negotiated” basket creations, what the AP actually delivers to the issuer to obtain fund units is the result of often protracted discussions between the two parties. Creations may also involve a mixture of index securities and cash.
And, far from the origins of the like-for-like exchange model used for the first ETFs, some creation baskets may contain only a much reduced version of the index, particularly in the market’s less liquid areas.
Smart beta isn’t smarter than cap weighting, but it is different, and that’s great for investors.
Trial by fire is one way to discover why ETF transparency matters.
Most people now realize leveraged ETFs can hurt you, but how, then, to use them?
What would a shift out of a mutual fund and into an ETF look like up close?