A Closer Look At: A Differentiated Algo Offering

Mauricio Sada-Paz, global head of e-FICC product and distribution at Barclays, discusses new algorithmic products available within the bank’s e-FX platform.

Profit & Loss: Barclays’ recently added a new algo, BARX Peg, to its e-FX platform. Can you explain the genesis of this product launch?

Mauricio Sada-Paz: The Barclays’ Gator execution channel was a pioneering product when it came out because it gave clients the ability to access liquidity available on a number of external venues using a smart order router, with the objective of intelligently looking at where it could get liquidity more quickly and efficiently with less rejects and more fills. Further, one of the main features of the Gator product was the ability to combine externally sourced liquidity with BARX principal liquidity, BARX principal liquidity hence contributing to the net liquidity available on Gator. Gator was extremely popular, it had its own keyboard, and was used by a lot of bank traders and hedge funds.

However, while BARX liquidity was a component of Gator from the outset, Gator did not, prior to the launch of BARX Peg, have an algo that interacted with our franchise liquidity in a manner that enabled trades to be filled entirely through the process of internalisation. FX market making from a principal perspective is a business that’s very important to us and we provide clients liquidity in this way via our single-dealer platform, multidealer platforms and FIX connections, but we also think that the Gator channel, typically used for hybrid agency trading, is extremely important as a means of accessing our franchise liquidity pools and that it will continue to grow at a fast rate.

That’s why we launched the BARX Peg algo, which taps into demand from our real money and corporate clients for an algo that seeks to minimise the amount of spread paid, that has less signalling risk and interacts with our unique franchise liquidity.

P&L: What is driving this demand for the Gator platform amongst clients?

MSP: Clients on the buy side are now willing to take execution risk to save money on execution. Algos allow them to access multiple pools of liquidity, including some of the bank’s internal liquidity pools which look to provide liquidity through attracting offsetting client flow rather than through access to external venues. This allows clients to be more passive for large transactions and attempt to beat the risk transfer price they would have gotten had they transacted through the e-desk or voice desk on a risk transfer basis.

P&L: So you see demand from corporates and real money clients for FX algos growing?

MSP: Absolutely. Five years ago corporates, and in particular real money clients, still did a lot of transactions by voice. But changes in both the FX e-trading marketplace as well as the regulatory landscape in recent years have helped push the electronification of FX flows and once firms start trading on e-platforms they naturally look to algos to save on their execution costs. A lot of buy side firms have organised their trading desks to have lowtouch and high-touch execution and they try to automate as much of their workflow as possible.

P&L: Given that all the major banks in the FX market now offer algos to their clients, is there still room for differentiation in this space?

MSP: Yes – the first thing that’s unique about the BARX Peg algo is that it interacts with the Barclays franchise liquidity in a manner that allows trades to be filled entirely through the process of internalisation – 35% of the offsetting client flow for this process comes through the single-dealer platform and is very benign – that is itself unique. The other interesting thing we’ve done with BARX Peg is that we’ve divided the liquidity pools of the Barclays franchise into five different sections, depending on the source of liquidity: the GUI BARX clickers, the multidealer venues that are full amount, the multidealer venues where there’s sweeping, the FIX connections and then finally the clients where there’s more clustering.

This means that the client has the choice of moving between those five liquidity pools, accessing all the liquidity from these subsets of our franchise liquidity.

P&L: Can they move between these liquidity pools while the algo is in-flight?

MSP: Yes, it can be changed in-flight, so if it’s going too slow it can be moved up. It’s basically deciding who you want to expose that axe to.

P&L: So the key aim here is reducing market impact?

MSP: The aim is to reduce market impact, reduce signalling risk and minimise the amount of spread paid. From a Barclays perspective, we earn a pre-agreed fee for the use of the Gator channel; this is built into the settlement price for the trades executed.

P&L: Market impact and signalling risk have been getting a lot of attention lately. Is this a new concern amongst the buy side or is it just getting vocalised in a more public way now?

MSP: Clients are increasingly hiring data scientists, they’re either building their own TCA or working with third-party TCA providers, they’re looking very carefully at market impact and I think they’ve realised that gone are the days of having 15 liquidity providers in your aggregated pool to sweep.

They’re also realising that not all liquidity providers are actually looking to manage risk through offsetting client flow and internalising, some are just recycling liquidity which might cause the other entities in your liquidity pool to widen their prices and the liquidity available to worsen.

The clients that still sweep are analysing their counterparties very carefully to see who is internalising risk and who is not, who is creating market impact and who is not, and then reducing the number of counterparties that they trade with accordingly. Additionally, some are moving from sweeping to full amount streams, which we do encourage.

In going through this exercise, clients have realised that they should have a tool in their arsenal that offers an alternative to going through the principal desk for a risk transfer, where this could potentially have market impact and give rise to signalling risk.

P&L: So looking ahead what’s the focus for you now on the algo side of your e-FX business?

MSP: Well, we recently launched a participation algo so we’ll be continuing to educate clients about that as well. This algo looks similar to the VWAPS you get in the equities markets because it looks at the estimated liquidity volumes in the interbank market and then gives clients the opportunity to target participation rates ranging from 5% of the market up to 25% of such market. The order duration will then adjust to how volumes are trading at the moment and how they evolve throughout the life of the order. This algo does primarily target external venues but it is an interesting option for some clients because, rather than spacing out their order over time, you space it as a targeted percentage of the volumes going through the interbank market, which comprises a blend of selected external liquidity venues.

In fact, just helping our clients understand the range and value of the different algos that we offer is an important focus. We have to make sure that our sales force has the best tools available to be liquidity consultants for our clients. Ten to 15 years ago, e-sales people were focusing on getting clients connected to the single dealer platforms, it was more like traditional platform sales.

That role has evolved to one of educating clients on the range of liquidity tools available to them, enabling them to assess what’s the best time to execute, what’s the best algo to use, whether they should be targeting passive fills, whether they should pull back and change their limit. Today, the e-trading process is a lot more interactive, so we have to make sure that the internal sales people have the same tools that are available to clients and that they’re are just as sophisticated as our clients are becoming from an execution and liquidity point of view.

Galen Stops

Share This

Share on facebook
Facebook
Share on google
Google+
Share on twitter
Twitter
Share on linkedin
LinkedIn
Share on reddit
Reddit

Related Posts in ,