Discussion Paper

No. 2012-18 | March 23, 2012
Cartel in the Indian Cement Industry: An Attempt to Identify It


This article is devoted to the problem of the detection of overt or tacit collusion equilibrium in the context of the choice of the appropriate econometric method, a choice that is determined by the amount of information that the observer possesses. The author addresses this problem in two steps. First, to provide a theoretical background, he uses a collusion marker based on structural disturbances in a price process’ variance. Then, he applies a Markov switching model with switching in variance regimes. The author considers this method adequate and coherent with the problem structure and the research objective, and useful for assessing the functionality of the collusion marker he uses. He uses the model to examine the Indian cement industry in the period 1994–2009 and finds some objective indications of collusion and competition phases. These phases are confirmed by certain historical facts as well as by numerous research articles.

JEL Classification:

C22, L13, L61


  • Downloads: 4128


Cite As

Sylwester Bejger (2012). Cartel in the Indian Cement Industry: An Attempt to Identify It. Economics Discussion Papers, No 2012-18, Kiel Institute for the World Economy. http://www.economics-ejournal.org/economics/discussionpapers/2012-18

Comments and Questions

Anonymous - Referee Report 1
March 27, 2012 - 11:29

see attached file

Anonymous - Referee Report 2
April 18, 2012 - 09:30

See attached file

Katharine Rockett - Decision Letter
June 12, 2012 - 11:48

Dear Author:
Thank you for submitting “Cartel in the Indian Cement Industry: An Attempt to Identify It”, MS 699. The paper uses a GARCH methodology to attempt to identify periods of relatively low price volatility, which may correspond to collusive phases in the cement industry. The data set in this ...[more]

... paper is of longer duration and higher frequency than previously used data with this method. Furthermore the results can be held up against regulatory investigations to check whether the technique predicts accurately. Some support for the methodology is found.
We now have two anonymous referee reports on your paper, and I have read the paper myself. While the subject matter is interesting and potentially important, we all have significant concerns about the rigour and accuracy of the treatment. Referee 1 is concerned that, for a general interest journal, this paper gives too little background to readers on why the methodology is appropriate to identify cartel behaviour. The referee also has a large number of clarifying questions about the data and techniques used. Both Referee 1 and Referee 2 are concerned that the paper does not adequately control for other causes of changes in volatility. Referee 2 also has concerns about econometric technique, in particular the handling of seasonality.
I share the referees’ concerns both about technique and about the significance of the results. Even though the results of this method correspond somewhat to the regulatory investigations, the methodology used without controls for other causes of changes in variability is not particularly robust. For both academics and regulators, identifying changes in variability without being sure of their causes is unsatisfying: for academics, it cannot be used to understand the industry better; for regulators, it cannot be used as a basis for a prosecution. Both adding controls and, as suggested by referee 2, applying this to data outside the cement industry to ensure that the test both rejects where it should and accepts where it should are important steps to follow. In terms of technique, there is a lot of “clean up” that needs to be applied before we can even be sure of the results within the model as stated. In short, there is considerable work on model specification and technique to be addressed before the results can be compelling.
In terms of publication, the work required to make a compelling case would take considerable time and more data. This means that timely acceptance, even if the work is undertaken, would be difficult. I would recommend that we leave the paper in “working paper form”, then. Should you decide eventually to undertake work to address the referees’ and my concerns, then we could consider a re-submission of a new paper at that point.
I hope that you have found this report and those of the referees helpful.


Katharine Rockett