High Frequency Trading Is Not Fast Enough

A new book by the original yellow journalist of Wall Street, Michael Lewis, initiated global coverage about the flaws of American capitalism. The culprit in Lewis’ new book is High Frequency Trading or “HFT.” There is no doubt that US capital markets are imperfect. New York Times DealBook writer Andrew Sorkin lays the blame at the feet of the stock exchanges of which there are so few remaining that the Federal Trade Commission could label them a monopoly.

Even defenders of HFT, like Tim Worstall at Forbes, have to admit that it has risks and problems. It pushes volatility when markets are under stress; programming errors and misuse of software packages have been known to bankrupt the trading companies. The argument in favor of HFT fails when its proponents bring in “free market” economic theories – primarily because the stock market is not “free” in any economic sense. There are a limited number of big players – 5 banks in the US control 85-95% of trading depending on which market you measure. That is still more like an oligopoly than a competitive market. There are barriers to entry set up by the SEC, the FRB, and state banking and securities commissions. Finally, the transaction costs are enormous. Anyone active in the market knows about trading commissions and management fees. DTCC took in over $1 billion in revenue in 2012 (latest available) and still lost over $25 million. You get the picture – there is no free market argument.

The programs used for high frequency trading are bastardizations of heat transfer dynamic equations. Those underlying equations are based on assumptions. First, they only hold true when time goes to infinity – but trades are executed in finite time. Next, they assume linear behavior – but markets are more like waves than straight lines. Finally, those equations require simultaneity of action. No matter how close the servers are located to the exchange, the computers are not fast enough to read the prices in one market and execute a trade in the next without some lag which violates the assumption. Richard Bookstaber called it A Demon of Our Own Design (Wiley, 2007). The university whiz-kids who built the programs knew they were violating the assumptions but they were under pressure from their Wall Street bosses so they decided to take the money and run the programs – warts and all.

Trading programs treat capital markets as if one security is indistinguishable from the next – and that defeats the purpose of having capital markets at all. The reason we have these markets is so that entrepreneurs can access capital to fund new opportunities. Instead of letting computer programs decide which stock has the best opportunity for a price change, investors should be deciding which business has the best opportunity for success. The funded opportunities create jobs that pay income to households who turn around and put some of those earnings into savings. Lots of little savings accumulate into a pool of loanable funds that become available to other businesses to fund other opportunities to create more jobs, etc., etc. The goal of high frequency trading is to make money – at any cost. And the cost is the ability of capital markets to serve their primary purpose.