Banca de DEFESA: CASSIO RODRIGO DA COSTA ALMEIDA

FINANCIAL POWER AND STATE ELECTIONS IN BRAZIL IN 2014, 2018 AND 2022

Political Geography. Electoral Geography. Accounting. State Elections. Brazil.

The research aimed to analyze the importance of financial aspects in the majority elections of
the Brazilian Federative Units (UFs) in the 2014, 2018 and 2022 elections. As for the specific
objectives, the aim was to: identify the sources of revenue for the majority electoral campaigns
of the UFs ; analyze the composition of the expenditure of the majority electoral campaigns of
the UFs; and analyze the influence of independent variables and controls on the success of the
majority electoral campaigns of the UFs. Our hypothesis is that the financial variable is a factor
that influences electoral campaigns, enabling greater reach to voters and the electoral success
of candidates. As for technical procedures, historical comparative analysis was adopted and the
sample was defined as the candidacies of those elected and not elected for the position of
governor of the 27 UFs in the 2014, 2018 and 2022 elections. Data collection took place through
DivulgaCand (TSE). To verify the spatial behavior of data through the location, interaction and
representation of phenomena, we used the open software Qgis version 3.32.0. To build the
statistical model, the Logistic Regression (Logit) technique was used. As for the dependent
variables, it was characterized as the elected and non-elected candidates for the position of
governor. For the independent variables, income and expenses, financial and non-financial,
were used. As for the control variables, they were: re-election, gender, candidate's assets, IVS
– infrastructure and per capita income. The results showed that the variable “total income”
received by the candidates did not reach a good level of significance, that is, it did not influence
the case of those elected. Likewise, the variable “financial income” did not reach a good level
of significance. Considering the variable “total expenditure”, it also did not reach a good level
of significance in the case of elected officials. However, the variable “financial expenditure”
showed that as people spend more, their probability of being elected increases by 1%.
Regarding the control variables, it can be identified that the “assets” of the candidates showed
that the greater their resources, the greater the odds ratio of success. Likewise, when a candidate
seeks “re-election”, his odds ratio increases. In percentage terms, this increase represents a
15.14% greater chance of being elected. Regarding “gender”, it was identified that when
candidates are male, their odds ratio of success is higher. As for the “IVS - infrastructure
dimension”, of the cities where the candidates lived during the electoral period, the higher this
index, the higher their odds ratio of success. Likewise, “per capita income”, because the higher
this indicator is in the states in which they competed, the odds ratio of success increases by 1%.
That said, the research hypothesis was answered, as we identified that the candidate who spends
more has a 1% greater chance of success. Thus, the hypothesis tested was considered positive,
combined with the other variables. It is concluded that the proposed model offered sufficient
subsidies to analyze the importance of financial aspects in majority elections in Brazilian states.