Streamlined dealing desk practices can steer brokers away from the brink.
Risk management is often perceived as mysterious and mystifying but an endeavour that must succeed for the good of the business. The necessity will always be true, but the perception shouldn’t be true at all.
Whether it be trading industry insiders, financial market participants, or even the financially unsavvy, most have a vague perception of risk and, typically, as a costly afterthought.
In retail trading, dealing desks are entrusted to risk-manage the firm’s solvency – assuming they do not rely on “magic” to conjure results. When it all doesn’t go to plan, dire consequences follow.
In 2015, dozens of millions were lost in seconds when the world’s third-largest retail broker – Alpari (UK) – mismanaged its market and counterparty risk during a sudden bout of Swiss franc currency intervention. The same wave affected several other brokers, including FXCM, the world’s largest at the time, while other firms, such as Effex Capital and Boston Prime, faced similar challenges.
Dealing has always been at the core of broking, with faulty offerings incurring significant financial losses. Yet, rumours continuously circulate that several brokers are within touching distance of a single large client causing irrevocable damage to their solvency.
In today’s markets, many brokers invest significant resources to demystify order flow management while inventing quirky new concepts like C-book and Value at Risk (VaR) variations to empower their dealing teams.
Ultimately, retail and institutional dealing desks are somewhat different despite doing the same thing.
Institutional dealing desks are usually stuffed with quantitative professionals with mathematics, engineering and programming backgrounds. Such individuals are required to construct highly complex models in Python and R.
In contrast, retail brokers typically employ individuals with market awareness yet lack the statistical and mathematical acumen to mitigate market risks from magnifying losses on their order books. In modern markets, a lack of statistical sophistication or a sloppy algorithm will ultimately lead to suboptimal performance and, more than likely, capital haemorrhage. Frequently, retail brokers mistakenly think they can outmanoeuvre market counterparties by predicting random flows or by keeping faith in the client’s propensity to make the wrong moves.
Traditionally, a retail dealing desk operates on what is known as a “B-book” model.
With a B-book, the broker assumes all the risk from its retail clients, meaning that clients trade directly against the broker. This model isn’t inherently unviable because even if the broker A-book’s the trade and offloads it to the market, it just means another counterparty will warehouse the risk, effectively turning one broker’s A-book into another’s B-book.
Specifically, the broker must accurately price trades, execute them with the best available prices, handle slippage in low liquidity markets, and replicate actual market conditions. Always using multiple Liquidity Providers (LPs) is crucial, and the dealing desk must always stay sharp to avoid delays. Specific instruments may have preferential sources compared to others, which requires regular monitoring.
Ideally, unlike the current status quo, the desk should operate a liquidity bridge to offset various orders to the broader market.
Institutional risk management is an entirely different animal from retail – the level of complexity is much higher, and so are the stakes. The moment a dealing desk misses crucial real-time information, they become instant arbitrage targets. Whether related to swaps, mark-up mispricing, partially hedged positions, or delays – any opening in the firm’s defences leaves it vulnerable to significant losses.
A different approach is required compared to the traditional management of retail flows. Typically, institutional desks hold risk for short periods or within specific timeframes before offloading it.
Their teams are typically staffed by quants, many holding PhDs from prestigious universities. They possess a deep understanding of various API feeds used by institutions. Employing logarithmic modelling, they scrutinise orders for potential arbitrage, lodging complaints with the offering party when discrepancies are identified. Quants are proficient programmers and maintain ECNs with price feeds from top-tier and second-tier institutions. Sometimes, they engage in flow arbitrage as far as the LPs allow them.
Depending on the service, they may adjust prices and discuss progressive concepts such as “last look” orders – a time-limited ability for LPs to reject an order. Some, but by no means all, utilise this feature as part of their usual operations. Additionally, highly quantitative teams have specific holding time preferences, such as maintaining Eurodollar positions for some prefix timeframes.
In all the years I spent on dealing desks, a question that always lingered with me was: is it better to operate one unified dealing desk covering both retail and institutional flow or to create two teams, each with a dedicated focus?
For a long time, I was convinced that two teams were the optimal choice. However, given the rapid development and deployment of fintech, the game has changed. Today, with the right tools, the proper training, and a rethink about what roles are needed in dealing, I firmly believe that one unified team is the best way to go.
For the first time in history, cost-effective brokers can use tech-powered and data-driven strategies to streamline their risk management and, thereby, emulate the operational prowess of their institutional peers.
Instead of perpetuating a perception of mystification and mystery, dealing desks should emerge from the wilderness by leading from the front with progressive risk management that truly de-risks their operations.
Ghostwritten by George Tchetvertakov in collaboration with Dr Demetrios Zamboglou
Whether it be trading industry insiders, financial market participants, or even the financially unsavvy, most have a vague perception of risk and, typically, as a costly afterthought.
In retail trading, dealing desks are entrusted to risk-manage the firm’s solvency – assuming they do not rely on “magic” to conjure results. When it all doesn’t go to plan, dire consequences follow.
In 2015, dozens of millions were lost in seconds when the world’s third-largest retail broker – Alpari (UK) – mismanaged its market and counterparty risk during a sudden bout of Swiss franc currency intervention. The same wave affected several other brokers, including FXCM, the world’s largest at the time, while other firms, such as Effex Capital and Boston Prime, faced similar challenges.
They all learned the hard way that risk management is where brokers should start, not finish.
Different folks different strokes
Dealing has always been at the core of broking, with faulty offerings incurring significant financial losses. Yet, rumours continuously circulate that several brokers are within touching distance of a single large client causing irrevocable damage to their solvency.
In today’s markets, many brokers invest significant resources to demystify order flow management while inventing quirky new concepts like C-book and Value at Risk (VaR) variations to empower their dealing teams.
Ultimately, retail and institutional dealing desks are somewhat different despite doing the same thing.
Institutional dealing desks are usually stuffed with quantitative professionals with mathematics, engineering and programming backgrounds. Such individuals are required to construct highly complex models in Python and R.
In contrast, retail brokers typically employ individuals with market awareness yet lack the statistical and mathematical acumen to mitigate market risks from magnifying losses on their order books. In modern markets, a lack of statistical sophistication or a sloppy algorithm will ultimately lead to suboptimal performance and, more than likely, capital haemorrhage. Frequently, retail brokers mistakenly think they can outmanoeuvre market counterparties by predicting random flows or by keeping faith in the client’s propensity to make the wrong moves.
Simply Retail
Traditionally, a retail dealing desk operates on what is known as a “B-book” model.
With a B-book, the broker assumes all the risk from its retail clients, meaning that clients trade directly against the broker. This model isn’t inherently unviable because even if the broker A-book’s the trade and offloads it to the market, it just means another counterparty will warehouse the risk, effectively turning one broker’s A-book into another’s B-book.
Specifically, the broker must accurately price trades, execute them with the best available prices, handle slippage in low liquidity markets, and replicate actual market conditions. Always using multiple Liquidity Providers (LPs) is crucial, and the dealing desk must always stay sharp to avoid delays. Specific instruments may have preferential sources compared to others, which requires regular monitoring.
Meanwhile, bonus campaigns should be carefully structured to prevent arbitrage in all market conditions. As markets get savvier, so do traders, which means any bonus-offer vulnerability will be quickly exploited. One solution is for dealing desks to incorporate Key Performance Indicators (KPIs) into their modus operandi to take the focus away from periodic profit and loss.
Ideally, unlike the current status quo, the desk should operate a liquidity bridge to offset various orders to the broader market.
Institutionally Intricate
Institutional risk management is an entirely different animal from retail – the level of complexity is much higher, and so are the stakes. The moment a dealing desk misses crucial real-time information, they become instant arbitrage targets. Whether related to swaps, mark-up mispricing, partially hedged positions, or delays – any opening in the firm’s defences leaves it vulnerable to significant losses.
A different approach is required compared to the traditional management of retail flows. Typically, institutional desks hold risk for short periods or within specific timeframes before offloading it.
Their teams are typically staffed by quants, many holding PhDs from prestigious universities. They possess a deep understanding of various API feeds used by institutions. Employing logarithmic modelling, they scrutinise orders for potential arbitrage, lodging complaints with the offering party when discrepancies are identified. Quants are proficient programmers and maintain ECNs with price feeds from top-tier and second-tier institutions. Sometimes, they engage in flow arbitrage as far as the LPs allow them.
Depending on the service, they may adjust prices and discuss progressive concepts such as “last look” orders – a time-limited ability for LPs to reject an order. Some, but by no means all, utilise this feature as part of their usual operations. Additionally, highly quantitative teams have specific holding time preferences, such as maintaining Eurodollar positions for some prefix timeframes.
The machinations of an institutional dealing desk can cascade down the pecking order to wreak havoc on a retail broker.
Unify or Multiply
In all the years I spent on dealing desks, a question that always lingered with me was: is it better to operate one unified dealing desk covering both retail and institutional flow or to create two teams, each with a dedicated focus?
For a long time, I was convinced that two teams were the optimal choice. However, given the rapid development and deployment of fintech, the game has changed. Today, with the right tools, the proper training, and a rethink about what roles are needed in dealing, I firmly believe that one unified team is the best way to go.
For the first time in history, cost-effective brokers can use tech-powered and data-driven strategies to streamline their risk management and, thereby, emulate the operational prowess of their institutional peers.
Instead of perpetuating a perception of mystification and mystery, dealing desks should emerge from the wilderness by leading from the front with progressive risk management that truly de-risks their operations.
Ghostwritten by George Tchetvertakov in collaboration with Dr Demetrios Zamboglou
Published by Finance Magnates
