"Algorithmic Trading - Economic Value?"

Sydney M. Williams

Thought of the Day

“Algorithmic Trading – Economic Value?”

February 24, 2010

Over the weekend, in Slate, Charles Munger, Vice Chairman of Berkshire Hathaway, published a parable about how one nation came to financial ruin. He writes that the large Pacific island was discovered in 1790 by Europeans and named “Basicland”. The country bears close resemblance to the United States.

He tells the story of a country that gradually descends into fiscal irresponsibility. Citizens worked hard, relied on government only for essential services and came to lead the world in terms of GDP. However, with economic success, government spending, as a percent of GDP, quadrupled over a few decades. An increasing number of people came to rely solely on the beneficence of government, while their affluent neighbors, with more and more leisure time, took to gambling, betting on security prices. In time, 25% of GDP and 22% of all earnings were derived from such activity. Highly talented mathematicians and engineers were lured to work in these casinos, making bets on what were “now called ‘financial derivatives’”.

Over time, increased competition from developing nations and rising energy prices disrupted the calm. Recommendations to discourage speculation, and the concomitant encouragement for the production of items that could be sold to foreigners, were ignored. A belief that one should not interfere with free markets, no matter how destructive they were becoming due to the rising importance of gambling, led traders to double down on their bets. Keynes admonition was ignored: “When the capital development of a country is the byproduct of the operations of a casino, the job is likely to be ill done.”

The end was a tale foretold. Basicland, Mr. Munger concludes his parable, is now under new management and has changed its name to Sorrowland.

The parable of Basicland strikes a chord with those of us who have watched the New York Stock Exchange grow from trading less than ten million shares a day to a billion and a half today. No one – certainly not I – want to return to those earlier days. Trading has benefitted as technology has grown concurrent with finance, especially in the last couple of decades. According to Carol Clark, writing in the most recent Chicago Fed Letter and citing the TABB Group, algorithmic high-frequency trading accounted for 70% of U.S. equity trading in 2009.

Supporters of high-frequency trading, which include an estimated 2% of 20,000 trading firms in the U.S. and who made an estimated $21 billion in 2008 using such techniques, cite benefits to investors – narrowing spreads and increased liquidity. Both benefits may be true, but are difficult to quantify. On the other hand, that the principals of those trading firms benefitted royally there is no question.

The risks of such strategies include the potential for systemic risk. The Chicago Fed Letter cites an example of a trading firm in 2003: “The firm became insolvent in 16 seconds when an employee who had no involvement with algorithms switched one on. It took the company 47 minutes to realize it had gone bust and to call its clearing firm, which was unaware of the situation.” It is the speed and size (in aggregate) of orders which pose risks. Latency is the term used to measure delays and latency is measured in micro seconds (millionths of a second). Credit Suisse was fined by the NYSE in 2007 for an incident in which a programmer changed the parameters on an algorithm which resulted in a message loop that sent 600,000 messages to the matching machine in twenty minutes, at least two thirds of them in error.

At this point we have been fortunate. Errors have been caught and losses have been relatively small. We may not always be so lucky. The size of trading, and the speed which participants demand, risk a financial fire bomb.

There is also the question of economic value. Derivatives, properly regulated, do provide value. They reduce costs of borrowing and allow market players to hedge risk, while providing an outlet for speculators. It is far more difficult to determine ways in which the economy has benefitted through high-frequency algorithmic trading.

Professor Edward Thorp of M.I.T., the godfather of today’s quantitative trading practices tested and honed his skills in Las Vegas in the early 1960s. If less than thirty percent of today’s trading activity is based on fundamental investing, it appears that our markets increasingly take on the mantle of a casino. Without proper regulation, a good dose of common sense and the means of avoiding systemic risk, we risk living the final chapter of Charles Munger’s parable.