Maintaining sufficient levels of competition ranks among the core interests of any national – and increasingly international – antitrust policy; however, the formal proof that a cartel really functioned economically and did not only exist in a legal sense is hard to deliver: market power is not identical to the existence of a legal cartel unless the monopolistic frontier is reached; the legal proof of a cartel does not imply that the market was harmed. From an economic point of view, focusing on legal proof of a cartel is fruitless unless collusion resulted in excess profits or excess revenues. This economic evidence, however, rests empirically on the proper definition of comparable markets, and a sound statistical methodology. When in spring 2003, the German Antitrust Agency (GAA) fined the German cement industry – € 661 million for having established quotas in each of the four market regions through the end of 2001, the legal issue seemed beyond doubt as, beside formal inquiries, two of the industry members had acted as key witnesses. However, the economic implications drawn by the GAA remain doubtful. In this paper, we use the quota agreement in the East German market, the region for which these allegations are undisputed by all major suppliers, as a reference case. We challenge the GAA’s computation of excess income of 10 €/ton on two grounds: (i), the comparative market period chosen, 2002, does not meet the requirements of a reference market, especially regarding a certain level of stability and converging prices; (ii) three parallel developments could have triggered the price decline: the openly announced end of the quota cartel, which generated general price-setting insecurity (ii-a), the price war triggered by one of the oligopolists, who desperately tried to improve poor utilization of capacity and squeeze out competitors (ii-b), and the general decline in construction activity (ii-c). Within the framework of an econometric model based on data of one German cement producer, we find that sufficient levels of competition prevailed throughout the cartel period. Furthermore, the demand structure did not change from 2001 to 2002 so as to suggest a fundamental change in competition. Finally, no excess income or profit can be computed. In fact, we show that the general demand regime estimated for the period 1995 to 2001, which is the period of alleged market power, equally well describes the market condition of 2002. Price war and a collapsing construction market lead suppliers to maintain levels of production and capacity utilization, thus sacrificing profits at the expense of the market shares of small and medium-sized suppliers independently from the cartel issue. This empirical finding of an agreed but ineffective cartel is supported by theoretical evidence on the conditions under which cartels can work effectively – which did not exist in the East: strong import competition, a high level of transparency limiting the effects of „cheap talk“ and spatial pricing that generates local market power in the absence of cartels. Furthermore, general supply-side conditions in the cement industry suggest that a considerable level of imperfect competition is structurally unavoidable; antitrust possibilities that in the short run enforce additional competition based on the wrong assessment of effective collusion may lead to exits and less competition in the long run. We conclude that the methodology described may be useful for antitrust policy as it offers a credible analytical tool to compute excess income and profit.
