A Normal Correction: February 27th Market Break

Includes: DIA, QQQ, SPY
by: Christopher Whalen
"For the market's getting creamed is a healthy thing. It shakes out the froth (that's you they're talking about) and sets the stage for another, bigger and stronger dash upward. So goes the post-meltdown mantra of the chronic bulls." (Alan Abelson in Barron's, March 5, 2007)

On Tuesday, 02/27/07, a 9% drop in the Shanghai market triggered a messy global sell-off, all this following a grim economic pronouncement by former Fed Chairman Alan Greenspan to an audience in Hong Kong. The market break and resulting turmoil in the Big Media verged on a systemic risk event and caused the usually indifferent White House to issue a statement, expressing confidence that the steep decline in global equity markets is "a normal correction."

For many in the financial markets, normal is no correction at all -- or at least one that lasts only a day or so and is painless in career terms. That is why we have things like the plunge protection team (aka the Working Group on Financial Markets) -- to maintain confidence, even as US economic fundamentals wane. But we start to resemble a conspiracy theorist...

The IRA noticed the market's gyrations at the time and, like most of us all, we went back to our chores when the market indices resumed their expected northward migration. But the following month, we got pinged by a fellow member of Professional Risk Managers International Association, who asked if we had noted the miscalculation of the Dow indices during the sharp market sell-off in February. Our colleague likened the event to "systemic information risk" and asked if this was not a precursor to a bigger, more messy event in the future.

Since that blog contact, we've spoken to a number of our colleagues in the risk management world and continue to hear tales of systemic stress coming out of the February market correction. By far the most surprising and remarkable reports, which come from senior operations officers of two bulge bracket firms, is a large jump in "breaks and fails" which apparently occurred during the two days of market carnage.

Last week, we circulated an informal questionnaire to our colleagues in the risk management world, asking a basic question: Did your firm see an increase in settlement breaks or fails during the 2/27-28 period, particularly fails for settlement of cash market equity trades?

One report of sharp increases in broken and failed trades described a situation where exceptions went from 20 pages to over 500 pages for 2/27. Colleagues were compelled to spend days working round the clock manually working through each fail and making a decision whether the firm need to cover a given transaction exposure.

"If 1987 was a 13 standard deviation event, February 27 was a 6," says one player involved. "We saw peak trading volumes almost double during that two days, but the number of breaks and fails expanded exponentially, far out of proportion to the volume increase." The same source claims it took his firm and others weeks to mostly dig out from the surge in bad trades, a process which continues even today.

The thing which catches our attention about anecdotal reports of back office blockages in February is that they are focused on the cash equity markets, not OTC derivatives where everybody thinks the BIG operational risk resides. One of the observations we've heard over the past couple of days is that back office systems only scale to a degree. When trades must be processed by hand or corrected, the throughput of the organization slows to the same pedestrian pace as that for OTC derivatives.

A big issue we hear repeated over and over concerns reduced head count in back office functions, this due to the wide spread of automation, leaving major financial institutions dependent on a relatively tiny cadre of specialists for any clearing mishaps. The vast expansion of the investment world has given experienced back office professionals choices when it comes to employment, leaving larger, productivity-obsessed organizations with a precious handful of people who can perform all of the tasks required to settle a trade manually.

"We manage all of our risk by exception, using automation to manage the event flags," one senior bank credit risk officer told The IRA last week. "But we have less than 20% of the people we did five years ago, managing much bigger trade volumes. Getting the marks [to market] and the systems right is crucial for us."

Another manager observes: "I am the only person on my floor who knows how to do everything in the settlement chain. When I started in the business, everyone in operations sat together and learned the process end-to-end. Now everybody is a specialist sitting alone in an office. When failed trades start to accumulate, we can only scale to a point."

The assumption is that modern cash securities markets are so automated and well-supported by electronic infrastructure provided by the private exchanges and clearing houses that an increase in volume is easily accommodated, but what if this assumption is false?

What if the surges in equity market volume observed over the past year or more are warnings on an impending breakdown in cash market clearing and settlement infrastructure? And what if, after the Street has spent billions improving and adding redundancy to trade clearing systems post 9/11, the weak link in the chain is not the hardware but a lack of skilled back office professionals?

So here's the question: Did you organization observe a "normal correction" in the final days of February? Did your firm observe increased numbers of breaks and fails on 02/27/07? Either as an investor or a dealer, did you see evidence of systemic stress on the cash settlement markets?

Disclosure: none