We investigate the practical use of Taylor-type rules in Trinidad and Tobago, which is in the process of implementing market based monetary policy and seeks to implement flexible inflation targeting in the presence of a managed exchange rate. This is motivated by the idea that normative Taylor rules can be shaped by the practical experience of developing countries. We find that the inflation – exchange rate nexus is strong, hence the country may be unwilling to allow the exchange rate to float freely. We contend that despite weak market development the Taylor rule can still be applied as the central bank is able to use moral suasion to achieve full pass through of the policy rate to the market rate. Our evidence rejects Galí and Monacelli’s (2005) argument that the optimal monetary policy rule for the open economy is isomorphic for a closed economy. Rather, our evidence suggests that the rule for the open economy allows for lower variability when the rule is augmented by the real exchange rate as in Taylor (2001). We also reject Galí and Monacelli’s (2005) hypothesis that domestic inflation is optimal for inclusion in the Taylor-type rule. Instead we find that core CPI inflation leads to lower variability. Additionally, our evidence suggests that the monetary rule, when applied to Trinidad and Tobago, is accommodating to the US Federal Reserve rate. Further, we expand the work of Martin and Milas (2010) which considered the pass through of the policy rate to the interbank rate in the presence of risk and liquidity. By extending the transmission to the market lending rate, we are able to go beyond those disruptive factors by considering excess liquidity and spillovers of international economic disturbances. We found that these shocks are significant for Trinidad and Tobago, but it is not significant enough to disrupt the pass through. As a result, full pass through was robust to the presence of these disruptive factors.

CoordenaÃÃo de AperfeiÃoamento de Pessoal de NÃvel Superior / The main objective of this study is quantify the impact caused by changes in federal tax on the tax policy of the states, considering the presence of horizontal tax competition between states. In federal systems, where various levels of government operate simultaneously in the same tax regime, setting tax rates on the same basis of taxation, there may be the existence of externalities. In that sense, the tax policy of a government level affects the size of the tax base of another level. Using a panel data covering 26 Brazilian states plus the Federal District for the period 1995-2009, it was estimated two models that represent the average reaction of state taxes in response to changes in federal tax rates, considering horizontal tax competition as a relevant variable in the model. The first model assumes that both levels of government act together simultaneously in setting its taxes rates (Cournot-Nash Model). The second model assumes that states react with delay to changes in federal taxes, as if this act as a Stackelberg leader in setting their tax (Model Stackelberg). The results indicate a positive response from the states to significant increases in the federal tax rate, in the presence of horizontal tax competition between states, resulting in a overtaxation tax base common to high taxes rates. The results of the J test proposed by Davidson and MacKinnon to choose the model indicate that state and federal adjust their rates both together as in a Cournot-Nash game. / O objetivo principal desse trabalho à quantificar qual o impacto provocado por mudanÃas no imposto federal sobre a polÃtica tributÃria dos estados, considerando a presenÃa de competiÃÃo tributÃria horizontal entre os estados. Em sistemas federativos, quando vÃrios nÃveis de governo operam simultaneamente em um mesmo regime fiscal, definindo alÃquotas sobre as mesmas bases de tributaÃÃo, pode ocorrer a existÃncia de externalidades. Nesse sentindo, a polÃtica tributÃria de um nÃvel de governo afeta o tamanho da base tributÃria do outro nÃvel. Utilizando um painel de dados que abrange os 26 estados brasileiros mais o Distrito Federal para o perÃodo 1995â2009, foram estimados dois modelos que representam a reaÃÃo mÃdia dos estados em resposta a mudanÃas nas alÃquotas do governo federal, considerando a competiÃÃo tributÃria horizontal como variÃvel relevante do modelo. O primeiro modelo considera que ambos os nÃveis de governo agem simultaneamente em conjunto no ajuste de suas alÃquotas (Modelo Cournot-Nash). Jà o segundo modelo considera que os estados reagem com atraso a mudanÃas nos impostos do governo federal, como se este agisse como um lÃder Stackelberg na definiÃÃo de seus impostos (Modelo Stackelberg). Os resultados apontam para uma resposta positiva e significante dos estados à aumentos na alÃquota tributÃria federal, na presenÃa de competiÃÃo tributÃria horizontal entre os estados, resultando em uma sobretaxaÃÃo da base tributÃria comum e alÃquotas elevadas. Os resultados do teste J propostos por Davidson e MacKinnon para escolha do modelo indicam que estados e federaÃÃo ajustam suas alÃquotas simultaneamente em conjunto como em um jogo de Cournot-Nash.