As more buy-side traders take on the routing and execution decisions for their trades, the demand for trade-support tools, such as transaction cost analysis, has grown in parallel.

“Ten years ago, the buy-side desk would never be involved in trading foreign exchange sets,” said Ian Domowitz, managing director and head of analytics at Investment Technology Group. “The FX would have been traded by the custodian.”

Since then, the percentage of FX trades executed electronically has grown by leaps and bounds.

“Although volumes have not necessarily grown, the amount that’s actually done through electronic means has definitely grown a great deal,” he said.

As of 2015, electronically traded FX comprised a bit more than half of all FX trades, according to industry data.

The challenge of creating a TCA benchmark for FX is the fragmented and over-the-counter nature of the market. There are no requirements to print trades to a consolidated tape and price quotes can from an individual bank on a one-to-one basis or across a pool of counterparties using one of the electronic communication networks.

When ITG decided to develop its TCA offering it arranged with 12 of the largest banks and six of the major FX ECNs to receive the same stream of actionable quotes, including no last-look quotes, that buy-side trading desks would receive.

This data gives ITG the cost of liquidity for every possible order size from $100,000 notional to $300 million notional, according to Domowitz.

“From that , we construct the analog of a global limit order book for currencies,” he explained. “From the limit order book, we can actually construct the cost of liquidity. From that, we actually can get reasonably precise measures of what people at least should be paying in terms of a benchmark.”

Working with its clients, ITG can see the actual experience of dozens of buy-side clients who trade FX electronically. “We can compare that to the cost of liquidity that we actually get from the banks,” said Domowitz. “As it turns out it actually squares up pretty well.”

ITG, in turn, stores the quotes from the banks and ECNs as well as their clients’ executions into a single data base to improve the future performance of its analytics.