Elections are the engine that drives democracy. The central question of this dissertation relates to the speed of that engine: How long does it take for elections to reflect changing preferences in the electorate? The findings presented in this dissertation suggest that electoral change is the result of a gradual process of natural selection in which the political environment, rather than district service activity, is the key variable. Comparing elections data across different types of district environment, I find evidence that the environment affects levels of competition and electoral outcomes. Utilizing an event history statistical model to examine various risk factors for electoral defeat, I find that the political environment of the district is the most important factor influencing the risk of defeat even when controlling for district service behaviors. Over time, the district environment operates as a self-correcting mechanism, purging political misfits and replacing them with representatives who better reflect the ideology of the district. Electoral change typically results more from evolution than revolution – it may not occur quickly, and it may not occur in every district, but it does occur when and where it is needed.

In developed democracies, incumbents are consistently found to have an electoral advantage over their challengers. The normative implications of this phenomenon depend on its sources. Despite a large existing literature, there is little consensus on what the sources are. In this three-paper dissertation, I find that both electoral institutions and the parties behind the incumbents appear to have a larger role than the literature has given them credit for, and that in the U.S. context, between 30 and 40 percent of the incumbents' advantage is driven by their "scaring off" serious opposition. / Government

The study of electoral defence and its stated advantages are an integral part of American political science. Post-war, much academic literature has emerged in an attempt to identify and explain rising re-election rates of congressional incumbents and the political consequences of such a phenomenon (Mayhew 1974; Fiorina 1977; Cain, Ferejohn & Fiorina 1987; Gelman & King 1990; King 1991). Conversely, the study of political incumbency in Britain can be attributed to a handful of scholars who tend to consider the repercussions at parliamentary level (Williams 1967, King 1981, Cain, Ferejohn and Fiorina 1984, Norton 1990 & 1994, Norris, Valance & Lovenduski 1992). Consequently, incumbency advantage at the local level remains a relatively under-researched topic in England, confined to the sub-chapters of Rallings & Thrasher (1997). The aim of this thesis is to research and present evidence in support of incumbency effects in English local elections and the extent to which they influence their outcome, in that, incumbent candidates fare better than less experienced candidates, to different degrees across the three major parties. It will do so using survey and electoral data collected by The Elections Centre at Plymouth University, drawing on established methods from the literature and demonstrating via a variety of data and methods, that incumbency advantage is indeed a real phenomenon effecting the outcomes of local elections in England. The research provides substantial evidence for Sophomore Surge and Retirement Slump effects throughout the period examined (1974-2010). These methods of estimation feature alongside a number of others, which are constructed to uncover the significance of defending, rather than challenging for a council seat. A number of influences on the advantage that defending councillors maintain are also presented, including district magnitude, ward size and rural/urban classification. Results reveal a modest advantage for Conservative and Labour incumbent candidates, whilst the effects are shown to be stronger for the Liberal Democrats, a finding that is in step with the existing literature on electoral trends and the local campaign strategy of the party (Dorling et al, 1998; McAllister et al, 2002; Russell & Fieldhouse, 2005; Cutts 2006).

No description available.