What if Oil was Less Substitutable?

The consequences of oil price shocks in the real economy have preoccupied economists since the 1970s and the absence of a reaction has stunned them in the 2000s. However, despite the huge literature devoted to the subject, no dynamic stochastic general equilibrium (dsge) model has been able to capture, all at the same time, four of the well-known stylized effects observed after the oil price increase of the 2000s: the absence of recession, coupled with a low but persistent increase in the inflation rate, a decrease in real wages and low price elasticity of oil demand in the short run. One of the reasons is that theoretical papers assume a high degree of substitutability between oil and other factors, an assumption that is not backed up empirically. This paper enlarges the dsge model developed in Acurio-Vásconez et al. (2015) by introducing imperfect substitutability between oil and other factors. The Bayesian estimation of the model over the period 1984:Q1-2007:Q3 suggests that the elasticities of substitution of oil are 0.086 in production and 0.014 in consumption. Furthermore, a sensitivity analysis of the estimated model points towards two main policy conclusions: (a) a stronger anti-inflationary Taylor rule can lead to a recession after an oil shock and; (b) wage flexibility could create a stronger increase in inflation and provoke a decrease in domestic consumption. This latter result contradicts the conclusions of Blanchard and Galí (2009) and Blanchard and Riggi (2013).