TheChek News Flow

It’s been a while since we looked at the behind-the-scenes mechanics of ETFs, or how market-making and algo-assisted arbitrage in the industry influences price action in stocks and indices more widely.

One big development in recent months has been the SEC’s new ETF rule, which came into effect on December 23, 2019. It was designed to lower barriers to entry into the sector and stimulate more competition.

The rule did get a fair bit of attention when it was first announced in September 2019, not least because it was supported by the US regulator’s notorious innovation champion Hester Peirce -- also dubbed “cryptomum” because of her support for cryptocurrency ETFs. But some important details -- notably how the rules will impact the industry practice of using customised baskets for creation and redemption -- probably failed to get the attention they deserved.

So what’s a customised basket anyway?

In the early days of ETFs -- before the universe was hugely expanded by the introduction of more and more niche index products, synthetics and active offerings -- the basic mechanics of how the system was managed were pretty simple.

ETFs tracked indices. To manage that tracking, they depended on a market-based arbitrage mechanism to regulate supply and demand of ETF units relative to the price of the underlying constituents of the indices they tracked. If the price of an ETF unit was trading at a premium or discount relative to the indicative value of its underlying constituents, a dedicated ETF market-maker (known as an authorised participant or AP) could exploit that mis-pricing until the two values were brought back into line with each other.

When ETF units trade at a premium to the so-called indicative net asset value (iNAV), APs have an incentive to buy the constituent stocks that make up a creation basket so they can receive ETF units. They can then sell these at a premium to the market capturing a spread. Such activity is known as creation and expands overall fund holdings as well as shares outstanding.

When ETF units trade at a discount to the iNAV, APs have the opposite incentive. They deliver ETF units to the manager in exchange for discounted holdings which they can then sell in the market at normal prices, capturing that differential for themselves instead. Such activity is known as redemption and it reduces overall fund holdings as well as shares outstanding.

All of which is simple enough.

Except, the reality on the ground was never that simple. Practitioners soon realised the above was an idealised process. While it was desirable in terms of risk management to faithfully replicate index weighings in baskets, in practice doing so didn’t account for the day-to-day frictions which could inhibit the smooth running of the creation and redemption and thus inhibit good ETF tracking.

One such friction was the variance in the availability of the underlying constituents and the potential difficulty in forging perfectly weighted baskets for delivery at short notice. This led many ETF managers to allow APs to deliver customised baskets, effectively permitting stocks to be delivered in weightings that suited APs rather than shareholders. In some cases, hard-to-find stocks or bonds were even allowed to be substituted with cash outright or other collateral (increasing the risk of index-tracking error even further).

The regulatory approach to this sort of customisation, however, was never consistent. As the SEC notes, in the early days of ETFs there were few explicit restrictions on basket composition. From 2006 onwards, however, the SEC began to place much tighter restrictions on baskets, requiring they generally correspond pro rata to the advertised portfolio holdings, allowing only limited circumstances in which ETFs could use non-pro rata baskets by way of applied exemptions.

As the SEC noted in its September ruling, the logic for discouraging customisation relates to the risk that APs could take advantage of their relationships with ETF managers to pressure them to construct baskets that favoured them -- to the detriment of ETF shareholders or the fund’s tracking objectives. The further fear was the system would encourage dumping of less liquid securities in creations, and the cherry-picking of more liquid ones in redemptions. The regulator also worried it could introduce a general liquidity mismatch risk to portfolios (our emphasis):

For example, because ETFs rely on authorized participants to maintain the secondary market by promoting an effective arbitrage mechanism, an authorized participant holding less liquid or less desirable securities potentially could pressure an ETF into accepting those securities in its basket in exchange for liquid ETF shares (i.e., dumping). An authorized participant also could pressure the ETF into including in its basket certain desirable securities in exchange for ETF shares tendered for redemption (i.e., cherry-picking). In either case, the ETF’s other investors would be disadvantaged and would be left holding shares of an ETF with a less liquid or less desirable portfolio of securities.

With respect to cash substitution the other risk was also that:

... during periods of market stress, an authorized participant may demand cash from the ETF instead of less liquid securities in exchange for ETF shares, impacting the liquidity of the ETF’s portfolio and the ability of the ETF to satisfy additional cash redemption requests from authorized participants

But the flip side of the argument was that discouraging customisation only added expenses and costs. ETFs without basket flexibility would typically require a greater number of individual securities within their baskets. This, the SEC noted, could lead to wider bid-ask spreads and potentially less arbitrage, creating a situation where ETFs with pro-rata basket rules would be competitively disadvantaged against those with more flexibility on composition. The disadvantage might then encourage them to turn to exemptions that allowed them to substitute for cash collateral, which the SEC worried would result in “cash drag” on ETFs’ performance and increase tax exposures.

Given the above, the SEC’s new rule aims to level the playing field by making basket customisation far easier for all ETFs. But -- in a bid to keep bad behaviour in check -- there’s a proviso that those who engage in custom baskets must adopt a formal policy or procedure governing their construction. They must also keep internal records, while appointing specific employees with the responsibility of ensuring custom baskets comply with the formalised policies. This is no doubt to prevent APs from exploiting the process by demanding overly bespoke adjustments that favour them over shareholders and other APs.

And yet, bizarrely, the SEC has stopped short of demanding public disclosure of what customised baskets end up being. This means it will still be very hard for the market at large to track or understand the potential correlation risk being introduced into portfolio holdings because of basket customisation.

The SEC had originally intended (and we think rightly so) to require ETFs to post information about such baskets on their websites each day so as to “facilitate arbitrage by providing APs and other market participants with timely information regarding the contents of a basket that the ETF will accept each day”. Also, to “allow market participants that do not have access to an ETF’s daily portfolio composition file (PCF) to compare the ETF’s basket with its portfolio holdings, assist in building intraday hedges, and estimate the cash balancing amount”. Both are strong arguments for disclosure.

In overturning the proposal it does lead us to wonder whose interests the SEC really ended up representing?

To support its non-disclosure reversal, the SEC cited a range of counter-arguments from industry commenters, ranging from the weak assertion that enforcing disclosure would be “irrelevant for secondary market investors”, to the idea it would risk confusion “particularly if the basket is mistaken for portfolio holding information”. The other contention was that public disclosure could “delay the process by which the ETF and an AP negotiate the contents of a custom creation or redemption basket” and that the basket composition information could be accessed through the National Securities Clearing Corporation (NSCC), an intermediary or the ETF itself.

But the real motivation for dropping the disclosure requirement probably related to the industry panicking that the act of forcing ETFs to publish basket compositions before accepting orders for creations or redemptions could, as the SEC put it:

... raise the risk that market participants front-run trades in basket securities or attempt to replicate authorized participants’ or other market makers’ trading strategies, particularly for those ETFs that have more frequent primary market transactions.

This is undoubtedly a real risk. But it’s also true that publicly publishing basket information was always going to disrupt the clear and obvious trading advantage afforded to market participants who have privileged access to PCF files in any shape or form at all.

What’s more, arguing the information can be easily obtained from other sources such as the NSCC or from ETFs directly hardly solves the transparency problem, since the sort of people likely to scrutinise the data for correlation or liquidity risk (retail investors, journalists and market-watchers) would still probably not have access.

As investment-research company Morningstar noted in its own comment to the SEC:

We believe that transparency around baskets is critical for investors. We understand that ETF sponsors need the flexibility to deviate from pro rata baskets. As such, all baskets that were utilized in a day should be disclosed, at a minimum, by the end of the day. Such disclosure would allow analysis of whether the baskets are facilitating appropriate market making and liquidity, particularly during times of market stress. We also believe that other information, such as whether the fund faces any impediments to creating new units and whether it is pending closure, should be disclosed on any day that this information is applicable.

And that really is the key point. Unless PCF information is made publicly available -- at a minimum -- on an end-of-day basis, those actors with privileged access to it will always get advanced insight into potential tracking-error risk and other impediments arising in the creation/redemption process.

And so we find ourselves in a situation where the SEC’s new rule will allow APs and ETF managers to mess around with deliverable/redeemable basket compositions like never before. And yet, the chances of analysts or media spotting any potential correlation, tracking or liquidity mismatch issues that arise from this freedom to deviate are more than slim, because none of that data is any easier to obtain by the public.

If any readers do have access to the PCF files of any prominent US ETFs, it would be great to see what impact the rule is already having, if any. Please do share!