It’s been quite a week in the HFT world with Navinder Sarao escaping jail time for his spoofing activities that the US authorities ridiculously claimed led to the May 2010 flash crash, and the UK’s FCA releasing a report stating that latency arbitrage in UK equity markets adds 17% to the cost of trading or $5 billion per year if extrapolated globally.
Sarao’s story will, I suspect, make for a very watchable, perhaps poignant, movie one day given how he clearly didn’t do it for the money and when he did make some cash he was apparently – according to the defence and prosecution – easily relieved of it by scammers and dodgy business decisions. Whether that will be a positive story remains to be seen, but clearly the authors of this week’s research paper released by the UK regulator have a negative story.
The paper itself states that it uses UK stock exchange message data to “quantify the negative aspect of high-frequency trading”, which we all know as latency arbitrage. The use of message data is crucial to the research as it contains failed attempts to trade or cancel a trade, instead of the more popular use of limit order book data, thereby highlighting not only the winners of the trade, but also the losers.
The research finds that latency-arbitrage races are “very frequent” (537 a day in FTSE 100 stocks – that’s one per minute, per stock), “extremely fast” (the modal race lasts 5-10 millionths of a second and the average race time is 81 microseconds), and account for about 22% of overall trading volume. Interestingly, the research finds that three unnamed firms account for over half of all race wins and losses – what’s the betting they are racing each other most of the time?
So, some thoughts. Well first up, I suspect the $5 billion number may underestimate matters because the US equity market is a lot more fragmented than the UK’s and that would lend itself to more latency arbitrage opportunities. Extrapolation is a very difficult process, but $5 billion is a good number on which hang one’s hat, however. It could also legitimately be asked, given the size of the global market, is this actually a significant number bearing in mind that the research finds the average win is worth GBP 2 (and no, there should not be a 0 there because yes, that is just two pounds). Of course, as the paper notes, add this up and the numbers become meaningful, but that also begs another question – who are they picking off and why are they trading?
It’s a question I have put to people for many years when they complain (often) about being picked off on an ECN for example; if you don’t want to buy at x then why did you bid there on an exchange/ECN? These platforms are “expression of interest” venues, there is no need to stream a two-way price, so why bid or offer if you don’t think you want to buy or sell there? More pertinently, if you do stream a bid/offer, don’t complain when the market goes away from you when you are hit!
I understand that in equities markets many exchanges operate a market maker function under which designated firms have to stream prices, but they get paid handsomely in the form of rebates and discounts to do that – if their pricing algos aren’t good enough, whose fault is that? What they are effectively saying is, if I don’t make money on every price I make, then the market is unfair – which is complete and utter…nonsense!
We have, therefore, to ask questions of the firms that are being picked off, but a second suggestion from the paper is that perhaps we don’t, because it finds that just six firms win 82% of races and lose 87% – in other words, it is probably the HFTs themselves that are getting picked off. Not only does this support one of my long term assertions that HFTs do indeed eat each other if left in the same liquidity pool, but it also throws some light on why exchanges in particular are anti-speed bumps in their equities markets. After all, would you threaten over 20% of your business?
I also think it worth noting that in a moving market there is always going to be a race – this is nothing new. Back in the day (yes, here we go again…) one of my voice brokers – he is still active in FX circles – was so used to being hit by more than one bank, his first email address was doublehit@… There have always been races to hit the best price, previously it was about speed of thought, then it was speed of hand, then it was about speed of technology, now it is about microwave towers, next…who knows? Either way, we can reasonably ask what the difference is between an HFT and an asset manager hitting a stale quote? Both are taking advantage of an information advantage in some shape or form, although I would always lean toward the player trading with a view longer than a few milliseconds when deciding who I want to hit that price. On which note, I think I have previously made my thoughts clear – I believe the concept of speed bumps, in whatever form they come, help make a market “fairer”, but that does not mean that every time the market moves after a trade we should be complaining.
To me, this is a market structure issue in equities, where matters have got out of control and speed is, actually, everything. If you take that advantage of extreme speed away, then inevitably you must end up with fairer markets, which, importantly, means that the “losers” in a trade can stop boring us with their complaints!
The prudent use of speed thresholds, along with minimum quote lifespans, should mean that the winners in markets are those who actually know where the asset price is heading. I understand this is a quaint, somewhat archaic, principle, but even if the market participant concerned is trading for a few seconds, they still have to get the direction right – which takes skill, be it human or technology-originated.
Foreign exchange markets seem to have had the debate and decided that latency arbitrageurs are not welcome in its midst, the same cannot be said for equities markets. This makes the fate of fixed income markets very interesting, for that asset class is on the brink of greater electronification. For what it’s worth I am not too hopeful for fixed income, mainly because the market structure is dictated by regulators – and they are the very people that created the conditions for the race to zero in equities. I could be wrong, however, it has been known, but either way, one final and valid question to ask is: when will some markets get away from their obsession with speed? It doesn’t seem to be adding anything to market quality.