A Dynamic Analysis of Collusive Action: The Case of the World Copper Market, 1882-2016

We advance a new framework for investigating the dynamic effects of collusion. In contrast to the standard reduced-form workhorse model, a structural vector auto-regressive model with sign restrictions allows us to endogenize cartel action and to distinguish unexpected market manipulations from other types of shocks. Utilizing a newly constructed monthly data set for the copper market from 1882 to 2016, we find that cartel action shocks have strong effects on price and output during collusive periods. More notably, these shocks have lessening, yet quite persistent impacts over the subsequent unwinding periods in which output damages dominate price damages.