Cette thèse s’articule autour de trois chapitres indépendants qui s’inscrivent dans les champs de la macroéconomie, de l’économie monétaire et de la finance internationale. Dans le premier chapitre, je construis un modèle néo-keynesien d’équilibre général sous incertitude pour examiner les implications de la production domestique des ménages pour la politique monétaire. Le modèle proposé permet de reconcilier deux faits empiriques majeurs: la forte sensibilité du produit intérieur brut aux chocs monétaires (obtenue à partir des modèles VAR), et le faible degré de rigidité nominale observé dans les micro-données. Le deuxième chapitre étudie le role de la transformation structurelle (réallocation de la main d’oeuvre entre secteurs) sur la volatilité de la production aggregée dans un panel de pays. Le troisième chapitre quant à lui met en exergue l’importance de la cartographie des échanges commerciaux pour le choix entre un régime de change fixe et l’arrimage à un panier de devises. "Household Production, Services and Monetary Policy" (Chapitre 1) part de l’observation selon laquelle les ménages peuvent produire à domicile des substituts aux services marchands, contrairement aux biens non durables qu’ils acquièrent presque exclusivement sur le marché. Dans ce contexte, ils procèdent à d’importants arbitrages entre produire les services à domicile ou les acquerir sur le marché, dépendamment des changements dans leur revenu. Pour examiner les implications de tels arbitrages (qui s’avèrent être importants dans les micro-données) le secteur domestique est introduit dans un modèle néo-keyenesien d’équilibre général sous incertitude à deux secteurs (le secteur des biens non durables et le secteur des services) autrement standard. Je montre que les firmes du secteur des services sont moins enclin à changer leurs prix du fait que les ménages ont l’option de produire soit même des services substituts. Ceci se traduit par la présence d’un terme endogène supplémentaire qui déplace la courbe de Phillips dans ce secteur. Ce terme croit avec le degré de substituabilité qui existe entre les services produits à domicile et ceux acquis sur le marché. Cet accroissement de la rigidité nominale amplifie la sensibilité de la production réelle aux chocs monétaires, notamment dans le secteur des services, ce qui est compatible avec l’évidence VAR selon laquelle les services de consommation sont plus sensibles aux variations de taux d’intérêt que les biens non durables. "Structural Transformation and the Volatility of Aggregate Output: A Cross-country Analysis" (Chapitre 2) est basée sur l’évidence empirique d’une relation négative entre la part de la main d’oeuvre allouée au secteur des services et la volatilité de la production aggrégée, même lorsque je contrôle pour les facteurs tels que le développement du secteur financier. Ce resultat aggregé est la conséquence des développements sectoriels: la productivité de la main d’oeuvre est beaucoup plus volatile dans l’agriculture et les industries manufacturières que dans les services. La production aggregée deviendrait donc mécaniquement moins volatile au fur et à mesure que la main d’oeuvre se déplace de l’agriculture et de la manufacture vers les services. Pour évaluer cette hypothèse, je calibre un modèle de transformation structurelle à l’économie américaine, que j’utilise ensuite pour générer l’allocation sectorielle de la main d’oeuvre dans l’agriculture, l’industrie et les services pour les autres pays de l’OCDE. Dans une analyse contre-factuelle, le modèle est utlisé pour restreindre la mobilité de la main d’oeuvre entre secteurs de façon endogène. Les calculs montrent alors que le déplacement de la main d’oeuvre vers le secteur des services réduit en effet la volatilité de la production aggregée. "Exchange Rate Volatility under Alternative Peg Regimes: Do Trade Patterns Matter?" (Chapitre 3) est une contribution à la litterature économique qui s’interesse au choix entre divers regimes de change. J’utilise les données mensuelles de taux de change bilatéraux et de commerce extérieur entre 1980 et 2010 pour les pays membre de l’Union Economique et Monétaire Ouest Africaine (UEMOA). La monnaie de ces pays (le franc CFA) est arrimée au franc Francais depuis le milieu des années 40 et à l’euro depuis son introduction en 1999. Au moment de l’arrimage initial, la France était le principal partenaire commercial des pays de l’UEMOA. Depuis lors, et plus encore au cours des dix dernières années, la cartographie des échanges de l’union a significativement changé en faveur des pays du groupe des BICs, notamment la Chine. Je montre dans ce chapitre que l’arrimage à un panier de devises aurait induit une volatilité moins pronnoncée du taux de change effectif nominal du franc CFA au cours de la décennie écoulée, comparé à la parité fixe actuelle. Ce chapitre, cependant, n’aborde pas la question de taux de change optimal pour les pays de l’UEMOA, un aspect qui serait intéressant pour une recherche future. / This thesis includes three independent essays in the fields of macroeconomics, monetary economics and international finance. In the first essay, I build a new Keynesian DSGE model to examine the implications for monetary policy of household production. The proposed theory helps reconcile the relatively strong response of output to monetary policy shocks as suggested by VAR-based evidence and the low degree of price rigidity found in micro data. The second essay analyzes the role of structural transformation (the reallocation of labor across sectors overtime) in shaping the volatility of aggregate output across countries. Finally, the third essay illustrates the importance of trade patterns in choosing between a single currency peg and a peg to a composite basket of currencies. “Household Production, Services and Monetary Policy” (Chapter 1) builds on the observation that consumer services (unlike consumer nondurable) have close substitutes at home. Households may therefore switch between consuming home and market service as the real wage (the opportunity cost of working at home) changes. To study the implications of this arbitrage for monetary policy, I embed a household sector into an otherwise standard two-sector (a nondurable good sector and a service sector) new Keynesian DSGE model. The fact that households are able to produce services at home makes service sector’s firms more reluctant to change their price. This translates into an extra endogenous shift term in the new Keynesian Phillips that is increasing with the extent of substitutability between home and market services. This increased nominal rigidity endogenously amplifies the output response to monetary policy shock, especially in the service sector, which is consistent with VAR-based evidence in the paper that consumer services are more interest-rate sensitive than consumer nondurables. “Structural Transformation and the Volatility of Aggregate Output: A Cross-country Analysis” (Chapter 2) is based on the evidence of a negative relationship between the employment share of the service sector and the volatility of aggregate output, which I obtain after controlling for several factors (including the level of financial development). This aggregate result is driven by sectoral labor productivity differentials: Labor productivity is substantially more volatile in agriculture and manufacturing than in services. Aggregate output would therefore become mechanically more stable as labor shifts away from agriculture and manufacturing, and toward the service sector. To quantify this conjecture, I first calibrate a model of structural transformation (secular reallocation of labor across sectors) to the U.S. economy, which I use to match the time path of labor shares in agriculture, manufacturing and services across OECD countries. The model is subsequently used to conduct a set of counterfactual experiments in which labor is endogenously constrained from moving across sectors. Computations suggest that the shift of labor toward the services sector is indeed volatility-reducing. “Exchange Rate Volatility under Alternative Peg: Do Trade Patterns Matter?” (Chapter 3) is a contribution to the literature on the choice of exchange rate regimes. I use monthly bilateral exchange rate and external trade data from 1980 to 2010 for the member countries of the Western African and Monetary Union (WAEMU). These countries have their common currency (the CFA franc) pegged to the French franc since the mid-40s and to the euro since its introduction in 1999. At the time of the initial peg arrangement, France accounted for most of the external trade of WAEMU countries. Since then, and more notably since the early 2000s, the trade patterns of these countries shifted briskly away from France and other Euro area countries and towards the BICs (China in particular). The chapter finds that a peg to a composite basket of currencies would have led to a less volatile effective exchange rate over the last decade compare to the current hard peg. This chapter, however, does not derive an optimal exchange rate for WAEMU countries, which is an important area for further research.

In dieser Arbeit werden die Folgen der Calvo-Annahme in dynamischen makroökonomischen Modellen untersucht. Dafür wird die Calvo-Annahme unter Anwendung des Konzepts der statistischen Hazardfunktion verallgemeinert. Ich untersuche zwei mögliche Anwendungen dieses Ansatzes innerhalb von DSGE-Modellen. Im ersten Artikel zeige ich, dass der Zugewinn an Handhabbarkeit, der aus der Calvo-Annahme für Neu-Keynesianische Modelle folgt, mit unerwünschten Folgen in Bezug auf die Inflationsdynamiken einher geht. Der zweite Artikel schätzt die aggregierte Hazardfunktion unter Verwendung des theoretischen Rahmens des ersten Artikels. Es zeigt sich, dass die Annahme einer konstanten Hazardfunktion, die aus der Calvo-Annahme folgt, von den Daten eindeutig abgelehnt wird. Im dritten Artikel analysiere ich die Implikationen der empirisch geschätzten Hazardfunktion für die Persistenz von Inflation und die Geldpolitik. Die Untersuchungen zeigen, dass mittels der empirisch plausiblen aggregierten Hazardfunktion Zeitreihen simuliert werden können, die mit der Persistenz der inflatorischen Lücke im US Verbraucherpreisindex konsistent sind. Anhand dieser Ergebnisse komme ich zu dem Schluss, dass die Hazardfunktion eine entscheidende Rolle für die dynamischen Eigenschaften von Inflation spielt. Der letzte Artikel wendet den selben Modellierungsansatz auf ein Real-Business-Cycle Model mit rigidem Arbeitsmarkt an. Unter Verwendung eines allgemeineren stochastischen Anpassungsprozess stelle ich fest, dass die Arbeitsmarktdynamiken von einem Parameter beinflusst werden, der das Monotonieverhalten der Hazardfunktion bestimmt. Insbesondere steigt die Volatilität des Beschäftigungsniveaus, wohingegen dessen Persistenz mit zunehmendem Parameterwert abnimmt. / The Calvo assumption (Calvo, 1983) is widely used in the macroeconomic literature to model market frictions that limit the ability of economic agents to re-optimize their control variables. In spite of its virtues, the Calvo assumption also implies singular adjustment behavior at the firm level as well as a restrictive aggregation mechanism for the whole economy. In this study, I examine implications of the Calvo assumption for macroeconomic dynamics. To do so, I extend the Calvo assumption to a more general case based on the concept of the statistical hazard function. Two applications of this approach are studied in the DSGE framework. In the first essay, I apply this approach to a New Keynesian model, and demonstrate that tractability gained from the Calvo pricing assumption is costly in terms of inflation dynamics. The second essay estimates aggregate price reset hazard function using the theoretical framework constructed in the first essay, and shows that the constant hazard function implied by the Calvo assumption is strongly rejected by the aggregate data. In the third essay, I further explore implications of the empirically based hazard function for inflation persistence and monetary policy. I find that the empirically plausible aggregate price reset hazard function can generate simulated data that are consistent with inflation gap persistence found in the US CPI data. Based on these results, I conclude that the price reset hazard function plays a crucial role for generating inflation dynamics. The last essay applies the same modeling approach to a RBC model with employment rigidity. I find that, when introducing a more general stochastic adjustment process, the employment dynamics vary with a parameter, which determines the monotonic property of the hazard function. In particular, the volatility of employment is increasing, but the persistence is decreasing in the value of the parameter.

Esta tese apresenta três ensaios sobre previsão de inflação e análise de dados em tempo real no Brasil. Utilizando uma curva de Phillips, o primeiro ensaio propõe um “modelo evolucionário” para prever inflação no Brasil. O modelo evolucionário consiste em uma combinação de um modelo não-linear (que é formado pela combinação de três redes neurais artificiais – RNAs) e de um modelo linear (que também é a referência para propósitos de comparação). Alguns parâmetros do modelo evolucionário, incluindo os pesos das combinações, evoluem ao longo do tempo segundo ajustes definidos por três algoritmos que avaliam os erros fora-da-amostra. As RNAs foram estimadas através de uma abordagem híbrida baseada em um algoritmo genético (AG) e em um algoritmo simplex de Nelder-Mead. Em um experimento de previsão fora-da-amostra para 3, 6, 9 e 12 passos à frente, o desempenho do modelo evolucionário foi comparado ao do modelo linear de referência, segundo os critérios de raiz do erro quadrático médio (REQM) e de erro absoluto médio (EAM). O desempenho do modelo evolucionário foi superior ao desempenho do modelo linear para todos os passos de previsão analisados, segundo ambos os critérios. O segundo ensaio é motivado pela recente literatura sobre análise de dados em tempo real, que tem mostrado que diversas medidas de atividade econômica passam por importantes revisões de dados ao longo do tempo, implicando importantes limitações para o uso dessas medidas. Elaboramos um conjunto de dados de PIB em tempo real para o Brasil e avaliamos a extensão na qual as séries de crescimento do PIB e de hiato do produto são revisadas ao longo do tempo. Mostramos que as revisões de crescimento do PIB (trimestre/trimestre anterior) são economicamente relevantes, embora as revisões de crescimento do PIB percam parte da importância à medida que o período de agregação aumenta (por exemplo, crescimento em quatro trimestres). Para analisar as revisões do hiato do produto, utilizamos quatro métodos de extração de tendência: o filtro de Hodrick-Prescott, a tendência linear, a tendência quadrática, e o modelo de Harvey-Clark de componentes não-observáveis. Todos os métodos apresentaram revisões de magnitudes economicamente relevantes. Em geral, tanto a revisão de dados do PIB como a baixa precisão das estimativas de final-de-amostra da tendência do produto mostraram-se fontes relevantes das revisões de hiato do produto. O terceiro ensaio é também um estudo de dados em tempo real, mas que analisa os dados de produção industrial (PI) e as estimativas de hiato da produção industrial. Mostramos que as revisões de crescimento da PI (mês/mês anterior) e da média móvel trimestral são economicamente relevantes, embora as revisões de crescimento da PI tornem-se menos importantes à medida que o período de agregação aumenta (por exemplo, crescimento em doze meses). Para analisar as revisões do hiato da PI, utilizamos três métodos de extração de tendência: o filtro de Hodrick-Prescott, a tendência linear e a tendência quadrática. Todos os métodos apresentaram revisões de magnitudes economicamente relevantes. Em geral, tanto a revisão de dados da PI como a baixa precisão das estimativas de final-de-amostra da tendência da PI mostraram-se fontes relevantes das revisões de hiato da PI, embora os resultados sugiram certa predominância das revisões provenientes da baixa precisão de final-de-amostra. / This thesis presents three essays on inflation forecasting and real-time data analysis in Brazil. By using a Phillips curve, the first essay presents an “evolutionary model” to forecast Brazilian inflation. The evolutionary model consists in a combination of a non-linear model (that is formed by a combination of three artificial neural networks - ANNs) and a linear model (that is also a benchmark for comparison purposes). Some parameters of the evolutionary model, including the combination weight, evolve throughout time according to adjustments defined by three algorithms that evaluate the out-of-sample errors. The ANNs were estimated by using a hybrid approach based on a genetic algorithm (GA) and on a Nelder-Mead simplex algorithm. In a 3, 6, 9 and 12 steps ahead out-of-sample forecasting experiment, the performance of the evolutionary model was compared to the performance of the benchmark linear model, according to root mean squared errors (RMSE) and to mean absolute error (MAE) criteria. The evolutionary model performed better than the linear model for all forecasting steps that were analyzed, according to both criteria. The second essay is motivated by recent literature on real-time data analysis, which has shown that several measures of economic activities go through important data revisions throughout time, implying important limitations to the use of these measures. We developed a GDP real-time data set to Brazilian economy and we analyzed the extent to which GDP growth and output gap series are revised over time. We showed that revisions to GDP growth (quarter-onquarter) are economic relevant, although the GDP growth revisions lose part of their importance as aggregation period increases (for example, four-quarter growth). To analyze the output gap revisions, we applied four detrending methods: the Hodrick-Prescott filter, the linear trend, the quadratic trend, and the Harvey-Clark model of unobservable components. It was shown that all methods had economically relevant magnitude of revisions. In a general way, both GDP data revisions and the low accuracy of end-of-sample output trend estimates were relevant sources of output gap revisions. The third essay is also a study about real-time data, but focused on industrial production (IP) data and on industrial production gap estimates. We showed that revisions to IP growth (month-on-month) and to IP quarterly moving average growth are economic relevant, although the IP growth revisions become less important as aggregation period increases (for example, twelve-month growth). To analyze the output gap revisions, we applied three detrending methods: the Hodrick-Prescott filter, the linear trend, and the quadratic trend. It was shown that all methods had economically relevant magnitude of revisions. In general, both IP data revisions and low accuracy of end-of-sample IP trend estimates were relevant sources of IP gap revisions, although the results suggest some prevalence of revisions originated from low accuracy of end-of-sample estimates.

本篇論文擴充Ang et al. (2007)之基本架構,分別建構台灣各式月資料與季資料的物價指數預測模型,並進行預測以及實證分析。我們用以衡量通貨膨脹率的指標為 CPI 年增率與核心CPI 年增率。我們比較貨幣模型、成本加成模型、6 種不同設定的菲力浦曲線模型、3 種期限結構模型、隨機漫步模型、 AO 模型、ARIMA 模型、VAR 模型、主計處(DGBAS)、中經院(CIER) 及台經院(TIER) 之預測。藉由此研究,我們可以完整評估出文獻上常用之各式月資料及季資料預測模型的優劣。 我們實證結果顯示,在月資料預測模型樣本外預測績效表現方面, ARIMA 模 型對 2 種通貨膨脹率指標的樣本外預測能力表現最好。至於季資料預測模型樣本外預測績效表現, ARIMA 模型對未來核心 CPI 年增率的樣本外預測能力表現最好; 然而,對於 CPI 年增率為預測目標的預測模型則不存在最佳的模型。此外,實證分析中我們也發現本研究所建構的模型預測表現仍遜於主計處的預測,但部份模型的樣本外預測能力表現則比中經院與台經院的預測為佳。 / This paper compares the forecasting performance of inflation in Taiwan. We conduct various inflation forecasting methods (models) for two inflation measures(CPI growth rate and core-CPI growth rate) by using monthly and quarterly data. Besides the models of Ang et al. (2007), we also consider some macroeconomic models for comparison. We compare some Monetary models, Mark-up models, six variants of Phillips curve models, three variants of term structure models, a Random walk model, an AO model, an ARIMA model, and a VAR model. We also compare the forecast ability of these model with three different survey forecasts (the DGBAS, CIER, and TIER surveys). We summarized our findings as follows. The best monthly forecasting model for both inflation measures is ARIMA model. For quarterly core-CPI inflation, ARIMA model is also the best model; however, when comparing the quarterly forecasts for CPI inflation, there does not exist the best one. Besides, we also found that the DGBAS survey outperforms all of our forecasting methods/models, but some of our forecasting models are better than the CIER and TIER surveys in terms of MAE.

Cette thèse s’articule autour de trois chapitres indépendants qui s’inscrivent dans les champs de la macroéconomie, de l’économie monétaire et de la finance internationale. Dans le premier chapitre, je construis un modèle néo-keynesien d’équilibre général sous incertitude pour examiner les implications de la production domestique des ménages pour la politique monétaire. Le modèle proposé permet de reconcilier deux faits empiriques majeurs: la forte sensibilité du produit intérieur brut aux chocs monétaires (obtenue à partir des modèles VAR), et le faible degré de rigidité nominale observé dans les micro-données. Le deuxième chapitre étudie le role de la transformation structurelle (réallocation de la main d’oeuvre entre secteurs) sur la volatilité de la production aggregée dans un panel de pays. Le troisième chapitre quant à lui met en exergue l’importance de la cartographie des échanges commerciaux pour le choix entre un régime de change fixe et l’arrimage à un panier de devises. "Household Production, Services and Monetary Policy" (Chapitre 1) part de l’observation selon laquelle les ménages peuvent produire à domicile des substituts aux services marchands, contrairement aux biens non durables qu’ils acquièrent presque exclusivement sur le marché. Dans ce contexte, ils procèdent à d’importants arbitrages entre produire les services à domicile ou les acquerir sur le marché, dépendamment des changements dans leur revenu. Pour examiner les implications de tels arbitrages (qui s’avèrent être importants dans les micro-données) le secteur domestique est introduit dans un modèle néo-keyenesien d’équilibre général sous incertitude à deux secteurs (le secteur des biens non durables et le secteur des services) autrement standard. Je montre que les firmes du secteur des services sont moins enclin à changer leurs prix du fait que les ménages ont l’option de produire soit même des services substituts. Ceci se traduit par la présence d’un terme endogène supplémentaire qui déplace la courbe de Phillips dans ce secteur. Ce terme croit avec le degré de substituabilité qui existe entre les services produits à domicile et ceux acquis sur le marché. Cet accroissement de la rigidité nominale amplifie la sensibilité de la production réelle aux chocs monétaires, notamment dans le secteur des services, ce qui est compatible avec l’évidence VAR selon laquelle les services de consommation sont plus sensibles aux variations de taux d’intérêt que les biens non durables. "Structural Transformation and the Volatility of Aggregate Output: A Cross-country Analysis" (Chapitre 2) est basée sur l’évidence empirique d’une relation négative entre la part de la main d’oeuvre allouée au secteur des services et la volatilité de la production aggrégée, même lorsque je contrôle pour les facteurs tels que le développement du secteur financier. Ce resultat aggregé est la conséquence des développements sectoriels: la productivité de la main d’oeuvre est beaucoup plus volatile dans l’agriculture et les industries manufacturières que dans les services. La production aggregée deviendrait donc mécaniquement moins volatile au fur et à mesure que la main d’oeuvre se déplace de l’agriculture et de la manufacture vers les services. Pour évaluer cette hypothèse, je calibre un modèle de transformation structurelle à l’économie américaine, que j’utilise ensuite pour générer l’allocation sectorielle de la main d’oeuvre dans l’agriculture, l’industrie et les services pour les autres pays de l’OCDE. Dans une analyse contre-factuelle, le modèle est utlisé pour restreindre la mobilité de la main d’oeuvre entre secteurs de façon endogène. Les calculs montrent alors que le déplacement de la main d’oeuvre vers le secteur des services réduit en effet la volatilité de la production aggregée. "Exchange Rate Volatility under Alternative Peg Regimes: Do Trade Patterns Matter?" (Chapitre 3) est une contribution à la litterature économique qui s’interesse au choix entre divers regimes de change. J’utilise les données mensuelles de taux de change bilatéraux et de commerce extérieur entre 1980 et 2010 pour les pays membre de l’Union Economique et Monétaire Ouest Africaine (UEMOA). La monnaie de ces pays (le franc CFA) est arrimée au franc Francais depuis le milieu des années 40 et à l’euro depuis son introduction en 1999. Au moment de l’arrimage initial, la France était le principal partenaire commercial des pays de l’UEMOA. Depuis lors, et plus encore au cours des dix dernières années, la cartographie des échanges de l’union a significativement changé en faveur des pays du groupe des BICs, notamment la Chine. Je montre dans ce chapitre que l’arrimage à un panier de devises aurait induit une volatilité moins pronnoncée du taux de change effectif nominal du franc CFA au cours de la décennie écoulée, comparé à la parité fixe actuelle. Ce chapitre, cependant, n’aborde pas la question de taux de change optimal pour les pays de l’UEMOA, un aspect qui serait intéressant pour une recherche future. / This thesis includes three independent essays in the fields of macroeconomics, monetary economics and international finance. In the first essay, I build a new Keynesian DSGE model to examine the implications for monetary policy of household production. The proposed theory helps reconcile the relatively strong response of output to monetary policy shocks as suggested by VAR-based evidence and the low degree of price rigidity found in micro data. The second essay analyzes the role of structural transformation (the reallocation of labor across sectors overtime) in shaping the volatility of aggregate output across countries. Finally, the third essay illustrates the importance of trade patterns in choosing between a single currency peg and a peg to a composite basket of currencies. “Household Production, Services and Monetary Policy” (Chapter 1) builds on the observation that consumer services (unlike consumer nondurable) have close substitutes at home. Households may therefore switch between consuming home and market service as the real wage (the opportunity cost of working at home) changes. To study the implications of this arbitrage for monetary policy, I embed a household sector into an otherwise standard two-sector (a nondurable good sector and a service sector) new Keynesian DSGE model. The fact that households are able to produce services at home makes service sector’s firms more reluctant to change their price. This translates into an extra endogenous shift term in the new Keynesian Phillips that is increasing with the extent of substitutability between home and market services. This increased nominal rigidity endogenously amplifies the output response to monetary policy shock, especially in the service sector, which is consistent with VAR-based evidence in the paper that consumer services are more interest-rate sensitive than consumer nondurables. “Structural Transformation and the Volatility of Aggregate Output: A Cross-country Analysis” (Chapter 2) is based on the evidence of a negative relationship between the employment share of the service sector and the volatility of aggregate output, which I obtain after controlling for several factors (including the level of financial development). This aggregate result is driven by sectoral labor productivity differentials: Labor productivity is substantially more volatile in agriculture and manufacturing than in services. Aggregate output would therefore become mechanically more stable as labor shifts away from agriculture and manufacturing, and toward the service sector. To quantify this conjecture, I first calibrate a model of structural transformation (secular reallocation of labor across sectors) to the U.S. economy, which I use to match the time path of labor shares in agriculture, manufacturing and services across OECD countries. The model is subsequently used to conduct a set of counterfactual experiments in which labor is endogenously constrained from moving across sectors. Computations suggest that the shift of labor toward the services sector is indeed volatility-reducing. “Exchange Rate Volatility under Alternative Peg: Do Trade Patterns Matter?” (Chapter 3) is a contribution to the literature on the choice of exchange rate regimes. I use monthly bilateral exchange rate and external trade data from 1980 to 2010 for the member countries of the Western African and Monetary Union (WAEMU). These countries have their common currency (the CFA franc) pegged to the French franc since the mid-40s and to the euro since its introduction in 1999. At the time of the initial peg arrangement, France accounted for most of the external trade of WAEMU countries. Since then, and more notably since the early 2000s, the trade patterns of these countries shifted briskly away from France and other Euro area countries and towards the BICs (China in particular). The chapter finds that a peg to a composite basket of currencies would have led to a less volatile effective exchange rate over the last decade compare to the current hard peg. This chapter, however, does not derive an optimal exchange rate for WAEMU countries, which is an important area for further research.

Esta tese apresenta três ensaios sobre previsão de inflação e análise de dados em tempo real no Brasil. Utilizando uma curva de Phillips, o primeiro ensaio propõe um “modelo evolucionário” para prever inflação no Brasil. O modelo evolucionário consiste em uma combinação de um modelo não-linear (que é formado pela combinação de três redes neurais artificiais – RNAs) e de um modelo linear (que também é a referência para propósitos de comparação). Alguns parâmetros do modelo evolucionário, incluindo os pesos das combinações, evoluem ao longo do tempo segundo ajustes definidos por três algoritmos que avaliam os erros fora-da-amostra. As RNAs foram estimadas através de uma abordagem híbrida baseada em um algoritmo genético (AG) e em um algoritmo simplex de Nelder-Mead. Em um experimento de previsão fora-da-amostra para 3, 6, 9 e 12 passos à frente, o desempenho do modelo evolucionário foi comparado ao do modelo linear de referência, segundo os critérios de raiz do erro quadrático médio (REQM) e de erro absoluto médio (EAM). O desempenho do modelo evolucionário foi superior ao desempenho do modelo linear para todos os passos de previsão analisados, segundo ambos os critérios. O segundo ensaio é motivado pela recente literatura sobre análise de dados em tempo real, que tem mostrado que diversas medidas de atividade econômica passam por importantes revisões de dados ao longo do tempo, implicando importantes limitações para o uso dessas medidas. Elaboramos um conjunto de dados de PIB em tempo real para o Brasil e avaliamos a extensão na qual as séries de crescimento do PIB e de hiato do produto são revisadas ao longo do tempo. Mostramos que as revisões de crescimento do PIB (trimestre/trimestre anterior) são economicamente relevantes, embora as revisões de crescimento do PIB percam parte da importância à medida que o período de agregação aumenta (por exemplo, crescimento em quatro trimestres). Para analisar as revisões do hiato do produto, utilizamos quatro métodos de extração de tendência: o filtro de Hodrick-Prescott, a tendência linear, a tendência quadrática, e o modelo de Harvey-Clark de componentes não-observáveis. Todos os métodos apresentaram revisões de magnitudes economicamente relevantes. Em geral, tanto a revisão de dados do PIB como a baixa precisão das estimativas de final-de-amostra da tendência do produto mostraram-se fontes relevantes das revisões de hiato do produto. O terceiro ensaio é também um estudo de dados em tempo real, mas que analisa os dados de produção industrial (PI) e as estimativas de hiato da produção industrial. Mostramos que as revisões de crescimento da PI (mês/mês anterior) e da média móvel trimestral são economicamente relevantes, embora as revisões de crescimento da PI tornem-se menos importantes à medida que o período de agregação aumenta (por exemplo, crescimento em doze meses). Para analisar as revisões do hiato da PI, utilizamos três métodos de extração de tendência: o filtro de Hodrick-Prescott, a tendência linear e a tendência quadrática. Todos os métodos apresentaram revisões de magnitudes economicamente relevantes. Em geral, tanto a revisão de dados da PI como a baixa precisão das estimativas de final-de-amostra da tendência da PI mostraram-se fontes relevantes das revisões de hiato da PI, embora os resultados sugiram certa predominância das revisões provenientes da baixa precisão de final-de-amostra. / This thesis presents three essays on inflation forecasting and real-time data analysis in Brazil. By using a Phillips curve, the first essay presents an “evolutionary model” to forecast Brazilian inflation. The evolutionary model consists in a combination of a non-linear model (that is formed by a combination of three artificial neural networks - ANNs) and a linear model (that is also a benchmark for comparison purposes). Some parameters of the evolutionary model, including the combination weight, evolve throughout time according to adjustments defined by three algorithms that evaluate the out-of-sample errors. The ANNs were estimated by using a hybrid approach based on a genetic algorithm (GA) and on a Nelder-Mead simplex algorithm. In a 3, 6, 9 and 12 steps ahead out-of-sample forecasting experiment, the performance of the evolutionary model was compared to the performance of the benchmark linear model, according to root mean squared errors (RMSE) and to mean absolute error (MAE) criteria. The evolutionary model performed better than the linear model for all forecasting steps that were analyzed, according to both criteria. The second essay is motivated by recent literature on real-time data analysis, which has shown that several measures of economic activities go through important data revisions throughout time, implying important limitations to the use of these measures. We developed a GDP real-time data set to Brazilian economy and we analyzed the extent to which GDP growth and output gap series are revised over time. We showed that revisions to GDP growth (quarter-onquarter) are economic relevant, although the GDP growth revisions lose part of their importance as aggregation period increases (for example, four-quarter growth). To analyze the output gap revisions, we applied four detrending methods: the Hodrick-Prescott filter, the linear trend, the quadratic trend, and the Harvey-Clark model of unobservable components. It was shown that all methods had economically relevant magnitude of revisions. In a general way, both GDP data revisions and the low accuracy of end-of-sample output trend estimates were relevant sources of output gap revisions. The third essay is also a study about real-time data, but focused on industrial production (IP) data and on industrial production gap estimates. We showed that revisions to IP growth (month-on-month) and to IP quarterly moving average growth are economic relevant, although the IP growth revisions become less important as aggregation period increases (for example, twelve-month growth). To analyze the output gap revisions, we applied three detrending methods: the Hodrick-Prescott filter, the linear trend, and the quadratic trend. It was shown that all methods had economically relevant magnitude of revisions. In general, both IP data revisions and low accuracy of end-of-sample IP trend estimates were relevant sources of IP gap revisions, although the results suggest some prevalence of revisions originated from low accuracy of end-of-sample estimates.

Esta tese apresenta três ensaios sobre previsão de inflação e análise de dados em tempo real no Brasil. Utilizando uma curva de Phillips, o primeiro ensaio propõe um “modelo evolucionário” para prever inflação no Brasil. O modelo evolucionário consiste em uma combinação de um modelo não-linear (que é formado pela combinação de três redes neurais artificiais – RNAs) e de um modelo linear (que também é a referência para propósitos de comparação). Alguns parâmetros do modelo evolucionário, incluindo os pesos das combinações, evoluem ao longo do tempo segundo ajustes definidos por três algoritmos que avaliam os erros fora-da-amostra. As RNAs foram estimadas através de uma abordagem híbrida baseada em um algoritmo genético (AG) e em um algoritmo simplex de Nelder-Mead. Em um experimento de previsão fora-da-amostra para 3, 6, 9 e 12 passos à frente, o desempenho do modelo evolucionário foi comparado ao do modelo linear de referência, segundo os critérios de raiz do erro quadrático médio (REQM) e de erro absoluto médio (EAM). O desempenho do modelo evolucionário foi superior ao desempenho do modelo linear para todos os passos de previsão analisados, segundo ambos os critérios. O segundo ensaio é motivado pela recente literatura sobre análise de dados em tempo real, que tem mostrado que diversas medidas de atividade econômica passam por importantes revisões de dados ao longo do tempo, implicando importantes limitações para o uso dessas medidas. Elaboramos um conjunto de dados de PIB em tempo real para o Brasil e avaliamos a extensão na qual as séries de crescimento do PIB e de hiato do produto são revisadas ao longo do tempo. Mostramos que as revisões de crescimento do PIB (trimestre/trimestre anterior) são economicamente relevantes, embora as revisões de crescimento do PIB percam parte da importância à medida que o período de agregação aumenta (por exemplo, crescimento em quatro trimestres). Para analisar as revisões do hiato do produto, utilizamos quatro métodos de extração de tendência: o filtro de Hodrick-Prescott, a tendência linear, a tendência quadrática, e o modelo de Harvey-Clark de componentes não-observáveis. Todos os métodos apresentaram revisões de magnitudes economicamente relevantes. Em geral, tanto a revisão de dados do PIB como a baixa precisão das estimativas de final-de-amostra da tendência do produto mostraram-se fontes relevantes das revisões de hiato do produto. O terceiro ensaio é também um estudo de dados em tempo real, mas que analisa os dados de produção industrial (PI) e as estimativas de hiato da produção industrial. Mostramos que as revisões de crescimento da PI (mês/mês anterior) e da média móvel trimestral são economicamente relevantes, embora as revisões de crescimento da PI tornem-se menos importantes à medida que o período de agregação aumenta (por exemplo, crescimento em doze meses). Para analisar as revisões do hiato da PI, utilizamos três métodos de extração de tendência: o filtro de Hodrick-Prescott, a tendência linear e a tendência quadrática. Todos os métodos apresentaram revisões de magnitudes economicamente relevantes. Em geral, tanto a revisão de dados da PI como a baixa precisão das estimativas de final-de-amostra da tendência da PI mostraram-se fontes relevantes das revisões de hiato da PI, embora os resultados sugiram certa predominância das revisões provenientes da baixa precisão de final-de-amostra. / This thesis presents three essays on inflation forecasting and real-time data analysis in Brazil. By using a Phillips curve, the first essay presents an “evolutionary model” to forecast Brazilian inflation. The evolutionary model consists in a combination of a non-linear model (that is formed by a combination of three artificial neural networks - ANNs) and a linear model (that is also a benchmark for comparison purposes). Some parameters of the evolutionary model, including the combination weight, evolve throughout time according to adjustments defined by three algorithms that evaluate the out-of-sample errors. The ANNs were estimated by using a hybrid approach based on a genetic algorithm (GA) and on a Nelder-Mead simplex algorithm. In a 3, 6, 9 and 12 steps ahead out-of-sample forecasting experiment, the performance of the evolutionary model was compared to the performance of the benchmark linear model, according to root mean squared errors (RMSE) and to mean absolute error (MAE) criteria. The evolutionary model performed better than the linear model for all forecasting steps that were analyzed, according to both criteria. The second essay is motivated by recent literature on real-time data analysis, which has shown that several measures of economic activities go through important data revisions throughout time, implying important limitations to the use of these measures. We developed a GDP real-time data set to Brazilian economy and we analyzed the extent to which GDP growth and output gap series are revised over time. We showed that revisions to GDP growth (quarter-onquarter) are economic relevant, although the GDP growth revisions lose part of their importance as aggregation period increases (for example, four-quarter growth). To analyze the output gap revisions, we applied four detrending methods: the Hodrick-Prescott filter, the linear trend, the quadratic trend, and the Harvey-Clark model of unobservable components. It was shown that all methods had economically relevant magnitude of revisions. In a general way, both GDP data revisions and the low accuracy of end-of-sample output trend estimates were relevant sources of output gap revisions. The third essay is also a study about real-time data, but focused on industrial production (IP) data and on industrial production gap estimates. We showed that revisions to IP growth (month-on-month) and to IP quarterly moving average growth are economic relevant, although the IP growth revisions become less important as aggregation period increases (for example, twelve-month growth). To analyze the output gap revisions, we applied three detrending methods: the Hodrick-Prescott filter, the linear trend, and the quadratic trend. It was shown that all methods had economically relevant magnitude of revisions. In general, both IP data revisions and low accuracy of end-of-sample IP trend estimates were relevant sources of IP gap revisions, although the results suggest some prevalence of revisions originated from low accuracy of end-of-sample estimates.

This thesis examines the measurement, applications and properties of consumer inflation expectations in the context of eight European Union countries: France, Germany, the UK, Spain, Italy, Belgium, the Netherlands and Sweden. The data proceed mainly from the European Commission's Consumer Survey and are qualitative in nature, therefore requiring quantification prior to use. This study first seeks to determine the optimal quantification methodology among a set of approaches spanning three traditions, associated with Carlson-Parkin (1975), Pesaran (1984) and Seitz (1988). The success of a quantification methodology is assessed on the basis of its ability to match quantitative expectations data and on its behaviour in an important economic application, namely the modelling of wages for our sample countries. The wage equation developed here draws on the theoretical background of the staggered contracts and the wage bargaining literature, and controls carefully for inflation expectations and institutional variables. The Carlson-Parkin variation proposed in Curto Millet (2004) was found to be the most satisfactory. This being established, the wage equations are used to test the hypothesis that the advent of EMU generated an increase in labour market flexibility, which would be reflected in structural breaks. The hypothesis is essentially rejected. Finally, the properties of inflation expectations and perceptions themselves are examined, especially in the context of EMU. Both the rational expectations and rational perceptions hypotheses are rejected. Popular expectations mechanisms, such as the "rule-of-thumb" model or Akerlof et al.'s (2000) "near-rationality hypothesis" are similarly unsupported. On the other hand, evidence is found for the transmission of expert forecasts to consumer expectations in the case of the UK, as in Carroll's (2003) model. The distribution of consumer expectations and perceptions is also considered, showing a tendency for gradual (as in Mankiw and Reis, 2002) but non-rational adjustment. Expectations formation is further shown to have important qualitative features.