Cutting edge introduction: Hidden models for hidden costs

Sequencing trades optimally to reduce costs is a 50-year-old problem that is still unresolved because the costs themselves are tricky to predict. Academics at New York University apply an intuitive Bayesian technique to simplify the process

Trading is not just about knowing which positions to take. Markets introduce friction in the form of transaction costs and larger trades have a tendency to push the market away – an effect called market impact. Sequencing trades optimally to reduce these costs is a computationally intensive process and most of the time firms end up doing it trade by trade instead.

In our first technical, Multiperiod portfolio selection and Bayesian dynamic models, Petter Kolm, a professor of mathematics at the