Abstract
This dissertation investigates the relationship between investment and military expenditure in the South African context between 1995 and 2016. The study uses a nonlinear autoregressive distributed lags (NARDL) model to test the existence of asymmetries in driving the military expenditure – investment nexus. Through positive and negative partial sum decompositions of the military expenditure, this approach permits one to test the nonlinear Cointegration using bounds test procedure. The NARDL model revealed the presence of a nonlinear impact of military expenditure on investment; suggesting that avoiding the intrinsic nonlinearities may provide skewed results. Particularly, it is found that an increase in military spending leads to an increase in investment while a decrease in military spending leads to a decrease in investment. In other words, 1 percent increase in military expenditure, holding other variables constant, will on average result in 3.2 percent increase in investment. On the other hand, 1 percent decrease in military expenditure, holding other variables constant, will on average result in 13.7 percent decrease in investment. From the increase in military expenditure point of view, the empirical results suggest that, to improve investment the South African government should add more funding towards military expenditure. The empirical results suggest that the South African government should consider prioritising funds towards the military expenditure particularly capital investment related expenditure in order to generate higher levels of investment and thus faster economic growth. With respect to the decrease in military expenditure, the empirical result suggests that reducing military expenditure may be detrimental to investment levels in the country.
M.Com. (Development Economics)