Manzoni, Elena
(2010)
Electoral campaigns with strategic candidates: A theoretical and empirical analysis.
PhD thesis, London School of Economics and Political Science.
Abstract
The main focus of this thesis is the analysis of political campaigns when candidates choose their statements in a strategic way. In the first chapter, 'Discretion and renegotiation in electoral campaigns', I present a model of electoral campaigning as a problem of competitive delegation. The chapter considers a situation in which there is uncertainty about what the optimal policy should be; in this environment voters may want to leave discretion to a candidate, in order to allow him to adjust his policies to the state of the world, once he is elected. The paper analyses how the ambiguity level of the political statements is influenced by the presence of uncertainty over the candidates' ideology, by the possibility of ex post renegotiation between the elected candidate and the voters and by several political variables. In the second chapter, 'Last minute policies and the incumbency advantage', joint with Stefan Penczynski, we investigate the timing of statements in political debates and campaigns. Early statements can influence the political agenda and signal competence and vision, late statements are based on more information about appropriate measures. We find that candidates speak early on issues they are better-informed about in order to signal relevance and move them up the agenda. Since opponents benefit from this revelation, however, candidates remain silent once their information is sufficiently precise and valuable. In the last chapter, 'Discretion and ambiguity in electoral campaigns: a look into the empirical evidence', I compare several models of ambiguity in electoral campaigns, including my own model which was introduced in the first chapter. I use the methodology of Campbell (1983) to have a proxy for ambiguity of the electoral statements, and the data from the American National Election Studies on Senate elections from 1988-1990-1992, to investigate which of the correlations predicted by these models seem to be present in the data.
Actions (login required)
 |
Record administration - authorised staff only |
Download statistics
Download statistics