Very early on in the life of TradeInformer, back in the heady days of ‘CFDs Weekly’ being on Substack, we did an interview with the first person to man the dealing desk at IG Group, bar the company’s founder.
The basic logic of managing risk in the early days of phone trading was that you had a manila folder for each instrument and you wrote down the client’s name and their position after they rang up to order. In the right hand column you had the running total, meaning you could keep track of your exposure. If that went over one lot then IG Index, as it was then known, went to the futures market and hedged, capturing a mark up on the price they passed to the client.
Over the decades, this model evolved from manilla folders, to Excel Spreadsheets to the modern trading systems we see today. But in many cases, due to the scale afforded by the electronification of trading, the process became far less personal. The individual rows and folders evolved into ‘books’ of business that banks, brokers and funds managed en masse in response to a drastic increase in the complexity of trading activities, and LP relationship management.
Fast forward almost fifty years and we see many banks and brokers that end up processing hundreds of thousands, if not millions, of trades per day. Writing order specs on a manila envelope is not going to work when you are dealing with an overly aggressive group of self directed traders. Even with drastic improvements in fraud protection, price validation and TCA, the concept of ‘toxic flow’ was not a phenomenon in quite the same way that it is in 2024.
However, the risk management processes used by newer players in the sector over the last two decades have mainly been either high risk or quasi-antagonistic towards liquidity providers. With regard to the former, providers internalise all risk, while the latter means providers treat LPs more like an exhaust pipe, sending out everything they don’t want and internalising everything they do.
My sense is that things could be changing for a few reasons.
- Time and expertise
- Technological change
- More informed and better educated traders
With regard to (1) it’s worth remembering that the large majority of players in the retail brokerage sector only really came into being in the last two decades. Keep in mind that CySEC itself was only launched in 2001.
The reality is that it takes a long time to build expertise in a given field and we are starting to see that pan out in the sector as a whole. To give one anecdotal example of this, I have had four conversations with individuals who either worked or still work on dealing desks in the last three months. All have integrated or are integrating some kind of ‘next generation’ risk management tool.
And that feeds into the second point which is that dealing desk technology has improved substantially in the last few years for retail brokers. To give one example of the sort of thing I’m talking about, you can look at oneZero, which recently launched a new product called ‘Portfolios’.
The basic way to think about these is that a ‘Portfolio’ is like an individual risk management strategy. You can create as many ‘Portfolios’ as you want and tweak each one, with the underlying strategy based on different criteria of your choosing.
This could be as simple as…
If my exposure becomes >X in instrument Y, then hedge after Zms with LP(s) ABC.
Or
If clients in subset A trade in instruments XY in volumes >Z then hedge with LP(s) ABC
Obviously it could be more complicated than this but as I am not a quant nerd, I can’t tell you what that would mean in practice. Other features mean you could compare both trader and LP performance across instruments, time frames (eg. do gold spreads blow out), and other factors.
The other interesting facet at play here is that third point, namely managing sharper clients. The funny thing is that just as brokers themselves have become more sophisticated, the pool of people in the world who are (1) very familiar with trading systems and (2) can do nerd programming stuff has risen substantially in the last decade. As noted when we spoke with ForexVPS Founder Will Thomas earlier this year, it’s plausible algo flow will be >50% of all volumes at retail brokers in the next few years. The manilla folder would not hold up.
This creates another layer of problems for brokers. Firstly, if you simply turn away traders who do not fit into your broadly defined risk categories, or LP relationships, they go and complain, damaging your reputation. But if you only send this client profile to a single LP, they will probably cut you off as well. Moreover, if you do things like widen spreads to stop them, you then end up risking losing clients or potential clients. In other words, there aren’t many easy solutions.
When you look at the last 18 months as well, when some providers have experienced insane swings in P&L, primarily from gold trading, then you have to wonder if there could be some sort of shift in the years ahead as to how retail brokers manage risk – even if it’s just so the company owners don’t go to sleep at night on the cusp of having a heart attack.
Arguably this would be like going back to those manila envelope days. Brokers net off flow where possible, internalise risk up to a certain limit, and then use LPs to hedge out the rest. The difference today is more in the details.
When you have something akin to the oneZero Portfolio, one could extrapolate how individual client flow can be better compartmentalised into known behaviors, matched with LP performance / expectations and put the intermediary in a better position to question not who you should do business with, but where the best fit for matching flow might be.
More sophisticated clients are also a push factor here. Though the traditional 1:1 ‘send us your exhaust’ model tends to fall over when ALL hedging flow is considered equal, through better relationship and risk profiling, the broker can now seek out opportunities where there are LPs interested in specific types of flow.
In turn this means – you would assume – that the broker/LP relationship becomes less strained and antagonistic. Brokers would be in a better position to ask for better conditions – higher rebates for example – relative to the past. This opens up the door to a variety of evolving models, which from surveying brokers in the market, are definitely top of mind coming out of the market dynamics in 2023: personalised hedging relationships, more systematic hedging approaches with defined and transparent guardrails, and a better understanding of the risk/reward of internalization.
Whether this will actually pan out remains to be seen. The big months of PnL always tend to make brokers cling to the higher risk model. But there are lots of push factors there for the shift to occur, especially as we see more players ask not only what their flow is worth, but what their firms are worth as well.