Essays on industrial organization.
PhD thesis, The London School of Economics and Political Science (LSE).
This thesis is composed by three essays and provides empirical and methodological contributions to the Industrial Organization literature. The ﬁrst essay (chapter 2) analyzes the welfare impacts of the Brazilian Biodiesel law. The second essay (chapter 3) develops an alternative estimator for dynamic games. The third essay (chapter 4) applies the methodology developed in chapter 3 to build a dynamic oligopoly model for Brazilian banking industry.
Chapter 5 concludes the thesis and provides some directions for future work.
Chapter 1 analyzes market eﬀects of Brazilian biodiesel regulations. Biodiesel was introduced in Brazil in 2005, mixed with mineral diesel to produce the BX blend (X stands for the percentage of biodiesel). Even in small quantities, the percentage of biodiesel has a positive impact on ﬁnal price of BX because the production costs of biodiesel are higher than those of mineral diesel. In order to analyze the welfare consequences of this price increase, I use a static partial equilibrium framework. The results show that the current proportion of biodiesel in the diesel mixture (5%) increases consumers' price by 1.7% and decreases the consumption by 1.5% compared to the scenario without biodiesel. Also, an increase in the biodiesel percentage to 10% would raise the price by 3.5% and reduce the consumption by 3%.
Chapter 3 provides an alternative estimator for dynamic games. Estimation of dynamic games is a numerically challenging task, in chapter 3 we propose an alternative class of asymptotic least squares estimators to Pesendorfer and Schmidt-Dengler's (2008). The estimator we propose is based on the equilibrium condition of the game when represented in the space of payoﬀs - in contrast with Pesendorfer and Schmidt-Dengler's (2008) that work in the probability space. Our estimator reduces signiﬁcantly the computational burden. This reduction is specially signiﬁcant under the linear-in-parameter speciﬁcation where our estimator has an OLS/GLS closed form that does not require any optimization. Also, we show that our estimator is asymptotically equivalent to Pesenrorfer and Schmidt-Dengler's (2008). This implies that there is no theoretical cost of using our estimator. Monte Carlo estimations show that our estimator has good small sample properties and provides signiﬁcant reduction in the computational time when compared to Pesenrorfer and Schmidt-Dengler's (2008) estimator.
Chapter 4 applies the methodology developed in chapter 3. We estimate a dynamic oligopoly model for the Brazilian banking industry. The results are used to build counterfactuals to examine the eﬀects of the privatization of public banks on the number of bank branches in small municipalities. We ﬁnd that public banks are not strategic and their presence generate positive spillovers on the private banks' proﬁts. The model however, is not able to disentangle the nature of this spillover. Also, the counterfactual shows that the number of branches operating in small markets would drop in a privatization scenario.
Chapter 5 is the conclusion, I discuss the limitations of the current work and provide some directions for future research.
Actions (login required)
||Record administration - authorised staff only