A-Book and B-Book Risk Management
Managing Risk Across A-Book and B-Book Models
Every broker offering CFDs or leveraged FX has to decide how client trading risk is managed. When a client places a trade, the broker may pass that risk to an external liquidity provider or retain the risk internally. These approaches are commonly described as A-book and B-book execution. In practice, many brokers operate a hybrid model, routing different trades according to client behaviour, instrument, trade size, market conditions and the broker’s current risk exposure.
This article explains how A-book and B-book models work, why brokers often combine the two, and how risk can be managed across a hybrid book.
A-book vs. B-book vs. hybrid
| Model | How it works | Where the risk sits | Where the risk sits |
A-Book | Client trades are passed to an external liquidity provider, ECN or exchange | With the liquidity provider - the broker earns spread/commission, not trading P&L | Larger, more sophisticated or consistently profitable clients |
| B-Book | Client trades are matched internally; the broker takes the other side | With the broker | Retail flow that is statistically unlikely to be consistently profitable over time |
| Hybrid | Trades are routed to A-book or B-book in real-time based on configured rules | Split dynamically, per trade | When different types of client flow and exposure are being managed by a broker |
Neither A-book nor B-book is inherently better. Each has a different risk and revenue profile. A-book execution reduces the broker’s direct market exposure on the routed trade, with profitability generally based on spread, commission or mark-up. B-book execution can be significantly more profitable when client flow is unprofitable in aggregate, but it exposes the broker directly to market risk if a client (or a cluster of clients) trades well.
Why do brokers use a hybrid model?
Using a single execution model across an entire client base can limit a broker’s ability to manage risk efficiently. Different clients, instruments and trading strategies can have very different risk characteristics. Flow that may be appropriate to internalise under one set of circumstances may need to be routed externally under another.
A hybrid model gives brokers greater flexibility by allowing execution decisions to be made at client or trade level. This can help brokers internalise suitable flow while externalising trades that fall outside their preferred risk profile.
The operational challenge is that these decisions cannot always be treated as permanent classifications. Client behaviour, market conditions and the broker’s own exposure can change, so routing policies need to be reviewed continuously, trade by trade, and adjusted accordingly.
How is risk managed at trade level?
Hybrid risk management typically involves three closely connected processes:
1. Classification - assessing the characteristics of a client or trade, including trading behaviour, profitability patterns and the potential for toxic or adverse flow.
2. Routing - based on the classification, deciding in real-time whether the trade should be internalised or passed to an external liquidity provider according to configured execution rules.
3. Monitoring - tracking exposure, P&L and client activity so that risk teams can adjust routing or hedge positions as conditions change. Even after a trade is routed, exposure needs to be tracked continuously.
This is where the difference between a liquidity bridge that supports hybrid execution and one that merely "allows" it becomes significant. Real-time classification and routing require both execution infrastructure (to act on the classification) and risk visibility (to inform it), which is why these two functions tend to sit in different but tightly connected systems. Routing decisions need to be applied quickly, while risk teams need sufficient information to understand how exposures are developing across clients, instruments and liquidity pools.
How do Gold-i’s tools support hybrid risk management?
Gold-i’s MatrixNET and Visual Edge are designed to work together to support different parts of this process. MatrixNET provides the real-time routing and execution layer, enabling brokers to apply configurable A-book and B-book routing rules as trades come in. Visual Edge provides the risk-management and monitoring layer, giving risk teams visibility into trading activity and exposure so that routing policies can be reviewed and refined. It can also support the identification and management of potentially toxic flow to to help brokers manage risk more effectively.
In practice, this means a risk manager can see exposure building on a specific symbol or client group as it happens, and either adjust routing rules directly or flag any that requires a closer review before that exposure becomes materially larger.
Getting the balance right
There is no universal A-book/B-book split that is appropriate for every broker. The right balance depends on the broker’s client base, instrument mix, risk appetite, liquidity relationships and regulatory obligations. More important than maintaining a fixed A-book/B-book ratio is that the broker has the right infrastructure in place to be able to monitor flow continuously and adjust execution rules as conditions change. For brokers operating a hybrid model, effective risk management depends on combining granular routing controls with clear, real-time visibility of trading exposure.
A-Book and B-Book Risk Management FAQs
A hybrid model allows a broker to use both A-book and B-book approaches, routing individual clients or trades externally while internalising others according to its execution and risk-management rules.
The decision may take account of factors such as client behaviour, instrument, trade size, market conditions, available liquidity and the broker’s current exposure.
Yes. Where the broker’s technology supports trade-level routing, execution rules can be adjusted as client behaviour, liquidity conditions and the broker’s own risk exposure change.