Stilianos Fountas, Ph.D., Economics, University of Macedonia, 156 Egnatia Street, Thessaloniki, 540 06, Greece, Donal Bredin, Ph.D., Banking and Finance, University College Dublin, School of Business, Blackrock, Dublin, 4, Ireland, and John Elder, Ph.D., Finance, North Dakota State University, College of Business, NDSU Dept 2420, PO Box 6050, Fargo, ND, 58108-6050.
There has long been an interest in the effects of uncertainty about energy prices on economic activity. Early theoretical foundations show that uncertainty about energy prices will induce optimizing firms to postpone irreversible investment decisions provided the expected value of additional information exceeds the expected short run return to current investment. Such shifts in expenditures can have large effects on aggregate output and employment if there are substantive frictions in the sectoral reallocation of labor and capital. An interesting feature of these models is that the effects of oil price increases and decreases are not symmetric: both oil price increases and decreases may be contractionary in the short-run.
In this paper, we investigate the output effects of oil prices and uncertainty about oil prices in G-7 using monthly data for the 1974-2007 period by utilizing a simultaneous equations model that accommodates both effects. The model is based on a structural VAR modified to accommodate multivariate GARCH-in-Mean errors. We measure uncertainty about the impending oil price as the conditional standard deviation of the forecast error for the change in the price of oil. This investigation is interesting and relevant for a number of reasons. First, our results provide direct evidence on role of oil price uncertainty and whether the response of output to oil shocks is asymmetric, which are issues of considerable recent interest (Kilian, 2008). Second, applying this empirical model to the G-7 provides a test of robustness of Elder and Serletis (2008), who find that oil price uncertainty adversely affects consumption, investment and output in the US. Third, the cross section of G-7 countries offers a diverse pattern of oil consumption, oil exports and economic conditions. For example, oil expenditures as a share of GDP for the US were 4.8% in 2003 and as high as 8% in the early 1980's. Such expenditures for both the US and Canada are considerably larger than for the remaining G-7 countries. Our sample also includes two countries that were oil exporters over much of our sample: Canada (since the mid 1980's) and the UK (prior to about 2005). Finally, applying this empirical model to the G-7 also provides additional insight into whether the apparent asymmetry in the response of US output to oil prices is actually due to domestic factors, such as US tax legislation, as suggested by Kilian (2008).
Our primary result is that oil price uncertainty has had a negative and statistically significant effect on output growth in four of seven countries (Canada, France, UK and the US). The results for three countries (Canada, UK and the US) are remarkably robust to various assumptions related to stationarity, as well as simplifications to our baseline VAR with Multivariate GARCH. Impulse-response analysis suggests that, in the short-run, both positive and negative oil shocks may be contractionary. Our result helps explain why the sudden collapse in oil prices in the mid-1980's failed to produce rapid expansion in the G-7, and why the steady increases in oil prices from 2003-2007 did not induce recessions.