Description

The shift to renewable energy and blockchain-based finance has created digital assets that have opened up sustainable energy markets to wider participation, but their place in the financial system and use in portfolio management are still unclear. This paper examines how energy tokens, primarily Powerledger, SunContract, and Energy Web Token, interact with a broad cross-section of 24 assets over the period December 2020 to January 2026, which includes major cryptocurrencies, crypto indexes, renewable energy equities, energy and metal commodities, the EU carbon allowance market, and green finance instruments. The locally stationary Bayesian time-varying parameter VAR framework is used to decompose connectedness into a transitory component of 1-5 days and a persistent component of 20-252 days, which allows shock transmission to be characterized as horizon-specific.
As a result of our network analysis, we find that energy tokens occupy a structurally peripheral position within a cryptocurrency-dominated core. At high frequencies, Powerledger and SunContract are shock transmitters; at low frequencies, they become shock receivers, while Energy Web Token acts as a persistent shock receiver across both spectrums. These findings suggest that energy tokens are primarily driven by cryptocurrency market dynamics rather than by the underlying energy fundamentals they are designed to represent.
Translating the network topology into portfolio implications, we evaluated equal-weight, global minimum-variance, and mean-variance-efficient strategies augmented with each token. Results are robust across the full sample, three crisis sub-periods (crypto market turmoil, the global energy crisis, and the recovery phase), and across short- and long-horizon decompositions. SunContract is the only token which consistently improves the out-of-sample Sharpe ratio of the global minimum-variance portfolio across all data splits; Powerledger yields broadly neutral effects, while Energy Web Token reduces long-run risk-adjusted performance. These results hold in a sustainability-constrained portfolio context, though the magnitude of the gain from SunContract is sensitive to the covariance estimation procedure employed.