This contribution deals with modelling of military expenditure and its potential security and economic determinants by an Autoregressive Distributed Lag (ARDL) model. This approach is applied to several NATO member countries over the period 2001–2016, namely to Visegrad group countries and Baltic states. Time series of military expenditure (database SIPRI), the risk of inflation, GDP growth, terrorism, foreign pressure, cross-border conflict and ethnic tension (database of Political Risk Service Group) are used in analysis.