Time-Series Properties and Empirical Evidence of Growth and Infrastructure

Abstract

After more than forty years studying growth, there are two classes of growth models that have emerged: exogenous and endogenous growth models. Since both try to mimic the same set of long-run stylized facts, they are observationally equivalent in some respects. Our goals in this paper are twofold. First, we discuss the time-series properties of growth models in a way that is useful for assessing their fit to the data. Second, we investigate whether these two models successfully conform to post-war U.S. data. We use cointegration techniques to estimate and test long-run capital elasticities, exogeneity tests to investigate the exogeneity status of TFP, and Granger-causality tests to examine temporal precedence of TFP with respect to infrastructure expenditures. The empirical evidence is robust in confirming the existence of a unity long-run capital elasticity. The analysis of TFP reveals that it is not weakly exogenous in the exogenous growth model. Granger-causality test results show unequivocally that there is no evidence that TFP for both models precede infrastructure expenditures not being preceded by it. On the contrary, we find some evidence that infrastructure investment precedes TFP. Our estimated impact of infrastructure on TFP lay roughly in the interval (0.19, 0.27).