Goldman Sach’s David Wilkins, global head of e-FX sales, and Ralf Donner, head of client FX algo execution, talk about a new algo launch that they claim offers a new way to trade FX.
Profit & Loss: So tell us about the new algo that you recently made available to your clients.
David Wilkins: It’s a Basket Algo which gives clients the ability to take numerous trades and then trade them as a basket. The advantages of this for clients are threefold: it offers increased automation, cost savings and greater control of your processes.
P&L: So can you break down each of these in more detail please….
DW: Sure. Looking at our client’s activity we could see individual trades coming through, but when you delve a bit deeper and get a better understanding of how their workflow operates you find out that, actually, at the starting point of these trades there’s normally something that looks like a basket.
And this is not just the case with asset managers, which was where we saw the obvious demand for this type of algo initially, but also with corporates, hedge funds and even banks. Any firm that has a large number of FX trades coming into them tends to split them down into their component pieces, but actually in the beginning it was probably something that was, or resembled,a basket.
And so with this algo we’re helping our clients automate these processes because they potentially don’t have to split up all those trades into their sub-components in order to trade them at different times, withall the workflow constraints that entails. They can instead do them as a basket.
The cost savings can be achieved because the algo offers genuine netting opportunities. We’re offering the ability to run dynamic hybrid algos on each leg of these trades and hence you get the benefits of internalisation and netting that come with these smart algos that you wouldn’t otherwise get if you were trading all of those legs individually.
Then it offers more control because clients have the ability to set limits that are applied to the whole basket and you can set controls such as “no worse than” that are applied to the whole basket rather than having to toggle between the various different legs of the basket. We also give
clients the ability to view the whole basket through our real-time order monitor so they have a real understanding during the execution of the basket and what all the legs are doing.
Not to overstate things, but we think that this is an entirely new way to trade FX.
P&L: Is this algo available internally?
Ralf Donner: It is, yes. We have plenty of internal use cases for it at our various desks around the world. For example, it also gives you the ability to do certain standardised baskets, like a DXY or an RMB basket. You can save down a basket which you typically trade and then load it up and adjust the scale and the various legs for the basket as needed and then execute the entire thing. So again, this comes back to the automation point.
P&L: So have you noticed the costs savings you’re touting internally?
RD: Yes – the cost savings come about in a very programmatic, formulaic way, so we generally know in advance what the cost savings are going to be.
Here’s a good example: if you take a case of a G10 basket that has Scandis in it – let’s say EUR/USD and USD/SEK and USD/NOK – and you’re both a buyer of EUR/USD and USD/SEK then you’re actually a net buyer and seller of EUR/USD in this basket. So this leads to a natural netting opportunity on the synthetic component. If you then have a basket that does have both a buy and sell in a given currency pair there would be a direct netting opportunity.
P&L: Were you surprised about the potentially broad applicability of this algo product beyond the asset management client segment?
RD: Well I think that the main point here is that any time that you have a model, like an equities model or a fixed income model, as it runs and rebalances it requires an adjustment to various FX positions. Then it naturally starts to look like a basket that needs to be executed.
One of the interesting things that we are discovering, and that post-trade TCA providers are beginning to recognise as well, is that traditional metrics for trade analysis are not as relevant any longer. If you imagine looking at a single execution and then looking at your performance
versus arrival price or an estimated risk transfer on that execution, it’s actually of limited value if you’re looking at the portfolio level. So new metrics are actually needed to be able to analyse the performance of these baskets.
P&L: Then are you working on developing such metrics for clients?
RD: Yes, we are looking at it ourselves. The most obvious ones are an indicator of what is the slippage that you have on the basket as a whole and what’s your progress through the basket looking at live liquidity on the various legs. So there’s a real-time order monitor aspect to it and then for post-trade TCA we will look to work with third-party providers as well as offering our own analysis.
Where I hope this ends up is in a situation where it becomes apparent what the cost is to clients of not using a Basket Algo. That should be the other thing that naturally emerges from these new metrics.
P&L: As you go through the process of developing and deploying an algo like this, what are the biggest challenges on your side?
RD: The biggest challenge by far is that this algo is very complicated on the back end, there’s no doubt that this has been our most complex algo product in FX to-date. The ironic thing is that baskets are so natural to FX because it’s the one asset class where you can trade synthetics, you can trade crosses, you can really break down legs into components and execute them. So on the surface it seems natural that a basket algo product should have existed for a long time – but the reality is that it’s so challenging to do that it’s taken us almost a year of work to golive with this product.
DW: I would add that part of the challenge is taking something with this level of complexity and then turning it into something that is very simple for the clients in terms of functionality and the ability to fit it into their workflow.
P&L: People have been talking about increased client adoption of algos for FX execution for some time now, with the starting point for many firms being the more basic algos that are widely available. Do you think that more firms are now getting to the point where they’re looking for something a little more complex on the algo side?
DW: Clients now have a much better understanding of the algos in and of themselves. What they’re looking for are algos that solve the real world problems that they have, and that junction hasn’t been made particularly well yet. For years the focus has been “here’s an algo to transact that ticket”, and I think that we’re moving to the next stage in the evolution because we’re expanding the boundaries of what they can do with the algo in the sense of doing multiple tickets whilst also fitting within their workflow.
RD: The other thing I would say about complexity is that there was a stage in algo development where it became quite popular to expose almost every feature of the algo’s functionality to clients to be able to modify. I think that complexity for complexity’s sake is no longer the way that people build an algo product, it’s only justified if it solves a problem. Even with this Basket Algo, we are counting the number of clicks that it takes a client to launch a ticket with the aim of making it as straightforward as possible.
P&L: So what’s next on the algo roadmap for you then?
RD: Well firstly, I would say that we’re by no means done with the development of our NDF algo that we launched previously. There are more currency pairs that are of interest, more types of algos and different liquidity sources that we’re looking at, while we also think synthetics for NDFs would be of interest. So there’s actually quite a long shopping list there right now, we’re really just in the beginning phase of development there.
The other thing that I think is very interesting to look at is the world of options trading. Options trading generates a lot of spot business, whether it’s through execution of delta or through managing a gamma strategy, so there’s lots we can potentially do there.
DW: I also think an area of focus that will lead to the expansion of the use of these algos is continuing to integrate into client workflows. It shouldn’t be underestimated how important this is – the more convenient and easy that you can make it for your clients to use these tools, the more likely they are to do so.