Fairness in the fast lane

In financial markets, fairness is easy to believe in but hard to define — especially when you're dealing with electronic trading systems that operate faster than the blink of an eye.

For most of history, fairness in trading meant access: If you could make it to the floor of the exchange or afford a broker, you could compete. In the electronic age, that definition shifted. Exchanges began equalizing physical cable lengths to ensure that every participant's signals had the same distance to travel. It was a visible, measurable safeguard — and for a while, it became the industry standard.

But technology never stays still. As more trading migrated to co-located servers and ultra-low-latency networks, the sources of competitive edge became harder to detect — and harder to regulate. The performance differences between market participants started coming down not just to where they were located, but how they were connected.

The Modern Fairness Problem

Today, cable length equalization is just table stakes. More attention is being paid to:

• Switch architecture: Different network switches, even from the same vendor, can have port-to-port variations in how they process and route data.
• Cross-connects: Physical connections between firms and exchanges inside data centers can vary subtly, introducing unpredictable micro-latencies.
• Multicast behavior: Distributing market data via multicast can result in messages arriving at slightly different times, depending on the topology.
• Queuing and matching algorithms: The exact sequencing logic inside matching engines — whether deterministic or randomized — can affect fairness in ways that aren't always obvious.

Some of these differences are measured in microseconds. Others, in nanoseconds or picoseconds. And in a trading environment where multiple firms are vying to be first in line, every edge matters. Being a few microseconds ahead can result in better fills, higher hit ratios, and outsized profitability.

This has pushed both firms and exchanges into a constant cycle of optimization — but not always in the same direction.

A Technical Arms Race

Firms are going to extraordinary lengths to gain an edge. Some examples from recent years include:

• Proprietary fiber paths with marginally shorter routes between exchanges.
• Microwave and millimeter wave links, which offer lower latency than fiber over long distances, despite weather-related risks.
• Packet pacing and shaping at the NIC (network interface card) level to control when data hits the wire.
• Clock synchronization with precision time protocol (PTP) and GPS-disciplined oscillators to timestamp events as accurately as possible.

Meanwhile, exchanges are under pressure to ensure that all of this innovation doesn’t compromise fairness for the wider market. This tension plays out in quiet but critical design decisions — like whether to use randomized or FIFO queues, or how much transparency to provide around internal infrastructure.

Is True Fairness Even Possible?

A recent report by the Bank for International Settlements confirmed what many in the industry have suspected: latency arbitrage is still highly concentrated. A small number of firms are consistently faster — and they extract a disproportionate amount of value as a result. Meanwhile, the broader market may be trading at prices influenced by activity they never had a chance to see in real time.

And that raises a deeper question: If you can’t remove the incentive to be faster — and you can’t prevent participants from innovating — can you ever truly level the playing field?

Some exchanges have introduced features like:

• Speed bumps that slow down all orders by a fixed amount to reduce the advantage of speed.
• Batch auctions or discrete trading intervals that eliminate race conditions.
• Multicast fairness algorithms, where messages are delivered in sync across recipients.

Each approach has trade-offs. Speed bumps can add latency for everyone, not just the fastest. Auctions can reduce continuous liquidity. Fairness algorithms may introduce complexity or new failure modes. There’s no perfect solution — only a set of evolving design decisions that reflect how exchanges balance innovation, access, and trust.

Fairness is therefore foundational to how markets function. If some participants consistently have better access or earlier data, confidence in pricing erodes. Liquidity providers may back away. Retail investors may feel disadvantaged. And regulators may start to ask difficult questions.

Ultimately, fairness is about trust — and trust is hard to win back once lost.

As electronic markets evolve, fairness will become more technical, more granular, and more contested. It’s not just about rules anymore. It’s about architecture. Design. Timing. Precision.

That’s why conversations about fairness today are happening in engineering meetings as much as in compliance departments.

The infrastructure is the policy.