Search

We use a general equilibrium model to show that a decrease in workers’ bargaining power amplifies the contribution to the output gap of adjustments along the extensive versus intensive margin of labour utilization. Under standard assumptions on the disutility of labour, this mechanism reduces the cyclical movements of inflation relative to those of the output gap. Micro-level evidence, based on a survey of Italian firms, provides support to the relationship between bargaining power and adjustments along the extensive margin versus the intensive one, as well as to attenuated price response when firms adjust labour input mainly through the extensive margin. A Bayesian estimation using Italian aggregate data for the samples 1970–1990 and 1991–2014 confirms that the decline in workers’ bargaining power has weakened the inflation–output gap relationship.

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.