Broker Risk Management in 2019

8 minute read

It’s not easy being a FX broker in 2019. Yields are under attack from regulations, increasingly predatory EAs and tighter spreads are a few of the major stresses.

shutterstock_520407727 (1)



ESMA leverage limits have dramatically pushed up the cost of ASIC licenses, which has seen huge volume growth for those with these precious pieces of paper.

The traditional B-Book play of watching over-leveraged clients eventually run into their margin buffer is now not an option for those still under ESMA scrutiny. Reduced leverage means clients struggle to max out their accounts and do not lose in the same amounts meaning the B-Book method of management yields very low and inconsistent yields.

Copy of How New Regulations Change B Book profitabilityWith huge volumes, the Australian brokers and their B-Books require constant hedging to reduce huge risk positions. Any broker with thoughts of Stock Exchange listing or sale cannot tolerate a $3m monthly loss and the regulator rightly states they cannot run infinite positions with a finite balance sheet.

This kills the very essence of a B-Book. As soon as a broker begins hedging, the link between profits and client losses is cut. The very point when client positions are largest are when the consensus is against a trend and as that trend continues the traditional B-Book broker cleans up. Having to hedge to reduce positions of course occurs when positions are at their largest, which down to no coincidence at all is when clients lose the most. Hedging when clients are most likely to lose therefore removes risk during the most fertile periods.

The scenario ends up that during periods when clients win, brokers lose. Then during periods where clients give back all their winnings, the broker is not in a position to swallow all these profits due to hedging away the positive expectation risk.



Predatory EAs are increasingly at work, every broker we have spoken to in the last 6 months is concerned about the damage done, with many mentioning the Chinese origination of these clients.

shutterstock_1096687181The basic premise is this, if enough clients follow an EA then it becomes a self fulfilling prophecy. 100 clients each buy EURUSD for 500k each, that’s 50m being bought, no big deal. Now 1k clients each buy 500k across multiple brokers that’s 500m and enough to have market impact. Replace the EUR in EURUSD with GBP and USD with something without the US Dollar’s strength such as JPY. Now instead of just pushing the EUR leg, there is the ability to push both the GBP leg and the JPY leg and have twice the impact. Now have 5k clients trading 500k and you have $2.5bn GBPJPY being bought across retail brokers and the prisoner’s dilemma kicks in. If all brokers sat on the risk they would be fine (sitting on a worrying amount of risk albeit!) but the optimal outcome is always to be the first to clear the risk before every other broker does causing GBPJPY to shoot up and brokers to face significant losses.

The actual EAs are becoming a little smarter, with multiple EAs in cahoots to mask the yields from brokers by not offering easy analysis of trading patterns.


Spread Compression

Tighter spreads are easier to understand. In the past a trading desk receiving excess spread could afford to only lock in 40% of this and demand a healthy bonus for profits made. Let us not forget that spread paid by clients is instant PnL into the brokers pocket. Reduce spreads and the PnL forgone has to be found from somewhere else.

The old thinking was that spreads did not matter. A client is eventually going to run into his/her margin buffer regardless of whether this is 50 or 51 pips away so if a client actually pays 1 pip spread or not is unimportant. But going back to my first point, now clients have reduced leverage (and better trading education) they are not running into these buffers so the spread on each trade is the major driver of PnL.


Is The A Book Brokering Model Still Valid?

The obvious solution is going back to a traditional A-Book model, meaning STPing the majority of flow to LPs. If the broker has sufficient capital to deploy then a range of LPs can create a fairly tight top of book. Risk positions are dramatically lower and the broker can have an easy existence with sufficient client flows.

This falls down from a PnL optimisation standpoint. If a broker pays $20/m in spread plus $8 commission to broker $5bn of flow a week the maths are as follows:

That’s over $7.28m over a year with which to pay for a trading operation and reap a large proportion of this number in additional PnL.

The other issue is passing on the EA flow I spoke about earlier. We have seen a number of bank LPs turn around and close a broker’s account after only one days worth of significant losses. This ‘dealing not full’ as we describe it, is considered by the major LPs to be the worst crime in FX. They can’t last look the flow away, they just sit on the risk and are steamrollered by the market moves. This A-Book model can lead to alienating LPs and damaging spreads permanently for the remaining business leading to increased hedging costs. Also, wider spreads when brokering mean less commission can be added by the broker which is now the only revenue source.


Efficient Risk Management

So, the only real solution is what I call ‘efficient risk management’. Risk management is not ‘I can see my risk positions on a screen and then click a button to trade’ because at no point are systems at work to optimise dealings with the external world at both a client and hedge level. More and more brokers are realising they need to treat risk in the same way as the big banks, with an exceptional risk management engine at the heart of it. Optimising PnL is maximising yield from client trades and minimising hedging costs WHILE keeping external parties happy. To borrow an analogy from the CEO of MahiFX, David Cooney;

“It’s about the window of time that you are looking to maximise PnL over. If you’re looking to maximise happiness in only in the very short term you would turn to heroin, with of course catastrophic long-term consequences. Passing on toxic flow may be great for commissions in the short term but damaging to the long term health of the broker.”

Creating a stable PnL for the long term therefore requires keeping your liquidity providers happy and building an environment where you can both co-exist and make profits. So hedging needs to be carefully monitored using analytics to make sure the broker is treating the LP fairly. At the same time no broker wants to pay excessive spreads or trade at in-opportune times. Random button clicking by humans cannot compete in 2019 with automated hedging scouring the market each millisecond for an efficient opportunity to pay as close to the ‘true mid’ as possible.

This idea of the ’true mid’ brings us back round to client pricing. Spread, as I have described crudely earlier, is not simply the width of price quoted by a broker. It is the distance from the ‘true mid’ to the bid or offer traded. Clients almost always trade on the weak side of the market (message me if you want an explanation of why this is true), which means if your mid is off, the actual spread you are being paid is less than you think. In exceptional cases this means a client buying at a price below the ‘true mid’ and an instant loss on revaluation.

The closer a broker’s mid price can approach the true mid at each moment in time, the more confident a broker can be in reducing spreads to be competitive, safe in the knowledge that a fair yield is being received from each client trade.

The right analytics are needed to highlight the damaging clients and EAs and find solutions to the issues they create, one of which is liquidity reduction. If multiple clients come forward and trade on a broker’s price before the feedback to the external market has come back, they have a right to adjust their outgoing liquidity to represent how much of it has been swallowed.

As an example, if clients sell 5m GBPJPY in 0.2s through many small clips, the broker has a right to remove 5m from its liquidity knowing if these trades were to go to external LPs then it would cause market impact in the underlying GBPJPY price. It follows that on the next price update, the broker should quote GBPJPY with 5m removed from its TOB to prevent it taking on significant positions at a price that is good for only 500k. When these type of client flows occur this is no longer retail style trading – this is institutional ow at non-institutional prices. I can see liquidity management/reduction becoming a big buzz word amongst the retail broker community in the months and years to come as a defence against EAs.



I believe 2019 will be the year more and more FX brokers look to efficient risk management solutions to maintain profitability. Consistent profits are the key to maximising earnings multiples and shareholder value and we believe here at MahiFX we have the very best solution to achieve this in the form of MFX Compass. More than ever before brokers need:

  • Advanced pricing to maintain client yields as spreads fall. Regulatory changes now mean spreads matter more than ever, at the same time as they are under attack due to competition.

  • Easy to use analytics to spot clients using EAs and be aware of their impact in the face of more advanced multi-account EA strategies. These analytics are also essential to make sure pricing is optimised and hedging strategies are both e client and non-toxic for LPs.

  • Efficient hedging framework to give back away as little of the profits earned by clients as possible while keeping risk under control.


At MahiFX we have the tools to achieve anything a broker wishes in terms of pricing and risk management with unrivalled flexibility, along with a team that is always by your side to help you achieve your goals.


Written by Alexander Ridgers, Director of Trading & Analytics, MahiFX.



Request a demo

If you would like a demo of MFX Echo, please contact: