Decarbonized, ESG-constrained portfolio optimization with CVaR : scenario-based investor profiles
Hawarihewa, Uthpalee Isuru Abhilasha (2025)
Pro gradu -tutkielma
2025
School of Business and Management, Kauppatieteet
Kaikki oikeudet pidätetään.
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi-fe20251217121553
https://urn.fi/URN:NBN:fi-fe20251217121553
Tiivistelmä
This thesis develops and empirically backtests a Conditional Value at Risk (CVaR) based portfolio optimization framework that integrates explicit ESG score floors and Scope 1 carbon intensity caps for a long only STOXX Europe 600 equity universe under realistic sector, single name, and turnover constraints. The study investigates how different investor risk profiles, risk taker, risk neutral, and risk averse shape the trade off between sustainability ambitions and risk return outcomes when ESG and decarbonization constraints are jointly imposed.
The framework is implemented as a linear programming model with a CVaR minimization objective, calibrated on rolling 36 month return scenarios using data from 2010 to 2024. Quarterly rebalancing, sector caps, and explicit turnover limits are included to reflect institutional grade implementation frictions and to generate out of sample performance statistics across the three investor profiles. A targeted sensitivity analysis is conducted around the strict risk averse configuration by systematically relaxing ESG score and carbon intensity caps to explore cost sustainability frontiers.
The results indicate that sustainability constrained CVaR portfolios systematically increase portfolio level ESG scores and sustainability reduce revenue normalized Scope 1 carbon intensity relative to an unconstrained CVaR benchmark, while incurring only modest return penalties. Across all investor profiles, constrained portfolios achieve lower CVaR and Value at Risk than the benchmark, with turnover and trading cost proxies remaining close to benchmark levels, suggesting enhanced downside protection without materially higher implementation costs. The sensitivity analysis identifies ESG carbon constraint combinations that nearly restore benchmark like performance while preserving improved sustainability metrics, showing that CVaR based optimization can deliver meaningful decarbonization and ESG enhancement at manageable and quantifiable expected return costs for institutional investors.
The framework is implemented as a linear programming model with a CVaR minimization objective, calibrated on rolling 36 month return scenarios using data from 2010 to 2024. Quarterly rebalancing, sector caps, and explicit turnover limits are included to reflect institutional grade implementation frictions and to generate out of sample performance statistics across the three investor profiles. A targeted sensitivity analysis is conducted around the strict risk averse configuration by systematically relaxing ESG score and carbon intensity caps to explore cost sustainability frontiers.
The results indicate that sustainability constrained CVaR portfolios systematically increase portfolio level ESG scores and sustainability reduce revenue normalized Scope 1 carbon intensity relative to an unconstrained CVaR benchmark, while incurring only modest return penalties. Across all investor profiles, constrained portfolios achieve lower CVaR and Value at Risk than the benchmark, with turnover and trading cost proxies remaining close to benchmark levels, suggesting enhanced downside protection without materially higher implementation costs. The sensitivity analysis identifies ESG carbon constraint combinations that nearly restore benchmark like performance while preserving improved sustainability metrics, showing that CVaR based optimization can deliver meaningful decarbonization and ESG enhancement at manageable and quantifiable expected return costs for institutional investors.