Information, Disinformation and Lobbying in a Median Voter Model

Candidates run for (a one-shot) election, adopting a platform of a single-dimensional policy variable p ∈[0,1]. Voters with heterogeneous single-peaked preferences over policies know candidates’ announced platforms, but perceive that the actual policy which would be implemented should that candidate be elected is a random variable (with mean equal to the announced platform) whose distribution is common knowledge. Under certain restrictions, when candidates can announce any platform, a median voter equilibrium results. When candidates can costlessly modify the perceived distribution of actual policies around the announced platform policy, a median voter equilibrium also results, but under certain circumstances, candidates may choose an uninformative distribution in equilibrium. When modifying the perceived distribution is costly, we assume it is undertaken with the support of lobbying groups with an interest in the outcome. In general, we show that with two opposed lobbying groups, one will wish to provide more information to voters (reduce uncertainty) while the other will wish to provide disinformation to voters (increase uncertainty). If the strengths of the two groups are symmetric around the default distribution, then both will expend lobbying resources but no new net information will be provided. Under some conditions, the equilibrium results in greater uncertainty for both candidates than the default distribution. * Professor, Public Policy and Management Department, Wharton School, University of Pennsylvania, Philadelphia, PA 19103. This research was conducted at the Institut d’Análisi Económica, Belleterra, Spain; the author is grateful for the Institut’s support. The author thanks Carmen Matutes, Joan Estaban, Giacinta Cestone and Nicolas Melissas of the Institut for their comments and suggestions. E-mail: [email protected]. INFORMATION, DISINFORMATION AND LOBBYING IN A MEDIAN VOTER MODEL