Regime-dependent performance of green and brown energy Exchange-Traded Funds under climate policy uncertainty: the role of investor sentiment

This study examines whether the returns of green and brown energy Exchange-Traded Funds (ETFs) are exposed to climate policy uncertainty (CPU), whether this exposure differs between low- and high-volatility market regimes, and whether investor sentiment moderates it. Using monthly data for five green and five brown US-listed energy ETFs from May 2015 to April 2026, a two-state Markov regime-switching model is estimated, with the AAII bull-bear spread employed as a measure of investor sentiment. The estimated regimes are persistent volatility states; for brown ETFs, the high-volatility regime spans the COVID-19 crash, the subsequent recovery and the 2022 energy-price shock. The most important finding is a divergence between the two groups, specifically, increases in CPU are associated with higher returns for all five brown ETFs in the low-volatility regime and for two funds in the high-volatility regime, whereas green ETF returns show no systematic relationship with the CPU index. This divergence is confirmed by a single-regime benchmark and by a green-minus-brown portfolio that accounts for the high correlation among brown funds, and it is robust to controls for extreme market episodes. However, when the November 2016 and November 2024 election months are controlled for, the low-volatility CPU coefficient remains significant at the 5% level only for IXC, so part of the brown exposure comes from these 2 months. Further, the difference in exposure between regimes is not statistically significant for any brown ETF, and investor sentiment does not significantly moderate the exposure at the 5% level, whether sentiment is measured by its change or by its level. Because the CPU index rises with uncertainty about both the tightening and the loosening of climate policy, the positive response of brown ETFs is consistent with expectations of deregulation as well as with uncertainty that delays the energy transition. The findings indicate that broad green and brown labels do not fully describe the exposure of energy ETFs to climate policy uncertainty and that this exposure should be assessed at the fund level.