Disaggregated defense spending: Introduction to data

Abstract

Theoretical and empirical research on causes and consequences of defense spending is plentiful. Most of this research uses ‘top line’ defense spending data, either as a share of GDP or as a raw monetary figure. Empirical research has been limited, however, by the ‘blunt’ nature of this data, which does not help to explain what countries are spending on. We introduce a dataset that provides information on disaggregated defense spending from 35 NATO and EU members over as many as 51 years. We discuss the main features of this data in the paper, and the replication files will enable other scholars to automate accessing it in the future. In addition to automating the extraction of NATO and European Defence Agency data on overall military expenditures, we make data on equipment, personnel, operating, and infrastructure spending available in a single dataset. We illustrate the utility of the disaggregated defense spending dataset by replicating canonical and newer analyses using both the overall data and its disaggregated components. The findings differ depending on which type of spending is considered. We found that differences in the relationship between national wealth and defense spending depended on the category of spending considered, as did the tendency toward ‘free-riding’. These exercises shed new light on seminal theories on burden-sharing and the political economy of security. Our initial analysis suggests that disaggregating defense spending is likely to improve the analysis of old and emerging research questions of considerable policy importance, and points to several opportunities to do so.