Last Updated: 15 May 2021
I will report on the former next week but I want to address the swaps vs. physicals issue while it is still fresh in my mind.
For those of you who don’t know, there is a major debate in Europe on the “best” way to structure an ETF. Some issuers, such as iShares, HSBC, ETFlab and Credit Suisse/XMTCH, use “physical replication” for most of their ETFs: if a fund wants to track the FTSE 100 Index, it will buy shares in each of the components of the FTSE 100. Others, such as Lyxor, db x-trackers, Source and Comstage, among others, use a “swap-based approach”, whereby the funds rely on swap contracts written by bank counterparties to track the performance of the index.
There are also some variants. Source and ETF Exchange, for instance, use a swap-based model but rely on multiple counterparties to provide exposure to the market.
The differences between the structures and the advantages and disadvantages of each are debated ad nauseam in the media. ETF issuers spend hours walking through the topic with investors, digging into the minutiae of counterparty risk and generally slinging mud at one another. It’s topic number one at conferences and a major source of content on IndexUniverse.eu.
But here’s my conclusion: it basically doesn’t matter.
As an investor, whether an ETF is physically backed or swap-based would be a long way down my list of concerns. Index choice, expense ratios, market-maker support and the liquidity of the fund are, to me, much bigger issues.
After all, this isn’t the U.S., where the difference between an ETF and an exchange-traded note (ETN) is stark. Nearly all ETFs in the U.S. are physically backed and come with limited counterparty risk. ETNs, by contrast, are debt securities: investors are 100% exposed to the credit of the issuing bank.
But in Europe, there’s really not that much difference in the counterparty risk between swap-based and physically backed securities. Swap-based ETFs are carefully managed under UCITS regulations so that the most an investor will lose in the event of a bank default is about 10%; in practice, it may be much less (if the ETF provider insists on fully collateralising swap exposure, for example).
Proponents of swap-based ETFs argue that physically backed trackers have their own counterparty risk if the funds lend out their underlying securities. But most of that securities lending activity is over-collateralised. You have to paint a very dark and, let’s be honest, far-fetched picture before a fund loses more than 5% of its value due to a securities lending meltdown. (It’s not clear to me, anyway, that swap-based ETFs have no securities lending risk on their collateral either.)
Herein lies the problem: I’m barely grazing the edge of the topic and already I’m getting deep into the weeds. The argument gets much, much more complex than all that. But no matter how far you dig, we’re talking about minimal losses in either structure, even under the worst-case scenarios.
UCITS works. Both fund structures work. Most of the time, the real risk you’re taking in the market is close to nil.
There are a few areas where I would pay attention. My understanding is that physically based ETFs holding UK shares are subject to a 0.50% stamp duty when creating new units, whereas swap-based ETFs tracking the same securities are not. In that case, why pay the tax if there is a liquid swap-based alternative?
But outside of these unusual tax situations, it’s six of one, half a dozen of the other.
Swap-based, physically backed, multiple counterparties − as long as they’re governed by the UCITS regulations, they are all good structures and investors can purchase them with confidence.