31.01.2012
By Simon Miller
In an industry where deals can be measured in micro-seconds, gaining an edge is a must for some companies that pride themselves at being the first to the line. And yet, is latency the be-all and end-all for investment strategy and should smaller firms catch up?
Bank of England’s executive director for financial stability Andrew Haldane described the chase for speed as a race to the bottom and it is a concept that is not lost on vendors.
With traders talking of speeds of pico-seconds – that’s one trillionth of one second – there is an obvious question of how fast can trading really go.
“The speed of light,” says Iran Hutchinson, software systems architect and product manager for US-based InterSystems. “I think we are just going to keep going and going until there are levels of latency architecture that you can implement and buy – that’s what we are seeing more and more of now and you will see hardware platforms focused around latency.”
He adds that as the market grows, latency-orientated systems will be
introduced into computers.
“You will have an inbuilt architecture design for low latency that will keep evolving,” says Hutchinson.
However, if getting latency is so simple, surely other companies would be doing so as well. Currently, in the US only 2 per cent of firms account for 70 per cent of all high frequency trades (HFT).
Hutchinson continues: “Most HFT now is definitely done by a small number of firms with the money resources and expertise. And the whole trickle-down theory is that it will get down to the smaller firms is less a theory than
a matter of time. Because processor costs reduce.”
And yet, despite lowering costs it may not be the case that all firms will be
chasing the lowest possible latency.
CEO of independent think tank JWG PJ Di Giamarrino points out that it is dependent on your investment strategies and uses the analogy of performance cars versus race cars.
“If you think about these like race tracks,” says Di Giamarrino, “the Formula One [of lowest latency] requires you to be hyper-competitive and treat it as a technology arms race.”
He continues: “But then again, there are plenty of Mercedes E class drivers out there that are perfectly happy not being in a Formula One car and their trading strategies can get along just fine with slightly higher latency.”
Rutger ter Hoeven, marketing manager at Interxion, agrees that it is dependent on requirements with firstly the high frequency traders who want the ultimate fastest latency involving the right location for their servers, best hardware and connections.
“The second group is probably clarified as automated users. People who want to be in the ball park but may not want to compete in the ultimate latency. The first group would count in micro-seconds while this groups counts in milliseconds,” ter Hoeven adds pointing out that this group were also looking at reducing operational risks by not chasing the fastest trades.
For ter Hoeven, the third group could be described as the “manual traders” of asset managers and the like. Although traditionally they have used smart routers and systems provided by their brokers, ter Hoeven says that there was a trend that the buyside is getting involved as well and are looking at becoming more active as well, “probably more for the requirements of reducing risks rather than beating the completion on latency, speed of execution”.
One of the issues for cheaper latency is the sheer physics of it. Despite the
advent of cloud systems, much still depends on copper wire and
proximity to exchanges to maximise speed, although it is not the be-all and end-all of trading systems.
The growth of cloud computing is continuing apace with virtual machine technology providing incredible speeds and it could be that will be the entry-level for those smaller firms that lack the scale to compete.
However, there are risks to the cloud, explains Hutchinson: “The inherent risks are not only in the cloud but in the technology that is around the cloud. A lot of the data management systems that are in the cloud are operating systems that aren’t completely tested.”
In addition, the cloud systems do not operate on a broad base, running the risk of selecting something for one thing but missing all the others. Other risks include using a technology that is completely unproven whatsoever or using different strategies in the cloud where a firm could run the risk of falling foul of regulatory issues.
So for the smaller firms trying to compete with the top dogs, risk will
become part of their innate strategies.
“Small firms have to be more innovative in order to compete because they can’t get into those exchanges to pay those fees to whoever’s hosting,” says Hutchinson.
“I think smaller firms will have to focus more on strategy than the bigger firms right away in order to get to that point where they can compete.”
Speedy regulation?
Like all parts of the financial world since the 2008 crisis, latency and trading
platforms have faced even greater regulation globally leaving firms struggling to keep with demands from various jurisdictions.
Be it Dodd Frank or MiFID, vendors and traders alike have to deal with vast streams of legalese which many believe are not set out with the interests of
customers or users in mind.
Hutchinson blames politicians for relying too heavily on academics to examine what-if scenarios rather than seeing the praticioners.
“Basically how does anyone start a new process that is going to be implemented? Logic stipulates that you start with the institutions, interview them, it is going to impact them and their customers and if the customers see the higher price and they know where it is coming from, someone is going to get blamed.” he says.
For Di Giamarrino, this is part of an on-going game between traders and regulators over how strategies are played in the market place.
“If a regulator puts in place a requirement to look at algorithms (algos), those regulators that have more technical nous and the resources to do that will make more of a meal of it than those that don’t,” he comments. “With firms having an obligation to their shareholders, they will put those algos into other places where the regulators will not be watching.”
Di Giamarrino continues: “It is a fascinating topic in that it speaks to the level of technological awareness and existence in the policy framework that are being pushed out globally, and of course those frameworks are more advanced across the pond, followed by the UK and then Europe in policy discussions.”
However, ter Hoeven points out that the focus on HFT ignores the fundamental issue that there will always be someone that will play the system, be it a rogue trader at UBS or the messenger for Rothschild’s racing back to bring the first news of the Battle of Waterloo to his bosses for commercial gain.
“There are always people that look to the edge of what’s possible and go over the edge at some stage but that is more to do with the philosophy of an organisation, then speed per se or advanced technology,” he says.
Taxing technology
The flash crash of 2010, where major equity indices in both futures and
securities markets crashed by 5 per cent to 6 per cent before recovering, shocked politicians and regulators alike with the feeling that events were out of control and turned their attention to HFT.
Further pressure was brought to bear in this arena by the accusations from
Europe that HFT had increased the pressures in the eurozone.
As a result, the idea of a financial transaction tax was introduced, at the time of writing, eurozone leaders were trying to get agreement to bring this so-called Tobin tax into being in Europe in the face of resistance from the UK.
Di Giamarrino points out that a financial transaction tax has been rising up the league table of regulatory issues to the point where it is now a credible threat and believes that it could have some effect on trading.
“The reality is, if you look at the size of orders over time, there is a lot more activity out there than before,” he says. “Now some of that activity may go away when it becomes uneconomic if you have to pay extra taxes on it. And I think that is where you fund a divergence of opinion on whether that’s healthy or not.“
However, Hutchinson believes that fees would simply be passed on to
clients and that trading would not fundamentally change: “You’ll increase execution prices for their customers, you won’t tax them, and they won’t feel a thing. Politicians don’t realise that anything they do to prevent, or show
prevention, of damaging organisations will be passed on and gotten around. They don’t get it and it is a fundamental flaw because they’re politicians.”
So is it actually possible to regulate against latency?
Hutchinson believes that regulation would just be gamed and transparency would only allow you to see the mayhem as it happened.
The answer appears to already been trialled in the States, where the
Securities Exchanges Commission is looking at circuit breakers where unusual trading for five minutes would cause trading to halt while human intervention can be sought.
Hutchinson comments: “As speed increases you can have bigger crashes faster, maybe even longer this time, so we need basic systematic circuit
breakers at certain levels maybe even targeted ones at certain organisations because we are in unknown territory you can’t take anything off the table.“
The one thing that is apparent is that the amount of regulation makes life
difficult for investment firms. Which compliance issue should they deal with first? Can a firm risk holding off on implementation strategy?
“So this isn’t one of those things that you should be waiting till 2015 to take a look at because by then it could be too late and your business strategy could no longer be valid. One way or another you have to make adjustments whether it is going to fight your corner or get out there directly and suggest what ought to be done or change your business model. It is a tricky time with a lot of problems for buy-side firms,” says Di Giamarrino.
For Hutchinson, there is a fundamental flaw in the way regulation is looking at the impact of industry. He says the industry will have to discuss the impact of low latency because technology is enabling financial impacts in ways it hasn’t got its heads around.
He adds: “It is not just about greedy bankers. Essentially they are just getting bigger guns to shoot themselves with.”
Despite the growth in technology, speeds and interfaces, at the end of the day investment is always going to depend on what strategy is needed.
ter Hoeven comments: “It depends on trading strategies, if you look at large block trades, it is not important to do those fast because the frequency of that type of trade is obviously much lower than the smaller high frequency trades.”
Hutchinson adds; “Let’s say we all have enough money to do that. The
virtualisation process gets faster and faster and you are going to have a
virtualised environment that can do that.You will have a lot more people having access to the exchange so you level the playing field. What do you do then?
"You focus on business strategy. You change strategy.”
