Despite the disparity in unemployment rates across gender, race and provinces, no study has captured the influence of military spending on unemployment rates among these groups in South Africa (SA). Thus, this study investigated the effects of military spending on total, gender, racial and provincial unemployment in SA over 2008Q1 to 2023Q4. The study applied Autoregressive distributed lag model (ARDL), Dynamic ARDL and Kernel-based Regularized Least Squares (KRLS), to predict the counterfactual shocks of unemployment rates based on a ±1 % change in military spending. From the ARDL, a rise in military spending reduced total, male, female, black race, and Gauteng, Eastern Cape, Mpumalanga Kwazulu Natal and Northwest provinces’ unemployment rates, in the short- and long run, but increased it among the Coloured, White and the Indian/Asian races and the Western Cape, Northern Cape and Free State provinces, in the short and long run. The DARDL simulation and the KRLS showed that a 1 % decrease (increase) in military spending increased (reduced) total, male, female, Black and North West, Gauteng, Eastern Cape, Mpumalanga, KwaZulu Natal and Limpopo unemployment rates in the short run before flattening in the long run. Conversely, a 1 % decrease (increase) in military spending reduced (increased) the unemployment rates in the Coloured, White and Indian/Asian Races, Western Cape, Northern Cape and Free State Provinces, in the short run while flattening it in the long run. The effect of military spending in SA are not homogeneous among the Races and Provinces. Therefore, government policies aimed at curbing unemployment should recognise the peculiarities of the races and provinces.