Logo image
Back
Price of service: A comparative study of money and military retention using time series and predictive analytics
Thesis   Open access

Price of service: A comparative study of money and military retention using time series and predictive analytics

Yvena Aristilde
University of West Florida Libraries
Master of Science (MS), University of West Florida
2026

Metrics

1 Record Views

Abstract

ANN Military budget Military retention Neural network RNN Time series models Military studies
The ability of the United States military to maintain a strong and ready force depends largely on retaining its most experienced personnel. However, predicting the reasons service members choose to stay or leave remains a significant challenge for policymakers. This thesis examines the effects of defense spending and regular military compensation on retention, with a specific focus on the differential responses of officers (rank O-3) and enlisted (rank E-5) personnel to financial incentives. The research employs two complementary approaches to investigate potential solutions. First, Vector Autoregression (VAR) models are used to capture dynamic relationships in historical data and evaluate how changes in military pay influence force size over time. Second, a Recurrent Neural Network (RNN) is used to model sequential patterns and forecast differential responses of officers and enlisted personnel to financial incentives. Both models suggested that the monetary incentives differ among officers and enlisted personnel. O-3 Officer personnel respond to greater sensitivity to immediate pay raises and bonuses, whereas E-5 Enlisted personnel respond to a broader set of long-term career incentives and overall defense spending. By understanding these differences, the military can move away from standardized policies and develop rank-based strategies that more effectively retain highly skilled personnel.
pdf
Price of service1.98 MBDownloadView
Preprint link to pre-print Open Access

Details

Logo image