Departing from the prior literature on the energy consumption-economic growth linkages, this paper aims to quantify the asymmetric impact of unpredictable variations in energy use on GDP (gross domestic product) volatility across the high- and low-volatility regimes in the United Kingdom. For this purpose, the paper utilizes quarterly and monthly data on energy consumption and real GDP, covering the period from 1960 to 2013. To obtain the regime-dependent conditional volatility series, we estimate the two-state Markov regime switching ARCH model. We find significant evidence of the presence of regime shifting in behaviour of both underlying volatility series. By estimating the Markov regime switching model, we find that not only the current (contemporaneous) but also the past (intertemporal) volatility shocks in energy utilization are significant in explaining the output dynamics. Yet, we show that the impact of energy consumption volatility is quite asymmetric, depending on the intensiveness of volatility. Specifically, we show that although the effect of energy consumption volatility on GDP volatility is statistically insignificant in the low-volatility regime, in the high-volatility regime, it appears highly significant. From a policy perspective, our findings suggest that controlling unwanted variations in energy usage would be useful in stabilizing real sector of the economy. (C) 2016 Elsevier Ltd. All rights reserved.