The first Chapter investigates Purchasing Power Parity (PPP) price convergence across U.S. states using a 1963-2018 panel dataset. We focus on precise measurements of the convergence rate towards PPP. The methodological sequence includes a benchmark AR (1) model, corrected for cross-sectional dependence and assessments for cross-sectional heterogeneity, yielding a more rapid rate of convergence via unbiased estimations of price level convergence. Specifically, in our major price-level indicator, we find a mean reversion rate of 3.29 years after correcting for cross-sectional dependence, as against 12.12 years before the correction. In addition, we examined the possibility of some states in certain region exhibiting faster price convergence relative to the others. We find a slower convergence rate in the Northern states relative to Southern states. Chapter two investigates the transmission of monetary policy into financial markets during the COVID-19 pandemic, examining the impact on interest rates, exchange rates, stock market indices, and long-term government bond yields across 17 advanced economies. Leveraging panel data spanning from 2002 to 2022, we employ interacted panel vector autoregression (PVAR) methodology to analyze differences in policy transmission between the pandemic period and normal times. Our findings indicate that the emergence of pandemic has weakened the transmission of monetary policy to financial markets to a large extent. Thus we see a notable difference in the effectiveness of conventional monetary policy during the pandemic, with policy rate changes exhibiting diminished impact on financial market variables compared to normal times. We attribute this divergence to heightened uncertainty, cautious investor behavior, and the unprecedented economic complexities brought about by the pandemic. Chapter three examines the persistence of productivity shocks across different U.S. states, employing the autoregressive coefficient to measure the extent to which past productivity levels influence current levels. The methodological sequence includes a benchmark AR (1) model, correction for cross-sectional dependence via the Dynamic Common Correlated Effects Model (DCCE), and assessments for cross-sectional heterogeneity. Our analysis reveals significant variation among states, with Pennsylvania displaying very high persistence and Virginia much lower persistence when using Minnesota as the numeraire. We also examine the half-life of productivity shocks, defined as the time required for the impact of a shock to reduce by half. Pennsylvania's half-life of 22.34 years suggests highly persistent shocks, whereas Virginia's 0.76 years indicates rapid dissipation. Minnesota's half-life is 1.93 years, providing a reference point for the comparison.

Made available in DSpace on 2016-03-15T19:31:02Z (GMT). No. of bitstreams: 1 Edward Bernard Bastiaan de Rivera y Rivera.pdf: 1871035 bytes, checksum: cf35ada408d6ca217674838231a332e2 (MD5) Previous issue date: 2014-04-09 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / The contribution of this thesis is the application of a new perspective in the assessment of the rational expectations model at the enterprise level, conducted in Nasseh and Straus (2004), Goddard, McMillan and Wilson (2008) and Rivera Rivera, Martin, Marçal and Basso (2012). These works investigate micro efficiency hypothesis in the sense presented in Jung and Shiller (2005) and adopt the restrictive hypotheses of independence in the error terms of the sample companies, as well as the absence of structural breaks. Common movements in stock prices characterize the systematic risk from unobserved common factors and structural breaks, such as financial crises, induce shifts in the relationship between prices and fundamentals. The objective of this investigation is the analysis of the rational expectations model at the enterprise level with constant and time-varying returns under the presence of multiple structural breaks and cross-sectional dependence. Recent econometric procedures of panel unit root and cointegration that contain properties of structural breaks and dependences are applied to Sao Paulo Stock Exchange (Bovespa) and S&P100 companies quarterly data covering the period of 1994-2012. Considering a multifactor error structure, results indicate a failure in the rejection of a unit root in (log) prices and (log) dividends. Panel cointegration analysis allowing structural breaks and cross-sectional dependence controlling to the presence of bubbles indicate that established procedures might not detect present structural breaks or generate oscilating statistics in size and power depending on the panel dimensions. A computational extention is developed allowing for enterprise-level structural breaks and cross-sectional dependence through bootstrap techniques. Results fail to reject the rational expectations model at the enterprise level and are consistent with the statistical significance and break dates detected in time-series cointegration routines. Evidences favor rationality and unforecastability of returns, where investors cannot profit consistently through speculation and active portfolio management. / A contribuição deste estudo é a aplicação de uma nova perspectiva na análise do modelo de expectativas racionais ao nível da empresa, conduzida em Nasseh e Strauss (2004), Goddard, McMillan e Wilson (2008) e Rivera Rivera, Martin, Marçal e Basso (2012). Estes trabalhos investigam a hipótese de micro eficiência analisada em Jung e Shiller (2005) e adotam hipóteses restritivas de independência nos termos de erro das empresas, assim como a ausência de quebras estruturais. Movimentos comuns nos preços dos ativos caracterizam o risco sistemático a partir de fatores não observados e as quebras estruturais, como crises financeiras, provocam mudanças nas relações entre preços e seus fundamentos. Neste sentido, o objetivo do estudo consiste na análise do modelo de expectativas racionais ao nível da empresa com taxas requeridas de retorno constantes e variáveis no tempo, admitindo-se a presença de múltiplas quebras estruturais e dependência nas seções transversais. Aplicam-se procedimentos econométricos recentemente desenvolvidos de raiz unitária e cointegração em painel que contenham as propriedades de quebras estruturais e dependências para dados trimestrais das empresas da Bovespa e S&P100 no período de 1994 a 2012. Os resultados indicam que, com uma estrutura de erro multifatorial, falha-se em rejeitar a hipótese nula de raiz unitária para (log de) preços e (log de) dividendos. Na análise de cointegração em painel com quebras estruturais e dependências que controlam a presença de bolhas racionais, procedimentos estabelecidos podem não detectar quebras presentes ou possuem estatísticas que oscilam em tamanho e poder conforme as dimensões do painel. Desenvolve-se uma extensão computacional que comporta quebras desconhecidas invididuais para as empresas e lida com dependências por meio de técnicas de reamostragens. Os resultados falham em rejeitar o modelo de expectativas racionais ao nível da empresa e são consistentes em significância estatística e datas de quebras detectadas em rotinas de cointegração para séries temporais aplicadas individualmente às empresas. Evidências favorecem a racionalidade dos agentes e imprevisibilidade de retornos, em que os investidores não poderiam obter lucros consistentes por meio da especulação e gerenciamento ativo de carteiras.