Welfare and the case of recent mergers in the iron ore industry

Abstract

This paper reconstructs the view of the European Commission regarding the allowance of a merger between two Brazilian iron ore mining companies, CVRD and Caemi, using data on the DRI pellet market. By using a simulation model, it is possible to directly simulate the total welfare effects from the merger and hence evaluate the merger from a new perspective. The results from the simulations suggest that the welfare effects are negative from the merger between CVRD and Caemi, which contradicts the final conclusion drawn by the European Commission in its decision. By performing different simulations between hypothetical merger candidates, our results show that only mergers between small candidates have the potential to be welfare enhancing.

Keywords:

• Welfare
• Iron ore
• Merger
• Simulation model
• Cournot

Notes

There may arise situations were the simulation procedure is not as straightforward as described. Mathiesen and Sorgaard2000 point out two such situations. First, if the merger is conducted on beliefs about future competition conditions, the present situation is irrelevant for calibration purposes. Second, as the model is presented, it is implicitly assumed that the firms are single‐goods producers and does not account for the possibility of multi‐goods producing firms.

Ericsson2001 points out the seemingly obvious fact that efficiency gains are difficult to obtain unless the merging mines are geographically close. Considering the way some mergers in this and related markets are dealt with by authorities this fact should be pointed out regularly.

Any market with an HHI over 1800 is considered highly concentrated. These guidelines have been revised twice since then, by the 1984 and 1992 guidelines. However, while the emphasis on the HHI is weaker, in the revised guidelines the numbers still remain as benchmarks.

The marginal costs for the larger Brazilian firm as calibrated in our model were approximately 0.4, 0.76 and 0.88 for the elasticities 0.4, 1 and 2.

As is noted in Froeb and Werden1996 in a model like the one used by Hausman et al.1994 where elasticities are kept constant, the welfare effect of the merger will be overstated.