🤖 AI Summary
This paper addresses three fundamental challenges in climate econometrics: model sensitivity to outliers, neglect of temporal dependence, and absence of principled model selection. Methodologically, it proposes a reconstruction involving robust data cleaning, nonparametric time-trend controls (e.g., splines and local regression), and a rolling-window out-of-sample forecasting validation framework spanning 700+ variables. Key contributions include the first empirical finding that mainstream climate variables—such as mean temperature—exhibit negligible predictive power, whereas humidity-related variables demonstrate superior robustness; this motivates a new evaluation standard centered on predictive validity. Results falsify several widely accepted climate–economic relationships and reveal that even optimal predictors explain only a limited fraction of variation, raising serious concerns about the empirical foundations of the field. The study advances climate econometrics toward a data-driven, validation-first methodological paradigm.
📝 Abstract
We assess empirical models in climate econometrics using modern statistical learning techniques. Existing approaches are prone to outliers, ignore sample dependencies, and lack principled model selection. To address these issues, we implement robust preprocessing, nonparametric time-trend controls, and out-of-sample validation across 700+ climate variables. Our analysis reveals that widely used models and predictors-such as mean temperature-have little predictive power. A previously overlooked humidity-related variable emerges as the most consistent predictor, though even its performance remains limited. These findings challenge the empirical foundations of climate econometrics and point toward a more robust, data-driven path forward.