Can a flash crash happen again?

By Scot Blythe | November 29, 2010 | Last updated on November 29, 2010
4 min read

For traders, the May 6 flash crash was the highpoint of a two or three year-period that has seen stocks become increasingly volatile. The mid-2000s, as well as the mid-1990s, were comparatively low-volatility regimes.

“The flash crash happened. It was one event. Could it happen again?” asks Bob Schwartz. “Sure it could. Does it happen everyday? Well, I don’t know about everyday, but it happens to a lesser extent.”

The Baruch College finance professor was speaking last week at a Strategy Institute summit on electronic trading and risk management in Toronto.

The low volatility of the mid-2000s and mid-1990s—as tracked by the VIX index published by the Chicago Board Options Exchange—may well have been deceptive. By one measure that Schwartz is investigating, along with Jim Ross and Sila Saylak, volatility has been on the rise since the late 1990s.

They are looking at intraday volatility in 27 Dow stocks from 1993 to 2010, including the flash crash. They examine volatility at half-hour intervals. There has been a marked increase in volatility on the markets open since 1993. It begins to change in 1998-2003 period, then steepens further in the 2003-1010 period.

Why choose the opening bell? Because that’s where price discovery is challenged as markets digest post-close news. By comparison, the rest of the trading day is relatively flat in terms of volatility.

“At various other times we can also have price discovery challenge,” Schwartz notes. “This is every day and what we’re finding, very significantly is that volatility is going up at the open and that’s the challenge of price discovery. It’s not clear cut,” he adds. “What happens is the noise in price discovery, errors in price discovery, leaving their footprints through the transaction record in the form of inflated variance and inflated volatility.”

Some of the factors that may lie behind the flash crash may be responsible, not separately, but in their combination. People originally pointed at a “fat-fingered” trade—a trade with an extra “0” attached, or the sudden sale of S&P futures contract worth $4 billion.

Among the other factors suggested were high frequency traders, toxic algorithms, the internalization of orders before they hit a marketplace, fragmentation among marketplaces—there are now 41 in the U.S. and dark liquidity pools where large block trades are executed. There are also changes in obligations of liquidity providers, unpriced market orders and the absence of good circuit breakers.

All of these represent a spatial and temporal fragmentation of the marketplace. Contributing to that is asset fragmentation through the increasing use of exchange traded funds, Schwartz says. No one factor is responsible for the flash crash; but together, they have profoundly changed the marketplace compared to the early 1990s.

While electronic trading has led to lower commissions and tighter bid/ask spreads, and potentially lower market impact costs for large block traders, it has also, Schwartz thinks, led to more fragile markets, something that can be tracked.

“In recent years, the two big exogenous driving forces of the markets are technology and regulation. I’ve listed four of the regulations that we all know about because I can see them, I can time-date them. Technology keeps evolving; I can’t give you landmarks on technology the same way.”

He singles out four measures: the order handling rules of 1997; regulation ATS in 1998, which opened the NYSE and the NASDAQ to electronic competitors; the introduction of decimalization in 2000; and regulation NMS in 2005, which set the rules for how orders are transacted across multiple marketplaces. With each new measure, there seems to be an increase in volatility at the open.

“You put these bullets together: commissions, spread and impacts and it looks like hey, we’ve entered a brave new world. We are doing better. But,” he adds “the markets have become more fragile.”

The flash crash is just one indicator. Volatility on open is another, though it’s hard to isolate the impact of technology, he says. That requires a multi-variate analysis. “But yes, technology kept reducing the connectivity between trades, trading’s frequency and that explains part of this, but it’s a complex process and I think it pays to look at these through the eyes of the regulation as well as technology.”

He concludes “that the financial markets are fragile, that they are complex, they should be better understood after all these years of study to date and everything else because markets are so complicated and there is more that I think that we could all understand about the dynamic behaviour of price formation in our markets. Market structure needs fixing.”


Scot Blythe