Investigating Time-Varying Causality Between Military Spending and Macroeconomic Indicators in the United States

Abstract Military expenditures constitute a large chunk of the United States’ annual budget and its macroeconomic implications had been modelled using the Granger causality test, which suffers power loss when variables are subjected to structural breaks. This study explored alternative approaches by applying both traditional VAR-based Granger causality and the time-varying causality test techniques to obtain new evidence on the causality between military spending and selected macroeconomic indicators (economic growth, investment, unemployment and inflation rate) in the United States. Relevant data covering 1972Q1–2021Q2 were analysed. The results of the VAR-based Granger Causality test are dominated by a unidirectional causality that runs from macroeconomic variables to military spending and the result are robust to alternative military spending measures. However, the results of the time-varying causality method show that bidirectional causality dominates the relationship between military spending and some macroeconomic indicators, especially economic growth, investment and unemployment. With the variance observed in the causality between military spending and macroeconomic indicators, policymakers need to moderate military spending to achieve desired economic outcomes.