In this paper we aim to improve existing empirical exchange rate models by accounting for uncertainty with respect to the underlying structural representation. Within a flexible Bayesian non-linear time series framework, our modeling approach assumes that different regimes are characterized by commonly used structural exchange rate models, with their evolution being driven by a Markov process. We assume a time-varying transition probability matrix with transition probabilities depending on a measure of the monetary policy stance of the central bank at the home and foreign country. We apply this model to a set of eight exchange rates against the US dollar. In a forecasting exercise, we show that model evidence varies over time and a model approach that takes this empirical evidence seriously yields improvements in accuracy of density forecasts for most currency pairs considered. / Series: Department of Economics Working Paper Series

O presente trabalho teve por objetivo investigar se há evidências de que o regime de política cambial brasileiro teria se alterado no pós-crise econômico-financeira internacional de 2008, além de captar insights acerca da eficácia dos instrumentos de intervenção recentemente aplicados sobre o mercado de moedas e acerca do poder explicativo dos fundamentos da taxa de câmbio. Três fatos estilizados do mercado de câmbio brasileiro incitam essa investigação. O primeiro deles encontra-se no histórico de mudanças de regime cambial do Brasil e das demais economias emergentes, as quais geralmente têm ocorrido em momentos de crises internacionais por impossibilidade dos governos em sustentar o regime vigente. O segundo assenta-se na dinâmica recente do real cuja variação tem se descolado da dinâmica das demais moedas commodities currencies. Por fim, o terceiro ponto está nas inovações recentes em política de intervenção das autoridades monetárias e fiscais visando à gestão da cotação da moeda com medidas tais como as modificações das alíquotas do IOF sobre operações cambiais. Para o alcance dos objetivos foi utilizado o modelo Markov Switching desenvolvido por Hamilton (1989) aplicado ao modelo estrutural de curto prazo para taxa de câmbio, no qual foram consideradas como variáveis explicativas os fundamentos e as intervenções na taxa de câmbio descritas na literatura atual. O modelo foi estimado com ambas as variáveis sujeitas a estados não observáveis, cujas probabilidades de ocorrência são geradas por um processo markoviano. Como resultado foi identificado que não há evidências representativas de modificação na dinâmica da taxa de câmbio no pós-crise que conduzam a interpretação de alteração no regime de política cambial adotado pelo Brasil. Verificou-se também que as intervenções das autoridades monetárias e fiscais não obtiveram eficácia em gerir ou direcionar a variação ou nível da moeda. Ademais, a aplicação do modelo de mudança de regime na série da taxa de câmbio permitiu compreender a sua dinâmica no período recente e o modo como os fundamentos e intervenção perdem e ganham influência na determinação da cotação ao longo dos ciclos da moeda. / This study seeks evidences of changes in the Brazilian foreign exchange rate regime after the international economic and financial crisis of 2008. Furthermore, captures insights on the effectiveness of the intervention measures recently applied on the currency market and the explanatory power of the fundamentals of exchange rates. Three particular facts of the Brazilian foreign exchange market stimulated this research. First, the historical changes of foreign exchange rate regimes in Brazil and other emerging economies. Generally, it has occurred in times of international crises due to the failure of governments to sustain the regime. Second, recent deviation of the Real, which has been detached from the dynamics of other commodities currencies. Finally, the third fact is the recent innovations on monetary and fiscal intervention policies by authorities in order to manage currency value, such as adjustments of IOF on foreign exchange transactions. This investigation uses Markov switching model, developed by Hamilton (1989), applied to the structural model for short-term exchange rate, considering the fundamentals and interventions on the exchange rate described in the current literature as explanatory variables. The model was estimated with both variables subjected to unobservable states, whose probabilities of occurrence are generated by a Markov process. As result, there is no significant evidence of changes in the dynamic of exchange rate in the post-crisis that explains changes in foreign exchange rate regimes adopted by Brazil. The intervention by monetary authorities was not efficient in managing or guiding the dynamic of foreign exchange rate. In addition, the application of the Markov switching model in the foreign exchange rate series allowed understanding the dynamics in recent periods and how fundamentals and interventions loses and gain influence in determining prices throughout the currency cycles.

O presente trabalho teve por objetivo investigar se há evidências de que o regime de política cambial brasileiro teria se alterado no pós-crise econômico-financeira internacional de 2008, além de captar insights acerca da eficácia dos instrumentos de intervenção recentemente aplicados sobre o mercado de moedas e acerca do poder explicativo dos fundamentos da taxa de câmbio. Três fatos estilizados do mercado de câmbio brasileiro incitam essa investigação. O primeiro deles encontra-se no histórico de mudanças de regime cambial do Brasil e das demais economias emergentes, as quais geralmente têm ocorrido em momentos de crises internacionais por impossibilidade dos governos em sustentar o regime vigente. O segundo assenta-se na dinâmica recente do real cuja variação tem se descolado da dinâmica das demais moedas commodities currencies. Por fim, o terceiro ponto está nas inovações recentes em política de intervenção das autoridades monetárias e fiscais visando à gestão da cotação da moeda com medidas tais como as modificações das alíquotas do IOF sobre operações cambiais. Para o alcance dos objetivos foi utilizado o modelo Markov Switching desenvolvido por Hamilton (1989) aplicado ao modelo estrutural de curto prazo para taxa de câmbio, no qual foram consideradas como variáveis explicativas os fundamentos e as intervenções na taxa de câmbio descritas na literatura atual. O modelo foi estimado com ambas as variáveis sujeitas a estados não observáveis, cujas probabilidades de ocorrência são geradas por um processo markoviano. Como resultado foi identificado que não há evidências representativas de modificação na dinâmica da taxa de câmbio no pós-crise que conduzam a interpretação de alteração no regime de política cambial adotado pelo Brasil. Verificou-se também que as intervenções das autoridades monetárias e fiscais não obtiveram eficácia em gerir ou direcionar a variação ou nível da moeda. Ademais, a aplicação do modelo de mudança de regime na série da taxa de câmbio permitiu compreender a sua dinâmica no período recente e o modo como os fundamentos e intervenção perdem e ganham influência na determinação da cotação ao longo dos ciclos da moeda. / This study seeks evidences of changes in the Brazilian foreign exchange rate regime after the international economic and financial crisis of 2008. Furthermore, captures insights on the effectiveness of the intervention measures recently applied on the currency market and the explanatory power of the fundamentals of exchange rates. Three particular facts of the Brazilian foreign exchange market stimulated this research. First, the historical changes of foreign exchange rate regimes in Brazil and other emerging economies. Generally, it has occurred in times of international crises due to the failure of governments to sustain the regime. Second, recent deviation of the Real, which has been detached from the dynamics of other commodities currencies. Finally, the third fact is the recent innovations on monetary and fiscal intervention policies by authorities in order to manage currency value, such as adjustments of IOF on foreign exchange transactions. This investigation uses Markov switching model, developed by Hamilton (1989), applied to the structural model for short-term exchange rate, considering the fundamentals and interventions on the exchange rate described in the current literature as explanatory variables. The model was estimated with both variables subjected to unobservable states, whose probabilities of occurrence are generated by a Markov process. As result, there is no significant evidence of changes in the dynamic of exchange rate in the post-crisis that explains changes in foreign exchange rate regimes adopted by Brazil. The intervention by monetary authorities was not efficient in managing or guiding the dynamic of foreign exchange rate. In addition, the application of the Markov switching model in the foreign exchange rate series allowed understanding the dynamics in recent periods and how fundamentals and interventions loses and gain influence in determining prices throughout the currency cycles.

Magister Commercii - MCom / Issues pertaining to the misalignment of exchange rate have become central in the analysis of open economy macroeconomics for developing countries. This is at least due to two reasons: first persistent overvaluation of currency is seen as a powerful early warning of potential currency crisis and second protracted periods of exchange rate misalignment are highly associated with poor economic performance in a number of developing countries. Owing to this fact, economists are in concession that aligning real exchange rates towards their equilibrium values is an important component of macroeconomic policy adjustments in order to achieve and maintain a sustainable development. For this purpose the estimation of the degree of the real exchange rate misalignment has become pivotal. However, despite the concession among economists regarding the need to minimize the frequency and magnitude of exchange rate misalignment, the estimation of the equilibrium exchange rate (hence the misalignment) has been among the most controversial and challenging issues in modem macroeconomics. For several decades, the Purchasing power parity (PPP) approach-which is based on the law of one price-has been the most widely used methodology for the estimation of the equilibrium exchange rate in both developed and developing countries. In South Africa some attempts have been made to estimate the misalignment of the rand against major currencies on the basis of the PPP approach. However, large numbers of empirical studies show that PPP does not hold except in the 'ultra' long run. In addition, PPP's assumption of a constant equilibrium exchange rate makes it ill-fitted to serve as a bench-mark for the analysis of the exchange rate in countries such as South Africa that experience substantial structural changes. As a result a number of macro-econometric models underlying on the macroeconomic determinants of exchange rate have been developed, albeit with little applicability in developing countries. In this study, we have used Edwards' (1989) intertemporal general equilibrium model of a small open economy in order to estimate the degree of the real exchange rate misalignment and its impact on the international trade competitiveness of the South African economy for the period 1985:1-2000:4. For this purpose a dynamic single equation error correction model of a first order autoregressive distributed lag model, ADL (1,1), and five years moving average technique have been employed to estimate the exchange rate misalignment. Whereas impulse response analysis and variance decomposition techniques of a cointegrated VAR (vector auto regression) have been established to assess the impact of the misalignment on trade competitiveness. The fmdings of the study reveal that the real exchange rate had been consistently overvalued during the period' 1988:3-1998:2 but undervalued during periods 1998:3- 2000:4. For most of the periods during 1985:1-1988:2 the rand had been undervalued. More over the study discloses that exchange rate misalignment debilitates South Africa's international trade competitiveness accounting for 20 percent of the variation in competitiveness.