The Gateway to Algorithmic and Automated Trading

High Frequency Trauma - What unintended consequences are in store as regulators target HFT?

Published in Automated Trader Magazine Issue 26 Q3 2012

Proposals to regulate automated trading have been coming thick and fast. Bob Giffords, with the help of a host of market participants, peers into a crystal ball to see what some of the side effects could be if any of these ideas were adopted, and whether the market has cause for concern.

shooting ducks

Justin Schack

Justin Schack

"There's a fair amount of actionism in the air," says Dennis Lohfert, fund manager at Ion Asset Architecture.

For Lohfert, the regulatory focus on high frequency trading (HFT) can be viewed through a psychological prism: in times of crisis, people want to be perceived as doing something. "History is filled with archaic constructs in legislation and regulation that were implemented without any real need and then happened to linger for literally decades."

The result is a broad and potentially threatening agenda for HFT traders. Some of the most talked-about ideas being bandied about include: a transaction tax, speed bumps for markets, maximum order-to-fill ratios, minimum order resting times and market making obligations for HFT traders.

Yet is the regulatory bark really worse than its bite? Certainly, many automated traders are feeling victimised. They demand to see evidence that automation coupled with competition is bad for markets. "Every empirical measure of market quality shows market quality has improved for investors," says Cameron Smith, president of Quantlab Financial, a Houston based quantitative technology and trading company.

Is such mistrust justified? What is the real impact of such regulation likely to be on the day-to-day automated world that many embrace, but others fear?

Financial Transaction Tax (FTT)

The driving force behind the FTT is mainly coming from continental Europe, where it is a moving target. "The UK, Sweden and Ireland are all prepared to veto an EU-wide tax, while France and Germany could move forward with their own taxes," says Justin Schack, managing director of market structure analysis at Rosenblatt Securities.

The French proposal comes in three parts. Chris Marsh, head of AES EMEA for Credit Suisse, says one part looks like stamp duty and might be anywhere from 10 basis points to 30 applied to net buys. Another part looks at HFT and will add 1 basis point, and a third piece will address CDS trading. In its present form, it would only affect HFTs in France, he says. "It will definitely happen, but may not have much impact on HFT. The UK already has stamp duty, and net buys will exempt market makers. The HFT piece is not yet defined but could be based on an order-to-fill ratio above some limit."

But while the effects of any French tax may be contained, there are broader proposals out there. Across the Rhine, a German-Austrian proposal has been advanced which would also introduce an FTT for some EU members.

If such plans go ahead, the key question will be implementation. Jan Dezort, a specialist in regulatory affairs at proprietary trading firm RSJ, says that if the tax were structured like stamp duty, the impact would be quite limited since traders would simply seek out untaxed markets to trade while lawyers got to work looking for ways around it. But he adds: "With strict rules, FTT would restrict liquidity and increase costs as many HFT traders would have to shut down. The consequences are potentially very negative."

Cameron Smith

Cameron Smith

One concern is that margins on securities transactions are razor thin. "So any new tax can have a negative effect on trading," says Mike Corrao head of equities compliance at Knight Capital. If Britain were to go ahead with such a tax, US lawmakers might consider something similar, but there has been no serious discussion in the US for some time, he adds.

"This is the most stupid idea since bungee jumping without a rope," says Peter van Kleef, CEO of Lakeview Capital Market Services. "This is a cost that would only hit institutional and small investors, as market makers will just widen the spreads. Volume will just move to markets without the tax."

Lohfert at Ion also notes the European Commission has described the tax as a way to force banks to pay for their own bailouts, but he too says in reality it would affect many other firms. "I fail to see why pension funds, private investors, and non-bank liquidity providers are being tapped for the same purpose."

What worries Lohfert most are the unrealistic assumptions about eventual revenue levels. "There are two scenarios that are possible: volumes decline by a factor of 80 percent or the market will find a way to avoid it," he says. "We may need new terms and conditions, but not taxes," says Hendrik Klein, CEO at Da Vinci Invest in Switzerland. "Taxes add costs, create avoidance opportunities, and could push markets off shore where it is more hassle to control them."

Speed bumps for the financial markets

When it comes to speed limits, much of the focus is in the US.

"After the Flash Crash," says Smith at Quantlab, "Europe realised the value of their existing and long-standing circuit breakers that provide protection against runaway electronic markets. The US equity market was the only major market centre that did not have them. Not any more. Circuit breakers eliminate the risk of any future flash crashes."

The Securities and Exchange Commission has now approved a limit-up-limit-down system to replace the existing single stock circuit breakers introduced after the Flash Crash in May 2010. The system should go live in February 2013. However, if the stock remains either limit-up or limit-down for some time, the market will halt and go to auction, as before.

Schack at Rosenblatt says the new system will probably prevent "fat finger" problems, technical glitches or other mistakes from causing wider dislocation. But it could create complications. "If there's significant news that legitimately moves a stock, this system keeps it from trading at the right price and could cause investors to make those trades elsewhere."

For this reason van Kleef of Lakeview prefers volatility interrupts, since they give people time for price discovery in the auction and then carry on. "Limit-up-limit-down can last all day and create uncertainty," van Kleef says. "Some traders will just offset their exposures in OTC or derivatives markets, while others can't."

Jan Dezort

Jan Dezort

In Europe such software protection is defined by each trading venue. "The only issue here is whether restrictions are applied consistently across markets," adds Marsh at Credit Suisse. "Different platforms already have such rules, but they work very differently and that can be a problem. Poorly calibrated rules might make that worse and add to market inefficiencies."

But if trading halts were to become market-wide in times of volatility, the market would need a consolidated tape, says Adam Eades, head of legal and regulatory affairs at BATS Chi-X Europe. "And that won't happen in Europe for a couple of years yet. It still might not be practical since different markets take different views."

Maximum order-to-fill ratios

While opinion was mixed on the virtues and vices of order-to-fill ratios, there was fairly broad agreement that decisions should be left to the trading venues.

"It all depends how well such limits are calibrated," says Smith at QuantLab. "In general, high messaging rates enhance price discovery and are a reflection of intense competition. At the same time, there can be excess." He argues that such limits should take into account the unique characteristics of each security and platform. "A one-size-fits-all standard will only harm investors."

Eades of BATS Chi-X Europe says he would prefer to work with participants and regulators to monitor activity and put pressure on members when they step out of line. "What is normal activity depends on the market model and keeps changing as technology gets faster," he says.

If set at the right levels, order-to-fill ratios would not be a problem, says Klein at Da Vinci Invest. "Only traders who are gaming the system would need to change their strategies."

Marsh at Credit Suisse notes the ratios people are talking about range from 75:1 to 400:1, numbers that would be unlikely to affect most traders.

"Many HFT traders will check with an exchange what traffic rates they are happy with and limit their flow accordingly," says Marsh. "Others don't have the control structure to do that, so they will incur an extra cost." However, he agrees that if the limit is too low it will force traffic back to the exchange and undo a lot of the benefits of competition. "It needs to be done on an exchange basis," he says, asking for an exemption for ETFs or illiquid products, where market makers quote all day with few trades. "Otherwise, liquidity will disappear."

Some markets already impose such penalties. A recent example is the anti-incentive scheme by NASDAQ for excessive order cancellation. "The rule is aimed at reducing inefficient order entry practices of certain market participants that place excessive burdens on the systems of NASDAQ and its members," says Corrao at Knight. He adds that a disincentives programme is generally preferred in such circumstances, provided it accomplishes the same goal as a limit.

Mike Correo

Mike Correo

Nonetheless, some still see such constraints as harmful. "Orders represent liquidity, it really is that simple," says Lohfert at Ion. "If you limit the amount of orders, you are also somewhat limiting liquidity. How much is obviously a function of how generous or restrictive these things are."

From the viewpoint of a systems architect, Lohfert also sees a tremendous amount of poorly implemented order submission strategies out there, which generate an excessive amount of traffic for no reason but poor design choices alone. "However, I would also point out that there are many exchanges and venues that have restrictions in place already," he says. "So again, depending on the venue, this is not necessarily a game changer."

Minimum order resting times

Some algo traders are relaxed about this one.

Lohfert of Ion says some venues already have minimum resting times, although there is the likelihood that longer minimum quote lives would require slightly wider spreads.

"We do think a reasonable number here is not necessarily a bad thing, in line with some of the poor messaging characteristics we see," he says. "This is obvious, and one just needs to decide which way they want to lean."

Corrao at Knight says US regulators have been discussing the idea publicly, though there was no proposal on the table right now. Knight is recommending a minimum order resting time of one second unless the order is executed.

But other algo traders remain sceptical.

"A small minority of people in the States are pushing for minimum resting times," says Schack at Rosenblatt, "but I don't think there's much of a chance of them being adopted."

Cancellation is all about risk management, he says. So if you cannot cancel, you take less risk, and that will mean wider markets and higher implicit transaction costs. "Yes, there may be people out there playing funny games with entering and cancelling orders without the intent for them to be executed," says Schack, "but that's against the law and should be prosecuted. You don't solve that problem by hurting the ability of all the legitimate players in the market to contribute to efficient price discovery."

Van Kleef puts it more bluntly: "This is a very bad idea. The longer my quote has to rest, the wider the spread. Liquidity will drain away and transaction costs rise. It's as simple as that. It's much better to have tight spreads with more competition and frequent updates."

QuantLab's Smith points out that studies show intraday short term volatility is falling and some suggest HFT reduces volatility. "So not only would resting times increase investor trading costs without a benefit, but they would also increase short-term volatility."

Peter van Kleef

Peter van Kleef

Da Vinci Invest's Klein also argues that for fragmentation to work, traders need to be able to cancel if they get filled elsewhere.

Meanwhile, Eades at BATS Chi-X simply wants flexibility for each platform to set the right level for their market model. "Hard parameters could negatively impact markets and competition," says Eades. "At the right level it shouldn't have a huge adverse impact. At the wrong one, it could impact liquidity and spreads."

Market making obligations for HFT traders

This one is all about definitions. One of the more controversial proposals under MiFID II has been the "continuous operation" requirement in Article 17.3.

"Trying to define algorithmic trading and HFT is problematic. What are the right parameters? What level of obligation should firms be under? What happens during technical outages or market disruption?" asks Eades.

While there is no specific regulatory proposal on this in the US, Schack at Rosenblatt sees the proposals from Europe as highly problematic. "If you force people who are using algos, or colocation, or direct data feeds to make two-sided markets all day, you don't just capture the HFT guys. All the big brokers are colocated and take direct quote feeds because there is no consolidated tape."

More broadly, Schack notes that history teaches that market-maker obligations rarely work when you need them most - during times of stress. "You can't force firms to go out of business for the greater good," he says. "Any new obligations would erect higher barriers to entry for smaller market-making firms, and it's not clear there would be an equal or better benefit to offset that blow to competition."

While there is no US proposal on this equivalent to MiFID II, Corrao says Knight is engaged in the debate about market making standards in general. "There has been some discussion by US regulators on enhanced obligations and incentives for market makers and we expect something will come out of it. We actually have recommended enhancements to rules on spreads, depth, capital adequacy etcetera. We are optimistic that the US regulators will address this issue soon and would expect that any standard considered in Europe would use similar criteria."

Smith at QuantLab, however, warns: "Regulators should not create a privileged class of market makers. Nobody believes they will stop a falling market and yet we know that creating a privileged class of traders will reduce liquidity and harm market quality." He believes that is why European markets generally do not have market making programs in their liquid securities.

Regarding the claim sometimes made that HFT traders abandoned the market in the Flash Crash, Smith says evidence shows most HFT traded through the crisis. "The few that did stop trading were only reacting to the severe conditions and did not cause them," he says. "The empirical studies of HFT during the high volatility days of 2008 show that HFT firms increase their activity during volatile market conditions."

Adam Eades

Adam Eades

There's a lot of uncertainty here," says Dezort at RSJ. "Clearly, it would raise entry costs for new competitors and could actually reduce liquidity and increase spreads once again if current liquidity providers exit the market. The technology and need to trade all hours would represent significant costs. It then might well shift liquidity to other marketplaces and asset classes, reducing market transparency."

At the same time, some say a wide definition that encompasses too many firms would penalise some groups and create unnecessary traffic.

"You can't force people to make two-sided markets," says van Kleef at Lakeview. "Everyone is HFT today to achieve best execution with fragmented liquidity. Should pension funds have to carry the risks of market making?"

Others see advantages for their firms in such rules. "We are happy to fulfil market making obligations," says Lohfert at Ion, "as will be many other firms. How this is supposed to help the whole transaction tax initiative then is beyond me." Since more market making will increase volumes, this would appear inconsistent with the other regulatory initiatives and FTT in particular.

Marsh at Credit Suisse says MiFID II doesn't specify what the obligations might be, but he notes a more restrictive definition might be welcomed by the larger HFT traders because it would limit competition and give them an advantage. "They can easily adapt their models to the new rules. Technology opened up the market to more competition. Do the regulators really want to close it down again?" Another example perhaps of regulatory dissonance?

Some final thoughts

Although there is not much enthusiasm among HFT traders except for speed bumps, all the tough regulatory talk may amount to very little indeed. Traders will probably adapt to all the proposals in a softer form. Whether market quality and competition suffer is, of course, another matter.

On the other hand, in each case the regulators could go too far in ways that seriously impair financial markets and actually make life worse for low frequency traders.

"There are an enormous number of variables here," says Marsh of Credit Suisse, "so just changing a few will have unexpected consequences. Every proposal creates loopholes that become their own problems. It then takes the regulators years before we get the next patch to the rules."

US & European Regulatory Proposals for HFT: A Quick Guide

Proposal impact Market Response Likelihood

Financial Transaction Tax: Also known as the Tobin tax, a charge levied on settled trades to discourage short- term trading strategies.

Higher trading and capital costs, lower liquidity and stock prices, potential drag on economy; if implemented like UK stamp duty with wide exemptions, then quite modest.

Traders impacted would widen spreads and shift liquidity to untaxed markets such as derivatives like CFDs unless also taxed.

EU: Highly likely in France, Germany and some EZ but not UK, with probably modest impact like UK stamp duty. US: No proposal, highly unlikely.

Speed bumps: Volatility interrupt auctions, limit-up-limit-down rules or other software circuit breakers to control volatility and sudden price swings. Widely implemented on European exchanges, but very few in US equities markets prior to the 2010 flash crash.

Limited aside from reducing risks of another flash crash; can in some cases add uncertainty if markets not free to absorb news, but should increase market confidence and reduce need for fat-finger halts or busted trades.

Traders will adapt algorithms
and may increase risk appetite in extreme volatility because of greater protection. Some may try to exploit trading restrictions to trade in other markets or asset classes (derivatives) during a limit stop.

EU: No proposal, already widely implemented by individual exchanges. US: to introduce limit-up- limit-down rules by 2013.

Maximum order: fill ratio limits & charges, possibly linked to cancels: Aimed at reducing excessive traffic on trading platforms beyond what is needed for a fair and orderly market. Many US and European trading platforms already do this with new proposals now emerging, possibly combined with the FTT.

Depends on the calibration and whether market specific or across the board; probably very limited given that current discussions are for high limits possibly with some discretion by platform. If too low could reduce liquidity and undermine the benefits of competition.

Traders will build in control structures to limit orders or cancels as required. Some strategies may need to be changed, if applied to less liquid products, like ETFs.

EU: likely to be part of French FTT, other countries still discussing. US: likely to spread by exchange
or ATS; for example, NASDAQ has just introduced a new 'disincentive' scheme.

Minimum order resting time: Aimed at ensuring all members have opportunity to access orders; discourages market abuse showing orders with no intent to trade. Some trading platforms already do this, notably the EBS FX platform.

Could have only a modest impact, if well calibrated, but volatility could rise, spreads widen, costs increase, & average size fall if not. If blanket rules adopted, might reduce both liquidity and competition.

Flow may move back to the primary exchange where it cannot be safely spread around. Trading models would have to be redesigned to reduce risk.

EU: Possible but unlikely. US: Probably not, but could be part of regulatory discussions on tightening the minimum standards for market makers.

Market Making Obligation for HFT: Attempts to trade off colocation and other HFT advantages for guaranteed liquidity at times of market stress. Regulatory minimal standards for market makers (MM) & liquidity providers: apply to HFT but aim to improve market quality more generally.

Potentially broad impact since so many traders have some HFT characteristics. Adds cost and market traffic, raises competitive barriers and may not improve liquidity in times of stress. Tighter minimum MM standards low impact, if well calibrated to trading venue.

While many larger HFT traders may become Mms some will not, and so will withdraw liquidity or change trading strategies. HFT traded through the flash crash or withdrew when markets became chaotic. So unlikely to lose money deliberately in volatile markets.

EU: Probable for HFT traders at least in some softer form as part of MiFID II US: No discussion for a blanket HFT obligation, but probably tighter minimum MM standards in general.