With increased interest in financial benchmarks generally, Colin Lambert talks to Jamie Walton, co-founder of Raidne, about Siren, a new FX benchmark that seeks to make things fairer for both buy and sell side.
Colin Lambert: What is Siren?
Jamie Walton: It’s an idea we had after a meeting with the Bank of England in 2018. The subject of an alternative FX benchmark came up – obviously there was the lingering scandal over manipulation of existing benchmarks and the MiFID benchmark rules were coming into existence – so we started looking at creating a benchmark fix that was both harder to manipulate and targets a fairer outcome for everyone.
The idea for Siren actually came from equities markets, where, if you want to trade the close and you don’t want to go through the auction process, you use what is called a target close algorithm. What that does is spread your trades out over a period of time leading up to the close to reduce market impact, and the volume of buying increases the nearer you get to the close. It’s a well-known algo in the equities space and we wanted to see if it could work in FX.
We have designed a benchmark that reflects how the order would be executed normally in the market – if it does that then it should be harder to manipulate.
CL: How does it work?
JW: We use a mathematical idea, which is optimal execution. Basically, you balance the market impact that is inevitable when you trade – and the longer you trade the less impact you are likely to have – against the risk you are exposed to by executing over a longer period of time.
From a practical perspective what we are saying is, ‘if you are going to pre-hedge a benchmark, then this is the way to do it’. If you take the Siren 4pm Fix as an example, you will trade ahead of time, gradually increasing your volume of trades as you approach 4pm.
We think this makes it harder to manipulate. If you try to buy aggressively at the start of the window you are exposing yourself to much greater market risk and would actually have less influence on the rate set.
CL: So this is effectively including that period of pre-hedging in the benchmark calculation?
JW: Yes. If you tried to trade large amounts in the five-minute window you are going to have market impact no matter how well you do it, everybody understands that. This way we can deliver optimal execution that provides lower market impact for a reasonable amount of market risk.
CL: You have a 20-minute window for calculation – how did you arrive at that?
JW: We ran a study with a pension fund’s trades, it involved over 3,000 fixing trades over three years, and looked for the sweet spot for reducing that market impact whilst not having too long a window. From an ideal point of view the longer the window the better of course, but we sat down with several of the largest funds last year and presented them with our findings and their feedback was they would prefer 20 minutes – so 20 minutes it is. We looked at fixing trades of all sizes, from under a million to hundreds of millions and this was the most practical solution that didn’t expose managers to higher risk.
We published a white paper that looks at a variety of windows, so from a designer point of view the window can be any length, but all of the tests conducted so far have indicated that Siren produces better outcomes, something like 60 dollars per million better than WMR.
Our analysis shows that most funds are 60-70% correlated to the markets, especially equity funds and that means they will be the same way with their fixing trades and will have market impact in a five-minute window. By executing over 20 minutes the correlations can be diluted and while obviously the market impact is higher at month and quarter ends, our analysis also shows it exists on a daily basis as well.
CL: How many benchmarks are you publishing and where does the all-important data come from?
JW: We are calculating and publishing in 71 currency pairs every 30 minutes from 10pm UK time on Sunday night to 10pm Friday night in the UK so we cover the entire trading week. The data is coming from our partnership with New Change FX (NCFX) – we have worked with them on our surveillance solution – Siren is basically our IP and NCFX’s data. NCFX calculate and publish independent mid-rate FX Benchmarks every 50 milliseconds and we take a one second snapshot over the 20-minute window, weight each point and calculate the Siren benchmark.
The underlying source data is available from NCFX and we will provide a live stream of each data point so that firms using the benchmark can see their tracking error in real time. We are live with Siren and the benchmark is published within a second of the fixing window closing, but even now people don’t have to wait to see their slippage. It’s a very transparent solution and that is proving popular with both buy and sell side. The buy side have the desired transparency and better execution, while the sell side can also benefit from the reduced tracking error.
CL: You recently also gained regulatory recognition for Siren as well.
JW: Yes, once the benchmark methodology was complete the next thing was to get regulatory authorisation, this we achieved in September 2019 – Siren is regulated by the UK’s FCA under its benchmark regime – and I think it’s important to point out that every one of our 48 daily benchmarks for each spot pair is regulated, it is not universally understood that only the 4pm WMR benchmark has regulated status.
CL: I would imagine that is quite important because awareness of market behaviour around the 4pm Fix is growing, people are aware that there is signalling risk and this would – you would think – prompt them to trade at other times?
JW: We hope so – we know we have a battle for hearts and minds to get people to look at an alternative benchmark. It’s all very well having that alternative and having it regulated, but actually getting people to use it is another thing, especially as so many managers currently reference WMR.
That said, we have been meeting with lots of asset managers and we have had firms say they will move to Siren if we can give them an executable benchmark and we now have several banks that have agreed to be an execution partner, which is a big step forward. Other banks are looking to come on board soon, it’s just a question of working through their internal processes.
We are also talking to index providers including RIMES, an independent RegTech specialist in benchmarking, it not so much about getting them to drop WMR and go to Siren as such, more it is about allowing managers to reference it if they wish, alongside any existing solution. We think our results speak for themselves so managers will want to reference and use Siren, but we know it will not be a quick or simple process. If we can get some of the bigger funds using Siren, however, and they do appreciate its benefits, then that will hopefully prove to be a tipping point for us. It’s going to be a challenge, but I am optimistic the data provides us with a strong selling point.
CL: What about opportunities for netting ahead of the windows? It looks to me like that would provide an extra benefit to managers?
JW: We certainly hope to introduce netting functionality to our process and we are talking with our preferred partner Siege FX who are working on just such a solution. A netting utility process ahead of a longer window can only improve execution outcomes because market impact will likely be reduced further, which will encourage people to post their interest because they can have confidence the benchmark is not going to be manipulated.
CL: Ultimately though, as you have hinted at, it will be the execution outcomes argument that is likely to win the day.
JW: Anyone who looks seriously at best execution knows that if there is a better way to execute then they should consider it – the fact that the benchmark is regulated and fully transparent in both methodology and underlying data is providing more comfort. Best execution is a process, but it is also a complex concept, especially if you try to judge signalling risk. We think Siren offers lower signalling risk and market impact and greater transparency and that is, hopefully, a very compelling argument for asset managers.