Solving the HFT problem: Abolish continuous trading

By Felix Salmon
July 31, 2009

Michael Wellman has an intriguing idea for solving all the issues surrounding high-frequency trading at a stroke: switch from a market with continuous clearing to a market which clears once per second.

Orders accumulate over the interval, with no information about the order book available to any trading party. At the end of the period, the market clears at a uniform price, traders are notified, and the clearing price becomes public knowledge. Unmatched orders may expire or be retained at the discretion of the submitting traders.

Even with a period as short as one second, the call market totally eliminates any advantage of HFT systems. It does not eliminate the opportunities for algorithmic trading in general–just those that come from sub-second response time. No party has privileged information about order flow, and no party benefits by getting a shorter wire to the “trading floor”.

According to Wellman, there would be other advantages to discrete-time trading too, including lower volatility.

Would this plan essentially give everybody in the market the advantages of being in a dark pool which only exists for one second? On its face, I think it’s a good idea. What would the downside be?

More From Felix Salmon
Post Felix
The Piketty pessimist
The most expensive lottery ticket in the world
The problems of HFT, Joe Stiglitz edition
Private equity math, Nuveen edition
Five explanations for Greece’s bond yield
Comments
16 comments so far

“Would this plan essentially give everybody in the market the advantages of being in a dark pool which only exists for one second?”
No, because a dark pool is a matching engine that depends on an outside reference price such as the midpoint of the best bid/offer spread of some data feed.

Excellent idea! Maybe even half a second.

In someways, HFT violates the free market principles that everyone gets the same information at the same time. I wonder how that got over looked.

JGT

Posted by John Thompson | Report as abusive

I think a fixing system might increase rather than decrease manipulation.

If I see a buy order in the market and I hit the price, then it’s my trade.

In a fixing system I will have to enter my order and for a split second – the time between my entering the order and the closing of the fixing – everyone will be able to see it – not enough for humans to get in the way perhaps, but it will allow every HFT system to decide whether or not to come between me and the buy order.

Posted by a | Report as abusive

John Thompson: Free market principles don’t seem to last long when they get in the way of making money.

Posted by Ken | Report as abusive

A commenter over on Cassandra Does Tokyo authoritatively reports that the Tokyo exchange matches trades at multisecond intervals due to technology issues. Are spreads and liquidity better there as a result?

JGT: I asked this question in comments on a previous post. Apparently the fact that access to advance information is “sold” is enough to satisfy many. Just because most traders can’t afford it doesn’t mean it’s not “public.” Apparently time really is money in this game. “Free” trade information on the web is 20 minutes delayed. Real time information is available at a cost, and faster-than-real time information is more expensive still. Seems a clear continuum to me!
(BTW, heavy sarcasm intended here!)

Posted by Rockfish | Report as abusive

Throw some sand in the gears. Impose a 0.1% transaction fee on all financial trades (stocks, bonds, derivatives) payable by the seller. Call it “Meltdown Insurance” or “Moral Hazard Insurance”.

Actually, I’d prefer 0.5% for positions held for less than 24 hours, 0.1% for up to 1 year, free after that.

Tax what you want less of. In this case frequent trading.

Posted by Tentakles | Report as abusive

“Orders accumulate over the interval, with no information about the order book available to any trading party.”

Sorry, I just noticed that the suggestion says there is no info about the order book. I’m finding this difficult to visualize; every second, I see the last price traded and the prices in the old order book? I’d have to think about this, but I think the disadvantage here would be greater volatility.

“A commenter over on Cassandra Does Tokyo authoritatively reports that the Tokyo exchange matches trades at multisecond intervals due to technology issues.”

Not an expert on the Tokyo system, but I believe that’s different from the present idea. The present idea is suggesting a fixing, where the market clears at a uniform price every second. I presume also the suggested system is one where orders do not have priority based on time (i.e. two orders at the same price will have the same priority even though one will have been sent before the other). The Tokyo system matches trades every few seconds, but (1) there may not be a uniform price at a single matching and (2) there is time priority. (To be honest I haven’t checked (2), but I’d be surprised if this is not the case.) Once you have time priority HFT works, because the point is getting the order in quicker than the others; when the matching actually occurs, instaneously or every few seconds, is not particularly important.

Posted by a | Report as abusive

As I remember it, Optimark used to execute every minute or two. I’m not sure that was a terrible idea, though it didn’t really work out for them, for one reason or another.

Lots of fairly difficult issues: How do we keep the matching in sync across tons of different endpoints? Are we banning dark pools or imposing the same regulations? Is the proposal just for US equities? Then what about futures and options markets? What happens to foreign markets? What exactly is the proposed matching algorithm, and how are we going to keep it from being manipulated? Many, many more.

Professional traders are simply going to have better information, whether it’s from their ability to respond quickly to data or through access to futures and options markets or a possible overseas single-stock futures market if the US imposed these regulations.

Actually, I honestly don’t understand what Felix terms the “HFT problem” actually is.

I do understand there’s a claimed threat that orders briefly displayed to dark pools or sets of professional traders can be front-run. But these order types must be actively chosen by the party submitting the order, so I doubt anyone would flash a market-moving order. I also don’t think big high-frequency liquidity providers would mind if these were banned, and in fact one (Getco) has been pressing for their abolition. It’s really not very exciting stuff or the source of big profits.

(Note that “flash orders” can also mean something different, the ability, recently introduced on several exchanges, to lock the NBBO for less than half a second–which benefits customers taking liquidity, who can either buy or sell at that price. I haven’t seen any complaints specifically about this practice, but a lot of confusion.)

Is the issue that 100 microseconds of speed responding to information shouldn’t affect your profits? Well, I guess, but what should affect your profits? Your status as a specialist? Your captive order flow? Your electronic access to other markets?

Posted by keikobad | Report as abusive

IF you want less high speed trading, tax it. The profit per trade in this game is trivial in absolute terms – 5c to 40c per 100 shares traded probably covers most high-volume players, and a tiny tax would destroy the entire economics.

Yet how much disincentive would a (say) 50c/lot tax be for anyone – institutional or individual – for someone with a year long horizon? A quarter? A month? A week? I submit: nearly 0 at any horizon longer than a day. (IMHO there is a lot of evidence that speaks to the elasticity of trading behavior vs costs and provides solid support for this claim. But a comment to a blog isn’t the place to dissect this.)

So: a somewhat valuable tax, from a politically vulnerable constituency (Wall St), that is tightly targeted at discouraging something we don’t want (?!?), having only tiny effects at the margin of regular investment behavior. If these are the ends we want, here are the means. Abandoning continuous trading work too, I suppose (I’d go for daily or weekly rather than /second
but to each his own) but the tax solution brings in revenue and is more easily implemented in a market of multiple trading venues.

There is just one more TINY issue … do we really want
to stop high speed trading? Maybe it isn’t so harmful?
Maybe it’s even helpful? Public discourse on this, including this blog, seems amazingly shallow and uninformed. The benefits and costs of this sub-industry are complex, but people seem to jump in with the most superficial knowledge – perhaps because the “villain” is Wall St that we don’t have to bother with research or understanding???

To dwell on one example… you may have heard about flash orders and how evil they are. And yes there’s something a bit dubious about them, at least on the surface. But before you weigh in, IMHO you should probably be able to answer yes to the following:

– do you know about the potential significant benefits to someone whose order is flashed? If not, you REALLY don’t understand enough to comment!!!

– do you know that no-one is involuntarily flashed? (well I mean: of course, you as an individual might be, but this is your broker or mutual fund or other agent making the trading decisions who lets their orders be flashed). I’ll repeat myself: if you or the person who submits your order doesn’t want your order to be flashed, it won’t be. Only about 10% of blogs/articles I have seen on this issue acknowledge this rather relevant point. Although (see previous point) you probably DO want to be flashed if you understand what is at stake.

– do you understand the economics of the industry to see why [insert certain major exchange names here] benefit hugely from a blanket ban on flash, even if they themselves have been forced to offer it for competitive reasons?

– do you understand the extent (not 100%, not even close, but > 0) of which the current furore is probably being orchestrated by certain major industry players for their own competitive benefit? (I mean, you might still believe they have a point, but I think you should understand enough to see what the world would look like after a flash ban and which of the current players are praying for such a day and why. You should know enough to see who might be speaking objectively and what is instead obvious propaganda).

Posted by byh | Report as abusive

Show me a world where everyone has access to the same information at the same time, and I’ll show you a world which is totally devoid of commerce. To argue “free market principles” in this case is incredibly myopic. Algorithms are nothing more than tools, and there are no restrictions preventing any interested party from using them. To lump all HFT into the same category is ludicrous, but in this age of uninformed political decisions seems to be par for the course.

If there is a particular problem, such as flash orders, and the preponderance of thoughtful and informed people feel the need to address it, then we should propose solutions for that particular issue. The financial markets are evolving, as they have always done, and computers and algorithmic trading now play an important role in them. I grow weary of hysterical finger pointing by uninformed people.

Posted by Thoughtful | Report as abusive

Wellman’s recommendation, coming from an academic, is surprisingly shallow in the motivation offered: how can we take such a recommendation seriously if there is no discussion of (a) what happens when there are sharp informational shocks in illiquid names or marketwide (b) what evidence there is that this won’t worsen spreads (as if I have not had the benefit of public disclosures of private value over the last second, I have more adverse selection to fear, so I’m going to build that margin into my limit, causing me and the contra party to become more conservative, thus widening spreads (c) there are 15 different dark pools in US equity markets, most of which are continuous time, and some of which are point in time. Any recommendation that doesn’t mention these and explain why none of these have tried to do a 1-second-clearing call mkt is weak.

Posted by ald | Report as abusive

Felix, Thanks for helping to air this idea. Yes, its conceptually like having everybody operate in short-lived dark pools. The comments in this thread raise some legitimate questions, but also many red herrings.

Commenter “ald” takes me to task for not thoroughly discussing a particular set of issues and comparisons s/he cites. These are relevant, and will be addressed in due course as I flesh out the proposal in more scholarly forums. The point of my blog post was just to get the concept out there in the current HFT interest thread. I have no illusions that anybody is going to adopt this tomorrow. In brief, though: (a) not much substantive usually happens in a second, except for order flow which is meant to be hidden from everyone; (b) you can withdraw or revise your order within the second if you do get information, so it’s strictly less risky than hanging a limit order in a continuous system; (c) existing dark pools run parallel to continuous trading, not as a replacement for it–but still it is a good question why these don’t adopt more granular clear times.

The most challenging issue is how to get there from here. The trading world inherently lives in a continuous-time mindset, and this is not easily adopted by lone actors. Those who would gain most are retail investors, who currently lack access to dark pools and HFT infrastructure. Those in the best position to implement a change are those with an advantage in the current system.

The downside to a call market would be you would pay higher spreads. Instead of the spread going to competing HFT dudes who drive prices down, it’s going to go to someone else. If your goal here is to strangle companies like GETCO and Goldman, you will fail: they’ll probably make more money, as they’re best positioned to do so. Dark crossing ain’t all that dark if you know what you’re doing.

I don’t see what all this emphasis on “fair” is. It’s not fair the consumer has to pay more for market access than a dealer. It’s not fair the consumer doesn’t have a team of 20 Ph.D.’s to make his trades for him, or transparent access to dark crossing networks, or even what the difference between a limit and fill or kill is. If you want to understand this “controversy” you need to understand who started it. The people who started it are the buggy whip manufacturers of the digital trading age. They want to go back to staring at technical patterns and yelling into a squack box like liquidity providers did in the 80s.

Post Your Comment

We welcome comments that advance the story through relevant opinion, anecdotes, links and data. If you see a comment that you believe is irrelevant or inappropriate, you can flag it to our editors by using the report abuse links. Views expressed in the comments do not represent those of Reuters. For more information on our comment policy, see http://blogs.reuters.com/fulldisclosure/2010/09/27/toward-a-more-thoughtful-conversation-on-stories/