"Alliance commitments are most vulnerable when a major change occurs after the treaty was signed, such as implementation of a new system of government or a weakening of the country's international power," said Ashley Leeds, associate professor of political science.
Leeds collected 218 primary documents for military alliances that were signed between 1816 and 1944, and she coded information about what the countries would be required to do under various conditions if war broke out. She then identified 143 instances in which a war required action by an ally and researched whether the country fulfilled its commitment to an ally and the possible explanations for a treaty violation if one occurred.
"New data analysis provides evidence that alliance commitments are fulfilled about 75 percent of the time," Leeds wrote in the abstract to her research paper titled "Alliance Reliability in Times of War: Explaining State Decisions To Violate Treaties." The paper was published in the fall 2003 issue of the journal International Organization.
Because of the high costs associated with negotiating and instituting alliances, national leaders are more inclined to be serious about their commitment if they go to the trouble of formalizing a treaty; consequently, Leeds argued that treaty violations can best be understood by analyzing the factors that reduce the costs of violation or increase the costs of fulfilling the commitment. Alliances are particularly vulnerable to violation when factors affecting these costs change after the alliance is formed.
"Changes in the power of states or in their policy-making processes are powerful predictors of the failure to honor past commitments," Leeds said. She cited as an example the Soviet leaders' 1917 declaration of all prior Russian international commitments null and void, which resulted in all existing alliance treaties being ignored.
By compiling historical data and creating a model scenario, Leeds found that if a country's power changes at least 10 percent, the probability that the country will violate an alliance commitment rises 35 percent. The probability of alliance violation for an average country that has not had a significant change in its international power is only 6 percent, but for a country that has experienced at least a 10 percent change in power, the probability is 41 percent.
"Non-democratic countries and major powers that suffer lower costs from reneging on agreements are more likely to violate treaties," Leeds said, noting that the probability of a democratic state breaking a commitment is 16.5 percent lower than for a non-democratic state. Major powers have an 11 percent greater probability of violation than minor powers, who are more likely to fear retribution from stronger allies if they violate commitments.
"Bluffing seems to be a fairly rare phenomenon in alliance politics," Leeds said. "It seems increasingly unlikely that states frequently form alliances that they are unwilling to fulfill."
Leeds noted that more research on the conditions under which leaders will or won't comply with alliance obligations can help scholars and practitioners influence and anticipate policy outcomes.
"Leaders should assume that under most conditions, allies will fulfill their promises," she said. "This means that engaging a state in war when allies have promised to intervene is a proposition that entails a significant risk of war expansion. It also means that the deterrent and compellent properties of alliances should be taken quite seriously."
Leeds' research was supported by the National Science Foundation, Florida State University and Rice University. Her paper is part of the Alliance Treaty Obligations and Provisions project, a broader effort to collect data on the content of military alliance treaties.
Journal
International Organization