Can We Get a Grip on the Machines that Move the Market?

Are the media lazy when it comes to reporting on market movements? No, they just don't know any better. When the markets fall or rise every day, editors ask for stories that explain what's going on. Reporters get quotes from analysts, and their stories get published. That casual approach to explaining price changes soaks into investors' awareness and yields false security about the stock market.

So when we have a 1,000 point intraday crash on Thursday and days later there's no clear explanation of what happened, despite the government and media looking into it, you'd think that people would be jarred awake.

Decades ago, when the October, 1987, 22.6% drop in the Dow was blamed on computer-program-based trading, the explanation for what moves markets tipped. But the media still like a simple human emotion model -- even though it's increasingly irrelevant since computers account for the vast majority of trading. Even more alarming, we know very little about such computer trading and their impact on the market.

Volatile Mix of Humans and Machines

Let's examine the state of play for explaining what caused Thursday's market massacre. According to The New York Times, the fat-finger explanation, wherein a trader accidentally hit billion when he meant to key in million, seems to be losing its explanatory power. The latest working hypothesis is that the 1,000 point drop was due to glitches in the hand-offs among human and machine-driven markets.

More specifically, the New York Stock Exchange (NYSE) allows people to look at trades before they get executed; whereas the Nasdaq OMX market does everything by machine. The Times suggests that what may have happened is that for some reason, as yet unknown, six individual stocks, suddenly dropped in value. An extreme example is Accenture (ACN) which plunged from $41at 2:30 p.m. to a penny a share at 2:47:53 p.m on Thursday afternoon, according to The Wall Street Journal.

The NYSE person in charge of Accenture stopped trading it for 90 seconds so the search for buyers of the stock switched via computer over to Nasdaq. Since Nasdaq didn't have price limits below which a trade would be stopped, the lack of demand for Accenture stock drove its price down to a penny where computerized bidders stepped in.

This price drop somehow cascaded to other computer-based traders. The Times quoted an official who said, "The pressure in the less-liquid markets was amplified by the computer-driven trades, which led still other traders to pull back. Only when traders began to manually respond to the sharp drop did the market seem to turn around."

Flash Trading's Murky Role

It's become something of a religion to look for fear and greed as the emotions that rule the market. But usually, there's little emotion involved at all.

That's because computers account for the vast majority of trading. These machines account for 70% of trading volume, and it's variously known as flash trading or high-frequency trading. Flash traders profit from legal "insidery information" -- by trading ahead of orders to buy or sell stocks on the way to being executed -- and may have contributed to Thursday's 1,000 point loss.

To be sure, emotion has not completely left the building. One way to measure that is through an index that prices fear levels -- such as the Markit iTraxx Financial Index of credit-default swaps (CDSs) on 25 banks and insurers, according to Bloomberg.

This index measures fear by setting the premium that an investor needs to pay to insure against the risk that a European bank bond issuer will stop making their payments. By that measure, the fear of owning European bank bonds has never been higher -- climbing 22% Friday to 223, according to Bloomberg. With European banks holding government bonds from countries like Greece, Spain, and Portugal as part of their capital bases, a default on those government bonds could force European banks to scramble to raise capital.

No Answers for Important Questions

With no clear understanding of how the machines that move the market will react to the rising risk in European banks, the humans whose 401(k)s are slammed by the resulting market plunges might naturally react with waves of nausea.

There are so many questions to which we deserve answers: Exactly how much capital is being traded using these computers? Who are the biggest operators? What are the criteria these computers use to buy and sell securities? How much money is borrowed by these operators as a proportion of their capital? How do the interactions among the computers affect prices and trading volume?

What we really need is a clear and compelling explanation of what's really going on in the market. Despite knowing for decades about the powerful role of machines in our markets, we still can't get a clear explanation.