Portfolio Construction Using Stratified Models

TL;DR

This paper introduces stratified models for asset return prediction based on market conditions, enabling adaptive trading policies that perform well out-of-sample and can be scaled to larger, more complex datasets.

Contribution

It develops Laplacian regularized stratified models for return mean and covariance, allowing for generalization to unseen market conditions and integration with a Markowitz-inspired optimization.

Findings

Models perform well out-of-sample on ETF data

Regularization enables borrowing strength across similar market conditions

Method scalable to larger datasets with proprietary data sources

Abstract

In this paper we develop models of asset return mean and covariance that depend on some observable market conditions, and use these to construct a trading policy that depends on these conditions, and the current portfolio holdings. After discretizing the market conditions, we fit Laplacian regularized stratified models for the return mean and covariance. These models have a different mean and covariance for each market condition, but are regularized so that nearby market conditions have similar models. This technique allows us to fit models for market conditions that have not occurred in the training data, by borrowing strength from nearby market conditions for which we do have data. These models are combined with a Markowitz-inspired optimization method to yield a trading policy that is based on market conditions. We illustrate our method on a small universe of 18 ETFs, using three well known and publicly available market variables to construct 1000 market conditions, and show that it performs well out of sample. The method, however, is general, and scales to much larger problems, that presumably would use proprietary data sources and forecasts along with publicly available data.