P&L Report Card
T his has become a very competitive field over the past couple of years and is showing little or no sign of slowing down, for just about every major FX bank has a suite of algo execution tools available for clients.
Achieving higher adoption rates remains something of a problem, however it is noticeable that the advent of third party mechanisms to analyse execution quality does appear to be giving clients some confidence in using algos – this is important if the impressive budget outlay is to be rewarded.
The big change over the past year in terms of clients’ attitudes to algos has been the desire to interact more with internal pools of liquidity – in short, the pure agency model has appeal, but not as much as something where the client can make decisions and have control.
This is interesting when it comes to the long time leader in this field – Credit Suisse and its AES (Advanced Execution Services) business, which has generally focused on accessing the best external liquidity that has included, importantly, other banks’ streams.
AES remains a very popular business but if there was one bank that seems to have been hit harder than the others by MiFID II it is Credit Suisse, which, with the agency model, undertook a heavy workload to meet the requirements necessary to meet the demands of the regulation. AES remains a popular business, however, and client satisfaction remains high, the bank has also enhanced aspects of the service, although in reality the calendar has worked against it this year as many of the enhancements are not due for release until later in 2018. The bank will remain a serious player and the enhancements will only make it a tougher competitor in years to some.
JP Morgan has also made great strides as we noted earlier and we would refer readers to the Best Execution section for a description of what makes the bank’s service so strong. The advances made by Deutsche Bank are also significant, any time the bank makes a move it reverberates around the industry.
Goldman Sachs is in the midst of reengineering its algo suite of products with much of the work likely to be available in the middle of this year, including adaptive algos, and as we have noted, Barclays has rolled out a new strategy as indeed has State Street. Elsewhere, the landscape remains unchanged – banks have a relatively small set of strategies that they can offer clients and the latter seem content with that.
In the past we have predicted more bespoke work on algorithms for more sophisticated clients, however we suspect that, not for the first time, we maybe wrong. More likely is the model pioneered by JP Morgan this year that allows the user to create a hybrid strategy using one or more of the bank’s existing strategies.
There seems little doubt that the basic TWAP strategy will remain the most popular for the near term, if that is the case then we believe it is important that banks have strategies that allow access to internal liquidity pools.
Winner – BNP Paribas
The overriding reason we have given this award to BNP Paribas is that the bank has indeed added a strategy that allows internalisation and this reduces market impact. In the past, the bank has been a close rival of AES – it continues to be, but unlike its Swiss rival it has had the budget available to not only spend time on regulatory matters, but also on services and functionality for the client.
The last year has seen BNP deliver more client specific controls around its algos, studying exactly what it was clients need and then codifying that in. This means that more have execution logic built into the algos specific to them, something that satisfies the all important compliance function at the client end.
The bank has also enhanced the STP aspect of its algo suite, allowing clients to launch on a third party application, manage the algo using BNP functionality and then have the end trade delivered into their own OMS.
The guided TCA tool on Cortex iX is simply superb with different layers of protection around the execution, such as maximum slippage, volatility barriers and limits for the “all in” rate. Clients can also look across a portfolio of trades when providing TCA (especially in the pre-trade). Real time TCA is provided so the user can monitor how the execution is progressing compared to the predicted path and the alerts functionality around the execution has been further enhanced.
BNP offers the flexibility for users to interact with the order as it progresses – as we have noted before this is important for those skilled execution teams who want to be able to leverage their own experience and knowledge – it also, importantly in this compliance obsessed age, registers every interaction with the algo.
The parameterisation of the algos, as well as their flexibility means the BNP suite of products is as close to replicating how a skilled human trader would act as possible. There is a “get me out” button that automatically provides the BNP Paribas risk transfer price in the balance of the order – an important step into the principal field by an agency business (with the principal/agency border being clearly defined). If the client doesn’t like the risk transfer price another click sees the algo resume.
There is also an internal match algo that only matches against other BNP algos that are operating in the market and clients can set parameters around the “all in” price they want, rather than just a fixed market level at which they want the algo to pause or stop. The bank has also built in Iceberg detection logic, something that will be helpful for clients as more and more participants use the Iceberg strategy on public platforms in an effort to hide their full interest.
The Execution Cockpit is the canister hosting the suite of products, it also includes the bank’s excellent pre-trade decision making tools with their drag and drop functionality and it is easy for clients to conduct post-trade operations such as rolls, splits and allocations in a more efficient fashion. The post-trade service also offers monthly reports that highlight trends in execution costs and, importantly, provides recommendations for adjusting execution policies that are tailored to the specifics of each client.
The data mining capabilities in the pretrade analytics are what really strikes the observer though for the bank has borrowed from its options business – indeed any options business – in providing “solve for” functionality. Under this guided pre-trade TCA model a client can enter an order for execution and then ask questions such as how long should the execution go for? What should the price constraints be? What is the benchmark to be used?
Again drag and drop functionality means the order can be moved easily to assess the likely outcome under different parameters and strategies, the platform automatically recalibrates the expected outcome while the order is being moved.
One final factor in BNP’s favour in this very competitive category is how it has transformed the post-trade TCA aspect of its service. The bank was one of the first to go with BestX for TCA reports, which left it with its own data and the search for some way for customers to use it.
The solution was BNP’s Liquidity Lab, which delivers real in-depth analysis of a client’s execution activity over different time horizons and in a totally empirical fashion. The analysis can measure the impact of a particular strategy over time and compare it to different execution methods available, thus really informing the client of where the potential shortfalls are in their strategies and how they can tweak for improvements.
This repackaged TCA is a great example of the different thinking that has allowed BNP Paribas to establish such a strong position in the algo world. Competition is undoubtedly growing, but the bank continues to roll just ahead of the curve – it is a go-to destination for algo execution.