Persistent Policy Uncertainty and Green Energy Valuation: A Long-Run ARDL Analysis of the CELS Index

Macroeconomic and Policy Uncertainty Effects on the Clean Edge Green Energy Index. Data Description
This study employs monthly data spanning from January 2008 to June 2025. The dependent variable is the Clean Edge Green Energy Index (CELS) obtained from NASDAQ, which serves as a proxy for U.S. renewable energy stock performance. Explanatory variables include:
• Climate Policy Uncertainty Index (CPU): U.S.-specific climate policy uncertainty index developed by Gavriilidis (source).
• Brent Crude Oil Price (OIL): Extracted from the Federal Reserve Bank of St. Louis (FRED: DCOILBRENTEU).
• BIS Broad Dollar Index (BIS): Monthly trade-weighted U.S. dollar index from the Bank for International Settlements.
• Global Policy Uncertainty Index (GPU): World-level policy uncertainty index constructed by Baker, Bloom, and Davis.
• 10-Year Treasury Yield (YIELD): Monthly U.S. government bond yield from FRED(FRED: DGS10).
• Volatility Index (VIX): Market-wide implied volatility, also from FRED.
To ensure stationarity and consistency with empirical finance literature, all asset price variables, including the NASDAQ Clean Edge Green Energy Index (CELS), Brent crude oil prices, and the BIS Broad Dollar Index, are transformed into returns. These returns are calculated as log differences of monthly closing values. Using returns instead of price levels reduces the risk of spurious regression because financial price series are often integrated of order one (I(1)), while their returns are usually stationary. This method also allows for interpreting results in percentage change terms, which provides a clearer view of market dynamics and relative changes across asset classes. Such a transformation is common in energy and financial econometrics. It ensures comparability and solid inference when examining the relationships among exchange rates, commodity prices, and renewable energy equity performance.
To capture extraordinary events, three dummy variables were introduced.
• DUM_PARIS (2016M12 onward): Represents the Paris Agreement implementation phase.
• DUM_COVID (2020M03–2020M05): Captures the initial market panic of the COVID-19 pandemic.
• DUM_UKR (2022M02-2022M06): Reflects pulse of geopolitical shocks from the Russia–Ukraine conflict.
These dummies are coded as 1 during the respective periods and 0 at other times. Their inclusion makes sure that results are not influenced by structural changes caused by these extraordinary