In the 1970s, large increases in the price of oil were associated with sharp decreases in output and large increases in inflation. In the 2000s, even larger increases in the price of oil were associated with much milder movements in output and inflation. Using a structural VAR approach, Blanchard and Gali (in J. Gali and M. Gertler (eds.) 2009, International Dimensions of Monetary Policy, University of Chicago Press, pp. 373–428) argued that this reflected a change in the causal relation from the price of oil to output and inflation. They then argued that this change could be due to a combination of three factors: a smaller share of oil in production and consumption, lower real wage rigidity, and better monetary policy. Their argument, based on simulations of a simple new-Keynesian model, was informal. Our purpose in this paper is to take the next step, and to estimate the explanatory power and contribution of each of these factors. To do so, we use a minimum distance estimator that minimizes, over the set of structural parameters and for each of two samples (pre- and post-1984), the distance between the empirical SVAR-based impulse response functions and those implied by a new-Keynesian model. Our empirical results point to an important role for all three factors.