Why is the issue of delayed settlement of ETF transactions generating such heated debate?
[This article previously appeared on our sister site, IndexUniverse.eu.]
The words “ETF” and “settlement fail” have been mentioned in the same sentence with increasing regularity in recent months. Is there anything to worry about? And—since the debate has so far focused primarily on the ETF market in the US—to what extent should European investors be concerned?
We’ll address questions of European ETF trading, clearing and settlement in a separate article next week. But why are ETFs being singled out as an example of market inefficiencies?
In a report released in March this year, Harold Bradley and Robert Fawls of the Kauffman foundation and Robert Litan and Fred Sommers of Basis Point Group, a consultancy, pointed out a broad-based trend towards rising settlement “fails” in US capital markets (a “fail” is defined as a trade that is not settled on the date agreed upon by the buyer and seller). According to the four authors of the report, ETFs (and mortgage-backed securities) are two types of financial instrument in which such fails are currently concentrated.
In 2010, according to the Kauffman foundation, settlement fails in US ETFs accounted for over half by value of all reported US securities market fails, even though ETF trades represent a much smaller proportion—less than 10%—of overall US equity trading volumes.
The trend towards rising fails, argued the Kauffman report’s authors, could be a “canary in the coal mine” for regulators, telegraphing that counterparties might be having trouble completing their trades on time and that systemic risk is on the rise.
ETF market participants have responded by saying that such assertions are exaggerated. For example, Noel Archard, North American head of product for the ETF market’s largest issuer, iShares, issued a blog-based rebuttal to the Kauffman report’s claims.
As far as the primary market is concerned (the market in which authorised participants or “APs” interact with ETF issuers to create and redeem fund units), said Archard, if an AP fails to deliver securities to the ETF issuer on the pre-agreed date (typically trade date plus three days, or “T+3”), the firm has to post collateral instead, protecting investors in the fund until delivery occurs.
In the secondary market, said Archard, the ability of market makers to settle with counterparties up to six days after a trade date means that an ETF trade can be reported as “failed” according to the standard equity settlement calendar (also T+3), even though the firm concerned intends to make good its obligations.
According to rule 204 of the US Securities and Exchange Commission’s regulation SHO, issued in July 2009, “subject to certain conditions, fails to deliver resulting from long sales or certain bona fide market making activity must be closed out by no later than the beginning of regular trading hours on the third settlement day after settlement date (i.e., T+6).”
Ben Johnson, European ETF analyst at Morningstar, made a similar point to iShares’ Archard in a report also published in March (“ETF Settlement Delays: No Cause For Alarm”). “While an equity trade that settles on a T+4 basis might be labelled a ‘failed’ trade,” argued Johnson, “it still is ultimately completed, just not within the stipulated time frame.”
Furthermore, continued Johnson, one reason for the fact that ETFs command a disproportionate share of reported trade “fails” is that they are simply traded more than many regular equities.
“The degree to which ETFs are affected by settlement failures is positively correlated with the velocity of their share bases,” said Johnson.
However, according to Fred Sommers, one of the Kauffman report’s authors, there is still cause for systemic concern when delayed settlements occur on a regular basis, and ETFs are a symptom of a disquieting trend.
“Since the 1960s there has been an effort to put rules, systems and processes in place to reduce settlement periods to minimise settlement risk exposure in the capital markets,” said Sommers in an email exchange with IndexUniverse.eu.
“We’ve moved from T+5 to T+3 in the equity market and to same-day settlement in the US Treasuries market. For ETFs, it seems we have reversed direction and actually increased settlement risk exposure for a security that is supported and managed by some of the most systemically important and influential [firms] in the capital markets,” Sommers argues.
“Further, individual investors must settle T+3; who gets the benefit of the 3 days between T+3 and T+6?” asks Sommers. “The data indicate that ETF systemic operational risk has a very different look from equity systemic operational risk. And this risk is concentrated in a small number of ETFs, imposing a disproportionate weight on market costs, processes, liquidity and possibly best execution.”
Sommers’ reference to the “benefits” available to some market participants from delaying settlement ties in with other studies on the subject.
Although operational difficulties can result in settlement fails, according to Federal Reserve Bank economists Michael Fleming and Kenneth Garbade in a 2005 publication, economic incentives may offer a better explanation for the phenomenon. “Surges in fails sometimes result from operational disruptions, but often reflect market participants' insufficient incentive to avoid failing,” said Fleming and Garbade.
Put another way, there may be profits (or reduced costs) available to those market makers who delay settlement beyond the pre-agreed date.
One obvious reason for market makers to delay settlement is that it may be preferable to cover a short ETF position or sell a long position in the secondary market rather than going (via an authorised participant) to the ETF issuer to create or redeem fund units. Going to the primary market carries costs, while waiting for an opportunity to trade in the secondary market, or to borrow the ETF, may be cheaper. If there’s no real penalty for delaying settlement, why not wait, in other words?
There may also be arbitrage opportunities available to traders. According to Morningstar’s Johnson, “there is a financing rate that is implicit to the settlement process. If this financing rate is lower than the cost of borrowing securities or cash to settle a transaction, there may be some small incentive for a party to temporarily delay delivery.”
Or, viewed from a trader’s perspective, a delay in closing out a short position in an ETF can potentially generate a profit, since the short position “earns” the fund fee.
It’s important to point out, argues Morningstar’s Johnson, that the average investor shouldn’t notice any of this.
“Investors gain (or cede) economic exposure to securities at the time a transaction is completed, at what we can describe as ‘T=0’,” says Johnson. “Whether a trade ultimately settles at T+3 or T+4 will have no impact on an investor's returns or access to cash proceeds. And, in fact, investors virtually never even see in their accounts that any trade ever "failed" to settle.”
But while that argument may be true, it’s missing the real point about delayed settlements, argues Basis Point’s Sommers. Beyond systemic risk concerns of the type regulators should be concerned about, says Sommers, institutional investors (and their clients) are paying the cost for operational inefficiencies, of which ETF fails are a symptom.
According to the Kauffman report's authors, “delays caused by settlement failures...result in hidden costs to beneficial owners of assets of as much as twenty-seven basis points every day, or about $300 billion in assets that cannot be reinvested. At a conservative annual interest rate of 3%, this estimate implies that institutional investors are losing US$9 billion annually to settlement failures. Investors are underwriting this transfer to trading profits on Wall Street as one part of this endemic fail-to-deliver problem being overlooked by regulators.”