Time series data are widely used to explore causal relationships, typically in a regression framework with lagged dependent variables. Regression-based causality tests rely on an array of functional form and distributional assumptions for valid causal inference. This paper develops a...