Defence Budget Forecast Tool
The International Institute for Strategic Studies
The IISS defense budget forecast employs a model based on a Bayesian network. The organization identified the key factors influencing defense budget allocations and, in conjunction with a data-driven model utilizing historical data, created an econometric forecasting model. This algorithm was developed in collaboration with data scientists from Objective. Initially, the data analytics team at Objective was assigned to create a data-driven algorithm to compare three distinct groups of countries, which involved reviewing and tidying data, engineering features, and exploring the datasets. Using statistical methods in R, three models were established, and after the IISS team's evaluation of their performance, the best model was chosen. Following promising results from the initial concept, a comprehensive analytics project to create a defense budget forecast using a Bayesian network model was authorized. This project included the construction, training, and testing of an econometric forecast algorithm that leveraged up to 45 highly correlated variables, such as macroeconomic factors, defense-related economic data, regime types, and dynamics of warfare and international relations. Expertise from IISS helped refine the model outputs for each country. With a soft launch in July 2020, the user interface was designed to analyze defense budget trends for selected markets: Australia, Russia, South Africa, South Korea, the United Kingdom, and the United States. Coverage has been broadened and will continue to expand to include most major markets globally. Furthermore, the tool allows users to adjust forecasts by applying percentage changes ranging from +25% to -25%, enabling them to modify the IISS projections for a more optimistic or pessimistic perspective, which serves as an overlay to the IISS projected figures.