| This dissertation is a study of the properties of macroeconomic variables in the short run and in the long run. Particular attention is given to national saving and investment in the United States. The approach developed has implications about the effect of international capital mobility on saving and investment hence poses an answer to the saving-investment "puzzle" which has been discussed in the literature over the past decade.;For nonstationary macroeconomic variables, there exist long run stochastic trends in the variables. In a single variable model, the primary emphasis is given to the understanding and estimation of the trend component. On the one hand, a general decomposition model is formulated based on an ARIMA representation using the Nelson-Beveridge decomposition where other unobserved components decompositions are of special cases of the general model. On the other hand, the persistence of the series, which is embodied in the representation, is also considered. In a model of two or more variables, the general decomposition model gives rise to a "common trend" if variables are cointegrated.;The short run properties are obtained through removing a common trend from saving and investment. The correlation of saving and investment is found to be substantially lower than 1. In the original studies about international capital mobility and saving-investment correlation, Feldstein and Horioka (1980) posed the puzzle that the empirically high saving-investment correlation suggests immobility of capital across countries. Subsequent research with time series analysis indicates similar results. The findings in this work suggest that saving and investment have a "common trend" in the long run, and therefore the transitory fluctuations of saving and investment indicate no inconsistency in the international capital mobility story. |
