"Multiple Equilibria and Identification" by Russel Cooper: A Comment

In his lecture at the Conférence des Annales, Russel W. Coope raises issues that certainly will become more and more important as models that exhibit multiple equilibria become more and more common in today economic literature. These models are appealing as they may allow economists to cast some light for instance on the mechanisms that govern the large swings in growth pace or on the reasons why similar institutions can lead to different economic situations. Russel made décisive contributions to understand some of these mechanisms. Nevertheless, a concern exists that due to their multiple equilibria, these models might not be falsified. Russel W. COOPER presents various empirical examples to illustrate that during the last ten years progress has been made. It seems possible to draw some empircal knowledge from these estimation exorcises. Notwithstanding, complete empirical investigations of these models are rare because their estimating remains difficult and economists have to resort to simplifying assumptions or calibration approach to avoid these difficulties. In this comment, at the risk of distortion, I will try to restate some of Russel W. COOPER'S points focusing on their statistical dimensions and emphasizing that recent progress in econometrics may help to tackle some of these issues.