A Spotlight on Automation and Algos

OK: What advice would you give to anyone starring their automation journey?

TT: Historically, if you trade directly with bank counterparties, you are transferring all your risk onto that sales desk; the question you have to ask yourself as you automate is; are you willing to take this execution risk back onto your own firm? For me, I was very happy to take on this risk, and
once we took ownership of this, we looked at the different avenues we have to trade, such as RFS (Request for Stream), streaming prices and algos. Finally, it is about the way your accounts are set up; the way we initially were staging trades and trading was very one dimensional. With a more
flexible approach to accounts, and approaching with multiple counterparties per accounts who we can trade with has enabled us to further enhance our automation capabilities.

OK: What’s your main motivation for increasing automation levels on the desk?

TT: Workflow efficiency without giving up best execution is key. It is so hard to find best execution through any TCA means if you are trading voice, as you have no idea where the risk is cleared, as the bank doesn’t provide that information unless traded via algo, therefore to me there is still an
argument today that everything should be traded through an algo just so you have the added transparency. Through algos, you know exactly where each fill was cleared if you have competent TCA and counterparties have to provide this information to clients. It is much more difficult to get
timestamps if you are trading via chat functions, streams, or RFS for example. Trading electronically especially via algos means you can have proper time stamps of every child fill along with the venue, and proper benchmarks in order to assess best execution. Once we were satisfied
that we could get the same or better execution without having to wait for the sell side sales person to give fills, we quickly realised voice was inefficient vs electronic, and once we understood which algos worked in which manner, it made it virtually impossible for us to go back to voice trading as the main method of execution.

OK: What are your top tips for buy side when looking to select an FX EMS?

TT: For those who have managed accounts, the allocation element becomes key. The most important thing when assessing an EMS is looking at how it integrates into your front to back-office workflow. Automation should encompass trade origination, trading, clearing and settlement, and
an EMS should consider all of these functions. If you take trading in isolation, you do not get the full picture, whereas if you work closely with your operations and tech teams, you can assess which platform is best suited for the entire life cycle of a trade from front to back. Think about what can help the firm, not just what helps the execution desk- it is about front to back automation, and if the entire life cycle of a trade cant be automated with an EMS, you are probably going down the
wrong path.

OK: How has increased automation on the desk changed your role? Have you found the experience to be positive or negative?

TT: Automation is without doubt a positive. In the future, you are still going to need FX and asset specific experts and experienced traders, but the set-up is going to be different. A trader will need to understand automation workflows, and the trader in future will be similar to a pilot in a cockpit,
monitoring to make sure all the trades are being executed the way they’re supposed to, but via automation through algos/API streams, and manually stepping in when needed, be it an algo that is not performing well, a trade that is better suited for voice, or an issue with the EMS. Having an experienced trader that can tailor the automation to your firms objectives, along with having the necessary skill set to trade manually should be a prerequisite for all firms. Simply filling the seat with an inexperienced trader that doesn’t possess the proper skills necessary to trade manually
would be a mistake.

OK: Is your FX algo usage increasing, and why?

TT: I feel, algo usage should always be increasing. Increased algos = increased data access, so it is a cycle in which more data is received which in turn should help tailor better automation. Of course, there is a fee to use agency algos, so sometimes a stream can be more cost effective.
Depending on how many counterparties can show axes on their screen, you can sometimes get better execution. For us It is size dependent and limiting the amount of trades you do, netting your managed accounts down, and once we start to net and do fewer trades, it then makes sense to
incorporate more algos.

OK: The average number of algo providers used amongst our FX community is 7. What is your algo selection rationale, and what do you think is the optimum number of algo partners to work with?

TT: It is all dependent on your account set up. Some funds could have 10-12 counterparties, which we do in some cases; but even when you do RFS, over the long haul there always tends to be 4-5 counterparties that come out on top. This is how it will work out with algo providers too, where you may work with 7-10 providers, but for the most part you will likely become most
comfortable with 5-7 of these. It’s about flexibility, having different accounts and options available for different types of trades and different times of day. Ultimately, there isn’t a massive difference
between the various bank algos, and it is more about the underlying venues where they are cleared. Now, there are some banks that are comfortable with you using their algo technology, but allow you to select the execution venue, this gives users the choice to access the liquidity they want for a certain trade, while still rewarding the banks for utilizing their technology.

OK: How can the buy and sell side work together to overcome integration, solution differentiation and transparency issues with algos?

TT: I don’t see transparency issues; when you trade an algo, they provide every fill when you have traded so this is as transparent as trading gets.

There is a disconnect in the ecosystem, and for each industry group there is a different motivation. The banks went through a time where they were pushing their single dealer platform, meaning they would not have to pay fees to an EMS. The market has now evolved towards an EMS as this
technology has become essential. The EMS’ sit in the middle and are a necessity in the market, and if the sell side are smart, they can motivate you and help you select the right EMS. On the buy side, you have so many different voices and opinions telling you what to do, which makes it challenging as to how the buy side and the sell side can work together, as you need to involve the EMS providers in this conversation. The buy side have a difficult job where you have so many different opinions. It is challenging to say how the buy side and sell side can work together, as it should also encompass an EMS in the conversation.

Once the buy side move on something, the sell side follow; it will take an understanding to find out what is mutually beneficial for all parties. The buy side need to ensure banks are still motivated to give the best business and not just compared on who gives the lowest pricing- you want to ensure you are paying for quality which in turn will incentivise your counterparties to continue to innovate. People need to realise that there is only a handful of clients that trade FX as an alpha generation asset class. For the others where FX can be viewed as a burden, they should look at automation here first. FX can be viewed as low value, so they need experts to come up with automated solutions here, where they trade enough FX so it makes sense in size and execution becomes meaningful.