Análises de impacto de intervenções para segurança viária

• Ano de defesa: 2020
• Autor(a) principal: Tabosa, Cristiane de Mesquita
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal da Paraíba; Brasil; Economia; Programa de Pós-Graduação em Economia; UFPB
• Palavras-chave em Português: Acidentes de trânsito; Séries temporais interrompidas com controle; Lei do farol aceso; Controle sintético generalizado; Programa vida no trânsito; Traffic accidents; Controlled interrupted time series; Headlights-on law; Generalized synthetic control; Life in traffic program; CNPQ::CIENCIAS SOCIAIS APLICADAS::ECONOMIA
• Link de acesso: https://repositorio.ufpb.br/jspui/handle/123456789/21173

This thesis develops two essays in health economics. The first essay seeks to assess the impact of law nº 13,290/16 (headlights-on law) on reducing traffic-accidents. Similar research in other countries has been controversial increasing uncertainty in the results of the law in force in Brazil. A controlled interrupted time series (CITS) design is used in order to perform a pre and post intervention analysis in the period from 2013 to 2019 of the monthly average of accidents during daytime, controlled by the monthly average of nighttime accidents that does not correspond to a treated time series. To increase the robustness of the result found, accidents in the South region that have more probability to accidents related to the lack of light during the day. As a result, changes after the law were found in accident rates on federal highways for 100,000 inhabitants and on average accidents per day period and severity. After accounting for the underlying trends and addressing autocorrelation problems, it was realized that the estimated results were not significant, therefore, statistically, the reduction in number of accidents was not associated with the headlights-on law. The second essay consist of an impact assessment of the Life in Traffic Program - LTP on the trajectory of road safety indicators, such as mortality rate and hospitalizations per 100,000 inhabitants, in Brazilian cities. The empirical application uses the Generalized Synthetic Control model, developed by Xu (2017). This method allowed the creation of synthetic capital adjusted to the trajectories of cities with LPT intervention in the period prior to the program. It was found that the estimates related to hospitalizations were not statistically significant, however, the results show positive impacts of LTP translated into a trajectory of reducing the mortality rate in Brazilian capitals. The average impact on the reduction over the analyzed period was approximately 2.55 deaths per 100,000 inhabitants.