Optimising Sustainable Financial Portfolios
What was the project about?
The goals of this project were to (1) to develop a technique that can help investors, including small wealth management offices, to take informed investment decisions that optimise a portfolio’s Environmental, Social, and Governance (ESG) impact, and, in particular, minimises the CO2 emissions of underlying portfolio assets and (2) to create a prototype implementation that, with sufficient computing power, accessed through massive-parallel programming technology targeting GPGPUs, provides decision makers with valuable tools for optimising and managing portfolio ESG factors, for realistic portfolios, while at the same managing portfolio risk and performance through classic portfolio optimisation techniques.
The project was realised partly through 9 BSc thesis projects (15 BSc students) and one MSc thesis project (one student). Whereas the BSc thesis projects mainly served to investigate the usability of utilising the combination of traditional asset time-series data and ESG data in decision processes, the MSc project investigated techniques, based on GPGPU computing, for efficient problem solving, utilising the Futhark programming language, a parallel programming language developed at DIKU.
Professor Martin Elsman served as project supervisor for all student projects and Associate Professor Omri Ross served as an additional project counselor providing expertise in modern portfolio theory.
The project was based on current research developed at the Department of Computer Science at the University of Copenhagen and done in collaboration with FinE and Invest Fondsmæglerselskab A/S. The industrial partners e.g. provided essential ESG data that was used in all student projects.
University of Copenhagen
Martin Elsman, Professor, Department of Computer Science
FinE an Optimal Invest Fondsmæglerselskab A/S
Kasper Unn Weihe: Convex Optimization and Parallel Computing for Portfolio Optimization, MSc thesis, DIKU (June 2023). Find here
Additional background material
Martin Elsman: Presentation at the 2023 Nordic Fintech Symposium. Find here
The Futhark Programming Language developed at DIKU. Find here
In the news
SCIENCE: Studerende skaber smart værktøj til grønnere investeringer (December, 2023). Find here