HFTRant: Much Ado About The Wrong Thing…
Everybody is getting energized about ‘high frequency trading‘ in the US, in which (as I understand it) brokerage houses use ‘bots’ to test whether a bidder on a stock is prepared to up their bid, by placing an order just above them and quickly removing it.
So imagine a stock is trading at $21.50 and you’re on the bid at $21.45; the last trade took place at the offer ($21.50). You want, say, 500 shares.
You’re watching market depth (so you’re not an ‘investor’ – you’re a trader) and you see a bid order for 100 shares enter the depth at $21.46.
Do you move your order?
If so, do you move it to $21.26 (i.e., into the queue behind the new order) or to $21.47 (to get in front of the new order)?
Let’s say you decide to jump above. (Why?).
You change your order to bid $21.47 and your order gets taken (doesn’t matter by whom). You now own 500 shares bought a little bit more expensive than you otherwise would have.
Then the 100-unit order disappears.
Bugger, you think. Coulda got ‘em for $21.45 like I planned.
Two things need to be said here.
First: this only matters in instruments where the bid-ask spread is large relatlve to the stock price. In the example above (which is based broadly on NAB – one of the largest and most liquid stocks in the Strayan market), the spread is 5c on a stock price of $21.45 – in other words the spread is about 0.2%. In jumping your order, you paid less than 0.1% more than you otherwise would have paid. Taken over a whole portfolio, that’s not going to change your investment outcomes by more than a poofteenth.
Second: it only matters to traders (and dumbass traders, at that… no sensible trader should be falling for anything so stupid). If you’re an ‘investor’ you should not be watching market depth. If you’re a trader, then in jumping your order you’ve signalled you really wanted that order filled. Why were you desperate to get that order filled? Did you think the market was going to run away from you? Then you’re a breakout-buyer… the nuffnuff who gets torched 80% of the time.
This ‘flash order’ malarkey happens all the time in the ES futures market during the overnight lull… you will see the bid or offer ‘blink’ during that period, with a single contract (or two, or three) flashing in and out of depth.
It’s not very interesting to watch.
I’ve actually done it myself from time to time when bored – I had my trade platform set up so that
- F5 was "Cancel all and Bid 2 at a tick above the current bid",
- F8 was ‘Cancel all and offer 2 at a tick below the current offer"; and
- F12 was "Cancel".
When the ES spreads got wide (they can get out to a full point in the lull), I would amuse myself by hitting F8-F8-F5…F12.
F5 was ‘dot’ and F8 was ‘dash’, and I would see if I could get my name spelt out in Morse Code before the order got filled.
So really? HFT is a non-issue (and it will absolutely not change the long run equilibrium price of a stock or index, as I made clear in a comment on ZeroHedge).
Those of youse who were reading my rantings back in the InvestorWeb days (and before – a letter to the Financial Review in 1998) will recall that I became highly annoyed with the issue of ‘U’ orders in the Australian market.
These were orders with an ‘Undisclosed Quantity’, which were put into market depth in order to try and get bidders (or offerors) of stock to jump the spread, by making it seem as if there was a big buyer (or seller) sitting just below (above) them. (A ‘U’order was always a big order, because to be ‘U’ it had to be at least $100k worth).
The ‘U’ was always likely to be a manipulative order – trying to get the market to move toward a counter-order on the other side of the market: if you want to buy $5k worth ofstock, you OFFER a ‘U’ above the market, and watch the nuffies on the offer run the price down to your $5k bid
Related to this was the refusal of the exchange to disclose broker details to the broader public – in other wirds, if there was a ‘U’ order, it would have been good to know whether it was from a large broker (who would be likely to have a client who actually wanted to buy/sell $100k of stock) or a small broker (where the ‘U’ might well be a manipulative order).
Daytraders were manic ‘U’ spotters. There were times when an absolute flood of traders would jump the spread in a stock when they saw a ‘U’ order, even if there was already tens of millons of dollars worth of stock on both sides of the depth.
Fortunately, both the broker depth issue and the ‘U’ order issue have gone away nowadays.
HFT is symptomatic of the corruption in the US system – it is unambiguously designed to ‘nickel and dime’ the market, in order to enrich a few favoured-access firms and their clientele. Since it’s corrupt it ought to be stamped out – otherwise people will simply vote with their feet (and exchange volume stats show that people are already doing so).
But what HFT is not, is a pox on a sensibly-constructed investment strategy (and it’s not even harmful to a sensibly-contrarian trading strategy).
In sum, there are far bigger issues when it comes to the rampant corruption in US and global equities markets. This is not so much a storm in a teacup, as it is an attempt to diligently remove a stray tea-leaf from a cup of tea, when the teapot has a gigantic turd in it.
It’s well motivated – and has been brilliant in bringing the issue to the fore – so it ought to be used as a launchpad for further work: in and of itself it’s neither important nor damaging to market functions.
This website uses IntenseDebate comments, but they are not currently loaded because either your browser doesn't support JavaScript, or they didn't load fast enough.