This paper explores hypotheses that could explain both the creation of independent regulatory agencies (IRAs) in Brazil, and the differences in the design of the Brazilian IRAs in the telecommunications and electricity sectors. To formulate specific hypotheses that make sense of the Brazilian case, the paper critically interrogates the "weak state" hypothesis and the "political bias" hypothesis. The first argues that countries with flawed governance structures, such as Latin American countries, are less likely to establish independent regulators than European countries. The second argues that "political bias" is a determinant factor in predicting the implementation of IRAs in Latin America. The first part of the paper uses these two general hypotheses as a basis to formulate specific hypotheses to explain the creation of IRAs in Brazil. The second part of the paper formulates specific hypotheses that could explain why institutional guarantees of IRA independence are stronger in the telecommunications sector, than in the electricity sector. In particular, the paper argues in support of a revised version of the "political bias" hypothesis to explain sectoral divergence, suggesting that bureaucratic resistance to reform may be the cause for the variations observed in Brazil between regulatory reform in electricity and in telecommunications.