O objetivo desse trabalho é, a partir de um modelo estrutural de economia aberta novokeynesiano, determinar se as flutuações econômicas no Brasil são resultados de algumas poucas fontes principais de choques ou de muitas fontes diferentes de choques bem como as respostas dinâmicas que estes geram nas flutuações econômicas, utilizando-se da análise de decomposição de erros de Cholesky. Para isso, a analise feita nesse trabalho inicia-se com a abordagem teórica do regime de metas de inflação, situando o regime dentro do novo consenso macroeconômico. Então, a abordagem empírica para o caso brasileiro é realizada estimando-se um modelo de vetores autorregressivos (VAR) contemplando proxies que melhor representam as principais variáveis macroeconômicas brasileiras, sendo essas escolhidas a partir do estudo dos trabalhos empíricos que trataram da política monetária no Brasil em período recente. A análise dos efeitos dos choques é realizada por meio das funções de resposta ao impulso e da decomposição do erro de previsão. Os principais resultados empíricos obtidos por esse trabalho sugerem que: i) a taxa de juros se mostra como um importante instrumento para a política monetária, sendo utilizada pelo Banco Central para conter pressões sobre o hiato do produto e da inflação; ii) surpresas inflacionárias não trazem ganhos temporários em termos de um produto maior do que o nível natural; iii) a taxa de inflação mostra uma razoável persistência; iv) a taxa de inflação mostrou-se sensível a variações no câmbio. / The mean goal of this dissertation is to determine if the Brazilian output fluctuation is due to few mean sources of shocks or to several sources of shocks, using Cholesky error decomposition analysis. In order to do it, the analysis starts with a theoric approach for the inflation targeting regime, placing it into the new macroeconomic consensus. After that, the empirical approach for the Brazilian case is done by estimating an autoregressive model using proxies that best represent the mean Brazilian macroeconomic variables. These variables were chosen after a careful study of the empirical Brazilian monetary policy working papers done recently. The analysis of the shocks effects is done through both impulse-response functions and error forecast decomposition. The mean results suggest that: i) the interest rate is an important monetary policy tool and its being used by the Brazilian Central Bank in order to avoid pressures on both inflation and output gap; ii) inflation shocks does not bring gains in terms of higher product in the sort run; iii) the inflation rate has a reasonable degree of persistence; iv) the inflation rate is sensitive to exchange rate changes.

A presente tese é constituída de três ensaios que abordam duas relevantes questões que estão intrinsecamente relacionadas em macroeconomia: política monetária e inflação. No primeiro ensaio, nós procuramos averiguar não linearidades na função de reação do Banco Central do Brasil (BCB) através da estimação de regressões quantílicas inversa, sugerido por Wolters (2012) e proposto por Chernozhukov and Hansen (2005, 2006). Este método nos possibilitou detectar não linearidades na função de reação do BCB sem a necessidade de fazer suposições específicas acerca dos fatores que determinam essas não linearidades. Em específico, nós observamos que: i) a resposta da taxa de juros ao hiato da inflação corrente e esperada foi, em geral, mais forte na parte superior da distribuição condicional da taxa de juros Selic; ii) a resposta ao hiato do produto apresentou uma tendência crescente e significativa na parte inferior da distribuição condicional da taxa Selic; iii) a resposta do BCB à taxa de câmbio real foi positiva e mais elevada na cauda superior da distribuição condicional da taxa Selic. No segundo ensaio, nós investigamos a existência de não linearidades na função de reação do Banco Central do Brasil (BCB) decorrentes de incertezas desse policymaker acerca dos efeitos do hiato do produto sobre a inflação. Teoricamente, nós seguimos Tillmann (2011) para obter uma regra de política monetária ótima não linear que é robusta às incertezas acerca do trade-off produto-inflação na curva de Phillips. Além disso, nós realizamos testes de quebra estrutural para avaliar possíveis mudanças na condução da política monetária brasileira durante o regime de metas de inflação. Os resultados indicaram que: i) as incertezas acerca da inclinação na curva Phillips implicaram em não linearidades na função de reação do BCB; ii) não se pode rejeitar a hipótese de uma quebra estrutural nos parâmetros da regra monetária ocorrendo no terceiro trimestre de 2003; iii) houve um aumento na resposta da taxa Selic ao hiato do produto e uma redução da reação ao hiato da inflação corrente no regime Meirelles- Tombini; e iv) o BCB também tem reagido à taxa de câmbio durante o regime Meirelles- Tombini. No terceiro ensaio, nós procuramos analisar os determinantes da inflação no Brasil através da estimação da Curva de Phillips Novo-Keynesiana (CPNK) proposta por Blanchard e Galí (2007) e a versão padrão proposta por Galí e Gertler (1999). Além disso, realizamos testes de quebras estruturais para avaliar possíveis mudanças na dinâmica da inflação brasileira durante o período de 2002 a 2014. Os resultados indicaram que: i) os testes de quebra estrutural apontam a existência de pelo menos uma mudança estrutural nos coeficientes da CPNK; ii) o componente forward-looking da inflação é dominante, embora sua relevância tenha sido reduzida após 2004; iii) a taxa de desemprego tem afetado negativamente a inflação, embora seja observado uma redução desse impacto nos últimos anos; iv) as mudanças na taxa de câmbio apenas tiveram efeitos sobre a inflação na primeira subamostra e tem perdido relevância no período mais recente; v) o efeito do hiato do produto sobre a inflação corrente diminuiu nos anos recentes; vi) em geral, nós rejeitamos a hipótese nula de uma curva de Phillips vertical no longo prazo a um nível de significância de 5%, mas não a 1%. / This thesis is composed of three essays to address two important issues that are intricately related in macroeconomics: monetary policy and inflation. In the first essay, we seek to investigate nonlinearities in the reaction function of the Central Bank of Brazil (CBB) by estimating inverse quantile regressions (IVQR), suggested by Wolters (2012) and proposed by Chernozhukov and Hansen (2005, 2006). This method enabled us to detect nonlinearities in the CBB’s reaction function without the need to make specific assumptions about the factors that determine these nonlinearities. In particular, we observed that: i) the response of the interest rate to the current and expected inflation was, in general, stronger in the upper tail of the conditional interest rate distribution; ii) the response to the output gap showed a growing and significant trend in the lower tail of the conditional Selic rate distribution; iii) the response of the CBB to the real exchange rate was positive and higher in the upper tail of the conditional Selic rate distribution. In the second essay, we investigate the existence of nonlinearities in the reaction function of the Central Bank of Brazil (CBB) arising from this policymaker’s uncertainties about the effects of the output gap on inflation. Theoretically, we follow Tillmann (2011) to obtain a nonlinear optimal monetary policy rule that is robust to uncertainty about the output-inflation trade-off of the Phillips Curve In addition, we perform structural break tests to assess possible changes in the conduct of the Brazilian monetary policy during the inflation-targeting regime. The results indicate that: i) the uncertainties about the slope in the Phillips curve implied nonlinearities in the CBB’s reaction function; ii) we cannot reject the hypothesis of a structural break in the monetary rule parameters occurring in the third quarter of 2003; iii) there was an increase in the response of the Selic rate to output gap and a weaker response to the current inflation gap in Meirelles Tombini’s administration; and iv) the CBB has also reacted to the exchange rate in Meirelles-Tombini’s administration. In the third essay, we proposed to analyze the determinants of inflation in Brazil through the estimation of the new Keynesian Phillips curve (NKPC) proposed by Blanchard and Galí (2007) and the standard version proposed by Galí and Gertler (1999). In addition, we perform structural break tests to assess possible changes in the dynamics of inflation in Brazil during the period 2002 to 2014. The results indicated that: i) structural break tests indicate the existence of at least one structural change in the coefficients of NKPC ; ii) the forward-looking component of inflation is dominant, though its importance has been reduced after 2004; iii) the unemployment rate has negatively affected inflation, although a reduction of this impact has been observed in recent years; iv) changes in the rate of exchange only had effects on inflation in the first subsample and losing relevance in the most recent period; v) the effect of the output gap on the current inflation has declined in recent years; vi) overall, we reject the null hypothesis of a vertical Phillips curve in the long term at a significance level of 5%, but not 1%.

Utilizando a survey junto ao consumidor da universidade de Michigan, obtivemos dados a respeito das expectativas dos agentes não só sobre inflação, mas também sobre nível de atividade (os estudos de Roberts (1997) utilizaram apenas os dados de survey sobre expectativas inflacionárias). Verificamos, então, que grande parcela do custo das estabilizações dos EUA foi antecipado pela maior parte dos agentes, o que nos levou a rejeitar os modelos de Taylor (1979, 1980) e de Calvo (1983), mesmo em suas versões com as hipóteses de falta de credibilidade e informação homogeneamente defasada. Em seguida discutimos como um modelo com fairness, pode explicar este quebra cabeças. Finalmente, examinamos, três possíveis fatores (mutuamente compatíveis): a hipótese de que parte da população tenha expectativas inconsistentes, incerteza knightiana e o modelo com fairness. Nossos resultados empíricos penderam a favor de uma combinação de pelo menos uma das duas últimas alternativas com a primeira. / Using the Michigan Universitys consumer survey, we obtained data about agents expectations of both inflation and output (the latter had not been used in Roberts (1997) studies). With this, we were able to verify that a great part of the sacrifice ratios of the US stabilizations were anticipated by common agents, rejecting the Taylor (1979, 1980) and Calvo (1983) models and, with it, the hypothesis that the only reasons underlying them are staggered contracts, homogeneous sticky information and lack in credibility. WE, then, discuss how a model with fairness can explain this puzzle. Finally, we examine three (mutually consistent) factors: the hipothesis that part of the population have inconsistent expectatitons, Knightian uncertainty and te model with fairness. The results favored the combination of at least one of the two latter alternatives with te former.

Dynamic stochastic general equilibrium models are derived from microeconomic principles and they retain the hypothesis of rational expectations under policy changes. Thus they are resistant to the Lucas critique. The DSGE model has become associated with new Keynesian thinking. The basic New Keynesian model is studied in this thesis. The three equations of this model are dynamic IS curve, Phillips-curve and monetary policy rule. Blanchard and Kahn's approach is introduced as the solution strategy for linearized model. Two methods for evaluating DSGE models are presented -- calibration and Bayesian estimation. Calibrated parametres are used to fit the model to Czech economy. The results of numeric experiments are compared with empricial data from Czech republic. DSGE model's suitability for monetary policy analysis is evaluated.

京都大学 / 新制・課程博士 / 博士(経済学) / 甲第24164号 / 経博第658号 / 新制||経||302(附属図書館) / 京都大学大学院経済学研究科経済学専攻 / (主査)教授 西山 慎一, 准教授 高橋 修平, 准教授 安井 大真 / 学位規則第4条第1項該当 / Doctor of Economics / Kyoto University / DFAM

The purpose of the thesis is to introduce novel measure of real marginal cost in the New Keynesian Phillips Curve (NKPC) and compares its performance with conventional mea- sures such as output gap and labour share of income. Real marginal cost is derived from a flexible function whereas labour share is based on restrictive assumption of Cobb-Douglas technology. Dynamic correlations and results of NKPC indicate that real marginal cost is better than ad hoc measure of output gap and labour share. Given the heterogeneity in price setting behaviour across sectors, cost functions and NKPC are estimated for the agriculture, manufacturing and other sectors of Pakistan's economy. Real marginal cost is derived from static and dynamic cost functions. In the presence of adjustment costs, dynamic cost functions that are consistent and integrated with their static systems are required. Such dynamic translog cost functions are estimated after testing the theoretical properties and existence of long term relationships in the static functions. Cost attributes, marginal cost, total factor productivity, technological progress, demand and substitution elasticities are derived from static and dynamic functions. Three specifications of forward looking and hybrid form of the Phillips curves are estimated with real marginal cost, output gap and labour share. Results indicate that hybrid specifications perform better than the forward looking models in terms of goodness of fit and statistical significance. Further, comparison of Phillips curves estimated with real marginal cost, output gap and labour share indicate that real marginal cost performs better in explaining inflation dynamics in Pakistan. The results indicate that forward looking behaviour dominates and high level of nominal rigidities persists in Pakistan. Finally, hybrid form of the NKPC is estimated for a panel of sixteen Asian economies. With the consideration of heterogeneity and aggregation bias, the mean group, random coefficient and weighted average coefficients are derived from individual estimates. The unobserved time variant common factors cause cross correlation in the errors that may lead towards inconsistent estimates. Therefore, cross section averages of the explanatory and the dependent variables are augmented in hybrid specification to capture the effect of latent variables. Findings suggest that the discount factor is almost 0.94, the nominal rigidities are 33% and the weights of expected and past inflation are 66% and 33% respectively. Nominal rigidities of the Asian economies are lower than the estimates for US and Euro areas. The weights of expected and past inflation of the Asian economies are consistent with the US but lower than the estimates from the Euro areas.

The dissertation contains essays concerning the linkages between macroeconomy and financial market or the conduct of monetary policy via DSGE modelling. The dissertation contributes to the questions of fitting macroeconomic models to the data, and so contributes to our understanding of the driving forces of fluctuations in macroeconomic and financial variables. Chapter one offers an introduction to my thesis and outlines in detail the main results and methodologies. In Chapter two I introduce a statistical measure for model evaluation and selection based on the full information of sample second moments in data. A model is said to outperform its counterpart if it produces closer similarity in simulated data variance-covariance matrix when compared with the actual data. The "distance method" is generally feasible and simple to conduct. A flexible price two-sector open economy model is studied to match the observed puzzles of international finance data. The statistical distance approach favours a model with dominant role played by the expectational errors in foreign exchange market which breaks the international interest rate parity. Chapter three applies the distance approach to a New Keynesian model augmented with habit formation and backward-looking component of pricing behaviour. A macro-finance model of yield curve is developed to showcase the dynamics of implied forward yields. This exercise, with the distance approach, reiterate the inability of macro model in explaining yield curve dynamics. The method also reveals remarkable interconnection between real quantity and bond yield slope. In Chapter four I study a general equilibrium business cycle model with sticky prices and labour market rigidities. With costly matching on labour market, output responds in a hump-shaped and persistent manner to monetary shocks and the resulting Phillips curve seems to radically change the scope for monetary policy because (i) there are speed limit effects for policy and (ii) there is a cost channel for monetary policy. Labour reforms such as in mid-1980s UK can trigger more effective monetary policy. Research on monetary policy shall pay greater attention to output when labour market adjustments are persistent. Chapter five analyzes the link between money and financial spread, which is oft missed in specification of monetary policy making analysis. When liquidity provision by banks dominates the demand for money from the real economy, money may contain information of future output and inflation due to its impact on financial spreads. I use a sign-restriction Bayesian VAR estimation to separate the liquidity provision impact from money market equilibrium. The decomposition exercise shows supply shocks dominate the money-price nexus in the short to medium term. It also uncovers distinctive policy stance of two central banks. Finally Chapter six concludes, providing a brief summary of the research work as well as a discussion of potential limitations and possible directions for future research.

This dissertation is divided into four essays, each of them having its own structure and methodological framework. Although each of the essays making the chapters of the thesis is self-contained, their topics are very closely related. Consequently, the reader will be able to follow the thesis in its unity. Essay I is a selective survey of the extensive, mostly theoretic, literature dealing with monetary policy rules. We aim at contextualization of the monetary policy rules in the existing monetary economics literature. We explain the logic, the inspiration and the history of the rules for the monetary policy conduct. We distinguish between instrument rules and targeting rules as two basic categories. Finally, we resume specific issues related to policy rules for small open economies. Essay II studies the logic of short-term interest rate setting pursued by 15 EU countries before and after the launch of the EMU. We employ econometric estimation of the augmented Taylor rule (TR) for individual 15 EU countries and the Euro area. Although a vast empirical evidence is available for the major economies like the US, the UK or Germany, there is an important gap in our understanding of the factors behind the short-term interest rate dynamics in smaller economies. We find that in the period preceding the euro adoption, the TR is a poor representation of monetary policy setting in most EU countries and that many central banks considered decisions made by dominant economies rather than their domestic macroeconomic developments. The analysis of monetary policy rule of the ECB features additional problems related to the heterogeneity of the EMU. We argue that results based on Euro-area aggregated series, commonly presented in empirical studies, are subject to diverse econometric problems. We provide some evidence that the ECB is concerned also with national information and propose quasi-panel analysis as a viable framework. Essay III explores the relation between the existing monetary policy and domestic price stability in small open emerging economies, in particular the 12 EU new member states. This work has three principal objectives. First, it aims at revealing the logic of interest rate setting pursued by monetary authority of each country. The linear specification of the Taylor rule, applied already in the Essay II, is accompanied by an extensive analysis of nonlinearities in monetary policy rules and the inference on their possible sources. We find that the official monetary policy is sometimes inconsistent with the empirical evidence on the short term interest rate setting. The second objective consists in revealing the determinants of the inflation process. We have found that inflation rates are driven not only by backward persistency but also by the forward-looking component. Third, we employ analysis of the conditional inflation variance so as to give account on the viability of the existing monetary policy setting for price stability. We conclude that the policy of inflation targeting seems to be preferable to exchange rate peg because it allows decreasing not only inflation rate but also its conditional variance. Essay IV seeks to shed light on inflation dynamics of four CEEC (Czech Republic, Hungary, Poland and Slovakia) and test when the predominant model of inflation, the New Keynesian Philips Curve (NKPC), is consistent with the data of these countries. According to the microfounded NKPC, the current inflation is related to inflation expectations and the real marginal cost. The empirical validity of this model has recently become a subject of major controversy in the monetary economics. Although we find some favorable evidence for the NKPC, it seems to be too restrictive model for small open economies. In particular, the failure of the NKPC to explain the inflation dynamics of these countries may be related to the assumption that inflation is related to forward-looking price setting of domestic monopolist firms while our evidence suggests that prices in CEEC have an important backward-looking component and the inflation is significantly driven by external factors like the exchange rate and the foreign inflation rate.

Une forte croissance économique en Chine et son rôle important dans le commerce mondial impliquent que nous devons analyser son inflation avec les pressions intérieures et extérieures. Tandis que les papiers récents ont concentrés leur travail sur les processus d'inflation en Chine en utilisant des données nationales (voir, par exemple Brandt et Zhu, 2000; Feyzioglu & Willard, 2006; Porter, 2010), très peu d'attention a été mise sur l'inflation entre les différentes juridictions Chinoises. Ainsi, la compréhension de la dynamique de l'inflation et ses interactions entre les juridictions Chinoise sont des questions importantes pour les décisions des banques centrales en matière de politique monétaire. Dans ma thèse, tout d'abord, en considérant les pressions intérieures et extérieures de l'inflation, j'analyse les dynamiques de l'inflation entre les juridictions Chinoises (Chapitre 3). Ensuite, grâce à la libre circulation des biens et les migrations intérieures en Chine, je m'intéresse aux effets d'interactions de l'inflation à travers le pays avec les pressions intérieures et extérieures (Chapitre 4). Enfin, nous conjuguons la variation provinciale dans la dynamique de l'inflation Chinoise et les caractéristiques d'économie ouverte en estimant la courbe de Phillips hybride dans l'économie ouverte pour les provinces chinoises (Chapitre 5). / China's swift economic development and share in global trade increasing rapidly imply a need to understand its inflation dynamics with internal and external pressures. While recent papers focus on inflation process analysis in the mainland of China by using a country-level data (see, e.g. Brandt & Zhu, 2000; Feyzioğlu & Willard, 2006; Porter, 2010), less attention has been paid to differences across China's jurisdictions. Thus, understanding of inflation dynamics and its interaction among the Chinese provinces are the important issues for central bank's monetary policy decisions. Firstly, considering the internal and external inflation pressures, I analyze the inflation dynamics among the CMU (Chapter 3). Secondly, because of the free flow of goods and internal migration across the country, I'm interested in analyzing the effect of inflation interaction among its provinces with internal and external pressures (Chapter 4). Finally, we combine the interest in the provincial variation in China's inflation dynamics with its characteristic of economic openness by estimating hybrid open-economy Phillips curves for the Chinese provinces(Chapter 5).

Este trabalho estima, utilizando dados trimestrais de 1999 a 2011, o impacto dinâmico de um estímulo fiscal no Brasil sobre as principais variáveis macroeconômicas Brasileiras. Na estimativa dos impactos permitiu-se que as expectativas dos agentes econômicas fossem afetadas pela existência e probabilidade de alternância de regimes (foram detectados dois regimes) na política monetária do país. Os parâmetros da regra da política monetária, nos dois regimes detectados, foram estimados através de um modelo - composto apenas pela equação da regra da política monetária - que permite uma mudança de regime Markoviana. Os parâmetros do único regime encontrado para a política fiscal foram estimados por um modelo Vetorial de Correção de Erros (Vector Error Correction Model - VEC), composto apenas pelas variáveis pertencentes à regra da política fiscal. Os parâmetros estimados, para os diversos regimes das políticas monetária e fiscal, foram utilizados como auxiliares na calibragem de um modelo de equilíbrio geral estocástico dinâmico (MEGED), com mudanças de regime, com rigidez nominal de preços e concorrência monopolística (como em Davig e Leeper (2011)). Após a calibragem do MEGED os impactos dinâmicos de um estímulo fiscal foram obtidos através de uma rotina numérica (desenvolvida por Davig e Leeper (2006)) que permite obter o equilíbrio dinâmico do modelo resolvendo um sistema de equações de diferenças de primeira ordem expectacionais dinâmicas não lineares. Obtivemos que a política fiscal foi passiva durante todo o período analisado e que a política monetária foi sempre ativa, porém sendo em determinados momentos menos ativa. Em geral, em ambas as combinações de regimes, um choque não antecipado dos gastos do governo leva ao aumento do hiato do produto, aumento dos juros reais, redução do consumo privado e (em contradição com o resultado convencional) redução da taxa de inflação. / This paper estimates, using quarterly data from 1999 to 2011, the dynamic impacts of a fiscal stimulus in Brazil on key Brazilian macroeconomic variables. The estimates take into account the effects of the existence and of the probabilities of occurrence of the switching monetary policy regimes (two regimes were detected) on agents expectations formation. The monetary policy rules parameters, in the two detected regimes, were estimated through a Markov regime-switching model composed only by the monetary policy rule equation. The fiscal rules parameters of the unique detected fiscal policy regime were estimated through a Vector Error Correction (VEC) model composed only by the variables pertained to the fiscal policy rule. The monetary and fiscal policy rules parameters were auxiliary in the calibration of a Dynamic Stochastic General Equilibrium (DSGE) model with regime-switching, nominal price rigidity and monopolistic competition (as in Davig and Leeper (2011)). After the DSGEs calibration the fiscal stimuluss impacts were obtained through a numerical routine (developed by Davig and Leeper (2006)) that solves a set of nonlinear expectational first-order difference equations and gives the dynamic equilibrium of the model. Our results suggest that fiscal policy was passive during the whole period and that monetary policy was always active, but they were more active at certain times and in others, less active. Overall, in both combinations of regimes, a government spending shock induces an increase in the output gap, increases in real interest rates, a reduction in private consumption and (contrary to the conventional wisdom) a reduction in inflation.