We investigate the role of execution quality in portfolio performance attribution. We show how conventional Transaction Cost Analysis (TCA) rewards behavioral trading practices that in some cases hurt rather than help portfolio performance. To align the incentives of the trading desk with optimal portfolio performance, we propose a new execution quality framework as an extension of Brinson Fachler. We show how to compute a baseline price by simulating the actions of a trading machine that sees the same instructions as the trader and executes the open shares following a pre-determined plan. The proposed baseline price is more realistic than TCA benchmarks because it processes the same stream of order instructions as the trading desk, including size changes, merging of overlapping orders from multiple clients and proportional allocation. Replacing the execution results with the baseline price in Brinson Fachler provides a measure of portfolio performance that is isolated from trader decisions; the offsetting terms attribute the trading desk’s role in portfolio performance and provide an unambiguous accounting of opportunity costs. We discuss the application of this framework to AI-assisted trading and estimate the impact of execution optimization on portfolio rankings.

Henri Waelbroeck, Ph.D., Portware
Carla S. Gomes, Ph.D., Portware