Uncertainty, ambiguity and volatility: taming information problems in sustainable finance

This thesis investigates various dimensions of uncertainty and information asymmetry in the context of sustainable finance and energy markets. The first chapter introduces a novel probabilistic framework for sustainable portfolio allocation, where portfolio weights are treated as random variables influenced by both financial and ESG criteria. Compared to a standard utility-based approach, this Dual Preference Utility (DPU) model demonstrates greater robustness to parameter changes, lower portfolio concentration, and superior out-of-sample performance, despite slightly higher variance. The second chapter focuses on ambiguity in ESG scoring systems. By modeling investor preferences under uncertainty and constructing a perceived informativeness index using Refinitiv data and the SASB materiality map, the analysis explores how policy shocks and score adjustments affect firm evaluations. Results highlight structural biases in current ESG ratings, particularly the underweighting of environmental and governance factors, and offer policy-relevant insights for improving ESG score credibility and alignment. In the third chapter, attention shifts to the energy and commodity markets, critical components in the green transition. Using high-frequency volatility estimators, the study compares HAR and GARCH models for forecasting the volatility of European Carbon Allowances, Dutch TTF gas, and Brent oil futures. While realized volatility improves model fit, only the simpler GARCH(1,1) model provides superior predictive accuracy in Value-at-Risk forecasts. The chapter also examines inter-commodity connectedness, revealing strong, typically negative correlations between carbon and gas futures, reflecting the supply-demand dynamics of the energy transition. Together, these contributions offer methodological innovations and policy-relevant findings for managing uncertainty in sustainable investment and energy market modeling.