Fractional Integration and Political Modeling

Description:

This dissertation investigates the consequences of fractional dynamics for political modeling. Using Monte Carlo analyses, Chapters II and III investigate the threats to statistical inference posed by including fractionally integrated variables in bivariate and multivariate regressions. Fractional differencing is the most appropriate tool to guard against spurious regressions and other threats to inference. Using fractional differencing, multivariate models of British politics are developed in Chapter IV to compare competing theories regarding which subjective measure of economic evaluations best predicts support levels for the governing party; egocentric measures outperform sociotropic measures. The concept of fractional cointegration is discussed and the value of fractionally integrated error correction mechanisms are both discussed and demonstrated in models of Conservative party support. In Chapter V models of presidential approval in the United States are reconfigured in light of the possibilities of fractionally integrated variables. In both the British and American case accounting for the fractional character of all variables allows the development of more accurate multivariate models.

Creator(s): Lebo, Matthew Jonathan
Creation Date: August 1999
Partner(s):
UNT Libraries
Collection(s):
UNT Theses and Dissertations
Usage:
Total Uses: 135
Past 30 days: 65
Yesterday: 3
Creator (Author):
Publisher Info:
Publisher Name: University of North Texas
Place of Publication: Denton, Texas
Date(s):
  • Creation: August 1999
  • Digitized: June 13, 2007
Description:

This dissertation investigates the consequences of fractional dynamics for political modeling. Using Monte Carlo analyses, Chapters II and III investigate the threats to statistical inference posed by including fractionally integrated variables in bivariate and multivariate regressions. Fractional differencing is the most appropriate tool to guard against spurious regressions and other threats to inference. Using fractional differencing, multivariate models of British politics are developed in Chapter IV to compare competing theories regarding which subjective measure of economic evaluations best predicts support levels for the governing party; egocentric measures outperform sociotropic measures. The concept of fractional cointegration is discussed and the value of fractionally integrated error correction mechanisms are both discussed and demonstrated in models of Conservative party support. In Chapter V models of presidential approval in the United States are reconfigured in light of the possibilities of fractionally integrated variables. In both the British and American case accounting for the fractional character of all variables allows the development of more accurate multivariate models.

Degree:
Level: Doctoral
Discipline: Political Science
Language(s):
Subject(s):
Keyword(s): fractional cointegration | multivariate models
Contributor(s):
Partner:
UNT Libraries
Collection:
UNT Theses and Dissertations
Identifier:
  • OCLC: 45054402 |
  • UNTCAT: b2228292 |
  • ARK: ark:/67531/metadc2229
Resource Type: Thesis or Dissertation
Format: Text
Rights:
Access: Public
License: Copyright
Holder: Lebo, Matthew Jonathan
Statement: Copyright is held by the author, unless otherwise noted. All rights reserved.