Banca de QUALIFICAÇÃO: HELMER ARAÚJO FREIRE DO NASCIMENTO

Uma banca de QUALIFICAÇÃO de MESTRADO foi cadastrada pelo programa.
STUDENT : HELMER ARAÚJO FREIRE DO NASCIMENTO
DATE: 23/10/2026
TIME: 09:00
LOCAL: Videoconferência
TITLE:

Government Financial Condition and Determinant Factors: An Analysis of Fiscal Stress in Brazilian States.

KEY WORDS:

GFC. Fiscal stress. Economic cycles. Brazilian states. Random Forest.

PAGES: 99
BIG AREA: Ciências Sociais Aplicadas
AREA: Administração
SUBÁREA: Ciências Contábeis
SUMMARY:

This dissertation comprises two articles. The first article presents a literature review on Government Financial Condition (GFC) covering the period from 2015 to 2026, based on a final sample of 955 documents. The objective was to identify key articles, trends, and contributions within the academic literature on this topic. To this end, a bibliographic Search was conducted in Web of Science and Scopus databases using R software, followed by the processing and analysis of identified Keywords and Key aspects of the scientific output. The second article analyzed the influence of macroeconomic, environmental, demographic, and fiscal factors on the occurrence of fiscal stress in Brazilian states between 2015 and 2023. The study examined these subnational entities over a nine-year period, resulting in a total of 243 observations. Variables identified and grounded in the GFC literature review were used for this analysis. More broadly, Business Cycle Theory was incorporated to enhance the understanding of how economic activity influences the financial condition of these entities and, consequently, the occurrence of fiscal stress. Empirical analysis involved multinomial logistic regression applied to panel data, tracking the federal units throughout the study period. The results demonstrate that the themes and Keywords identified in Article I provided the foundation for the factors investigated in Article II, enabling an analysis of relationships regarding the determinants of financial condition that have been little explored in the literature. Among the variables analyzed in the multinomial model, export dependende (DEXP), population, Gross Domestic Product (GDP), and interest rate (SELIC) yielded significant results, highlighting their importance for understanding and their potential for future research and modeling regarding Government Financial Condition. As a complementary methodology, the Random Forest algorithm was employed, with Variable importance assessed via Mean Decrease Accuracy (MDA) and Mean Decrease Gini (MDG) metrics. Among the variables analyzed, the Aging Index stands out as a significant factor within the model. Thus, the results reinforce the potencial of the Random Forest algorithm as a complementary approach to multinomial logistic regression, offering and additional perspective on the relevance of the explanatory variables to the study.
KEYWORDS: GFC. Fiscal stress. Economic cycles. Brazilian states. Random Forest.

COMMITTEE MEMBERS:
Interno - 3220688 - ALEXANDRO BARBOSA
Externo à Instituição - CARLOS RENATO THEOPHILO - UNIMONTES
Interno - 1149632 - VICTOR BRANCO DE HOLANDA