The more FX markets move to electronic trading, the more algos are deployed. And while any discussions of algorithmic trading continue to be dominated by latency, at this point speed has become mostly commoditized.
“Everybody is always trying to get faster, and closer to the physical limits that any given infrastructure can offer them,” said James Dalton, Global Head of Algorithmic Execution at CitiFX.
The innovation is that we start to see algos on all these derivative products...
Six or seven years ago, he added, there was a gap between many sell-side firms market connectivity compared to the ultra-high frequency trading community, but that gap has narrowed: “There has been a lot of convergence in terms of capabilities there.”
In the FX markets, Quant Hedge uses high and medium frequency trading strategies for the most liquid pairs. On the research side, the firm is looking at artificial intelligence applied to capital markets for price forecasting.
Victor Lebreton, Director at Quant Hedge, said that, alongside spot FX, he sees a surge in demand for derivatives - NDFs and futures - to hedge an uncertain economic environment.
“The innovation is that we start to see algos on all these derivative products. On FX futures there is already lots of algos running, but on swaps electronification there is some opportunity to accelerate the development of new tools, and algos on these tools, making them more opportunistic and bringing Liquidity in the market,” Lebreton added.
Multi-Product, Multi-Asset
Each product develops a bit differently from the spot market, he explained. For example, NDFs are becoming extremely popular to hedge emerging markets currency positions.
“If you are doing a deal and there’s a lot of volatility in the market, you want to be hedged. So it’s not only a single trade but also the overall portfolio,” he said.
People who used to trade primarily Forex , (now) arbitrage against futures, against equity...
There is a question mark over the rate of adoption, however: “There is a lot of change for FX people who use it to hedge their long-term and short-term positions.”
One specific trend Lebreton pointed to is outsourcing. In the market, he finds himself trading with an increasing number of independent execution providers using the same kind of algos.
“Most people don’t want to buy tools and software directly, only large banks and (asset managers) are doing that. Most FX managers don’t want to manage the technology and personnel. Also, they can push transaction cost analysis towards the broker,” he said.
About four years ago, cloud provider Beeks Financial started providing low latency for retail traders running algos on popular platforms such as MT4. They’ve since moved into the institutional space for FX and futures, hosting for some 100 funds and brokers.
CEO Gordon McArthur said that a lot of the bigger hedge funds, money and asset managers are moving into multi-asset class trading: “People who used to trade primarily forex, (now) arbitrage against futures, against equity…that is the biggest change we see at the moment.”
That’s partially because as single asset classes or futures markets mature, latency arbitrage opportunities become fewer and farther between. So algo funds are checking out new markets.
Source: Bank for International Settlements, January 2016
FX Into the Cloud
There’s also fewer barriers to entry. McArthur said setting up in data centres has become cheaper, while testing algorithms or systems on different venues is easier.
Although Beeks’ clients tended to be on the buy side, McArthur noted that more sell side is coming into the cloud. Security remains a big hurdle to overcome, particularly in the wake of a number of Denial-of-Service (DDoS) attacks late last year and Beeks caters to these concerns with a specialized service, he added.
Most people don’t want to buy tools and software directly...
Still, the low cost is trumping concerns in some aspects. “We are seeing people take the first steps in using cloud environments to do high performance computing – taking a pile of data, and back test data, running some scenarios on it,” he explained.
It’s a product Beeks is keen to stay on top of, but there’s still a lot of buzz words around next generation technology such as machine learning.
Portware’s Head of FX Trading Services, Chris Matsko, said that artificial intelligence developments in equities have already gathered significant momentum. A number of clients are using it to make decisions between different algos – how aggressively to trade, for example, given market conditions.
“We can take that and we can apply it to our algo suite for FX, however, I don’t think that the community is quite ready for that yet, given that institutional FX, or at least tier one global asset managers, just don’t have that appetite just yet,” he said.
