The New York Stock Exchange, owned by Intercontinental Exchange Inc. (ICE, Financial) and listed on its own exchange, just got permission to implement a speed bump on its exchange, which lists 370 companies. The purpose of the speed bump is to prevent high frequency traders, or HFTs, from clipping pennies off front-running trades. These trades are, as the nomenclature suggests, very high frequency, loading and offloading huge positions in fractions of a second.
The idea of doing that rubs traders of all kinds the wrong way, but should it?
The speed bump delay will be 350 microseconds per trade, or 350-millionths of a second. That means HFTs looking to hold positions for shorter than that amount of time will not be able to. As reviled as HFTs are in investment culture, though, the notion that they rob people of profits is questionable at best.
In order to trade anything, whether high frequency or long-term hold, one obviously needs to find a willing partner, either a buyer if you’re selling or a seller if you’re buying. For argument’s sake, let’s say that an HFT wants to buy 1 million shares in stock A and sell it all within 100 microseconds to make a profit of one-tenth of a cent per share. In that case, the HFT will make a profit of $1,000 on the trade in 100 microseconds. This is supposed to fuel righteous anger over the little guy who can’t trade 1 million shares of anything in microseconds.
But why should this be an injustice? It’s not like the HFT is magically pulling $1,000 out of thin air like, say, the Federal Reserve when it inflates the money supply to buy $1,000 in U.S. Treasurys. (Or $85 billion a month as the case may be.) The HFT is actually performing a service, albeit one that is not often noticed. What service is that? It is increasing liquidity in the market. Liquidity, defined as how much of a security one can buy or sell without affecting the price of the security, is a good thing. Companies want liquidity in their stocks. HFTs increase liquidity by shrinking the increments by which stock prices move up or down in any given transaction.
Keep in mind that HFT can only make money, just like anyone else, if it can buy low and sell high. In the case of the HFT the low and high are extremely close together, but the principle is the same. Even an algorithm cannot buy if there are no willing sellers at a particular price, nor can it sell if there are no buyers at a particular price. An algorithm can only match buy and sell orders, not create them lacking a willing partner.
Opponents of HFT would counter that HFT is used to front-run trades, and that is true. This means that a high frequency trade will usually take place only when a big guru-type investor, say Warren Buffett or Carl Icahn (Trades, Portfolio) or whoever, has placed a large order for stock A for, say, $10 per share. The HFT, usually a firm that executes the trade itself, will then take a large position for itself before executing the trade order for its whale client. Since the HFT knows about the order before anyone else, it will quickly look to buy the same amount of shares for a tiny fraction less than $10, and then sell to its client for $10 and pocket the difference.
Had it not done so, that tiny fraction of a difference would have been pocketed by the initial sellers rather than the HFT. This does sound unfair, being that one party has knowledge that others don’t. But how is front-running in this way any different from a broker charging a commission? It’s just a different kind of commission, a capital gains commission rather than a fee commission. Practically, it is the same thing.
An artificial delay is meant to make front-running trades like this more difficult to do. Let’s assume that it makes it impossible. What happens then, when an HFT can no longer clip pennies like this? What ends up happening is that the price of fee-based commissions rises instead to meet the market price of executing trades, so buyers and sellers end up losing the same amount of money they gained by cutting out the HFT middleman. They just lost it in a different way through fees rather than through the actual prices of securities.
So while a 350-microsecond delay will have critics of HFT cheering, it won’t change much in terms of trading costs, practically speaking.
Disclosure: No positions.
Start a free seven-day trial of Premium Membership to GuruFocus.
