Bayesian spatial voting model to characterize the legislative behavior of the Colombian Senate 2010-2014

TL;DR

This paper employs Bayesian methods to analyze Colombian Senate voting behavior from 2010-2014, revealing a latent opposition trait and identifying key legislators, thus enhancing understanding of legislative dynamics in unbalanced parliaments.

Contribution

It introduces a Bayesian ideal point estimation approach tailored for unbalanced parliamentary contexts, specifically applied to the Colombian Senate.

Findings

Identified a latent opposition trait influencing votes.

Revealed the political space dimensionality of the Senate.

Pinpointed pivot legislators critical to legislative outcomes.

Abstract

This paper applies Bayesian methodologies to characterize the legislative behavior of the Colombian Senate during the period 2010-2014. The analysis is carried out through the plenary roll call votes of this legislative chamber. In addition, parliamentary electoral behavior is operationalized by implementing the one-dimensional standard Bayesian ideal point estimator via the Markov chain Monte Carlo algorithms. The results contribute mainly to two points: political space dimensionality and the identification of pivot legislators. The pattern revealed by the estimated ideal points suggests a latent non-ideological trait (opposition - non-opposition) underlying the vote of deputies in the Senate. Thus, in addition to providing empirical evidence for a better understanding of legislative policy in Colombia during the period under analysis, this work also offers methodological and theoretical tools to guide the analysis of roll call vote data in contexts of unbalanced parliaments (as opposed to the U.S. parliament), taking the particular case of the Colombian Senate as a reference.