This dissertation consists of three essays. The first essay looks at the impact of biomass burning on health. It exploits variation in sugarcane burning in Brazil to identify the effect of pollution from biomass burning on respiratory morbidity. The second and third essays investigate how politics play a role in defining the quality and quantity of local health and school services. In particular, the second essay looks at whether variation of representation in the Brazilian Chamber of Deputies can explain differences in local output of local public goods and services. The third and last essay investigates whether term-limits affect the performance elected politicians in terms of the quality of local health care services and schools provided. Connecting all three essays is the fact that they all relate to Brazil and analyze different factors that influence the quality of health locally. Below are the title and individual abstracts for each of the three essays.Chapter 1: Biomass Burning and Health: The Case of the Brazilian Sugarcane IndustryBrazil has a long tradition in growing sugarcane, a feedstock used to produce both sugar and ethanol. Over the past decade, mainly due to the advent of flex-fuel cars, production has increased sharply. However, this monoculture has brought several concerns, one of which is air pollution. Although harvesting technology has advanced, the practice of pre-harvest burning is still common. Using geographic information system and remote sensing data on sugarcane burning together with Brazilian public health information on respiratory morbidities at the municipal level, this paper estimates the effects of sugarcane biomass burning on health. Two key sources of variation in sugarcane burning are explored: the variation between harvesting periods, and the variations due to a recent policy in the State of São Paulo that has inhibited the practice of pre-harvest burning. It is found that a one unit increase in fire occurrences on burned sugarcane fields nearby leads to a 0.23% increase in the number of children’s respiratory morbidities. Moreover, during harvesting months, municipalities that have sugarcane in their surrounding on average present more cases of respiratory morbidities than municipalities further away from sugarcane plantations. This research contributes to the growing literature on ambient air pollution externalities. Furthermore, by exploring GIS, it shows how such tools can be incorporated in order to overcome problems in measuring air pollution when precise data are not available. Lastly, this study adds to the debate on the social and environmental impacts of biofuels.Chapter 2: Representation, Pork-Barrel Politics, and Local Provision of Public Goods and Services in BrazilThis paper investigates the impact of political representation on the local provision of public goods and services. By exploiting variation across terms in the number of elected deputies in the Brazilian lower chamber associated with municipalities, this research investigates whether this variation partially explains differences in local provision of public goods and services. Brazil’s unique electoral system, together with varying political ambition among federal deputies, create three connected features that are of importance to identify variation in municipal representation in the lower chamber. First, seats in Congress are allocated to states, so the number of deputies elected associated to each municipality within the state can vary. Second, historically, deputy re-election rate is around 50%, which implies that there is variation in the composition of deputies in the lower chamber across terms. Third, a considerable portion of the seats are taken by newly elected federal deputies that are not necessarily connected to the same municipalities as their counterparts that were not re-elected. A fixed effects model that captures variation in representation within municipalities, but controls for all unobservable time invariant characteristics, shows that there is a positive partial effect of political representation on health and school outcomes. For instance, in health, the results indicate that one additional seat in congress representing a municipality decreases the share of low birthweight newborns by 0.06 %-points. When looking at school outcomes, the share of schools with a sports court increases by 0.5 %-points. These findings are also observed across other school and infant health outcomes. Moreover, the results suggest that pork barrel is the mechanism explaining the improvement on public goods and services, as discretionary federal transfers to local authorities are positively correlated with our measure of political representation.Chapter 3: A Note on the Effects of Term Limits on Brazilian Elected Politicians’ Provision of Health and Education GoodsThis study exploits the “quasi-experiment” of close elections to identify the effect of term limits on elected politicians’ performance. The prediction is that term-bounded incumbents shirk or engage in opportunistic behaviour. Klašnja and Titiunik (2017) investigate Brazilian mayoral elections and find that term-bounded mayors in weak parties spend less on health, education and social assistance programs. Using data from Brazilian local elections since 2000 and a rich set of health and education outcomes, I find no support for electoral accountability models. The difference in findings is due to both differences in the sample of mayoral elections considered and choice of outcomes measuring local provision of health and education goods.
PDF
download
878987797
028-85220240
