Product differentiation models: extensions and applications.

Summary In his theory of monopolistic competition Chamberlin, 1933, postulated that products by firms are not perfectly substitutable. Since Hotelling, 1929, many models of product differentiation with "address" have been developed. We show in this thesis how these models differ and how they can be applied to contexts as different as industrial economics and political science. In the first part, we analyze horizontal integration and show that the existence of product differentiation leads to incentives to merge when the merged firms continue to produce their full set of qualities. The freedom to enter the market appears to be the only disincentive to horizontal integration. In the second part we use location models to understand the behavior of voters and candidates in the run-up to an election. We then show how it is possible to estimate the distribution of the electorate on a left-right axis and to locate the candidates on this same axis.