Why are dark pools of liquidity becoming more meaningful for algo FX execution teams?
The short answer is that we are seeing increased participation, a surge in matching success and trade volumes, so the flywheel is accelerating. With match rates in excess of 30%, the Siege MidPool is now a meaningful source of liquidity and value. What’s even more exciting is that we have a healthy pipeline of both Sell-side algo providers and Buy-side traders who want to join to rest orders in the dark. In a market that has seen periods of substantial spread widening with increased spot volume on all platforms, dark matching in the MidPool has delivered great value for existing participants. Especially in EM and “Scandies” big cost savings were realised for our participants through spread savings and reduced market impact.
Why is mid-matching in dark pools a great way to mitigate market impact in FX trading?
The problem with trading larger amounts in the ‘lit’ markets is that your activity is a signal to other traders: some of whom will get out of your way while others will jump ahead of you. This signalling information is distributed through the ‘lit’ market data feeds used by liquidity providers, other algos and of course high frequency traders. Importantly Siege does not sell your data back to other traders, so resting orders in the MidPool does not result in information leakage.
Not all dark pools are created equal. The Siege MidPool uses the independent FCA-regulated mid-rate from New Change FX, protects all participants with anti-polling controls and prints trades against central counterparties to provide anonymity. This means that the act of trading in the MidPool does not influence your next price, so there is no market impact.
In what ways has the growing use of TCA and execution analytics in algorithmic FX trading helped to drive interest in dark pools and the benefits they can deliver?
The advance of algo-focussed TCA is helping Buy-side traders to compare outcomes from their algo providers and risk transfer alternatives in terms of total cost including market impact. This is something that Sell-side traders have probably had access to earlier through in-house quant analysis. We conduct our own mark-out analysis against NCFX mid-rates and can see that MidPool order submission does not create impact. This fact is key to the notion that once you have decided to trade over time, there is no opportunity cost to aim for dark matches ahead of other risk outlets. Although we cannot know how our participants trade elsewhere, we do observe that MidPool trade activity has an attractive mark-out profile which motivates participants to engage with us more and more.
How successful has the Siege MidPool been so far in attracting liquidity and what factors are influencing this?
At the outset we were lucky to have a core group of Buy-side and Sell-side supporters who were committed to the idea of all-to-all dark matching and had both the time and resources to work with us. The turning-point was a more targeted approach to Sell-side algo providers that has developed into strong partnerships with the banks. Algo strategies offer a great access route for both Buy-side and Sell-side traders as they wrap credit provision with order routing, risk rebalancing and trade ticket aggregation. Ease of connectivity and workflow integration is also of huge importance for our partners. Now that we have critical mass and data driven proof of value, the flywheel has gained its own momentum.
…and the flywheel is now turning faster and faster
What steps can buy side firms take to help maximise the dark pool opportunity?
The first and easiest step is to engage with your algo providers to discuss venue access and of course to ask for the Siege MidPool by name. The next step is to find out which strategy can operate at longer resting times and send larger clip sizes, because resting time really drives matching success. If your algo just sends and retrieves order clips within seconds, your queue position is lost and dark matching opportunities are minimised. Traders know that different market conditions and currency pairs will have different optimal execution durations. This is the trade-off between market impact and market drift. We see a lot of activity in EM, where longer durations are more common and dark matching at mid is even more attractive.
Are there other services available to Buy-side traders that compete with dark matching?
Bank internalisation offers similar benefits to Buy-side traders, but we see that as complimentary to dark matching on the Siege MidPool. Banks can create offset opportunities from their eFX liquidity and algo flows that many clients value highly. These Buy-side traders told us that the ideal sequence for lowering execution impact is bank internalisation, followed by the MidPool and only then to show any interest into ‘lit’ markets. Some of our bank partners have implemented the MidPool that way and it is indeed very successful for clients. However other clients choose not to access bank internalisation and some algo providers distinctly offer more ‘agency’ style execution. For these clients and bank algos, Siege works best as the first port of call.
How optimistic are you about the future growth prospects for the use of dark pools in algorithmic FX trading and where do you see this coming from?
We are very optimistic as we continue to see great growth in Spot FX and we expect this to multiply in the next few years with ever greater participation. As markets are increasingly automated new liquidity providers enter, venues proliferate and algo trading increases even more. As Algos have learned to place large clips in our MidPool for the whole duration of the order, increased matching reduces the algo time path with better results: Shorter market risk combined with lower market impact are the clear benefits of dark matching Our experience in EM suggests that there is value to extend the offering to NDFs. Even for FX swap trading we have some very exciting conversations with our partners on how to crack that nut!
