Ensaios em Macroeconometria

• Autor(a) principal: Oliveira, José Sérgio Casé de
• Data de Publicação: 2017
• Tipo de documento: Tese
• Idioma: por
• Título da fonte: Biblioteca Digital de Teses e Dissertações da UFPB
• Texto Completo: https://repositorio.ufpb.br/jspui/handle/123456789/12511
• topic: Política monetária; Modelo DSGE; Preços de ativos; Estimação bayesiana; Independência do banco central; Política monetária; Regressão quantílica; Modelo de regressão beta inflacionada em um; Curva de Phillips Novo Keynesiana híbrida; Métodos robustos à identificação; Conjuntos de Confiança; Instrumentos Fracos; Monetary policy; DSGE model; Asset prices; Bayesian estimation; Central bank independence; Monetary policy; Quantile regression; One inflated beta regression model; Hybrid New Keynesian Phillips curve; Identification-robust methods; Confidence sets; Weak instruments; CNPQ::CIENCIAS SOCIAIS APLICADAS::ECONOMIA

This thesis develops three essays in macroeconometry. In the first essay, we investigated whether the Central Bank of Brazil takes into account the stock market to determine the Selic interest rate. Based on Nisticò (2012) and Funke, Paetz e Pytlarczyk (2011), we proposed a Dynamic Stochastic General Equilibrium model (DSGE) for a small open economy with asset market, and we used Bayesian methods to estimate the parameters of this model. The results indicated that the Brazilian monetary authority responds to inflation and to output gap, but it does not react to stock prices. These results are in accordancetothe theoretical view,whichsaysthecentral bank shouldreactonlyindirectly to asset prices, but only when those prices imply changes in inflationary expectations and current inflation. We also observed that shocks in asset prices has little influence on macroeconomic variables. In the second essay, we investigated the determinants of the central bank independence for a set of 89 countries. As the index of the central bank independence belongs to an interval of [0,1], we used parametric and semiparametric models, which consider the limited nature of this variable. More specifically, we propose to model the central bank independence index by quantile regression using link functions based on Cauchy and Logistic distributions. Furthermore, we modeled the central bank independence index using one-inflated beta regression proposed by Ospina e Ferrari (2012). The results show that the participation on the European Union, and the variability of the unemployment rate, positively affects the central bank independence. Moreover, political variables, such as the federalism and the democratization, can affect directly the central bank independence. In the third essay, the parameters of hybrid New Keynesian Phillips Curve(NKPC)areestimatedwiththeobjectiveofevaluatingthedeterminantsofBrazilian inflation.Forthis,weusetechniquesderivedfromAndersoneRubin(1949),adaptedtothe MinimumDistance(MD)methodbyMagnussoneMavroeidis(2010)anditdoesn’trequire any hypothesis about the identification. Confidence sets of the hybrid CPNK structural parameters are estimated using the parametric specification worked by Magnusson e Mavroeidis (2010). In addition, a robustness analysis of the results obtained using different proxies for actual marginal cost is made. The results are compatible with the literature, indicating a high degree of price rigidity and a considerable degree of indexation of prices in the Brazilian economy, indicating the rejection of the hypothesis that the NKPC is purely forward-looking.