Abstract:
In the face of wars and a geopolitically challenging environment, military expenditures have once again become political focal points in developed countries. However, the scientific literature remains inconclusive regarding their impact on economic growth. This paper conducts a meticulous meta-analysis, examining 405 estimates from 67 studies and incorporating over 30 variables to account for variations in their characteristics. The meta-analysis reveals a consistently negative average effect of military expenditures on economic growth, coupled with an absence or mild presence of publication bias. Both Bayesian and Frequentist model averaging highlight the diversity among individual estimates, attributing this variation to the data characteristics of individual studies. Notably, factors such as the panel structure, number of observations, number of countries, and time span emerge as crucial contributors to this diversity. The pivotal influence of data originating from the 1990s suggests the significance of de-escalation periods and hints at potential non-linearities within the observed effects. This paper makes notable contributions to prior meta-analyses by adopting an updated dataset, a more robust approach to publication bias analysis, and providing a more refined solution to addressing model uncertainty in the heterogeneity analysis.
