Essays in Development and Labor Economics in Brazil

Abstract

In my dissertation I studied the political economy of economic development as well as the the urban wage premium, all using data from Brazil. In the first chapter, which is joint work with Carlos Varjao, we analyze the effect that increased political opposition in the city council has on corruption and public service provision at the local level. In the second chapter I study the sources of the high wage premium observed in cities, including firm sorting, firm and occupational matching, and compensating wage differentials. Finally, in the third chapter I study what happens to the provision of public education when a school teacher is elected to the city council (which actually occurs quite frequently). More detailed summaries of each chapter follow below. Chapter 1: In 'Political Opposition, Legislative Oversight, and the Performance of the Executive Branch', we study the effect that increased political opposition has on corruption and other measures of the mayor's performance in Brazil. The separation of powers between the executive and legislative branches is a cornerstone of democracy. This system of checks and balances, however, can be circumvented by partisan loyalties if legislators strategically avoid exerting oversight when their own party controls the executive branch. It is thus an empirical question whether the separation of powers prevents the abuse of power in practice. We answer this question by measuring the extent to which members of political opposition parties in a city council effectively check the mayor's performance in Brazil. We employ a regression discontinuity design to estimate the causal effect of an additional politically opposed legislator, and we find that political opposition increases oversight action and decreases corruption, with the effect fully concentrated on mayors facing reelection pressure. We trace the impact of oversight, via a reduction in healthcare spending irregularities, all the way to impacts on healthcare service delivery and health outcomes. Chapter 2: In 'Decomposing the Urban Wage Premium in Brazil: Firms, Matching, and Compensating Wage Differentials' I study the sources of the high wage premium observed in cities. In this chapter I used detailed employer-employee matched data from Brazil to understand 3 important elements of the urban wage premium: (1) the role of firms sorting into cities, (2) the role of firm and occupational matching in creating agglomeration economies, and (3) the role of compensating wage differentials. I first exploit identification from multi-city firms to show that positive selection of high-wage firms into larger cities accounts for 44% of what is often considered `agglomeration economies'. Then I show that improved firm and occupational matching together account for 87% of agglomeration effects. I then turn my attention to compensating wage differentials--- a possible explanation for the high-wage firms in cities. I estimate revealed-preference valuation of jobs, and show that jobs in cities in fact have better non-wage characteristics, and so high urban wages cannot be due to compensating wage differentials. This evidence together suggests that in Brazil, cities exist because they provide thick labor markets where high-wage firms and high-wage workers can go to find productive matches. Chapter 3: In 'Teachers in Politics: Teacher-Politicians, Gender, and the Representation of Public Education' I study what happens to public education in a city when a school teacher is elected to the city council, and I find that it depends on the gender of the teacher. Using a regression discontinuity design that exploits close elections, I find that when a female teacher is elected to the city council, the city hires both more teachers and more qualified teachers, and pays them more. Having a female teacher on the city council also increases the likelihood that the city's schools have necessary teaching resources, books, and financing, and possibly increases student test scores. No significant effect is found for male teachers elected to the city council. This difference may be due to different political career concerns for men versus women, a simple amplification of existing gender policy preference differences, or some mixture of the two.