Flash Boys aren't rigging the market

Michael Lewis is a great storyteller. This is probably why he's the on economics journalist to have had three of his books made into Oscar nominated movies. He has a real knack for turning wonky topics like Credit Default Swaps and the Eurozone debt crisis into simple, compelling stories. But as Tyler Cowen warns us, we should be suspicious of simple stories.

That's certainly the case with his recent(ish) best seller Flash Boys: A Wall Street Revolt, it looks at the relatively new phenomenon of algorithmic high-frequency trading. Lewis takes a dim view of the practice, seeing it as little more than an exercise in rent-seeking. His big concern was that High Frequency Traders were providing very little in the way of societal benefit, but were competing in a zero-sum game for rents. According to Lewis, this competition for rents had become an incredibly costly arms race. He points to firms spending hundreds of millions to build underground cable networks between stock exchanges with the not-so-lofty aim of shaving milliseconds off data connections between New York and Chicago.

What are the rents on offer? Lewis was worried about HFT firms being able to buy faster direct access to exchange quotation data. HFT firms could then use this data to predict changes in the data that most normal trading firms typically use to price trades. Whenever a discrepancy arises between the two, HFT firms can make risk-free trades - easy money. If that's the case, then there's a risk that firms will engage in socially costly arms races, buying faster connections at great prices, while providing little in the way of societal benefit. This has prompted many to call for increased regulation and even financial transaction taxes to deter any arms race. A sort of Nuclear Non-Proliferation treaty for finance.

Whether we need new taxes or regulations is ultimately an empirical question. A new paper from the University of California, Berkeley suggests that the answer is no. Robert Bartlett and Justin McCrary looked at a massive sample of time-stamped trades representing around $4tn in value over a whole month. Analysing that data they found that of the $4tn traded liquidity providers would have only saved around $11m had they switched from the standard data everyone uses to the faster direct access. In other words, it's just not worth it to engage in these expensive arms races that Lewis worries about.  It seems that the controversial rent seeking strategies described in Flash Boys are a thing of the past. 

As my colleague Ben Southwood points out. Once you look past Michael Lewis' unfounded fears about arms races, HFT firms are providing a social good. Attempts at regulating them ended up hurting retail investors and research from the ECB suggest they facilitate price discovery especially at times when markets are especially volatile.