This paper reviews some of the theoretical and econometric issues involved in estimating growth models that include military spending. While the mainstream growth literature has not found military expenditure to be a significant determinant of growth, much of the defence economics literature has found significant effects. The paper argues that this is largely the product of the particular specification, the Feder–Ram model, that has been used in the defence economics literature but not in the mainstream literature. The paper critically evaluates this model, detailing its problems and limitations and suggests that it should be avoided. It also critically evaluates two alternative theoretical approaches, the Augmented Solow and the Barro models, suggesting that they provide a more promising avenue for future research. It concludes with some general comments about modelling the links between military expenditure and growth.