This paper – which can be found here – was written by a number of economists at University of Indiana, and it looks at the incentives of competition authorities to pursue international cartels.
A benevolent competition authority with a mandate to maximize national welfare might prefer to delay the prosecution of domestic firms when those firms will benefit from anticompetitive conduct at the expense of foreign consumers. To demonstrate this, the paper develops a simple two-country model of collusion and antitrust policy where there is an industry comprised of two multinational firms that operate in countries with competition agencies.
In a first variant of this model, the competition agency of the country where the effects of an anticompetitive practice are felt (country B) will discover this practice only if the agency of the country where the companies are based (country A) starts an investigation. The authors show that country B always prosecutes the colluding firms as soon as there is information of prosecution, while country A: ‘only prosecutes the firms when the gain in consumer surplus for country A’s consumers exceeds the reduction in profits (including fines) from domestic and foreign markets for country A’s firms.’ These results hold even when variations are added to this baseline model regarding: ‘the type of collusive schemes available to the firms (global versus market specific collusion); the punishment that firms can use to sustain collusion (in each market separately or in both markets); and the information prosecution spillover (prosecution in country A is indefinitely or only temporarily informative for country B).’ Under a second version of this model, country B can detect collusion independently of country A. This is a more realistic scenario, and the model demonstrates that: ‘country A will be more willing to prosecute the firms as soon as collusion is detected, reducing the international bias of the prosecution decision. ’
They then add the possibility of competition agencies investing resources in the detection of anticompetitive practices to the model. In this context, a: ‘competition authority with a mandate to maximize national welfare selects a higher detection probability if the market is served only by foreign firms than if it is served by purely domestic firms, and it selects an even lower probability if the market is served by domestic firms that also operate in foreign markets’. In common English: ‘Sectors operated by foreign companies are expected to be prime targets for antitrust cases, while sectors dominated by domestic companies with large foreign operations are expected to receive a more lenient treatment’.
A last variant of the model adds the possibility that countries may specialise in the different industries. In this scenario, each country can investigate companies from the other country. The model reveals that: ‘the multi-industry extension opens the door to mutual gains from the integration of competition authorities (…) With integration, each country loses the collusion profits in the foreign market, but it gains the increase in consumer surpluses in both domestic markets. When countries are of different sizes, this might not be enough to avoid winners and losers from integration, but in the case of perfectly symmetric countries, we prove that both countries gain with an integrated competition authority.’
As is often the case, I’m not qualified to comment on the methods (and hence, the results) of this paper. However, the results align with my intuitions on the topic – and with the existence of rules on export cartels in many jurisdictions. I also wonder whether the analysis about the ‘integration of competition authorities’ could extend to instances of international cooperation which, as we know, is a more realistic scenario in practice.