By Ian Fraser
Date: 16 April 2012
High-speed traders use powerful computers to identify orders as they emerge and instantly trade ahead of them, hoping to earn a small crust on each trade. They use complex algorithms to churn out thousands of trades, and to lesser extent orders, on multiple markets in fractions of a second. And if they can do this in huge numbers, the rewards can be immense. It’s a ‘zero-sum’ game that is being won by those trading outfits with the best computers, the smartest ‘quants’ and access to the best algos, but lost by the rest of us.
Not only is HFT legalized front-running. It is also a socially worthless activity that amplifies market movements, increases market fragility, inflates asset price bubbles, and naturally worsens market crashes. And as we saw with the ‘Flash Crash’ of May 2010, it can also fuel market mayhem.
One of the high speed traders’ most outspoken critics is Charlie Munger, vice-chairman of the US insurance group Berkshire Hathaway. In an interview with CNN in May 2011, Munger said: “I don’t think the rest of us have anything to gain in having massive trading between computers which try to outwit one another with their algorithms. To the extent that one succeeds, the rest of us are all paying …”
And in a speech at the Institute for New Economic Thinking’s Berlin conference on April 14, the Bank of England’s Andy Haldane said that high speed trading can create a “mirage of liquidity“:
Accompanying this shift in speed has been a dramatic change in the composition of trading and market-making. During this decade, High Frequency Trading (HFT) has come to dominate. It now accounts for between a half and three-quarters of trading volume on the world’s major equity markets… In some markets, HFT firms have become the de-facto liquidity providers or market-makers…. This emerging topology of trading has had some clear benefits. This is manifest in narrower bid-ask spreads for trading. For equity markets, these have fallen by an order of magnitude since 2004. This is typically taken as evidence of improved market liquidity and price efficiency.
But this dominant role for HFT, in their race to zero, has also had some potentially less benign side-effects. Take order cancellations. All market participants cancel some of the orders they place before they are executed. HFT firms have taken this to new levels, submitting many more order messages than they are willing to execute. A decade ago, ten orders might be cancelled for every one executed. Today, that order cancellation ratio can be closer to 60. There are some good reasons for order cancellation, including the arrival of price news after an order has been placed. But there are, too, some potentially jarring side-effects. With HFT firms serving as de-facto market-makers, but with the vast majority of their orders cancelled, many on-screen quotes may not actually be executable. This creates a potentially misleading picture of market resilience, a mirage of liquidity.
Second, because bandwidth is finite, submitting multiple quotes may slow down activity by competitor traders. This practice of clogging competitors’ screens is known, accurately if inelegantly, as “quote stuffing”. It is a classic congestion externality or commons tragedy.
These externalities are not hypothetical. They have already shown their face in the extra-ordinary dynamics in market prices for an hour on 6 May 2010, the “Flash Crash”. Market dynamics during that period are yet to be given a fully-convincing explanation. But some aspects are documented. One is the disappearance of liquidity and market-making by HFT firms for a period during the Flash Crash. The mirage of liquidity proved just that. That evaporation appears to have played a key role in propagating stress during the Flash Crash.
So too did message traffic congestion. One side-effect was to slow-down price discovery across exchanges. Many traders firms found themselves observing stale prices. As a result, identical stocks traded at different prices across exchanges. In principle, this represented an arbitrage opportunity. In practice, arbitrage relies on costless trading and, at that time, trades could not have been executed at any price.
The Flash Crash was short-lived. Some have argued the lessons have been learned. Yet it appears to have been anything but a flash in the pan… The uncertainties and externalities associated with the race to zero are one candidate explanation. Liquidity mirages and message traffic congestion are nibbling away at the common good of market stability. The competitive quest for individual speed risks a fragile evolutionary equilibrium. If there were a benign, enlightened regulatory planner, able to co-ordinate the actions of traders on a lower-velocity equilibrium, the Flash Crash might have been forestalled. Unfortunately, there was not.
A joint report published by the Securities & Exchange Commission and Commodities Futures Trading Commission (CFTC) in October 2010 concluded that the May 6, 2010 ‘Flash Crash’ was fuelled when HFT algorithms “began to quickly buy and then resell contracts to each other — generating a ‘hot-potato’ volume effect as the same positions were passed rapidly back and forth.” In other words, the algorithms ran amok, and caused the market to tank.
Given this background, it’s perhaps unsurprising that legislators, central bankers and regulators have been looking at HFT more skeptically since the crisis. New regulations that would cramp speed traders’ style are already in the pipeline in Europe and the US. In March the Wall Street Journal reported that the SEC is investigating whether HFT firms are abusing their links with the operators of computerized stock exchanges they trade on such as BATS Global Markets, in order to manipulate markets and cheat ordinary investors.
The Journal’s Scott Patterson also recently lifted the lid on how high-frequency specialists Pipeline Trading Systems LLC and Millstream Strategy Group developed a “dark pool” licensed from Fidelity Investments, seemingly with the specific goal of fleecing the uninitiated.
Stock exchanges have much to lose from any regulatory crackdown and are desperately seeking to preempt one with various measures. The London Stock Exchange Group, Deutsche Boerse and Nasdaq OMX are taking various steps to slow trading down, in the hope of persuading regulators they deserve to be allowed to continue to self-regulate. Andrew Bowley, who leads the computerised trading unit at investment bank Nomura International, told Reuters: “The exchanges are looking to push self-regulation rather than have regulation imposed upon them.”
The LSE Group’s Borsa Italiana is introducing penalties for firms that exceed an order-to-trade ratio of 100:1, with a sliding scale of fines ranging from 0.01 to 0.025 euros per trade above that, depending on the severity of the breach. Under the Italian exchange’s new rules, firms sending 101 orders before producing a real trade each day would be punished, whereas those that have a ratio of 99:1 or less would avoid censure.
According to Investopedia, HFT gained traction in the US after exchanges like the NYSE started offering incentives and rebates to market participants such as Goldman Sachs who added liquidity to the market. The ‘supplemental liquidity providers’ rebate offered by the NYSE is in the region of $0.0015 per trade. That adds up to a stack of cash if you multiply it by millions of transactions per day. There are clearly a lot of vested interests at play here.
Given regulators’ apparent determination to root out some of speed trading’s sins, it will be interesting to see if the exchanges’ pre-emptive strikes actually work.
HFT does have some defenders, most of whom work in finance or are academics with close ties to the industry. One of their main contentions is that HFT narrows the bid/offer spread, which actually helps ordinary buy-and-hold investors. In a recent Bloomberg op-ed piece Prof Bernard S. Donefer, associate director of the Subotnick Financial Services Center at City University of New York (CUNY), stood up for the practise. He argued: “Automated market makers (or AMMs), a subset of HFTs, are liquidity providers. Their quote-and-cancel rates may be high, but unless they offer the best price in the market, they won’t get order flow… AMM systems automatically stop trading when market data appear out of normal bounds or when regulatory capital reaches prescribed limits. These are reasonable actions.”
Donefer conceded that certain HFT practices may require further scrutiny since they may lead to market manipulation. He said regulatory responses should focus on stamping out bad practices, not cracking down on HFT per se. Joe Saluzzi of Themis Trading published a powerful riposte to Donefer’s claims, highlighting omissions in his arguments.
This article was first published on QFINANCE on April 16th, 2012