ABSTRACT
Arms dependency is typically framed as a security issue that states seek to avoid. Dependency creates an opportunity for an exporter to attempt to exert influence over the importer’s foreign and domestic policy. However, the arms trade is a trade and influence attempts to create economic costs for exporters by damaging relationships with current and potential customers. Thus, heavily dependent states do not necessarily need to change suppliers to avoid the threat. Additionally, as arms transfers are a signal of political support, dependency may be a sign of a mutually beneficial relationship rather than one that is potentially dangerous. This article evaluates these arguments using logistic regression models to evaluate changes in suppliers of major weapons systems. It finds that the relationship between dependence and arms transfers is more complex than previously argued where the nature of the relationship depends both on the type of exporter and the type of arm being exported.
Acknowledgments
I would like to thank Scott Gartner, Caitlin Milazzo, Jeannette Money, Randy Siverson, Shaina Western, and an anonymous reviewer for their comments. All errors remain my own.
Disclosure statement
No potential conflict of interest was reported by the author.
Supplementary Materials
Supplemental data for this article can be accesed here.
Data Availablity
Replication data is available at www.richard-johnson.net.
Notes
1. Due to data limitations, only the level of dependence on weapon systems can be tested. While transfers of spare parts and ordnance are important, data are not available.
2. I thank an anonymous reviewer for highlighting this argument.
3. The motives I assign are the primary motive, not sole motive, of the supplier type.
4. As these producers rarely get the opportunity to use their goods in true international conflict (obviously a good thing) there are also economic motives to not exert influence to restrict importers – displaying battlefield competency of the goods, particularly by a comparatively less trained military, helps increase the reputation of the good.
5. For example, during the War of Attrition, the USSR agreed to provide Egypt with SA-3 surface-to-air missiles, but this complicated weapon system required six months of training for the operating team to become independent of their training staff (Maoz Citation2006). If the Russian advisers had not remained during the war to help Egyptian troops use the weapon system, the outcome may have been even worse for Egypt.
6. Ship-based systems are excluded from the analysis.
7. A long-term external threat measure is an importer’s strategic reference group (SRG) (Maoz Citation2009). This measure is the ratio of the importer’s capability level with capability level of the importer’s enemies and the enemies’ allies. Data from the ATOP alliance data set (Leeds et al. Citation2002), Thompson’s (Citation2001) rivalry data, and the Correlates of War CINC scores are used to create this measure. However, these data have a limited date range until 2000 and the coefficients in the tests are consistently insignificant; thus, I do not include the variable in the models in order to extend the date range through 2007. These models are in the online appendix.
8. This measure may seem extreme, but there are few data sets on civil conflict that cover pre-1980.
9. Alternatively, in a monadic research design a measurement of number of defense pact partners has been used before as a proxy for capability (e.g. Siverson and Johnson Citation2016), which does not account for capabilities.
10. Models without importer capability are presented in the online appendix.
11. Tests using fixed effects are presented in the online appendix, as are tables with correlation matrices and summary statistics.
12. No graphs are presented for self-propelled guns and support helicopters as the different types of dependency have no substantive effect on the probability of change.