IE 11 is not supported. For an optimal experience visit our site on another browser.

Creature from Black Monday . . .  alive

Program trading was blamed as a culprit in the 1987 stock plunge, when it comprised around 10 percent of trading volume. Now, however, it makes up 30 percent. But is there a danger?
/ Source: a href="http://www.washingtonpost.com/wp-srv/front.htm" linktype="External" resizable="true" status="true" scrollbars="true">The Washington Post</a

In the late 1980s, whiz kids on Wall Street were experimenting with computerized trading programs, using new technology that allowed them to trade millions of shares of stocks and related options simultaneously. By 1987, this "program trading" accounted for about 10 percent of all buying and selling in the U.S. stock market and was making big profits for big players, such as university endowments and brokerage firms.

Then came Black Monday. On Oct. 19, 1987, investors worried about inflation and rising interest rates suddenly lost confidence in the market and share prices began to fall. Hundreds of program traders responded by hitting the sell button, and the Dow Jones industrial average took the largest one-day plunge in its history -- 508 points, or 22.6 percent. Within hours, commentators and regulators had fingered computer-assisted trading as the villain in the crash. The New York Stock Exchange swiftly adopted new rules to control program trading, and Congress even talked about making it illegal.

That was then.

Today program trading stands at about 30 percent of weekly volume, according to the NYSE, which defines the practice as a bundle of trades involving 15 or more securities and worth more than $1 million. Pension funds, mutual funds and hedge funds all rely on computers to buy and sell big baskets of investments. Sometimes they're trying to cut trading costs and pass the savings on to their small investors; sometimes they're trying to match their holdings to a particular stock index.

"People are buying and selling portfolios rather than buying and selling individual stocks," explained Hans Stoll, a finance professor at Vanderbilt University. "It's a more efficient way of assembling a portfolio."

Now that the market has entered another period when investors are worried about inflation and whether rising interest rates will trigger a recession, market observers are wondering how program traders would respond to a serious price decline. Some worry that they have already seen a preview, in May and June of this year. That's when hedge funds -- high-risk, high-fee investments that cater to the wealthy -- were blamed for deepening the worldwide slide in emerging market stocks because they used computerized trading to dump entire countries and sectors simultaneously.

"In the course of the normal range of events [program traders] allow markets to normalize and stabilize," said Adam L. Schneider, a partner at Deloitte & Touche who consults on trading systems. But he added that computerized trading could have the opposite effect during extraordinary times. "You still can't [withstand] a situation where every investor runs for the exits at the same time."

What scared market observers in 1987 was program traders' tendency to buy and sell stocks without much regard for the quality or achievement of individual companies. Instead, they use elaborate math equations known as algorithms to compare the prices of various investments and then buy or sell particular stocks or sectors if they appear under- or overvalued compared with historical norms. That means the programs tend to be most active in markets that are already moving and can accelerate a steep decline. (The programs can also inflate a rise, but that upsets far fewer people.)

Two varieties of program trading, known as portfolio insurance and index arbitrage, were considered by some to be culprits in the 1987 crash and another "mini-crash" two years later, in October 1989. In both those instances, big investors were hedging their stock holdings with futures contracts. But the rapid fall of stock prices and market indices triggered sell orders in many programs, which exacerbated the slide.

Though NYSE rules and many computer programs now have built-in stopping points designed to limit huge losses, the sheer amount of money and shares that change hands carry some risk, analysts said. "Program trading is so large and so fast that it can have a multiplier effect," said Peter C. Stockman, a financial services consultant with PA Consulting Group, which works with both brokers and institutional investors.

In the middle of this spring's market slide, program traders bought or sold, on average, more than 1.3 billion shares every day on the NYSE, which estimates that it handles about two-thirds of such trading. Nasdaq, the other major American stock marketplace, does not track program trading, its spokesman said. Since Nasdaq on average trades more shares each day than the NYSE, this further adds to the uncertainty about program trading's effect on the broader markets.

Defenders of program trading and some market experts argue that the practice was unfairly blamed in 1987. They argue that program traders were simply responding to a broader loss of investor confidence and that small investors panicked -- sending the market down even further -- after a ham-handed comment by then-Securities and Exchange Commission Chairman David S. Ruder suggesting that trading might be halted .

They argue that today's computer programs and their underlying equations are far more sophisticated, using options, futures contracts and derivatives to protect investors against falling markets. "Our markets have gotten very, very good at managing risk," said Mark Gurliacci, NYSE vice president for research. "People don't have to run for the exits. They're more hedged than in 1987."

If that is true, computerized trading could actually stabilize a falling market rather than speed up the decline. These commentators credit hedging programs for sending stock market volatility to record lows for the first part of this year and keeping it lower than historical norms.

Gurliacci points out that index arbitrage -- the practice at issue in the 1987 and 1989 market slides -- now accounts for a much smaller percentage of program. This summer, index arbitrage, in which an investor buys a basket of stocks and simultaneously sells futures contracts for the index they represent, or vice versa, has accounted for less than 10 percent of all program trades, down from as high as 50 percent in the 1990s. The identity of program traders has also changed. In the early 1990s, investment banks and brokerage houses trading on their own accounts dominated the ranks. But now large institutional investors -- including mutual funds, hedge funds and pension plans -- account for nearly two-thirds of the volume.

That means small investors benefit, albeit indirectly, in two ways. The mutual funds they rely on are paying lower commissions for their trades, increasing long-term returns; and the constant, rapid-fire trading by computers also keeps the spread between buy and sell prices small, making it easier for small investors to get the best price on individual stocks. "Program trading has taken away a lot of the mispricing that used to exist in the marketplace," Stockman said.

Most program traders and observers argue that the computer programs are here to stay and that in normal markets they do more good than harm. But other experts say they have a nagging worry: Because the algorithms that most traders rely on use historical norms to recommend when the computers should buy and sell, program traders and the market in general could be left vulnerable to huge losses during events that don't fit previous patterns.

That's one of the things that happened in 1987. Back then, many big investors relied on a kind of program-trading called "portfolio insurance" that was supposed to use options and futures to hedge against steep declines. On Black Monday, those programs helped trigger the crash.