Dynamics and causality in industry-specific volatility

This paper presents comprehensive empirical evidence on the dynamics and causality within 30 US industry-specific volatilities during July 1963 and June 2008. We find that linear trends are present in 17 of the 30 industry volatilities. Granger-causality tests reveal that the industry of business supplies and the industry of finance are the most important lead indicators of industry volatilities. To uncover contemporaneous causal relationships in the market, we implement an emerging data-driven method of directed acyclic graphs. The results suggest that volatility shocks originating from business supplies, machinery, and consumer goods industries are sources of risks that affect most other industries. By contrast, volatilities in the two traditionally important industries, oil and autos, do not appear to have a substantial influence on other large industries in the contemporaneous time. Finally, business equipment and services, both containing information technology components, are the most important driving forces of the industry volatility surge in the late 1990s.