We’ve been keeping an eye on the high-frequency trading story since Matthew Goldstein of Reuters broke the story of a computer programmer at Goldman Sachs stealing a chunk of its automated-trading software code.
One of the most intriguing and worrisome parts of the high-frequency trading debate is the idea that the algorithms behind the automated trading will somehow converge, latching onto the same trend and crashing the market in a kind of feedback loop. It’s not an unreasonable question when automated trading accounts for the majority of all stock trades.
Goldstein hits on this point in a column yesterday and further legitimizes by talking to a source who oughta know: Michael Durbin, the guy who set up a high-frequency trading desk for Citadel Investment Group to trade stock options. Citadel is a mega hedge fund, and Goldstein reports its options HFT desk is the industry’s biggest. Durbin is now director of HFT research for a smaller firm. Here’s what he says:
Durbin says it’s reasonable to wonder whether Wall Street’s unfettered embrace of algorithmic automated trading could be setting the stage for a future meltdown.
“You have multiple HFT trading firms and sometimes their agendas are complementary and sometimes they’re not,” explains Durbin, director of HFT research with Blue Capital Group, a small Chicago-based options trading firm.
“There could be a time where these HFT programs unintentionally collaborate and you have a two- or three-minute period where the markets are going crazy. Then other traders respond to it and it simply gets out of control.”
Something similar happened in the great crash of 1987, as Paul Wilmott pointed out in an excellent New York Times op-ed the other day:
It has been said that the October 1987 stock market crash was caused in part by something called dynamic portfolio insurance, another approach based on algorithms…
By 1987, however, the problem was the sheer number of people following the strategy and the market share that they collectively controlled. If a fall in the market leads to people selling according to some formula, and if there are enough of these people following the same algorithm, then it will lead to a further fall in the market, and a further wave of selling, and so on — until the Standard & Poor’s 500 index loses over 20 percent of its value in single day: Oct. 19, Black Monday.
Goldstein has the correct impulse here:
It’s high time for the Securities and Exchange Commission, the Commodity Futures Trading Commission and overseas securities regulators to start working together now to assess the potential systemic risks posed by high frequency trading before a problem occurs.
Indeed. I’d be shocked if the SEC knew diddly squat about what’s really going on here. The problem is that this stuff is a black box, there’s little accountability, and it’s something that could have devastating effects for us all.
Goldstein writes that this potential risk may be all for nought, despite high-frequency traders’ claims that it provides liquidity to markets, making them more efficient. Well, all for nought for us, anyway. It does enrich the traders:
In 1999, some 445 million stock option contracts were traded, according to The Options Clearing Corp. A year later, when the International Securities Exchange opened for business, annual trading volume jumped to 672 million contracts.But stock options trading really began to surge after Citadel, Interactive Brokers and a handful of other brokerages and hedge funds jumped head-first into high frequency trading in 2003. So by 2004, a little more than 1 billion contracts were traded. And in 2008, some 3.28 billion contracts were traded, according to the OCC.
That big increase in trading came from a rather exclusive club.
The focus on potential feedback loop is not to mention the other serious questions raised about whether high-frequency trading programs are allowing firms to manipulate the markets and “front run” via flash orders.
We’ve called for more drumbeat reporting from the press, which too often has a great story that’s just a one-off and moves on to the next topic—something Audit Chieftain Dean Starkman has compared to “corks bobbing on a news Niagara.”
- 1
- 2
I've posted some funky stuff on this (even a poem!) but your more serious readers might like to dig into this transcript of an '08 podcast an algo trading, with links to a half dozen more podcasts/transcripts.
http://housingdoom.com/2009/07/22/aleynikov-origins/
#1 Posted by John McLeod, CJR on Fri 31 Jul 2009 at 04:37 PM
FYI good article on Ars Technica about this
The Matrix, but with money: the world of high-speed trading
http://arstechnica.com/tech-policy/news/2009/07/-it-sounds-like-something.ars
#2 Posted by me, CJR on Sat 1 Aug 2009 at 06:53 PM
more follow up
Computer-trading worries grow as NYSE builds new datacenter
http://arstechnica.com/tech-policy/news/2009/08/nyse-builds-computer-trading-mothership-worries-abound.ars
#3 Posted by me again, CJR on Mon 3 Aug 2009 at 01:12 AM
Good article explaining how it works
Goldman Sachs, the lords of time
By Julian Delasantellis
http://www.atimes.com/atimes/Global_Economy/KH05Dj03.html
#4 Posted by me, CJR on Tue 4 Aug 2009 at 10:54 AM
Honestly, I worry about HFT. I think it has too much potential to add volatility to the market by artificially altering the trading price of stocks in high speed sculpting trades and falling into the program trading trap that we did in '87. It's going to happen, since technology is not something that can be stopped (nor should it), but the regulations around it have to be strict and heavily enforced.
Chris
http://bizconnectionsnow.com/blog/business-funding/fast-trading-potential-for-disaster/
#5 Posted by Chris, CJR on Tue 4 Aug 2009 at 10:41 PM
A mathematical economist acquaintance [not quite the same as econometrician or pure quant] at George Mason Univ. a few years ago pointed out that most of the 'sophisticated' credit risk models being used by pools, hedge funds, rating agencies and 'banks had come to depend on the same set of relations based around what is called a gaussian copula* that, in effect, created an [unintentionally] standardized, flawed model.
*"Li's breakthrough was that instead of waiting to assemble enough historical data about actual defaults, which are rare in the real world, he used historical prices from the CDS market. It's hard to build a historical model to predict Alice's or Britney's behavior, but anybody could see whether the price of credit default swaps on Britney tended to move in the same direction as that on Alice. If it did, then there was a strong correlation between Alice's and Britney's default risks, as priced by the market. Li wrote a model that used price rather than real-world default data as a shortcut (making an implicit assumption that financial markets in general, and CDS markets in particular, can price default risk correctly).
It was a brilliant simplification of an intractable problem. And Li didn't just radically dumb down the difficulty of working out correlations; he decided not to even bother trying to map and calculate all the nearly infinite relationships between the various loans that made up a pool. What happens when the number of pool members increases or when you mix negative correlations with positive ones? Never
mind all that, he said. The only thing that matters is the final correlation number—one clean, simple, all-sufficient figure that sums up everything.
The effect on the securitization market was electric. Armed with Li's formula, Wall Street's quants saw a new world of possibilities. And the first thing they did was start creating a huge number of brand-new triple-A securities. ..
...
The damage was foreseeable and, in fact, foreseen. In 1998, before Li had even invented his copula function, Paul Wilmott wrote that "the correlations between financial quantities are notoriously unstable." Wilmott, a quantitative-finance consultant and lecturer, argued that no theory should be built on such unpredictable parameters. And he wasn't alone. During the boom years, everybody could reel off reasons why the Gaussian copula function wasn't perfect. Li's approach made no allowance for unpredictability: It assumed that correlation was a constant rather than something mercurial."
http://www.wired.com/techbiz/it/magazine/17-03/wp_quant?currentPage=all
Hopefully the long quote drives home a message of what can develope as we move into a world dominated by ultra high speed trading performed by relatively few very large entities - 'feedback' becomes inherent and 'a future meltdown' all but certain as each struggles to maximize its particular global gain across a range of asset classes. The term uncontrollability becomes ever more appropriate.
#6 Posted by Juan , CJR on Sat 14 May 2011 at 04:52 AM