Market Making: A Practical Perspective

Introduction

This article provides an analytical overview of the principles and challenges of market making, focusing on the practical risks and inefficiencies observed in real-world trading environments. While academic models often attempt to formalize optimal behavior, empirical experience reveals that these models frequently fail to capture the microstructural and competitive realities of modern markets.

Defining Market Making

From a practical standpoint, a market maker is an entity that submits a substantial portion (typically more than 1%) of all make orders to an exchange or broker. Contrary to the common perception that market makers simply profit from the bid-ask spread, the actual mechanics are far more complex. The spread itself is rarely under the direct control of the market maker and is shaped by competitive forces within the order book.

Order Book and Spread

• Order Book: The list of buy and sell orders created by makers, which takers can accept.
• Spread: The price difference between the best bid and ask. For instance, if an exchange buys USD at 89.9 and sells at 90.0, the spread is 0.1.

While this basic structure seems straightforward, risk emerges as soon as market prices begin to fluctuate.

Inventory Risk

The largest risk faced by market makers is Inventory Risk, the exposure resulting from holding an imbalanced position.

Example:
A market maker posts a sell quote for USD at 90.0. A trader buys 1,000 USD, leaving the market maker short. Later, the market price rises to 95.0, forcing the maker to adjust quotes and potentially repurchase USD at a loss.

Such risks can be statistically modeled. For example:

• Average traded volume per 8 hours: 1,000 USD.
• Price variance per 8 hours: 500 units.

The question becomes: is the expected spread large enough to justify exposure to such volatility?

The Illusion of Spread Control

In order book–based markets, the spread is not a controllable variable. Market makers can only narrow spreads to gain queue priority, not widen them arbitrarily. As a result, profitability depends on precise evaluation of whether reduced spreads compensate for elevated inventory risk.

In the majority of markets—where directional uncertainty dominates—reducing spreads is often unprofitable. The expected gain from being filled rarely offsets the risk of adverse price movement.

Limitations of Academic Models

Models such as Avellaneda–Stoikov (2006) attempt to estimate optimal spreads by assuming that prices follow a Brownian motion and that risk can be quantified purely through volatility. While elegant, these models fail in real environments for several reasons:

1. Market makers do not control spreads; competition dictates them.
2. Realized losses from adverse price movements frequently exceed modeled expectations.
3. Market microstructure, including queue dynamics and latency competition, is not adequately represented.

Therefore, spread optimization based purely on theoretical variance models is ineffective in nearly all real-world markets.

Additional Market Making Risks

Beyond inventory exposure, several other risks play critical roles in profitability:

• Adverse Selection Risk: Occurs when faster or more informed traders exploit stale quotes.
• Queue Position Risk: In order-driven markets, being displaced from the first few levels of the order book drastically reduces fill probability.
• Bid War Risk: Aggressive competition among market makers can compress spreads to unsustainable levels, eroding profit margins.

These risks interact dynamically, and their cumulative impact often outweighs the theoretical benefits suggested by simplified stochastic models.

Conclusion

The practice of market making involves continuous exposure management in highly competitive environments. While academic literature provides foundational insight, it often fails to capture the real complexity of microstructure behavior, latency races, and feedback dynamics.

Future analyses will address each risk type in greater depth, focusing on how empirical data and adaptive algorithms can improve resilience and profitability in market making systems.