Using causal graphs to test for the direction of instantaneous causality between economic policy uncertainty and stock market volatility

This paper uses causal graphs to test for the direction of instantaneous causality between economic policy uncertainty (EPU) and stock market volatility. The paper derives simple test regressions from a causal graph, applies the regressions to data for 23 countries, and shows how causal graphs can be used to examine the robustness of the results. The empirical results imply that stock market volatility is not a cause of EPU and that domestic EPU and US EPU are often instantaneous causes of stock market volatility. A comparison of the graph-based approach with a VAR-based causality analysis demonstrates that both approaches yield consistent statistical results. However, in contrast with the VAR approach, the graph-based approach utilizes the additional information provided by the causal graph to infer the direction of instantaneous causality.