This morning’s NYT tells of prosecutors finally catching up with the alleged perp behind the “flash crash” in 2010, when the Dow fell almost 600 points in a few minutes. It bounced back soon after, but not until everyone’s nerves were seriously jangled and confidence in the soundness of the equity markets took a big hit.
US prosecutors allege that the crash was the work of a guy in a London row house a few miles from Heathrow. That’s right, some schlub in his bathrobe supposedly tanked the markets by “spoofing:” algorithmic trading that executes tens of thousands of buy and sell orders only to cancel them milliseconds later. Once the market reacts to the spoofer’s large bets for or against the stock, she moves in and cashes in on the price change. Basically, a high-frequency variation of “pump-and-dump.”
As the Times notes:
The case also played into worries that have swirled around the increasingly automated and complex financial markets, where regulators have struggled to keep up with nimble new participants like high-frequency trading firms that use sophisticated networks to make money in milliseconds using rapid-fire trades.
Regulators have various ideas of how to regulate against spoofing, front-running (where flash traders get information on trades milliseconds before the public), and other such high-frequency fun and games; I’ve written about them before. They generally work by creating speed bumps in the trading process, say by moving from continuous trading to “batch trading,” thereby taking away the millisecond advantages of the flashers.
That might work, but it might not. You ask me, an arms race against quants who live to write regulation-beating algorithms is a recipe for more of the same. For example, suppose batch trades across different exchanges are not perfectly synchronized. That’s an opportunity for high-frequency arbitrage.
A better, simpler way—and one with numerous positive externalities—is a financial transaction tax, a small excise tax on the security trades, typically a few basis points (hundredths of a percent) on the value of the trade. A three basis points FTT is scored as raising over $300 billion over 10 years, a score that includes its dampening impact on trades.
Of course, that last bit is a feature, not a bug. We’d have to try it to find out, but it is widely believed that an FTT, even one of the tiny magnitude just noted, would wipe out most high-frequency trading. Though the flash boys can generate huge payouts, the volume of trades they must execute to do so quickly becomes too costly once they’re taxed.
In that regard, the FTT is a Pigouvian tax: a tax that offsets the significant, external costs imposed on the larger society by activities like smoking or polluting. And it does so while generating much needed revenue.
There are, of course, arguments against the FTT—by reducing trading, it dampens liquidity; it pushes traders to other exchanges to escape the tax. I deal with some of these concerns here, as does Dean Baker here. I take these concerns seriously, but my strongly held belief is that the likely benefits outweigh potential costs.
Perhaps the most valid concern in the Pigouvian context is that an FTT would fall on everyone, not just the flash mob. However, that’s not unique to the FTT. It takes me about two years to go through a bottle of whiskey, but I still have to pay the “sin” tax. It’s the heavy drinkers who bear the brunt of the tax, as would the high-speed traders in the case of the FTT.
So instead of a technical workaround that the quant jocks could probably beat before the regulators had their shoes on, the smart move here is to introduce a small FTT. I’d love to see this idea surface in the forthcoming campaign…any takers???