Streetwise Professor

October 1, 2012

The Euros Want to Fix Computerized Trading In The Worst Way, and are Succeeding

Filed under: Economics,Exchanges,Financial Crisis II,Politics,Regulation — The Professor @ 11:20 am

HFT is under continued assault, both in the press and from regulators and legislators.  Europe is taking the lead on this, and Mifid II will impose some restrictions on HFT.  The Germans are particularly hot on restricting it, and are proposing other measures as well.

One goal of these regulations is to increase liquidity, and to improve the quality of liquidity.  Perversely, however, certain of the regulations intended to have these effects will do the exact opposite.

In particular, as I’ve argued since soon after the Flash Crash when many of these ideas were first proposed, things like minimum resting times for limit orders will reduce liquidity and make sudden reductions of liquidity more likely precisely when order flows become more toxic.  Limit orders are like options granted by the market maker, and extending the lifetime of the option increases its cost-particularly during times of high volatility and high order flow toxicity.  If you increase the costs that market makers, including those using HFT strategies to make markets,  incur to quote, they will quote wider, they will quote less, and their quotes will become particularly wide and particularly scarce during periods of elevated uncertainty.

In other words: there is no free lunch.  You raise the costs of market making, and liquidity costs go up.  Period.

Larry Tabb makes the point quite well:

So what will happen? If we mandate longer time in force periods, lower cancellation ratios, and higher market maker participation rates, liquidity providers will just widen their spreads to compensate them for the greater risk.

Won’t this force real investors to come in and bid? They may, however investors don’t quote on both sides of a market. They either buy or sell and not both. So many more quoting investors would be needed to make up for fewer market makers and HFT.

But, will HFTs actually leave? Market makers may leave but high-speed traders probably won’t; they will just change their stripes. Liquidity providers will flip to liquidity takers. Given a speed advantage, if it no longer serves a high-speed trader’s purpose to provide liquidity, they can just as easily take it. And since market makers and quoting investors are locked into providing liquidity for at least 500 milliseconds, HFTs will be the first to pick off every stale quote. And with a half second quoting mandate, there will be plenty of stale quotes to go around.

This last point is particularly powerful.  The regulations will penalize liquidity suppliers not just directly, but indirectly, by actively encouraging predatory/opportunistic algorithmic/HFT strategies that look for and pounce on quotes that are stale and not reflective of current information because of the restrictions on the time quotes must be enforced, and on cancellation and participation ratios.  (Note: these last restrictions are likely to create some rather complex time dependencies that can create some unanticipated sources of instability.  A trader’s current quoting behavior will depend on his past cancellation and participation rates, not just on current market conditions.  That adds a complex new feedback to an already complex process.)

The elimination of tiered pricing (e.g., differential pricing for quote makers and quote takers) will also tend to reduce liquidity.  Why don’t exchanges have the incentive and the information to get that right?  They are much better incentivized and informed than regulators.

So as Larry indicates, the regulations will reduce liquidity by raising the cost of quoting directly, and indirectly as well by encouraging predators to feast on those brave enough to quote:

So let’s combine points. Market makers will leave, spreads will widen, investors will post quotes, but HFTs instead of posting will take quotes. With investors’ quotes being frozen, HFTs will just pick off those investors. And once these investors learn this game they will stop posting, and then there will be little incentive for anyone to post quotes as the value of the limit order option grows just as the compensation to the option poster falls.

But won’t HFTs run afoul of HFT rules, you ask. Not really, since many of these rules focus on liquidity posting and not liquidity taking. If you take liquidity, there is no need to post or cancel, so cancellation rates decline to zero, and time-in force rules become immaterial.


Mifid II will also restrict those scary dark pools, in particular by banning crossing networks.  Which, as Tabb notes, will expose those seeking liquidity to predatory HFT, and other predatory strategies that raise execution costs.  For dark pools exist precisely to reduce the execution costs of those who are uninformed liquidity demanders. They swim in a (dark) pool because they don’t want to swim with the sharks in the open waters of the public markets.   Mifid will deprive them of that safe harbor, with the predictable result that Larry describes:

If all of this comes to pass, spreads will widen, depth will evaporate and there will be fewer places to hide as broker dark pools will face restriction. And the life of the buy-side trader, instead of becoming easier, will become much harder. And who suffers? The people who always suffer: the investors.

In other words, the Sorcerer’s Apprentices are at it again.  They claim to work magic that will make investors’ lives easier, but will in fact unleash forces that will wreak havoc.  And, no doubt, they will use the resulting havoc as the justification for yet more magical regulatory spells.

Not all HFT is good.  Not all algorithmic trading is efficiency enhancing.  It would be worthwhile to explore ways of curbing the inefficient forms of computerized trading without harming good forms.  There is justification for some regulation that can reduce the likelihood and impact of bad algos.   Mifid II is unlikely to have these effects.  Indeed, in my opinion-and in Larry Tabb’s too-it will likely have the exact opposite effect, and will disproportionately damage the beneficial forms of computerized trading and off-exchange trading.

Like the old joke says.  The Euros want to fix computerized trading in the worst way.  And they are succeeding.

Print Friendly, PDF & Email


  1. To be blunt, what we are seeing is a combination of two deadly diseases that most regulators are prone to: “Don’t just sit there, do something”, and “If I don’t do something NOW, people will realize I am useless, and will lose my job”. What is needed here is an outbreak of “Masterfull Inactivity” as Disraeli(?) put it, until some time has been given to trying to see the full effect of what they will propose.

    It is sad, because if they could just use the third vice the Eurotards suffer from especially – “A problem delayed is a problem solved” (see Debt Crisis), things would be ok, at least for a while.

    To request that the regulators differentiate between good and bad HFT assumes they really understand it, and have the judgement to implement their decisions in a sophisticated way that can be explained to their political masters. While this has happened before (see Alfred Kahn, and more suprisingly, the former New York State Insurace Department, which pulled off a number a really good moves in the past), it is unusual. Should something go wrong, it would be harder to explain a sophisticated policy than a stupid but simple one.

    Comment by Sotos — October 1, 2012 @ 3:16 pm

  2. It seems like the mental exercise people concerned about HFT want to go through is to create as set of rules that will replicate the cost structure of trading before technology allowed HFT. In effect, they want to “un-invent” HFT institutionally. Since there is no reasonable way to apply “arms control” at the technology level by banning fast algorithms or hardware, they are trying to slow down trades by rule. If I understand the post, the reason why the Mifid regulation won’t do this is because its behavioral sway is only partial–it stops liquidity providers from moving fast, but not liquidity consumers. It’s like gun control that differentially disarms the average citizen while leaving the criminal unencumbered.

    My guess is that the regulators’ response to this critique would be to try to increase the scope of their rules to cover the behaviors they’ve “left out” of the party. I’m not sure how they would do that in this case, but I wouldn’t be shocked by something even more intrusive and blunderbuss in the future.

    Comment by srp — October 1, 2012 @ 7:17 pm

  3. Professor, Larry’s and your analysis is compelling. I think it leads to the next question, though. If the primary victim of the set environment is becoming the investor, what is the investor going to do about it? Not invest or get armed?

    In my view, one of the solutions of the emerging situation is that the investors/asset managers should have the same tools and access to market at their disposal as the trading operations.

    Of course, it might make no sense for a small investor to do so, which leads me also to a conjecture that new types of investment pools or funds of funds might be merited.

    Comment by MJ — October 1, 2012 @ 10:16 pm

  4. Personally, I think the biggest problem in the US markets is the abuse of the liquidity rebates or more generally payments by the marketplace to people posting quotes that lead to execution. Specifically, the problem is that the executing brokers post too many quotes, pocket this revenue and don’t pass it to the customer, and as a result the customer gets inferior execution relative to just crossing the spread. Brokers have the incentive to direct the customer order flow to venues that pay the highest liquidity provision rebates and to rest orders there for the high-frequency traders to exploit. It’s yet another corrupt arrangement in trading.

    Comment by ptuomov — October 2, 2012 @ 6:45 am

  5. @MJ-there are substantial scale economies in HFT. Large scale asset managers do use algorithmic strategies, and avail themselves of dark pools, in order to economize on execution costs.

    @ptuomov-one issue is that the size of the spread depends on the maker rebates, so it’s not clear that your claim that investors get inferior execution vs. crossing the spread would remain true once that effect is taken into account. I’d also push back on why competition doesn’t lead to brokers shading commissions or other fees if they pocket the revenue from the maker fees. That is, it’s not trivial to determine the ultimate effect of these fees on investors. Finally, exchanges set the rebate. They have no incentive to enrich brokers. If, as you suggest, brokers just pocket the rebate, then trading volumes would not increase if exchanges provide rebates-so what’s in it for them?

    Not saying that you’re wrong, just that (a) economics suggest that you might be missing something, (b) it is an empirical issue as to whether you are in fact missing something, and (c) it’s not trivial to do the relevant empirical testing.

    Your argument suggests that exchanges are giving money away. Maybe so-but I’m skeptical.

    The ProfessorComment by The Professor — October 2, 2012 @ 12:32 pm

  6. SWP —

    Competition only works well under certain conditions. One of them is that the customer understands the true price. Currently, the customer does most assuredly not understand the true price. Instead, the customer is fed various kinds of drivel about best execution and not crossing the spread, when the motivation for posting quotes is the rebate revenue captured by the broker that 95% of institutional and 99.99% of retail customers don’t know about. This leads to a situation in which the brokers do compete, but they compete on who will most blatantly abuse the customer by posting quotes while ostensibly quoting low commission rates.

    The problem is not that the exchange pays for posting quotes that lead to trades. That’s fine, and there’s enough automated trading systems for the competition to work ok there. The problem is that the middleman, the broker, steals these rebate payments and doesn’t pass them on to the customer. This creates wrong incentives for the competition among brokers.

    No amount of free market or competition is going to create a good outcome if stealing is allowed.


    Comment by ptuomov — October 2, 2012 @ 1:56 pm

  7. Professor, apparently some are already capitalizing on small asset manager market:

    Comment by MJ — October 5, 2012 @ 4:12 am

  8. There is a lot of pontificating about how to best organize markets, SWP included. How do you honestly think you can stamp out the “bad” HFT and keep the “good” HFT? True the Euro regulators are wrong on first order- clueless in fact, but all the rest of your in-depth analysis will be bitten by second-order, third and n-order unintended consequences. You are not smarter than the aggregate crowd of market-makers, and certainly not fast enough to adapt to changing economic and technological condition. Whenever their are arbitrary rules of engagement, the advantage goes to the bad guy or incumbent. The regular guy and innovator is dis-advantaged. A fair-market is one of private negotiation and evolution. Governments do not create markets- they generally destroy them. Trading market arise spontaneously under leafy trees and are best nurtured by the profit/loss incentives of their members.

    Will be there be treachery, fraud, corners, panics and major losses with a non-government regulated market? Of course, but there will be fewer of those events with competition and private rules. Don’t fear crises, instead manage your risk based on their expected probability. In fact- it is the history of all civilizations to improve the management of risk (famine,wars, etc), not just financial history.

    I hardly shed a tear for the mutual funds and sleepy asset managers who are taken advantage of by the HFTs. These asset managers generally provide little value to their clients, and then go screaming to the regulators when their steady cash-flow is threatened by technology. Perhaps they should leave their second home in Forida (or Monaco) and spend more than a 2 days a month in the NY/Chicago/London office.

    Comment by scott — October 6, 2012 @ 3:30 am

  9. @scott. I don’t think we disagree at all. In general, I believe that it is never optimal to stamp out all deleterious activity because it is costly to do so.

    Moreover, if you read what I’ve written on the subject over the past months and years, you’ll see that I have favored “competition and private rules.” I have argued that exchanges have better incentives and better information and better tools to mitigate the costs of “bad” HFT. Competition and the profit motive are far more likely to result in better rules regarding HFT than top-down regulation.

    I encourage you to look back at my several posts on HFT and I think you’ll see we’re on the same page.

    The ProfessorComment by The Professor — October 6, 2012 @ 7:09 am

RSS feed for comments on this post. TrackBack URI

Leave a comment

Powered by WordPress