Inferring Causal Interactions in Financial Markets Using Conditional Granger Causality Based on Quantile Regression

Granger causality analysis emerges as a typical method for inferring causal interactions in economics variables. Yet the traditional pairwise approach to Granger causality analysis may not clearly distinguish between direct causal influences from one economic variable to another and indirect ones acting through a third economic variable. In order to differentiate direct Granger causality from indirect one, a conditional Granger causality measure is derived based on the parametric dynamic quantile regression model. The proposed method can characterize the causal interactions on the entire distribution in more detail. Simulations are carried out to illustrate that the proposed tests have reasonable size and power against a variety of empirically plausible alternatives in finite samples. An economic application considers the causal relations between the stock trading volume, the stock returns, and the exchange rates for both domestic and cross-country markets by using the daily data of three financial markets: China, Japan, and South Korea. The results demonstrate that the causal effects of stock returns and stock trading volume that exert on the exchange rate are not direct at high quantile in China; besides, no direct causality is found running from stock returns to stock trading volume at high quantile in South Korea. In contrast to Granger causality in mean and Granger causality in quantile, our results provide a comprehensive and theoretically consistent extension of Granger causality to capture causal patterns in general multivariate systems, financial markets in particular.