Saturday, June 18, 2011

Millisecond mayhem

The terrifying Flash Crash of 6 May 2010 has long dropped out of the news. The news cycle more of less ended with the release of the SEC's final report on the event in October of last year which concluded that...well... the event got kicked off by a big trade in E-Mini Futures by Waddell and Reed and played out in two subsequent liquidity crises exacerbated -- and crammed into a very short time-sale -- by high-frequency traders. In essence, the report concluded that A happened, then B happened, which caused C to happen, etc., and we had this Flash Crash. What it didn't explore is WHY this kind of this was possible, WHY the markets as currently configured should be prone to such instabilities, or WHY we should have any confidence similar things won't happen again.

I'm not sure what triggered my interest, but I had a quick look today to see if any similar events have taken place more recently. Back in November of last year the New York Times reported on about a dozen episodes it called "mini flash crashes" in which individual stocks plunged in value over a few seconds, before then recovering. In one episode, for example, stock for Progress Energy -- a company with 11,000 employees -- dropped 90% in a few seconds.  These mini whirlwinds are continuing to strike fear into the market today.

For example, this page at Nanex (a company that runs and tracks a whole-market datafeed) lists a number of particularly volatile events over previous months, events in which single stocks lost 5%, 17%, 95% over a second or five seconds before then recovering. According to Nanex, events of this kind are now simply endemic to the market -- the 6 May 2010 events simply seems larger than similar events taking place all the time:
The most recent data available are for the first month and three days of 2011. In that period, stocks showed perplexing moves in 139 cases, rising or falling about 1% or more in less than a second, only to recover, says Nanex. There were 1,818 such occurrences in 2010 and 2,715 in 2009, Nanex says.
 A few specific examples as reported on in this USA Today article:
•Jazz Pharmaceuticals' stock opened at $33.59 on April 27, fell to $23.50 for an instant, then recovered to close at $32.93. "There was no circuit break," says Joe Saluzzi, trader at Themis Trading, because Jazz did not qualify for rules the exchanges put in place after the flash crash for select stocks following extreme moves.

•RLJ Lodging Trust was an initial public offering on May 11. It opened at $17.25 its first day, then a number of trades at $0.0001 took place in less than a second before the stock recovered. The trades were later canceled, but it's an example of exactly what is not supposed to happen anymore, Hunsader says.

•Enstar, an insurer, fell from roughly $100 a share to $0 a share, then back to $100 in just a few seconds on May 13.

•Ten exchange traded funds offered by FocusShares short-circuited on March 31. One, the Focus Morningstar Health Care Index, opened at $25.32, fell to 6 cents, then recovered, says Richard Keary of Global ETF Advisors. The trades were canceled. "No one knows how frequently this is happening," he says.

•Health care firms Pfizer and Abbott Labs experienced the opposite of a flash crash on May 2 in after-hours trading. Abbott shares jumped from $50 to more than $250, and Pfizer shot from $27.60 to $88.71, both in less than a second, Nanex says. The trades were canceled.
 Apparently, according to the Financial Times, something similar happened just over a week ago, on 9 June, in natural gas futures.

I haven't seen anyone who has explained these events in some clear and natural way. I still see a lot of hand waving and vague talk about computer errors and fat fingers. But it seems unlikely these tiny explosions in the market are all driven by accidents. Much more likely it seems to me is that these events are somehow akin to those dust devils you see if driving through a desert -- completely natural if rather violent little storms whipped up by ordinary processes. The question is what are those processes? Also -- how dangerous are they?

The best hint at an explanation I've seen comes from this analysis by Michael Kearns of the University of Pennsylvania and colleagues. Their idea was to study the dynamics of the limit order mechanism which lies at the mechanical center of today's equity markets, and to see if it is perhaps prone to natural instabilities -- positive feed backs that would make it likely for whirlwind like movements in prices to take place quite frequently. In other words, are markets prone to the Butterfly Effect? Their abstract gives a pretty clear description of their study and results:
We study the stability properties of the dynamics of the standard continuous limit-order mechanism that is used in modern equity markets. We ask whether such mechanisms are susceptible to "Butterfly Effects" -- the infliction of large changes on common measures of market activity by only small perturbations of the order sequence. We show that the answer depends strongly on whether the market consists of "absolute" traders (who determine their prices independent of the current order book state) or "relative" traders (who determine their prices relative to the current bid and ask). We prove that while the absolute trader model enjoys provably strong stability properties, the relative trader model is vulnerable to great instability. Our theoretical results are supported by large-scale experiments using limit order data from INET, a large electronic exchange for NASDAQ stocks.
The "absolute" traders in this setting act more like fundamentalists who look to external information to make their trades, rather than the current state of the market. The "relative" traders are more akin, at least in spirit, to momentum traders -- they're responding to what just happened in the market a split second ago and changing their strategies on the fly. Without any question, there are indeed many high-frequency traders who are "relative" traders -- probably most. So mini-flash crashes -- perhaps -- are merely a sign of natural instability and chaos in the micro dynamics of the market.

No comments:

Post a Comment

Post a Comment