Germany is an economic juggernaut, with a huge stable of exporting heavyweights, some of the world’s biggest banks and a corporate tradition of technological sophistication. So it stands to reason that Germany’s buy-side would be among the more enthusiastic clientele when it comes to using algorithms for forex execution. But in fact, large asset managers and corporate treasurers have only been slowly coming around to using them. FXAlgoNews speaks to analysts, platform providers and others in the industry to understand why Germany has been tentative about adopting these tools and whether that may change.
A TOE IN THE WATER
For the FX algo community, there seems to be one clear fact and a lot of open-ended questions when it comes to algo adoption in Germany. On the factual side, the buy-side community in Germany is clearly interested in using algorithms for trading foreign exchange. Interviews by FXAlgoNews with a variety of market participants indicate that asset managers and corporate treasurers have been asking lots of questions. But will that interest translate into greater usage? What is behind the hesitancy to date and what are the chances that it will change in the near future?
The answers to these questions are not so clear-cut. The general trend, as is the case elsewhere in the developed world, is that algos are generating interest from sophisticated buy-side players who are keen to reduce market impact and demonstrate best execution. But Germany is a special case.
To get a sense of how much real trade is generated by Germany – the world’s fourth biggest economy behind the United States, China and Japan – it is instructive to look at the hard numbers. Germany’s companies exported a record 1.13 trillion euros worth of goods and services in 2014, while imports inched up to 916 billion euros. Exports, according to the country’s statistics office, were 3.7 percent higher than in 2013.
If a sizeable order in FX terms constitutes anything more than $50 million, Germany’s corporate sector alone in aggregate will have generated FX flows worth far more than 40,000 major orders last year. Those are orders that every algo provider would be keen to be able to execute. Of course, FX business does not slice up nearly so neatly. But the point that Germany offers a rich vein of potential clients for FX algo providers is not lost. The need is there, but the reality is that many firms on the buy-side are cautious.
While Germany appears to be slower than other geographical regions in moving towards algos, some large corporates are looking at the use of algos for best execution. One issue is they typically have much fewer transactions to make than other buy-side segments and as such they may not get used to using algos in the same way as other firms. Institutional funds, on the other hand, typically have people who trade FX as a job so their comfort levels will be higher.
“The trend is definitely going to be more algorithmic. Most of these algos are in-built within pricing engines, to create pricing and hedge in the back end in an algorithmic way,”
Alfred Schorno, managing director at 360 Trading Networks, said corporates, banks, hedge funds and asset managers have widely different processes for using algos.
Schorno said that asset managers have started to ask more questions about using all sorts of execution algos, particularly in the light of discussions about fixings. He added that Germany does not have as many hedge funds as other countries.
“If we take the corporate world for example, only in the last half year have we had requests to support more advanced execution algos,” Schorno said. “I understand that some of them had some initial trial phases in single bank algorithmic executions and thought they were interesting, especially for special purposes, large amounts for example.”
He added: “They are not widely used in the corporate and asset manager world so far.”
Schorno clarified that he was speaking here about more complex algos than standard TWAP and VWAP execution algorithms. “We know that a few of our clients use single bank algos and proprietary algos of banks at this moment and only recently, in the last two or three months, have we had requests to actually incorporate some of them in our multi-bank portal.”
One point that has come up at times as being an impediment to adoption of FX algos has been the lack of agreed standards for transaction cost analysis and what constitutes “best execution”. Not everyone shares this view, however, and some believe the real issue is the degree to which the buy-side learns about the value proposition algos offer.
Germany has been one of the more proactive countries when it comes to regulating algorithmic trading. The German parliament passed a law in 2013 that requires any high frequency trader, regardless of where it may be based, to have a license if it is trading in German markets. The law also introduced conduct and business rules for these firms.
While the law differentiates between HFT and other uses of algorithms, such as using them to source liquidity and achieve best execution, some have argued that the political and regulatory climate in Germany may have discouraged firms from experimenting even with execution algos.
Dr Alexandra Dreibus, a co-author of a paper on German regulation by Freshfields Bruckhaus Deringer (www.freshfields.com/uploadedFiles/SiteWide/Knowledge/35608.pdf), said German banks need to collect certain information from any of their clients who are providing algos to the bank for the execution of their orders. “German banks are also subject to special organisational duties for normal algorithmic trading,” she told FXAlgoNews.
That said, there seems to be little question that corporates and so-called “real money” firms are enquiring about algo usage, albeit not to the extent seen, say, in the United States, where usage is more common.
Dr Dreibus said Germany has wanted to be ahead of the curve when it came to financial regulation because politicians believe that may give the government more influence at a European level.
Germany’s general inclination to be seen to be out in front on regulation certainly can be seen in the language of a report from BaFin, the German financial regulator, last year when it noted that steps were being taken at a European and global level to address FX and interest rate fixing procedures. “From BaFin’s point of view, the draft regulation falls into the same category as all previous reform projects: they are headed in the right direction, but do not go far enough,” it said after initial reforms were proposed.
Dr Dreibus said that she was not aware of any other additional regulation initiatives at the moment concerning algorithmic trading in Germany.
WHY ALGOS COULD GAIN TRACTION
There have been signs that volatility is returning to the world’s foreign exchange markets, particularly in the wake of the Swiss National Bank’s stunning policy reversal which led to huge FX swings earlier this year. While extreme volatility such as that makes any kind of trading problematic, a general uptick in volatility is expected to encourage algo adoption as it means greater volumes.
Schorno said that client interest from Germany, at least on the 360T platform, had been quite limited over the past couple of years. “Those who use proprietary algos of banks do so in very specific situations, mainly if they have to do very large amounts,” he said. “On the other hand, these large institutional clients, like corporates and asset managers, get super liquidity for quite sizable notionals just by asking an RFS standard request which reduces the need for execution algorithms.”
But for the hedge fund community it is a different story. “In the hedge funds space we see a big need of course, the ones who do market making use execution algos to hedge themselves. That’s a totally different ballpark.”
Schorno said one instance where German corporates do tend to use execution algos is when they have acquisition business and need to hedge large amounts.
But even then there are those corporates who prefer voice dealing. “We have seen corporates, splitting whatever their requirement would be – two to three billion euros against dollars for example –spreading requests over a week by just asking every few minutes for 50 million euros to two or three banks, executing and then doing that again just some time later, always with different banks.”
Schorno said 360T hears similar feedback from Switzerland to what it experiences in Germany, and even from the corporate world in the UK. “I don’t think it’s just this country.”
He adds that clients in Germany are not so focused on being at the cutting edge. “Definitely they are very professional on one hand, but on the other hand their main target is not only to have the best pricing, but also to consider other parameters,” he said. In that respect, relationships to their banks take on a greater role.
“Of course they have a duty and requirement to have best price execution, but this best price definition is different from somebody who runs his own risk under his own position, like a hedge fund. There it is clear what you want.”
Schorno noted that many companies typically would have credit lines and credit commitments with these banks and that they will feel the need to remunerate them with business out of other areas such as FX.
Another detriment to algo adoption is the lack of clarity as to where the FX business may go. For instance, if a large client uses a sophisticated execution algorithm, a portion of the liquidity may end up with various venues, including dark pools, but it won’t necessarily go to the bank that the firms wants to maintain a relationship with.
“I think the drivers for the corporate and real asset managers are different than for banks and hedge funds,” Schorno said. All of that said, the expectation is that algo adoption will continue to increase.
Germany’s lower tier banks are likely to adopt algo usage much more in future, Schorno said. “The trend is definitely going to be more algorithmic. Most of these algos are in-built within pricing engines, to create pricing and hedge in the back end in an algorithmic way,” he said. “The requirement there is definitely clear for more, not exotic, but generally more algorithmic execution.”
Schorno added: “One of the trends in the last one or two years is definitely that the asset manager real money world becomes more aware of the underlying factors that influence FX execution. They are all experts in equity and most of them experts in bonds execution, though not so much in FX.”
PROP TRADING PROSPECTS
“I don’t think it’s that much known, but in Germany, at least on the equities side, there’s a lot of quantitative trading,”
Looking beyond the issue of real money flows and their demand for the products offered by sell-side providers, there is a question as to whether proprietary funds will show demand for algos, either for alpha signal generation or for execution.
Balazs Klemm, a Chartered Financial Analyst, is CIO at Almax Capital, an event-driven fund in Germany. He is focused on equities; while that does involve dealing with some FX exposures, he is quick to point out that his firm relies on brokers rather than algo-based trading.
But while Klemm may not have a direct need for algo trading, he says that the trading scene in Germany is a lot more sophisticated than people may realise, a point that must resonate with sell-side providers looking to get a foothold in this market.
“I don’t think it’s that much known, but in Germany, at least on the equities side, there’s a lot of quantitative trading,” Klemm said. “You don’t hear that much about it in public.”
And while to date that trading has almost exclusively been on the equities side, there are signs that FX is getting more attention.
“We hear about people who are trying to position FX as an independent alpha source, which I personally think is absolutely true because it’s not necessarily correlated to other asset classes and if you do it right I think there can be alpha signals,” Klemm said.
But he added that for all the talk he was not so sure about how much real business was taking place. “As far as my experience goes and what I have seen, there has been more discussion than actual serious allocation to FX as an independent alpha source asset class.”
For his part, Klemm doesn’t see himself branching out into FX, but that’s because the use of news-driven or event-driven algos for finding alpha is more suited to equities trading, where a single piece of information can dominate the trading trend for an extended period.
Pierre Witthus, a Munich-based quantitative researcher for technology provider Haawks, said his company does have some clients in Germany but they are probably not representative of the majority of German market participants. Haawks specialises in complex event processing and plans to release an FX news and sentiment product this year.
“If you compare it to other European countries, there are not so many proprietary traders in Germany,” Witthus said.
So ultimately this all adds up to a glass-half-empty/glass-half-full picture. Germany sees huge real money flows but it has a cautious asset management community. It has some of the most sophisticated financial practitioners in the world, but its politicians have created a climate that does not provide much encouragement for algo adoption. For the providers looking to tap into this market, it may be a long, hard campaign. But that doesn’t mean it won’t be a successful one.