We propose a polynomial logit model to quantify the price effects of mergers in a static Nash setting. The proposed model is parsimonious in parameters and is shown to have excellent predictive power, rivaling the in-sample and out-of-sample predictive accuracy of the widely-used AIDS model. The analysis, using actual scanner data on bread sales, demonstrates that a linear logit model is likely to over-estimate the merger price effect.

Download this paper.

“Predicting the Price Effect of Mergers with Polynomial Logit Demand,” A. Saha, P. Simon, International Journal of the Economics of Business, Antitrust Special Issue, 7(2) (2000), 149-157.