Infrastructure, long-run economic growth and causality tests for cointegrated panels

We investigate the consequences of various types of infrastructure provision in a panel of countries from 1950 to 1992. We develop new tests which enable us to isolate the sign and direction of long-run effects in a manner that is robust to the presence of unknown heterogeneous short-run causal relationships. We show that while infrastructure does tend to cause long-run economic growth, there is substantial variation across countries. We also provide evidence that each infrastructure type is provided at close to the growth-maximizing level on average globally, but is under-supplied in some countries and over-supplied in others.