The Price of Protection: Explaining Success and Failure of US Alliance Burden-Sharing Pressure
The Price of Protection: Explaining Success and Failure of US Alliance Burden-Sharing Pressure (ISBN: https://doi.org/10.1080/09636412.2021.2018624) was written by Brian Blankenship in 2021. It was published in Security Studies (vol. 30, no. 5).
The author examines burden-sharing in military alliances, especially NATO. They note a persistent argument in U.S. domestic politics that the country overpays for collective security. The predominant theories suggest that smaller powers free-ride because withholding their (relatively smaller) contributions does not undermine the alliance's success. The author however identifies several countries that contribute more than would be expected of them (i.e., Estonia, Latvia, and Lithuania).
The author proposes a patron-coercion model. Patrons are able to use threats of abandonment to coerce members into sharing more of the burden. Members that face greater exposure to risk are more likely to comply. The author here defines 'sharing more of the burden' as domestic military expenditures.
Analysis begins with the set of U.S. alliances identified in COW. Annual data covering 1950 to 2010 is collected for the level of national military spending in these countries. Countries in the Americas are excluded however; the author offers several explanations:
- these countries were all far removed from adversaries
the Rio Treaty covering almost all of these countries differed in important structural ways from all other military alliances
- covariates are nearly constant for these countries
Countries are coded as sharing a land border with a common adversary or not. Such adversaries are largely defined by the Soviet bloc (up to 1989). North Korea is a common adversary for the U.S. and South Korea throughout, and China is a common adversary for the U.S. and Japan from 1996 onward. An alternative coding is explored on the basis of having less than 400 miles over water between the country and a common adversary.
Countries are coded as being contiguous with a strategic maritime chokepoint or not. Only chokepoints within 400 miles of a common adversary are identified, such as the Greenland-Iceland-United Kingdom (GIUK) Gap which serves as a maritime chokepoint on Russia.
The count of nearby U.S. allies is also collected. Nearby here is defined as a land border or less than 400 miles over water.
The author fits an OLS model on this dataset predicting military expenditures. Country clustered errors are always used, and an alternative model uses annual and region fixed effects. Controls include:
- GDP level
- annual GDP growth
- regime type (binary democracy indicator)
- number of wars ("militarized interstate disputes" or "MIDs", as the author describes them) involved in
- number of adversaries not common with the U.S.
- whether the country had a colonial empire or not
- In some models only, also control for number of U.S. troops stationed in the country or for the country's distance from the U.S. These may be colinear with the indicator for bordering an adversary.
Countries sharing a land border with an adversary spend more, while countries with fewer nearby U.S. allies spend less. These suggest a complex pattern of geopolitical positioning; the immediate threat of an adversary is salient but threats of abandonment are less credible when the U.S. has no alternatives in the region.
Countries near a naval chokepoint spend less, again suggesting that threats of abandonment are less credible given certain geographies.
An additional model subsets to countries that do share an adversary with the U.S., and then adds descriptive controls based on the adversary's capabilities. These controls are:
- A rolling three year average of the number of wars the adversary was an initiator of.
- COW's Composite Indicator of National Capability (CINC) scores.
The relations are largely unchanged in this model.