Dynamic Exposure Targeting

Abstract

Achieving desired exposures to information sources is critical to successful active investing. If a portfolio manager, for example, wants to increase her portfolio’s exposure to a signal—how long will it take? By how much should she change the signal’s weight? How will exposures to other signals change? If she turns off a signal, for how long will the exposure to it persist? Can she accelerate these shifts? These challenges can present themselves whenever markets or business conditions change. Rule-of-thumb answers can mislead as exposures are affected not only by the sequence of weight changes but also by signal speeds and correlations and by the levels of risk and turnover. Sneddon and Yukhymuk solve for the combined effects of portfolio and signal dynamics, derive explicit predictions of the resulting exposure dynamics, and provide the intuition behind these results. The solutions include cost control as well as expectations over all realizations of the signals, both current and prior. A simple but central result describes exposure shift speeds. The authors show that interactions between different features can lead to unexpected results. The information presented in this article can help a portfolio manager to predict and manage the timing and extent of a portfolio’s exposure to its information sources.

TOPICS: Manager selection, statistical methods, in portfolio management