Military expenditure and economic growth: using causality, cointegration and missing variables

There is considerable interest in the relationship between military expenditure and economic growth, especially since Benoit's (1973, 1978) seminal work. Most studies of this linkage before our present work use a simple causality test. These results are unreliable or may be flawed, due to missing–variable bias and other limitations. This paper overcomes these deficiencies by deploying dynamic multivariate–causality tests applied to data from 22 countries over the period 1988–2012. Our novel panel–data estimation methods allow for more robust estimates by utilising variation between countries as well as variation over time. We find equilibrium relationships between military expenditure and economic growth and significant causality links deriving uniquely from our innovations.