In the Terminator series, machines rise up to replace humans. On Wall Street, fiction now appears to be reality, as complex algorithms and super-fast computers conspire to wrest control of market liquidity.
Recently, the New York Times brought attention to the ins-and-outs of high frequency trading, or HFT. (Read here.) And with Goldman Sachs producing eye-popping trading results, many observers – looking to identify the next weapon of mass destruction – are crying foul. (Read here, Goldman's profits in HFT.)
Sen. Charles Schumer (D-NY), pushing to ban components of HFT, is increasing visibility even more.
HFT, basically, is rapid-fire buying and selling of securities, in milliseconds. Proprietary trading desks and hedge funds orchestrate it for own-account gains, trading for pennies or less literally in the blink of an eye. By constantly churning, these funds' trading activity often constitutes the market for many different securities. (Read explanatory diagram here.)
Here's where problems arise. When a mutual fund submits a buy order with a limit price, HFT funds can buy-and-sell the difference between the current price and the limit price. If, for example, a dealer looks to execute a buy-order with a $25 limit and the price is $24.90, HFT funds could theoretically profit from the entire 10c difference.
As the mutual fund portfolio manager, you worry that your best price should have been something less than $25. You know that pennies on millions of shares represent a material cost, and that your shareholders bear this.
In this case, the market isn't achieving true price discovery for the security in relation to its liquidity. A substantial amount of liquidity is addressing price movement alone (irrespective of the security), while only a portion is takng into account the security's actual value.
HFT is controversial because it potentially creates information asymmetry and false liquidity. Its proponents argue, however, that barriers to entry are low (once you get past the set up costs), so pricing is actually transparent; and some liquidity (albeit low quality) is better than no liquidity.
As with derivatives, the problem – if we want to call it that – in HFT isn't the product itself, but the typical inefficiencies of a marketplace that can never be completely uniform. Some folks are just really clever at exploiting those inefficiencies, and market functions take time to keep pace.
Whatever response lawmakers and regulators take, as long as there's money to be made, the folks who brought you CDOs and HFT will spawn something new and different.
That is, until the machines actually do rise up, and all humanity is finally gone.
How to Understand High Frequency Trading (Atlantic Online)
High-frequency Trading (Marginal Revolution)