The Widening Profitability Gap between Renewable and Fossil Power Firms in Europe

Battle Count: 52

The paper is primarily focused on energy policy and corporate finance rather than direct quantitative trading strategies. However, findings on the structural profitability divergence between renewable and fossil portfolios could inform sector rotation strategies, ESG-themed portfolio construction, and long-term positioning in utility stocks. The identification of crisis-driven windfall profits as temporary anomalies (rather than durable advantages) is relevant for mean-reversion trading in energy sector equities. The geographic heterogeneity analysis could inform regional allocation within energy sector ETFs. The paper does not provide trading signals, backtests, or direct alpha generation strategies.

Moderately complex implementation. DTW hierarchical clustering requires careful preprocessing of time series portfolio data and selection of distance metrics and linkage methods. BMA over 2^20 (~1 million) model specifications requires MCMC sampling (birth-death algorithm) and careful prior specification (g-prior, beta-binomial). The rolling window approach adds computational burden. However, established R packages (bms, dtw) handle most computational requirements. The main complexity lies in data preparation (merging Orbis financial data with Capital IQ asset data at the direct asset-owner level) and ensuring proper panel structure with time fixed effects. Interpretation of BMA results (PIPs, posterior densities, credible intervals) requires Bayesian statistical expertise.

Code will be made public upon publication. However, primary data (S&P Capital IQ Pro-Energy, Moody's Orbis) are under licensing restrictions and cannot be publicly shared. Supplementary data not under licensing restrictions will be shared. The BMS package (v0.3.4) in R is used for BMA computations. Methodology is well-documented with detailed parameter settings (200,000 MCMC iterations, 20,000 burn-in, 2^20 model space). Robustness checks with alternative priors (UIP, BRIC) and alternative profitability measures (ROE) are provided.