Cartel detection is one of the most basic and most complicated tasks of competition authorities. In recent years, however, variance filters have provided a fairly simple tool for rejecting the existence of price-fixing, with the added advantage that the methodology requires only a low volume of data. In this paper we analyze two aspects of variance filters: (i) the relationship that can be established between market structure and price rigidity, and (ii) the use of different benchmarks for implementing the filters. This paper addresses these two issues by applying a variance filter to a gasoline retail market that is characterized by a set of unique features. Our results confirm the positive relationship between monopoly and price rigidity, and confirm the variance filter's ability to detect non-competitive behavior when an appropriate benchmark is used. Our findings should serve to promote the implementation of this methodology among competition authorities, albeit in the awareness that a more exhaustive complementary analysis is required.
|Number of pages||26|
|Journal||Review of Industrial Organization|
|Publication status||Published - Oct 2012|
- Competition policy
- Gibbs sampling
- Variance filter