Measuring the effectiveness of anti-cartel interventions in the shadow of recidivism¹

It is important to measure the effectiveness of Competition Authority (CA) anti-cartel enforcement, recognising that the total effect may be larger than the direct effect routinely measured by CAs because of the indirect/behavioural effects of interventions—primarily on deterrence. However, existing measurement methodologies assume that prosecuting and penalising cartels brings price fixing in an industry to an end forever. It is increasingly recognised that following successful prosecutions, collusion may re-emerge and that the extent of such recidivism depends on the structure of post-prosecution interventions. Failure to allow for possible re-emergence could produce biased measures of CA effectiveness. We develop a framework for measuring the effectiveness of anti-cartel interventions that admits that recidivism could arise depending on the nature of post-prosecution interventions. Our general model nests the no recidivism assumption as a special case and, hence, improves upon the existing methodologies. The new framework enables us to measure the extent of bias arising from the failure to allow for recidivism, and we show that it can be significant. We make a number of other significant extensions to the existing frameworks. In particular, we allow for indirect price effects as well as indirect deterrence effects and analyse the marginal effects of CA interventions.