Ian Daniels, executive director, head of e-FX distribution, EMEA, at Nomura, talks about algorithmic trading trends in the FX market.
Profit & Loss: Since you joined Nomura, you’ve been working on developing the bank’s algo offering, what are the latest developments there?
Ian Daniels: Our algos are now live on Bloomberg and will soon be live on a number of other major third party venues. I think that one of the benefits of being a later entrant into this space is that you have the experience of previous roll-outs to draw upon. So we knew that access to liquidity and the customisability of our algos would be important characteristics for our clients and that’s why we created algos that enable clients to execute in a default mode or that can be tailored to meet their specific needs.
P&L: So clients will be able to change how the algo operates once it’s already in-flight?
ID: Yes, clients have flexibility to change parameters to enhance their experience, which I think is key. Needing to prove best execution is very important for real money clients, especially since the introduction of Mifid II. By using algos, they can look back after the trade and then use that post-trade information to inform their pre-trade decision next time by looking at how the algo performed to determine what fashion they want to use it in. The desk then has evidence that it’s performing well, can show compliance that they’re adding value and can go to the end client and show the algo execution against a given benchmark.
P&L: You mentioned a late mover advantage, what specifically were you able to incorporate into your algos as a result of this?
ID: Well it’s clear how important it is to have full TCA upon execution showing a full range of benchmarks and getting that information out to clients quickly.
Another thing we’ve observed is that in the past, people thought that you needed lots of different types of algos, but there’s been a realisation now that most clients don’t need thousands of algos, they tend to just focus on using a few.
Then it’s also about providing the flexibility that I mentioned before, people really like that they’re able to use the default mode or interact with the algo themselves to add certain parameters.
P&L: What kind of algos do clients want? In most cases are they really looking for something sophisticated or are they happy with a simple TWAP or VWAP?
ID: I think people are looking for more sophisticated algos that interact with liquidity and with the liquidity you have. Yes, you can have a simple TWAP, but the question is then: how does that interact with the market? Are you able to internalise? What’s the market impact of that algo?
It obviously depends on each particular client, but really the algos need to be simple and easy to use and then have a layer of complexity on top of that which a client can utilise. Simple TWAPs and VWAPs are very popular and are often the go-to choice initially, but I think that people tend to want to use something that’s going to lessen their market impact and allow them to interact with the right liquidity at the right time, and to do this effectively, they might need that layer of additional complexity.
P&L: Can you go into a little more detail about what you mean by interacting with the right liquidity at the right time?
ID: Of course. We have a lot of discussion with independent TCA providers about benchmark performance and they talk about how the access to liquidity that algo providers has is then utilised by their algos.
So when I refer to interacting with the right liquidity at the right time, I was talking about whether the liquidity provider has got good access to ECN “X” and does ECN “X” cause market impact. It’s about digging down into deeper detail regarding the performance of the algo, which I think has become very important.
P&L: And what do you view as the key drivers for algo adoption amongst clients?
ID: There’s a mixture of drivers. The ever-increasing electronification of the market is one. For some people, this driver could be purely workflow related as they get better cost savings with more electronic execution. Regulation is another. This has pushed more people to look at their execution and, even if the price is good, question instead whether the fill was good. It may well be that even with the cost of an algo, that the overall performance is better and therefore it is more cost-effective to use one. The adoption of algos is also a result of the changing market dynamic. Liquidity has changed over the years and being able to access the various pools of liquidity as a client is important.
P&L: The last point about liquidity changing seems to be a constant refrain in the market. Do you see the reduction in average trade size as another driver of algo adoption?
ID: I think it is. It’s become harder to get a large ticket done, the market takes much longer to absorb liquidity and that tends to lend itself to more algo adoption. When someone wants to do a large size trade and they don’t want to get rejected above a certain amount they either do a full amount risk transfer or use an algo, so the latter becomes another important tool in the toolkit for these firms.
P&L: So as a result of this adoption trend, do you see buy side firms, either willingly or reluctantly, taking on more risk themselves?
ID: To an extent, yes. I think that as a bank and an algo provider you need to be cognisant of the fact that you are effectively transferring the risk away and back to the customer to take care of, so you need to be fully confident that they’re happy with those risks and it’s explained in detail how these algos work. The good thing about this is that it does end up creating a good dialogue between the bank and the customer to help build their understanding of how various algos work. We’ve found that clients are keen on the consultative approach on how to use the algos and as they talk about their experiences and we talk about ours, it becomes a good two-way dialogue that helps deepen relationships.
P&L: So do you see clients becoming more sophisticated in their understanding of algos then?
ID: As algo adoption continues to increase, clients are continuing to understand them better. It’s not a rapid upward slope, but the understanding will definitely increase further because, from a compliance perspective, the users of these algos need to know how they work and are interacting with the market. So it’s incumbent upon both the algo user and provider to keep having these dialogues.