The stock market isn't rigged, but it is taxed.
It always has been. As Justin Fox points out, for as long as people have been trading stocks, there have been middlemen taking a cut of the action. Now, that cut has gotten smaller as markets have gotten bigger and more technologically-advanced, but it's still there. It's the implicit fee that intermediaries charge for making sure there's a buyer for every seller, and a seller for every buyer—for "making markets."
But there's a new kind of middleman today. They don't work at stock exchanges or banks. They work at hedge funds, and trade at whiz-bang speeds. These "high-frequency traders" (HFT) use computer algorithms—a.k.a., algobots—to arbitrage away the most infinitesimal price discrepancies that only exist over the most infinitesimal time horizons. You can see just how small and how fast we're talking about in the chart below from a new paper by Eric Budish and John Shim of the University of Chicago and Peter Cramton of the University of Maryland. It uses 2011 data to show the price difference between futures (blue) and exchange-traded funds (green) that both track the S&P 500. These should be perfectly correlated, and they are—at minute intervals. But this correlation disappears at 250 millisecond intervals, a little more than half the time it takes to blink your eyes. This is the "inefficiency" that HFT makes less so.
This rise of the robots certainly seems to have helped ordinary investors. Bid-ask spreads—the difference between what buyers want to pay and sellers want to be paid—have fallen dramatically the past 20 years. Part of this is because, since 2001, stock prices have gone from trading in fractions to pennies—which has allowed them to be increasingly precise. Another part is that electronic trading, though not super-fast, has made markets more liquid. And the last part is that HFT has added even more liquidity, eliminating bid-ask spreads that would have been too small to do so before. Indeed, researchers found that Canadian bid-ask spreads increased by 9 percent in 2012 after the government introduced fees that effectively limited HFT.
That doesn't mean, though, that HFT is unambiguously good. It's not. In fact, it might not even be ambiguously good. As Noah Smith points out, we just don't know enough to do any kind of cost-benefit analysis. Now, we do know that smaller bid-ask spreads, which cut the cost of trading, are one benefit. But how much of one is it? Bid-ask spreads are down to around 3 basis points today—from 90 basis points 20 years ago—so even if curbing HFT increases them, say, 9 percent like it did in Canada, we're not talking about a big effect. There might be diminishing returns to liquidity that we've already hit, and then some.
Then there are the costs. Michael Lewis' new book, Flash Boys, describes some of them. In it, there's Lewis' requisite group of plucky outsiders—is there another kind?—taking on a rotten status quo. Except this time, they're not really outsiders; they're big bank traders. And they've figured out that the market doesn't work like it should for big investors, like pension and mutual funds, because of the algobots. But it's a little bit more complicated than that. Here are the three biggest, though hard to quantify, costs of HFT.
1. Market-taking, not market-making. Lewis' protagonist, a trader named Brad Katsuyama, had a problem. Every time he tried to buy stock for a client, he could only get a little bit of what was supposed to be there at the price he saw. Now, oddly enough, he could get all the stock he saw at one particular exchange, but he had to pay more at all the others. What was going on?
Well, he was being front-run. HFT firms pay public and private exchanges to see their incoming orders. That's why Katsuyama was getting all of his order filled at the exchange closest to him—that is, as the fiber optic cable lies—but nowhere else. The HFTers were seeing his order at the first exchange and then racing to buy all the rest of the stock he wanted everywhere else, so they could sell it to him for more. This happens all the time: Nicholas Hirschey of the London Business School found that HFT funds only tend to buy aggressively right before everybody else does.
It's not too different from what HFTers do when they buy early access to public data. Again, they're paying for a trading advantage that isn't really adding liquidity. It's what Barnard professor Rajiv Sethi calls "superfluous financial intermediation." HFT firms aren't connecting buyers and sellers who might not find each other. They're jumping in between buyers and sellers who would have found each other anyways in a few milliseconds. It's not making markets more efficient. It's cheating.