Presidencialismo e governabilidade no Brasil: formação de gabinetes, distribuição de poder através do controle orçamentário e os impactos sobre a base de apoio ao Executivo (2003-2018).

Abstract

This master’s dissertation aims to understand the governability of Brazilian presidents (Lula, Dilma, and Temer) between 2003 and 2018 through an empirical investigation of incentives provided to the Executive and Legislative branches for mutual support — based on political parties’ theories. This work is accomplished by analyzing the impact of the distribution of power to political parties through control over a portion of the public budget and the allocation of ministries in the presidential cabinet. Governability is measured by the variable known as "governismo" (or “legislative success”), represented by the correspondence between the voting guidelines in roll-call votes and the votes of federal parliamentarians in the Chamber of Deputies that follow the government's direction. The purpose is to identify patterns of political actors' behavior through descriptive and inferential analyses. The last one is conducted to test the research hypothesis that the length of a presidential term reduces the governability of the president. The data shows that the governability of the president was consistently high, except in specific periods. The Lula administration achieved superior governability compared to other administrations. Party dissatisfaction with the distribution of power — control over ministries and their associated resources — was a frequent reason for leaving the government's allied base, impacting the “governismo”. The research presented particularities in relation to some parties and certain ministries controlled by them. One parameter that proved valid for measuring governability was the variable "difference of governismo," used to indicate the level of importance, to the government, of the parties that comprise its support base. The comparative measure is the difference between the governability of all parties and the governability of those parties that occupy the presidential office. The time course variable (“decurso do tempo”), which measured whether governability is reduced as the presidential term progresses, did not have a well-defined pattern of behavior; that is, there is no specific moment during the presidential term when party resistance is greater or lesser. In three of the five presidential terms analyzed (Lula's first term, Dilma's second, and Temer's), the least resistance occurred during the first term of office. Although the null hypothesis was rejected – which suggests that the greater the distance in days until the end of the presidential term, the greater the governability (“sucesso legislativo”) –, the explanatory power of the variable was low in all cases. The findings may contribute to a better understanding of the functioning of Brazilian presidentialism and the obstacles in the decision-making process that hinder the presidential administration's governability.