Research

End-to-End Portfolio: Estimation, Allocation and Risk Management - In Progress

Work with Carlos Castro-Iragorri

Portfolio optimization in finance is typically addressed through a two-step process: first, the attributes of the assets are predicted, and then these attributes are employed as parameters in an optimization problem to determine the optimal portfolio. To streamline this process, we have integrated these two steps by developing a neural network that predicts the optimal portfolio based on the available data. Specifically, we leverage a hidden layer in the network to perform the attribute prediction, while solving the optimization problem in an implicit layer using the cvxpylayer. 

ESG Scores and Portfolio Optimization - In Progress

The aim of this study is to review the concepts related to sustainable finance and climate change, and their relevance to propose new techniques for portfolio optimization. Specifically, we investigate the applicability of ESG Scores and Climate risks in addressing portfolio optimization and risk management problems. Notably, this concept is easily computable, which makes it a promising solution for these computational tasks.