Quantile Granger causality between US stock market indices and precious metal prices

Abstract

The current paper examines the dynamic causal relationships between US stock market indices and precious metal prices across all quantiles of the conditional distribution covering the period 1987M1-2021M3. For empirical analysis, we employ the newly developed quantile unit root test, quantile autoregressive distributed lag (QARDL) model, and Granger-causality-in-quantiles testing approach. The QARDL analysis indicates that the relationship between precious metal prices and US stock market indices is quantile dependent. Further, the empirical findings indicate that Granger causality is quantile-dependent and differs according to each precious metal. We find evidence of bi-directional Granger causality between the US stock market indices and gold prices at the lower-and-middle quantiles of the distribution. Results also support bi-directional Granger causality from changes in the US stock market indices to silver prices at the extreme lower quantiles of the distribution. Lastly, we find evidence of bi-directional Granger causality from changes in the US stock market indices to changes in platinum prices in all tails of the conditional distribution. These empirical results may have important implications for investors, commodity market players and regulators, and policy makers.