The Gateway to Algorithmic and Automated Trading

Flash ban wallop

Published in Automated Trader Magazine Issue 15 Q4 2009

Who loves ya baby? Well, if you’re an automated/algorithmic trader (and especially if you trade high frequency), right now it probably feels like nobody does. A grubby summer of media drive-by shootings haven’t done much for your popularity. David Dungay sets the record straight.

David Dungay

David Dungay

To judge by this summer's events, confusion, misconception and misunderstanding are just three of the printable words one could use to describe the mainstream media's grasp of the automated/algorithmic trading world. But that's OK, because in response to this mangling of reality, Automated Trader readers have been happy to provide us with all the unprintable ones we could ever need…

The holiday season saw a lot of non-specialist journalists step into uncharted territory; unfortunately, instead of buying a map, they reached for the scattergun. While some of the resulting inaccuracies may have been initially entertaining, there is a real risk that the end result will be the opposite of the much-claimed objective of protecting the interests of retail investors. To say nothing of the damage it may do to the perfectly legitimate interests of professional traders…

Where it all started - flash trading

With summer came the rise of a media storm around flash trading. First discussed by a group of financial websites, the New York Times (NYT) then picked up the story.

On 24th July the NYT ran a piece titled "Stock Traders Find Speed Pays in Milliseconds". The allegation was that an exclusive group of traders could see information before other market participants, via flash services. They could thereby obtain a better price at the expense of the retail investors. Various others in the mainstream media also tagged along with their own variations on this theme.

US Senator Charles Schumer then stepped into the fray, calling for the SEC to impose a complete ban on flash trading. Schumer claimed that flash orders allowed certain members of Direct Edge, NASDAQ and BATS to buy/sell order information prior to that information being made available to the public. Schumer's intent was the protection of retail investors. (Remember this point for later.)

The fight back

The main issues raised in the blizzard of flash order stories, together with paraphrased industry responses, were:

• Flash orders created unique front running opportunities. (err, no they didn't)
• Flash orders were used solely by high frequency trading firms. (no, they weren't)
• Flash orders were intended to disadvantage retail investors. (nope)
• Exchanges intended to flash quotes solely to a select group of traders. (no…wrong again)

Joe Ratterman

Joe Ratterman

Individual industry figures also kicked back at the flash allegations. Joe Ratterman, BATS CEO, denied that flash trades created an unfair advantage. In dealing with the various misconceptions above, Ratterman pointed out that BATS' BOLT orders (an order type used for flash trading that BATS suspended on September 1st) did not create a disadvantage for retail traders. "They provided price improvement to traders who could execute against them, whether they were retail or not," he said.

Ratterman also pointed out that BOLT quotes were made as readily available as the regulatory structure legally allowed and the only place they were not found was on the consolidated tape. "This enabled firms to publish legitimate quotes quicker than the consolidated tape, thus merely representing a more reflective picture of market conditions than the consolidated tape allowed," he said.

Evan Wies, CTO at quantitative investment manager Athena Capital Research, took a similar line: "The New York Times never added the disclaimer that flash orders are optional (you can opt out of them), so it sounded like whenever you place an order it's automatically a flash order and therefore susceptible to evil automatons," he said. "The other inaccuracy was that it was made to sound like only a small group of people could actually see and react to the flash orders, which isn't the case. In fact it is democratised, so for example in the case of a NASDAQ flash order, any NASDAQ participant could access it; it's not like a dark pool that might only have a select group of participants."{comment}

Paul Lynch

Paul Lynch

Blurring the lines - high frequency trading

The original NYT article also discussed high frequency trading (HFT) in less than specific terms. "Powerful computers, some housed right next to the machines that drive marketplaces like the NYSE, enable high-frequency traders to transmit millions of orders at lightning speed and, their detractors contend, reap billions at everyone else's expense." Oh and precisely how many billions are they reaping? Some hard numbers and the odd specific example would have been nice.
The "millions of orders at lightning speed" bit didn't go down too well with some of our readers. "Leaving aside the fact that light doesn't propagate through fibre at the speed of light, there isn't an exchange in the world that can handle more than two or three thousand orders a second, so how can people claim to trade quicker than that?," said Vladan Jovanovic, CEO of proprietary trading firm Communicating. "If you try firing just a couple of thousand orders a second at most exchanges for any length of time, one of two things is likely to happen. The exchange will disconnect you or the exchange will crash."

The NYT article continued: "These systems are so fast they can outsmart or outrun other investors, humans and computers alike. And after growing in the shadows for years, they are generating lots of talk."

All this appears to be based on the very large and very false premises:

• That every other market participant, including retail investors, is an idiot…
• …that also wants to trade high frequency.

Plenty of Automated Trader readers do just fine with lower frequency automated strategies thanks. What makes high frequency trading inherently "smart" anyway? And these strategies weren't hiding anywhere, we've been writing about them since Q1 2006.

The NYT then managed to imply that high frequency trading and flash trading were inextricably linked and failed to make the distinction between them. (While flash trades may make up a certain percentage of HFT volume, just how much is unclear.) Things then really started to run off the rails: "It is the hot new thing on Wall Street, a way for a handful of traders to master the stock market, peek at investors' orders and, critics say, even subtly manipulate share prices."

The "peeking" the NYT refers to is specific to flash orders (and only assuming, as Wies points out, you haven't opted out) and not of course HFT as a whole. The article doesn't draw this distinction. Unfortunately, by this time the Gray Lady's rather creaky bandwagon was already out of the barn and rumbling down the track. So there were plenty of "me too-ers" on other papers, magazines, sites, blogs and twits quick to scramble aboard and run variegated repetitions of the same misconceptions. To keep mangling the metaphors, the snowball (yes, thanks AT readers, we know you think it was another type of "s" ball) was rolling, gathering speed and size in the process.

Vladan Jovanovic

Vladan Jovanovic


A second bite

The NYT was at it again on September 4th with another HFT article titled "Secrets of Alleged Oil Price Manipulation Exposed". This time it focused on a Dutch company called Optiver, which was suspected of manipulating the price of oil with its high frequency market making strategies. On this occasion, instead of blurring the distinction between flash and high frequency trading, the NYT was blurring the lines across HFT, market making and market manipulation.

Phrases such as "super secret" and "cowboy trading tactics" certainly set the sensationalist tone, but the article also managed to drop a few clunkers along the way. One classic was the reference to the SEC's investigation of: "…the ability of some of the most powerful computers to jump to the head of the trading queue and - in a fraction of a millisecond - capture the evanescent trading spread before the rest of the market does."

Um, capturing trading spreads is what market makers do. That's the deal; you provide liquidity and in return have the opportunity to sell at the offer and buy at the bid. And getting there first isn't a crime either; turn back the clock to before electronic trading started and look at what used to happen in the futures exchanges' trading pits. He or she who shouted loudest and fastest got the trade and made the spread. That's how it works - it's competitive, it's a market - not a social engineering project.{comment}

Peter Green

Peter Green

Spill over

To be fair, the September 4th NYT article doesn't say: "All high frequency traders are thieving cheats, screwing the faces of widows and orphans into the dust." Unfortunately, by not drawing its distinctions clearly, it does create the impression that HFT = manipulation. The concern here is not individual rights/wrongs, but that what started as incomplete coverage of flash trading (repeated ad infinitum elsewhere) is spilling over to create an inaccurate and negative picture of high frequency traders and, by association, automated and algorithmic traders as well.

Plenty of AT readers take exception to this blurring of terminology. "Flash trading is a very different story from everyone being able to buy the fastest computers available to be the quickest into the market," said Kyte Group CEO Peter Green.

Already there are high profile financial blogs that have made this leap of blame by connecting HFT with algorithmic trading. A popular theme is that HFT has the potential to: "spark a market meltdown because so many computer programs employ the same algorithmic strategies". UAL's price collapse after a very out of date news story about a bankruptcy filing (which was accidentally run on certain news services) is often cited as an example of this.

This conveniently overlooks the fact that human traders can precipitate mayhem just as spectacularly, if not more so. Remember our friend at Mizuho in December 2005 who sold a thousand times the correct order size and caused carnage in the Nikkei 225? Or the October 2002 fat finger artiste who managed to shut Eurex for three hours with his handiwork? Or the bright spark who went elbow trading in Premier Oil in '99? So while high frequency trading algorithms may go haywire (and the smarter trading venues have measures in place to neutralise the systemic risks of this), they certainly don't have a monopoly. {comment}

Getting it wrong - regulatory repercussions

The biggest concern surrounding this muddle is the possible legislation/regulation that may result.
The chain reaction isn't hard to trace:
In many respects this situation could hardly have arisen at a worse time. After the massive fall out of the credit crisis, pretty much anyone professionally involved in financial markets is perceived as already half guilty of something. As a result, the general media are looking hard for skeletons in financial cupboards and if they can't find a complete one, building a composite from a selection of random bones will do just fine.

That gets the ball rolling nicely, but the political class is also pretty well sensitised after mostly failing to grasp the consequences of sub-prime until too late, so has plenty of motivation to be hyperactive in this area. Finally, most regulators didn't exactly cover themselves in glory over the credit horror, so they also have plenty of motivation to appear dynamic this time round. (Incidentally it's worth mentioning that the speed with which the SEC responded to Schumer's initial flash ban call was almost in the realms of, um, high frequency.)

Sang Lee

Sang Lee

Ultimately there's a measure of convenience for various interested parties here. "The global economy is in a mess and everybody is looking for a bogeyman," said Paul Lynch, managing partner of algorithmic trading specialists PE Lynch. "It is far easier to point the greedy finger at 'algorithmic traders' than it is to question why people on modest salaries were allowed to purchase 5 'buy to let' new build flats 'off plan' which have subsequently gone sour."

The SEC banning flash trades could be just the tip of the regulatory iceberg. Already other regulators are considering regulation of HFT with some, such as the FSA, openly stating this. Various rumoured measures include:

• A specified maximum ratio between message volumes and completed trades
• A cap on the percentage of total traded volume a single participant can execute in a given time window
• A minimum period a quote must be exposed for
• Restrictions on cancel/replace volumes

A large part of the motivation for any such regulation is that retail investors are being ripped off, which isn't actually the case. "Often people assume that what benefits the institutional market must be damaging to the retail market," said Sang Lee, Managing Partner of Aite Group. "In fact, there is an argument that retail investors have never had it so good. With the advent of online trading venues execution costs have almost been completely dispelled and much wider dissemination of data means that they enjoy a level of transparency as never before."

There is always the view that things could be made even better for retail investors, but any resulting imbalance could prove disadvantageous for all. "Once you mix in the potential threat to the retail investors, things could get dangerous," said Lee. "Regulators may over-focus on the retail market and come up with regulation that may not be so good for the institutional market, which could in the long run make everyone's lives harder."

Restrictions on HFT and algorithmic trading could be a classic example of this law of unintended consequences. "Governments want their citizens to have cheap access to investment products such as index trackers and capital guarantee bonds, so that they can save for retirement," said Paul Lynch. "Products like these require very liquid markets with tight bid ask spreads, so that they can be accurately priced in real time and traded in a speedy and efficient manner. High frequency market makers provide quotes on the underlying constituents of such products therefore allow people to add to or reduce their investments at very short notice. This is typically done by using an algorithmic trading strategy."


Transaction Tax

In addition to the imposition of regulation, automated/algorithmic traders may also have the small matter of a transaction tax to contend with. While the original Tobin tax (suggested by US economist James Tobin) has been kicking around since the early 70's, countries such as France and Germany have recently been highly vocal in their support for reviving such a global transaction tax.

However, there is a strong feeling in some quarters that proceeding with a Tobin tax wouldn't just affect high frequency traders but also retail investors and pensioners through higher cost investment products. "If governments want this to continue then they have to accept that market makers are an essential part of an efficient market," said Paul Lynch. "If they are removed and a transaction tax is introduced then liquidity will decrease, bid ask spreads will increase and therefore the transaction costs of the investment product go up."

Others see a transaction tax as unlikely to come to fruition anyway, because of the difficulty of imposing such a tax on a highly mobile group of traders. "If a tax was introduced in one jurisdiction and not in another you would find all sorts of tax arbitrage taking place," said Kyte's Peter Green. "With the order routing systems that exist around the world it's not difficult to route orders from one country to another. Unless you could guarantee the Tobin tax being applied globally, which is unrealistic, you will find as much opposition to it by politicians on one side of the house as the other."

Some US traders have a similar view. "There has been talk of a transaction tax, but I think like a lot of these things once the probable consequences are understood, it will be dropped," said Bill Braman (see this issue's First Person). "The tighter market with a better bid/ask spread created by HFT benefits everyone and in any case a tax would be a big incentive for traders to simply shift location." {comment}

Bill Braman

Bill Braman

The future

So what does the future hold? Right now, it doesn't look good. There's a pile of misinformation out there, along with a stack of twitchy fingered politicians and regulators. Not a great combination.

High frequency traders, as well those trading automatically or algorithmically at lower frequencies, have a market call to make here. Batten down the hatches and hope the storm will pass over as the general media, politicians and regulators move onto something else? Or try to explain slowly and clearly exactly how everything really works and fits together?

The former is the tempting low-effort option with no immediate cost, but this time it just may not work. The second alternative is probably the better trade, but it is a labour intensive one. We'd like to do our bit here, so we're putting together a guide to the auto/algo trading industry that will be available on to anyone looking for a quick factual introduction. So if there are any misconceptions that affect your business you've seen floating around, please fire them over to - and we'll do our best to nail them. {comment}