Use this URL to cite or link to this record in EThOS:
Title: Political interactions and voter responses
Author: Tucker, Luc
ISNI:       0000 0004 5348 3042
Awarding Body: University of Warwick
Current Institution: University of Warwick
Date of Award: 2014
Availability of Full Text:
Access from EThOS:
Access from Institution:
This thesis investigates empirically the way in which agents in political bodies can influence their peers, as well as the ways in which voters respond to the behavior of legislators in their electoral choices. These relationships are fundamental in trying to comprehend the way in which political decisions are made. Economists should take particular interest in these topics, given their importance in understanding the incentives faced by legislators. Questions as to the possibility of peer effects between political agents are of huge importance in democratic governments. Political debates play a central role in many legislative bodies, where the assumption is implicitly made that opinions can be influenced by the other debate participants. This fundamental assumption is tested in Chapter 1, which is the first to measure the extent of peer influence regarding reported political opinions in an explicitly political environment. This has previously not been possible, given that discussions and debates in legislative chambers take place between participants with particular characteristics and political interests, making it hard to separate the role of peer effects in determining their preferences. This thesis makes use of experimental data, which offers a unique opportunity to distinguish these effects and quantify the degree to which peer effects can influence political preferences. In particular, Chapter 1 uses data from an experiment conducted in Australia in 2009 to consider whether participants showed evidence of having influenced one-another during political discussions. Each of the models used exploits the fact that table allocations were randomized in this experiment and controls for agents’ characteristics, which were also recorded. The key finding of this chapter is that when asked to assign weights to eleven criteria for an effective political system, agents who sat on the same table during the experiment reported preferences that were more similar than those who did not share a table. The effect is small at 4.8% of a standard deviation but is statistically significant and of larger magnitude than other pairing characteristics which could have been expected to influence the differences between weighting choices, such as whether the two players were of the same gender. One year after the Citizens’ Parliament, participants were asked to report their political positioning on the ‘left-right’ scale. It is not found to be the case that the table allocations influenced these reported positions. Having demonstrated that participants in legislative bodies can influence one-another’s reported political preferences, this thesis goes on to analyze the relationship between legislators and the constituents they represent, by considering the question of whether politicians who are more active in parliament are rewarded with a higher probability of being reelected. The particular parliamentary behavior analyzed is the asking of parliamentary questions. The UK House of Commons uses a ballot system to determine which members are selected to ask a question from those who expressed an interest in doing so. This chapter is the first in the literature to exploit this randomization to show that the asking of such questions increases a member’s chances of being reelected by their constituents. It is shown that while the ordering of parliamentary questions is determined at random, the practicalities of conducting debates introduce a potentially endogenous element to the determination of which questions receive oral answers (particularly the speed at which questions are answered). This chapter uses a matched sampling approach to cope with such non-random cases, but also includes alternative results, to show that the findings are not reliant on the use of this technique. Chapter 2 exploits a natural experiment to show that Members of Parliament who are selected to ask parliamentary questions are more likely to be reelected in forthcoming elections. It was necessary in this study, however, to drop certain observations as a result of the fact that the Speaker in the House of Commons, who chairs debates, has some influence over the number of questions reached in each debate, which could undermine the randomization in these cases. Chapter 3 of this thesis goes on to consider this process in more detail. This chapter shows that in fact questions posed by older and more experienced members, as well as those from opposition parties, are more likely to receive oral answers than should be expected under a true randomization. Chapter 3 offers the first opportunity to consider the Speaker’s role in parliamentary debates under the conditions of a ‘natural experiment’. Results presented here point to the role of the Speaker in controlling the speed at which debates progress as contributing significantly to the findings listed above, for example by acquiescing to pressure from more senior members by allowing them to ask their questions in debates where time constraints would otherwise prevent them from doing so. The finding is also an important consideration for future studies which aim to exploit such randomizations as natural experiments relating to parliamentary activity. Such a finding is potentially significant in the context of the UK political system, where the ballot system is in place precisely to ensure that all members of the House of Commons have an equal opportunity to ask questions, regardless of their levels of seniority. The final chapter of this thesis continues to examine the link between legislators and the citizens they represent. In particular, Chapter 4 makes use of the large Behavioral Risk Factor Surveillance System (BRFSS) dataset from the USA. While this dataset has been extensively used to study health outcomes, this chapter represents the first attempt to use the dataset to study the link between political outcomes and the economic prosperity of constituents. This is achieved by matching survey respondents to their representatives in Congress and restricting attention to cases where members of the lower house seek election to the upper house. Members of the US Senate (the ‘upper house’ in Congress) are elected to serve a state as a whole, whereas members of the House of Representatives (the ‘lower house’) serve a district within one of those states. This chapter shows that members of the House of Representatives who seek election to the Senate (without necessarily being successful) tend to have previously served in districts with permanently higher incomes. Furthermore, incomes are found to be temporarily higher in districts where the representatives are successfully elected to the Senate than those where the representatives were unsuccessful in their attempt to be elected. This is interpreted as showing that in Senate elections, voters reward legislators who served districts where average incomes were seen to increase under their tenure. These chapters use a diverse range of datasets to consider the impacts of political behavior. It is shown that the behavior of agents in political environments not only influences their peers, but is also recognized and rewarded by the voters they represent. Voters are found to respond to political behavior by both reelecting legislators who are more active (by asking more parliamentary questions) and by electing those legislators who have previously served districts where average incomes increased under their tenure.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available
Keywords: JA Political science (General)