The flash crash of May 6, 2010 occurred exactly a year ago, and while a repeat hasn't materialized, a number of less broad, but similarly startling, price dislocations have occurred in several ETFs in the past few months.
First there were 10 FocusShares ETFs that became unhinged in late March-on their second day of trading. At least one of the funds was driven down to a penny a share in the episode, inviting comparisons to the actual flash crash. Then, about two weeks later, the iPath Optimized Currency Carry ETN (NYSEArca:ICI), a security as thinly traded as FocusShares' new equity ETFs, shot up into the stratosphere. The exchanges subsequently canceled the "clearly erroneous" trades, just as they did after the flash crash.
The affected exchange-traded products in both of these episodes had low trading volume, and the irregular trading began with market orders. Both of these facts are important, and I'll return to them later.
What's rarely mentioned in the media accounts about these ETF "blow-throughs" is that they occur regularly for common stocks as well. I'll give credit to the New York Times for publishing an article in September discussing "mini flash crashes" in Apple Computer (Nasdaq:AAPL) and the Washington Post ( WPO ), among others.
The article, in fairness, didn't mention ETFs, simply because such mini flash crashes are related to nuances of electronic trading and market-structure regulation rather than to some intrinsic flaw with the ETF structure. This is likely to be bad news for the anti-ETF pundits because they'll have to find another product to misrepresent. The simple truth is the ETF structure remains one of the greatest financial products ever developed.
If we look more closely at the development of electronic exchanges, the rules that govern them and how investors and financial intermediaries interact through various order types and processes like high-frequency trading and algorithms, it's easy to see how these periodic "blow-throughs" occur.
There's nothing in the structure or use of ETFs that causes these events, but rather, there are systemic issues with equity trading that have existed in the markets for decades. These events may at times be unintended or inevitable because of new regulations, new technologies or human error. While exchanges and regulators have made great strides to reduce the occurrence and impact of these events, completely preventing them is difficult. That said, more work needs to be done, especially on ETFs.
Every Convenience Brings Its Own Inconvenience
The U.S. equity markets continue to evolve and, by definition, toward greater efficiency. This evolutionary process in the stock market is due to advancements in technology and communication coupled with regulatory changes.
This was evident as markets moved off of floor-based exchanges and into electronic exchanges. The move wasn't made because we suddenly found new technology, but because financial events and the obligation of regulators and exchanges to protect investors created the need for change.
The stock market crash in 1987 is a great example. A large concentration of sell orders in the futures market create a massive selling in the equity market, and the Dow Jones industrial average fell more than 600 points in the last few hours of trading on Oct. 17, 1987.
Investors were frustrated that phone calls weren't answered; thus, orders weren't taken or filled. And, if they were filled, there was very little price discovery or dissemination of prices, as the quoted markets for stocks were "locked" or "crossed"; that is, when the bid side of the market is equal to or larger than the offer side. Some specialists on the floor of the NYSE halted their stocks and didn't open them again until late afternoon on the following day. Some took longer.
In the end, it took the markets almost a week to clear all the trades and get all the stocks trading again at proper bid/ask ratios. The event prompted the development of electronic solutions for small orders to reduce the phone call traffic. The DOT system (designated order turnaround) on the NYSE and the SOES (small order execution system) on Nasdaq were born, making the electronic trading world we now inhabit pretty much inevitable.
Human error can also cause price dislocations. In the mid '90s, a Salomon Brothers clerk on the floor of the NYSE entered a large buy program just before the close of trading that sent the market racing up 1 percent; the problem was that the broker's client entered a sell program.
It took the markets more than two days to reconcile the canceled and replaced trades and for the indexes to incorporate those changes. We spent a whole day not sure what the value of the Dow Jones industrial average or S&P 500 Index was. Changes to order entry and order management systems were quickly developed.
Now, let's fast-forward to a year ago. We heard lots of complaints that the flash crash on Thursday, May 6, 2010 was a combination of systemic and human error not too different from the events I described above. But a big difference is that with the electronic nature of our markets and the ability to execute a trade in one second or less, the impact of unintended and inevitable events are cut dramatically.
On the same day, the market rebounded and found fair value in a matter of minutes-not days or weeks as we had seen in the past. Investors were able to quickly re-enter the market with confidence. This is the convenience of electronic trading. On the flip side, the inconvenience of having "fast markets" is the toll they can take on less liquid securities.
Illiquid ETFs And The Not-So-Unintended Consequences Of Reg NMS
Reg NMS (national market system), in my opinion, is one of the greatest financial regulatory feats of all time. It balanced a need for regulators to catch up with the fast-paced changes and fragmentation that technology was imposing on the markets while balancing investor protection and the cannibalization of existing Wall Street trading desks and the exchanges.
It effectively allowed the U.S. market to become completely electronic, and brought with it the death of the floor-based exchanges. By rule, Reg NMS mandates that exchanges have to deliver a trade to each other in less than one second. Thus algorithms and high-frequency trading became common and necessary market practices. As Reg NMS was being developed and implemented, it was clear that this new rule set was perfect for highly liquid securities like IBM and DELL, but less liquid securities were going to be challenged.
As an example, there was a stock with the ticker symbol "INTL." Many investors assumed that it was the ticker for Intel Corp, which is actually "INTC." INTL was a $3 stock that had small average daily volume, and when INTC-sized orders hit INTL, the volatility was excessive; trades made by investors who had bought or sold the wrong stock had to be canceled.
The big issue wasn't just the ticker but also the use of market orders. Again, it was an unintended consequence of new regulation. The Reg NMS rule that caused trading to occur in subseconds can cause tremendous damage to a security that doesn't have enough liquidity to hold up to excessive order flow.
So, how did regulators account for this in their deliberation for a new national market system? They didn't. They recognized the challenges of illiquid securities and decided to let the free market come up with the solutions.
Unfortunately, the exchanges and broker-dealer community had very little incentive to adjust market structure. Illiquid securities account for a small portion of the revenues for the exchanges, and changes to better accommodate illiquid securities would be a significant cost in technology and human resources.
To be sure, there was no shortage of ideas-among them were proposing the use of nickel increments in quoting rather than pennies, volatility bands and changes to order entry systems in the use of market orders.
But, as I said, the broker-dealer community didn't want to see significant changes, as illiquid securities bring larger spreads and more volatility, both of which produced greater profit opportunities for traders. Thus, the status quo for illiquid securities remains. The exchanges did produce pretrade analytic tools and opening and closing cross-processes to help traders and brokers predict the effect that their trades may have on the market. These are useful tools, but not everyone uses them, and thus human error still exists.
Since a large majority of ETFs are illiquid due to the incubation period that new ETFs need, they are increasingly subjected to these issues of market structure.
Order-Driven Vs. Quote-Driven Securities
ETFs are known as quote-driven securities; there is little need for an order book at prices away from the quoted "top-of-book." This is because ETF pricing is a reflection of the value of the underlying index, and not of any price discovery process based on fundamental analysis and supply and demand, like what we see in common stocks.
The arbitrage process inherent in ETFs keeps the ETF and the index value in line. So, the pricing mechanisms for ETFs and common stocks are significantly different. Also, the creation/redemption process is the main source of liquidity for ETFs and can be accessed at any time. That obviates the need to layer the book with orders, because liquidity in ETFs is not displayed but is accessible.
The current market structure doesn't account for this; Reg NMS has produced an order-driven market. All orders under 10,000 shares must be displayed and there's no real residual source of liquidity outside of the quote prices.
So, in times of stress and with trades executing in subseconds, the liquidity in some ETFs can't be accessed fast enough. Thus mini-flash crashes occur. The same lack of liquidity can be found in various common stocks. But, those stocks are not followed by ETF pundits with an agenda, and thus ETFs get more negative press. It's really unwarranted.
Reducing The Impact
Since the ETF industry can't make changes to either Reg NMS or the current market structure, there are other ways to reduce the impact from unintended and inevitable events.
The first thing that ETF providers did when ETFs moved off of the specialist system and into electronic exchanges was to begin educating their clients on the use of market orders. This is probably the single most important step that can be taken. It could be taken even further to suggest that in ETFs with a low average daily volume, market orders won't be accepted by the exchanges.
Only limit orders would do, but that's something the exchanges and the Securities and Exchange Commission would have to sign off on.
Also, reducing the size of creation/redemption units would allow liquidity to be more easily accessible. This is a more challenging project because the cost of trading would increase.
There have also been discussions at various times to build an ETF-only exchange. This would create the means necessary to write a rule set that is complementary for quote-driven securities. It would include incentives to market makers that would produce better-quality markets, and it may even lead to the development of new financial products.
Once again, there's no shortage of ideas on how to reduce the impact of mini-flash crashes for ETFs. What's missing is an industrywide discussion with the exchanges and regulators.
There's plenty of private discussions with the major ETF providers and the regulators, but no collective ETF industry voice, which actually may be the one flaw of the ETF structure. Who really, in the end, is defending ETFs? The anti-ETF pundits are defining the conversation, and that needs to change.
Richard Keary is the founder of Global ETF Advisors LLC, a New York-based firm that offers advisory services that help clients bring ETFs to market. Keary also spent seven years at Nasdaq, where he helped build Nasdaq's Market Intelligence Desk, and created and built Nasdaq's ETF listings business.
Copyright ® 2011 Index Publications LLC . All Rights Reserved.