Asymmetric volatility in industrial production: some international evidence

Abstract

This letter reports details of tests for the presence of conditional variance asymmetries in the growth rates of monthly international industrial production series using exponential-GARCH and threshold-GARCH generalizations. We find evidence of asymmetries of EGARCH form for US industrial production, TGARCH asymmetries for German, Japanese, and US industrial production, and such that the variance of output is always greater following negative shocks to production than following positive shocks of the same magnitude. However, residual diagnostics for estimated models indicate unmodelled structure in several cases. Notable exceptions are provided by a TGARCH specification for Japan and an EGARCH model for the US, the latter confirming previous results.