Maximally Machine-Learnable Portfolios

50 Pages Posted: 28 Apr 2023 Last revised: 27 Nov 2023

See all articles by Philippe Goulet Coulombe

Philippe Goulet Coulombe

Université du Québec à Montréal - Département des Sciences Économiques

Maximilian Göbel

Bocconi University

Date Written: April 24, 2023


When it comes to stock returns, any form of predictability can bolster risk-adjusted profitability. We develop a collaborative machine learning algorithm that optimizes portfolio weights so that the resulting synthetic security is maximally predictable. Precisely, we introduce MACE, a multivariate extension of Alternating Conditional Expectations that achieves the aforementioned goal by wielding a Random Forest on one side of the equation, and a constrained Ridge Regression on the other. There are two key improvements with respect to Lo and MacKinlay’s original maximally predictable portfolio approach. First, it accommodates for any (nonlinear) forecasting algorithm and predictor set. Second, it handles large portfolios. We conduct exercises at the daily and monthly frequency and report significant increases in predictability and profitability using very little conditioning information. Interestingly, predictability is found in bad as well as good times, and MACE successfully navigates the debacle of 2022.

Keywords: Portfolio optimization, Machine Learning, Statistical Arbitrage, Nonlinear time series

JEL Classification: C12, C14, G11, G12, G17

Suggested Citation

Goulet Coulombe, Philippe and Göbel, Maximilian, Maximally Machine-Learnable Portfolios (April 24, 2023). Available at SSRN: or

Philippe Goulet Coulombe (Contact Author)

Université du Québec à Montréal - Département des Sciences Économiques ( email )

PB 8888 Station DownTown
Succursale Centre Ville
Montreal, Quebec H3C3P8

Maximilian Göbel

Bocconi University ( email )

Via Sarfatti, 25
Milan, MI 20136


Do you have negative results from your research you’d like to share?

Paper statistics

Abstract Views
PlumX Metrics