Volatility Post Mortem: Was Market Structure to Blame?
Whipsawing markets, massive sell offs, and VIX fund implosions! After a year of record-low market volatility, February 5 was a wake-up call for investors, particularly those who were betting that volatility would stay low for the foreseeable future. On that Monday, the Dow Jones Industrial Average saw its largest ever one-day point drop while the CBOE Volatility Index (VIX) experienced its largest ever one-day percentage gain. Time will tell if this was a healthy correction (likely), or the beginning of an inflation linked bear market (unlikely). 

It’s worth acknowledging what this dip wasn’t: a market-structure event, like the flash crashes of 2010 and 2015. This time, instead of a market malfunction that triggered a fast, steep decline that disproportionately impacted ETFs, trading proceeded in an active but orderly fashion. Exchange traded products (ETPs) were again disproportionately affected, but this time around it was primarily VIX-linked exchange traded notes (ETNs) rather than ETFs.

Why was this time different? Through coordinated efforts, exchanges addressed deficiencies in the Limit Up/Limit Down (LULD) plan that surfaced during the August 2015 flash crash. Because of these changes to how exchanges handle extreme, transient volatility, orderly trading was able to continue with few pauses, even though the volatility-linked ETNs experienced steep declines and, in some cases, were dissolved by their issuers.

Understanding Limit Up/Limit Down, and its Limits

Created in the wake of the 2010 flash crash, the LULD mechanism acts like a circuit breaker for exchanges during periods of extreme volatility. It is designed to keep stocks and exchange traded products (ETPs) trading within specified price bands and to pause trading when a security’s price experiences a sudden, significant move outside its band. The LULD plan promotes fair and orderly markets and seeks to protect investors from sudden declines, which can be exacerbated by disjointed trading.

During the August 2015 flash crash, however, a confluence of factors resulted in chaotic trading and allowed ETPs to trade outside of the LULD-determined price bands. Market makers, who are instrumental in helping pair buy and sell orders, had stepped away from the market in reaction to an unusual early-market-open scenario that prevented them from attaining price transparency for many underlying securities in ETFs. Within 15 minutes of the market open that day, volatility triggered nearly 1,300 trading pauses in 471 securities, 83% of which were ETPs.

NYSE Arca, the home of more than 90% of ETF listings, reopened many ETPs within a tight collar, without pairing off all of the open market orders. Those unpaired orders were released into the continuous trading session and triggered new LULD trading pauses. Ironically, 60% of the LULD trading pauses that day were “limit up” halts, which means that trading was stopped because the security prices were rising through the pre-set price band. This prevented securities from staging an immediate price recovery. Further, a trading glitch caused “leaky bands,” which allowed trades to occur outside of the boundaries of the price bands for a short time upon resumption of trading, exacerbating the declines.

Exchanges Unify to Amend LULD

Following the 2015 event, the major exchanges came together to agree on coordinated procedures for pausing and reopening trading during periods of elevated volatility. The exchanges proposed an amendment to LULD, which was implemented in 2017, and received its first market test in early February. Key features of the LULD amendment were:

  • Harmonization of exchanges: The pace of the 2015 crash accelerated in part because, when trading was paused on an ETP’s primary listing exchange, it continued on other exchanges, which allowed prices to continue to fall. Now, when volatility triggers a trading pause—and until trading in the security can be reopened in an orderly way—all orders are redirected to the ETP’s primary exchange. This prevents securities from trading outside their designated price bands and allows the LULD plan to function as intended.
  • Dynamic trading collars: When a security rises or falls through a trading band, it triggers a five-minute trading pause. During that time, the exchange attempts to pair off buy and sell orders within a set trading collar based on the median price from the previous five minutes. If trades can be paired off, trading can resume. If there aren’t enough offsetting trades within the price band, it triggers another five minute pause and the price collar is widened by 5% in the direction of the imbalance. The collar continues to widen by 5% every five minutes until orders can be paired off and orderly trading can resume.

These changes help increase transparency and create a safe haven during periods of extreme volatility that encourages market makers to remain engaged in the market. This is important given the role of market makers in facilitating trades of ETFs. When they step away from the market, as they did in the 2015 flash crash, the market can lose its equilibrium and prices can decline more quickly.

The Results Are In

On August 24, 2015, the S&P 500 finished down 4.5%, but prices for 19% of ETFs fell 20% or more, closing well below the market value of their underlying securities. This recent bout of volatility impacted far fewer securities: The week of February 5 saw 25 LULD pauses, impacting only 1.2% of the total number of ETPs. This represents less than 8% of the total number of pauses that occurred during the August 2015 flash crash. This time around, 74% of the LULD halts were in volatility-related securities, most of which were ETNs which, unlike ETFs, are unsecured debt instruments that carry counterparty risk.

It is likely that the changes made to the LULD plan helped the markets remain on stable footing despite active trading, and potentially prevented another flash crash.  Investors concerned about ETP-related market structure issues may take some comfort in the results of this first test of the LULD amendment. While it won’t prevent steep market declines, it can help ensure that inadequacies in market structure are not to blame.