This paper is motivated by the large fine the EuropeanCommission imposed on Volkswagen in January 1998 forobstructing sales to foreigners. The first partdiscusses the competitive framework in which the carsector operates, especially the exclusive dealershipsystem. This is followed by a detailed examination ofthe degree of price-differentiation across nationalmarkets by 24 major manufacturers. The evidencereveals VW and Audi not to be special cases. Thisfinding implies that the Commission should reconsiderits own block exemption if it wants to establisha more competitive car market along the principles ofthe Common Market.
We examine the effect of Amnesty Plus on dynamic cartel formation in a multimarket setting. The Amnesty Plus program has been adopted as part of the US Corporate Leniency Policy in 1999 and is now vigorously advertised as one of the most compelling tools in generating successful cartel investigations. Amnesty Plus is aimed at attracting amnesty applications by encouraging firms, convicted in one market, to report their collusive agreements in other markets. We show that Amnesty Plus has two opposite effects on firms' ability and incentives to collude which are both increasing in the fine reduction granted. First, Amnesty Plus can reduce the deterrence of cartel activities by making it easier for firms to sustain multimarket collusion through strategies involving self-reporting after a first cartel detection. Second, it can reduce the expected duration of cartels or delay their formation by increasing firms' incentives to self-report after a first cartel detection.
This paper assesses the performance of common estimators adjusting for differences in covariates, like matching and regression, when faced with so-called common support problems. It also shows how different procedures suggested in the literature to tackle common support problems affect the properties of such estimators. Based on an Empirical Monte Carlo simulation design, a lack of common support is found to increase the root mean squared error (RMSE) of all investigated parametric and semiparametric estimators. Dropping observa¬tions that are off support usually improves their performance, although the amount of improvement depends on the particular method used.
Based on the Canadian National Population Health Survey we estimate the effects of individual sports and exercise on individual labor market outcomes. The data covers the period from 1994 to 2008. It is longitudinal and rich in life-style, health, and physical activity information. Exploiting these features of the data allows for a credible identification of the effects as well as for estimating dose-response relationships. Generally, we find positive long-run income effects. However, an activity level above the current recommendation of the WHO for minimum physical activity is required to reap the long-run benefits.
