Published online by Cambridge University Press: 04 January 2017
Theories of voting tell us that party vote dynamics have two core components, long-run equilibria and more or less rapid returns to equilibria after short-run deviations. Statistical models used to represent these theories, however, tend to emphasize one component or the other. The unbalanced emphasis leaves the theoretical ideas underspecified and produces biased estimates of both long- and short-run components of vote dynamics. We specify the two components in the familiar form of an error correction model and demonstrate its advantages in terms of its theoretical consistency, simplicity, and precise prediction of whether something is wrong. We illustrate its usefulness by applying it to two sets of analyses reported in the literature and showing that it usually changes the conclusions reported in regard to both the equilibrium levels of party competition and the strength of restoring forces.
Authors' note: Ian Budge provided comments that sharpened our analyses and arguments in many ways. We are grateful to Yuliya Tverdova for pointing us to important insights about dynamic processes and to Silvia M. Mendes and Harvey Palmer for advice that honed our arguments and improved our presentation. Suggestions from referees helped us to clarify important points and improve our argument and presentation in several ways. Data for replication are available through the Center for Democratic Performance, Binghamton University.