Time Series Analysis of Macroeconomic Conditions in Open Economies

Abstract

The novelty of the present dissertation research is in the combination of time domain time series methods with bootstrap techniques. The former is widely used in the econometrics literature for estimation of dynamic models. The latter is used here for hypothesis testing purposes. The combined technique is here used to analyze three macroeconomic issues of current interest where accuracy in inference, rather than the lack of macroeconomic theory, has increasingly become the central concern. Each of the three issues is examined and discussed in a separate self-contained study. The first study tests the business cycle theory with application of an enhanced Augmented Dickey-Fuller test on the U.S. time series of real gross national product. Unlike previous studies, the null hypothesis of a unit root is rejected. The second study tests for IS-LM conditions in the U.S. during the post-Bretton Woods era by combining the Johansen's approach to cointegration with bootstrap algorithms. The estimated model produces a dynamic version of the IS-LM that permits short-term evaluations of fiscal and monetary policies. The third study seeks to explain the observed persistence in the Bolivan dollarization process. It is found that dollarization is now an irreversible process, with the Bolivian economy in transition toward equalization with U.S. prices and interest rates.