Debate on High-Frequency Trading: A Formula for Liquidity or a Recipe for Meltdown?

Meeting Date: 24 May 2011

Report

High-Frequency Trading: A Formula for Liquidity or a Recipe for Meltdown?

Meeting Date: 24 May 2011

Con Keating, Head of Research, BrightonRock Insurance Group & Dave Cliff, Professor of Computer Science, University of Bristol

Written by Michael Mainelli

On behalf of the Real Time Club, Maury Shenk, Chairman, welcomed 100 low frequency (sic) members and guests to the National Liberal Club for an evening’s debate on the merits of trading near the speed of light. Both speakers gamefully defended their positions, though it was equally clear that they could see the other’s point of view. Dave Cliff kicked off with a number of telling points in defence of high frequency trading (HFT). Technology always speeds things up; authorities haven’t found a culprit for the Flash Crash, so innocent until proven guilty; machines really just do what people do, but faster. In short, Dave pointed out that the trend to high frequency trading is inevitable, another stage in the evolution of financial markets.

Con Keating followed with an argument that focused on the importance of liquidity and our need to understand liquidity before judging HFT. Con noted that, “if liquidity didn’t have a cost, every market would have it” and that “one big failing in finance is the assumption of liquidity”. Markets, on their own, don’t create or destroy liquidity. He pointed out that for investors volatility is bad, for speculators volatility is good, leading the audience to question whether highly volatile, high speed markets function in the interests of investors or speculators. Con also tossed out a challenge as to whether all this trading showed any evidence of new information being added and whether it actually indicated a degradation of market quality by adding unnecessary noise. Con would appear to agree with Huw Gronow of Principal Global Investors who has said, “A closer look shows that many HFT practices have, at best, a neutral effect on liquidity and, at worst, reduce liquidity.”

There were some fun spots too. Dave explained the four “F”s of humans, feeding, fighting, fleeing and reproduction. Con relating a tale of the FSA official, clearly uncomprehending market data, casually asking him to email 7 terabytes of data to explain something. Andy Low bursting into poetry (nuff said). After a short break, the ever-vigorous Real Time Club interchange began. Some of the highlights included:

• is HFT a sign of a bad economic system? Where does the money come from to provide the profits? Pensioners.
• if HFT cannot be banned, then so what? It doesn’t matter if it’s good or bad, just another part of the “death of distance” through increasingly global telecommunications. Further, markets need to explain away losses – we need explainable disasters – and HFT conveniently provides this. If HFT is slightly bad, perhaps it’s just slightly too much of a good thing?
• HFT induces more risk as everyone starts to use the same mathematical models. Diversity in approaches is necessary to avoid increasing instability.
• a later, contrary, point was that the limits of modelling actually provides that stability and that the human domain was to reset models from time to time – though those resets may become less frequent.
• the need to distinguish between informed and uninformed trading was enlivened with a flying analogy. On a commercial flight, the pilot is just the passenger with the best view. His or her presence simply reassures the passengers. Piloting is boredom spiced with 0.1% terror. So what financial markets need is reorganisation where the financiers are mixed with the investors.
• does HFT distort markets? Sure, but not as much as regulators.
• finance is still wild. Safety critical systems approaches should be applied to markets where we want stability.

In summary, a vibrant, fun, good-natured evening pulling out some good nuggets for discussion. Sitting in the corner taking notes, your correspondent mused on the 30 September 2010 SEC and US Commodity Futures Trading Commission (CFTC) report on the 6 May 2010 Flash Crash – “high trading volume is not necessarily a reliable indicator of market liquidity”. Money is society’s principal risk management tool and it’s important to get it right in order to make good decisions. It’s not clear whether HFT is actually helping us make better decisions. Perhaps we should restrict decision-making on decisions about value to humanity, to humans. But then we would need a Turing Test for Trading. How would you be able to divine whether you’re trading with a human or a cyborg?