A Lesson From the NYSE Outage: Fragmentation Works

Last Wednesday, for a few hours if you paid attention to Twitter, you might've thought that the United States was being brought down one prominent institution at a time.

First, United Airlines downed all of its flights thanks to a computer glitch. A little later that day, the Wall Street Journal's website went down for a little while.

But for this audience, there was the Big One: New York Stock Exchange (NYSE) suspended trading for over three hours due to a new software release that was intended to test new requirements for trade timestamps, according the Financial Times. The New York Times quoted a trader saying that a whopping 700,000 trades had to be manually cancelled.

Senator Barbara Mikulski (D-MD), who serves on the US Intelligence Committee, wasn't quite buying that all these events weren't intertwined in some way, perhaps thanks to a coordinated cyber assault on America's infrastructure. FBI director James Comey dismissed such speculation. John McAfee isn't so sure.

Glitches Are Inevitable 

No one skipped a beat. They had to stop routing orders to NYSE, but there was plenty of liquidity in the marketplace. So as a capital markets question, the system worked.

But let's assume that NYSE is to be believed and it was a simple software glitch and United's incident was a coincidence. I have a few friends who work as traders or in operations on Wall Street, and they said that this outage didn't seem like that big of a deal. Taking away the cyber-security speculation, none of my sources hammered NYSE on the glitch, though some have taken to blogs with pitchforks in hand.

Basically, these things happen, and the market reacted in fine form.

I was speaking with one source whose opinions I value very much. He gave me the equivalent of a shrug when I asked him about what he's heard from clients since the outage.

"These are very complicated systems and pieces of software," he says. "No matter how much testing you do on this stuff, an odd event can occur that no one could've anticipated or tested for. You build it into your test cases going forward, but it's still a piece of software. It's inevitable that these things will happen, but you build the system to survive these inevitabilities."

Fragmentation Worked ... In This Instance, Anyway

While this was a very embarrassing event for NYSE and for the Intercontinental Exchange (ICE) ─ which now owns the venerable exchange ─ it was more dramatic rather than harmful. All these outages happening in succession led to people thinking the worst, but the market performed well, broker-dealers continued to process orders and reroute flow, and there was liquidity in the market. If this was two decades ago, that would not have been the case, says the source.

"The good part is ─ and this goes right to the heart of what makes the markets better ─ is that NYSE went down but everyone else was open," he says. "Twenty-five years ago if NYSE went down, all the derivatives markets would've closed. But no one skipped a beat. They had to stop routing orders to NYSE, but there was plenty of liquidity in the marketplace. So as a capital markets question, the system worked.

"These events are valuable," he adds. "It's like a brushfire: It cleans the bad stuff out and allows more life to grow. Everybody learns from these events."

The greater concern is something that's more systemic. Repeated outages aren't tolerated. If these types of outages and glitches prove endemic at NYSE, then ICE CEO Jeffrey Sprecher will have a problem on his hands.

As an example, the Singapore Exchange has seen an overhaul at the top of the company and plans to improve its systems after the operator suffered three trading stoppages last year.

Moscow Exchange recently announced that it will "implement new precautions to ensure uninterrupted failover to its backup systems in the event of any service interruption, following the third issue with market data distribution in as many weeks," according to Sell-Side Technology sibling publication Inside Market Data.

And regulators are also focusing on ensuring that exchanges and trading participants are bolstering their infrastructures as more and more trading is conducted electronically. Last year the Securities and Exchange Commission (SEC) introduced the Regulation Systems Compliance and Integrity (Reg SCI) mandate in an effort to strengthen the US securities markets' systems' compliance and integrity.

For those in favor of fragmentation, the NYSE outage served as evidence that more options help the market to recover when there's a bump. If this would've happened at the CME in the futures market, or if this would've occurred during NYSE's opening or closing auction ─ when the venue sucks in added trade flow ─ the results could've wrecked more havoc.

Maybe I'm being naïve or am ill-informed, but NYSE's outage isn't all that concerning, so long as this serves a blip on the radar, rather than the first quake that yields a series of aftershocks.

Disagree? Should the markets be worried? Give me a call (646-490-3973) or shoot me an email.

A Few Random Thoughts

● Looking for some young talent to hire on, how about this 15-year-old kid who uncovered a math error in the golden ratio at a 34-year-old exhibit.

● You absolutely must read this story from The New Yorker about the Cascadia fault line in the Pacific Northwest. According to some estimates based on very good science, there's a one-in-three chance that an earthquake of 8.0-8.6 will hit the Pacific Northwest in the next 50 years, and one-in-10 that a quake of 8.7-9.2 will strike in that timespan.

Most homes and buildings near the Cascadia subduction zone are not quake proof, as it wasn't understood until the 1980s and early 90s that the region had already suffered one devastating 9.0 earthquake in 1700, before the land was settled. Here's one graph that really jumped out at me:

"We now know that the Pacific Northwest has experienced forty-one subduction-zone earthquakes in the past ten thousand years. If you divide ten thousand by forty-one, you get two hundred and forty-three, which is Cascadia's recurrence interval: the average amount of time that elapses between earthquakes. That timespan is dangerous both because it is too long-long enough for us to unwittingly build an entire civilization on top of our continent's worst fault line-and because it is not long enough. Counting from the earthquake of 1700, we are now three hundred and fifteen years into a two-hundred-and-forty-three-year cycle."

● RIP to Nintendo's president Satoru Iwata.