On a panel discussion entitled “The Twists and Turns of FXPB”, speakers at Profit & Loss Forex Network Chicago discussed the possibility for technology to radically re-shape the prime services ecosystem.
Technology’s impact on prime services was the jumping off point for the last panel at Profit & Loss Forex Network Chicago. Peter Plester, head of prime brokerage at Saxo Bank A/S highlighted the impact that technology had already in terms of risk management in this segment, pointing out that the traditional plumbing for starting up a prime broker was to connect to NEX Traiana and the various ECNs and have STP for tickets, but that the central risk system internal to the PB was fairly manual.
“You would type out a designation notice and send it to the liquidity providers for that client, but if you had to stop that client from trading with a particular liquidity provider, you would have to literally phone up that LP and ask them to stop quoting that client and tell them that you’ll send a cancellation for the designation notice. That was the traditional plumbing and model,” he said.
By contrast, Plester said that by the time Saxo set up its prime services business five years ago, the risk management could all be automated, the margin engine could be plugged into vendors with risk management tools and there were kill switches that could be implemented. As a consequence, he said, the firm was able to support clients that traditional FXPBs perhaps couldn’t support due the client’s lack of balance sheet.
However, some of these risks tools are not without their drawbacks, as PierreEmmanuel Pomès, service head for risk and documentation at NEX Traiana, pointed out.
“So let’s just state the problem clearly here first. A prime broker has to manage both a client’s overall limit but also a flurry of liquidity limits, both on the liquidity provider and the ECN side, which means that for a given client, a PB may have to manage 50+ limits. So where technology can help is by providing a link between the overall client’s limit, which is what the PB cares about, and those liquidity limits, which is the only thing that can stop the client from trading,” he explained.
Kill switches are one way of doing this, and have been around for some time. These kill switches are effectively on/off buttons for trading that can either be placed in the client’s technology stack, to create a pre-trade limits systems, or in the post-trade technology stack at the PB, so that once a client has executed and reached their limits, the PB can stop them from executing further.
“Both have value,” said Pomès. “The pre-trade kill switch might seem to be the right approach to credit checking, but it adds latency and, because it lives in the client stack, it may pose some problems for prime brokers. However, the kill switch in general is actually a very blunt instrument and usually stops clients trading at the worst time, and even as prime brokers, you want your client to actually reduce their exposure when they hit their limits rather than just stop trading.”
One solution for this that exists today, according to Pomès, is to have a dynamic credit distribution solution that functions as a “dimmer” rather than as an on/off switch.
“So once a client gets closer to their utilisation, you basically take credit out of their lines rather than turning them off and turning them on,” he said.
Pomès then added: “Ideally, instead of offering some automatisation of the complexity, we’d like to actually just wipe this out and have some form of distributed network where the same limit is applied to all of the nodes of the network. But I think we’re still quite a few years away from such a solution at the moment.”
Noel Singh, head of e-FX business development at Sucden Financial, agreed that the current binary nature of most kill switches, where clients can either trade or not, needs to be refined going forward.
“I think that if we start to understand the makeup of a client’s NOP better then we can produce better solutions. It could be that a client has hit their NOP limit when in fact selling EUR/USD would have no impact on their NOP, so instead of having a binary solution let’s look at what currencies and in what direction clients should still be allowed to trade and make the process a little more dynamic,” he said.
Providing a client perspective on this issue, Matt Fetta, director of trading operations, global fixed income and macros strategies, at Citadel, said that kill switches can be a helpful mechanism in the marketplace, especially in times of distress.
But he then added: “There needs to be a mechanism that alerts market participants as they approach a credit limit, so that they can divert the next trade away from that counterparty. Being in a position where an externally operated kill switch prevents you from putting through your next trade and affects your ability to deploy capital or manage risk is problematic. It is important that kill switches are implemented in areas that make them effective and not disruptive.”
Sanjay Madgavkar, managing director and global head of FXPB at Citi, agreed that kill switches are useful in certain use cases, with systematic traders being a clear example of where they can be important, but stated that the risk profile for macro hedge funds that are well capitalised is different than the smaller, more thinly capitalised entities where kill switches need to be prioritised.
“I think that a good way to think about this is to imagine if you had to create a prime brokerage ecosystem from scratch,” said Madgavkar. “You’d create different utilities: you’d create a risk utility which has designation notices and all the credit allocations that are currently done through emails and bilateral communications, you’d have a utility that would follow kill switch rules which the entire industry has agreed to, and you’d have more protocols and a central utility for novations and other post-trade activity, which has become very important given the current environment.”
He continued: “So I think what we’re doing at the moment is strapping on these tools to a market that already operates in a different legacy manner. But if you had to start from scratch, I think it would look quIte different.”
Building on this idea of designing a whole new prime services ecosystem, Singh highlighted the credit risk methodology on a lot of the different platforms as one particular friction point that he would like to see addressed in the future.
“I know there’s technology out there that normalises a lot of this risk, but fundamentally, that’s just a fudge. I think that we need to start looking at building a solution from scratch and I find it encouraging that there are initiatives out there that are using distributed ledger technology where there is a connection between a client’s available credit and their existing positions and the limit that’s provided to those execution venues. And so it’s the dynamic allocation of risk directly to the venue so that the limit check is done at the venue according to available credit, rather than being done with a pretrade risk limit on a third party venue. That’s where I really want to see this going,” he said.
Fetta, meanwhile, said that the first thing he would like to see made available in the prime services space is a robust process for checking available credit across every executing counterparty that hedge funds trade with through their prime brokers.
“The prime brokers that we work with are constantly on top of monitoring limits, but having a real time mechanism and a pre-trade process that allows us to see exactly how much NOP we have available with a given counterparty and updated to include the last transaction that we’ve done would be extremely helpful,” he explained.
Fetta added: “Looking further down the transaction lifecycle, I think what’s really important now is how market structure is evolving to support post-trade market events in the FX space. Historically, because the FX market has been short dates. there hasn’t been demand for posttrade processes that support active management of portfolio size. However, given the increase in capital charges and balance sheet costs, the ability to manage the gross notional of the portfolio has become more important. These post-trade processes and supporting technology don’t really exist yet.”