Canadian industry level production and energy prices

Highlights: • We examine which Canadian industries are most strongly affected by energy price volatility. • We estimate a heteroskedastic VAR with multivariate t distribution to account for excess kurtosis. • Mining and extraction, manufacturing and transportation are strongly affected by energy uncertainty. • Accounting for the excess kurtosis significantly improves the fit of the empirical model. • Results suggest that energy price uncertainty arises from both energy markets and macroeconomy. Energy price volatility has a negative effect on economic activity in the United States, Canada and other countries. In this paper, we use disaggregated production data to examine which Canadian industries are most strongly affected by energy price volatility. We also generalize a heteroskedastic vector autoregression with a multivariate t distribution to account for excess kurtosis in conditional energy returns. We find that Canadian industries related to mining and extraction, manufacturing and transportation are most strongly affected by energy volatility; and that accounting for the excess kurtosis significantly improves the fit of the empirical model. We include several tests of robustness, including measures of energy volatility that incorporate stochastic volatility and asymmetries in the variance process. Our results suggest that uncertainty about energy prices reflects both macroeconomic uncertainty and uncertainty about energy markets.