Arms Negotiation, War Exhaustion, and the Credibility of Preventive War

Abstract

Why do some states agree to suspend their weapons programs in exchange for compensation while others fail to come to terms? I argue that the changing credibility of preventive war is an important determinant of arms construction. If preventive war is never an option, states can reach mutually preferable settlements. However, if preventive war is not credible today but will be credible in the future, a commitment problem results: the state considering investment faces a “window of opportunity” and must build the arms or it will not receive concessions later on. Thus, agreements fail under these conditions. I then apply the theoretical findings to the Soviet Union’s decision to build nuclear weapons in 1949. War exhaustion made preventive war not credible for the United States immediately following World War II, but lingering concerns about future preventive action induced Moscow to proliferate.

KEYWORDS:

• formal modeling
• negotiation
• power transition

Notes

¹ See Van Evera (1999) and Copeland (2000) for other recent treatments.

² One might alternatively conceptualize A as the coalition of states against B.

³ The results are the same if costs are only paid in the period of fighting. Moreover, the proof is identical except that substitutes for , where .

FIGURE 1 The baseline model. All payoffs listed are for the period, though the war outcomes lock in their respective payoffs every period for the rest of time.

⁴ Similar results would obtain if the costs of war changed in post-shift periods.

⁵ See the online supplemental appendix for complete proofs of this lemma and all the propositions. Throughout, I assume to avoid corner solutions.

⁶ The fact that preventive war does not occur here should be unsurprising since the game has complete information and power shift is observable and endogenous (Chadefaux 2011).

⁷ This extension nests the original model, with

⁸ The results presented would be similar if the reluctance to engage wore off nondeterministically.

⁹ Indeed, as the following case study discusses, Stalin was fully cognizant that the United States would apply greater coercive pressure to the Soviet Union as the years passed.

¹⁰ Thus, it must be that ; if , then A can always credibly threaten preventive war.

¹¹ To wit, American proliferation just a few years earlier required 130,000 workers (the size of America’s automobile industry at the time) to construct the first nuclear weapon (Hugues 2002:9). While secrets stolen from the Manhattan Project eased the Soviet effort, the Soviet Union lagged behind US efforts due to an inferior industrial base. This meant that the Soviet Union had to pay a greater cost in diverting industrial resources to the Soviet bomb project.

¹² This concept of war exhaustion dates back to at least Richardson (1960:232), who wrote that “a long and severe bout of fighting confers immunity on most of those who have experienced it; so that they no longer join in fights.” Many issues might cause the exhaustion, though political reluctance, manpower shortages, and limited military resources (Treisman 2004) are particularly problematic.

¹³ At the end of the war in Europe, Churchill commissioned a contingency plan, entitled Operation Unthinkable, which called for a surprise attack on the Soviets on July 1, 1945. Advisors ultimately scrapped the idea as infeasible; the best Britain could hope for was fleeting change in Poland, as an invasion of Russia would have been prohibitively difficult for the reasons outlined subsequently. Still, when collecting German arms, Churchill required British troops to organize the weapons in a manner such that they could be easily redistributed to the Germans, in case Britain needed German soldiers for the offensive (see Reynolds 2006:249–251).

¹⁴ Quoted in Harrington 2012 (101). See Harrington (99–118) for an overview of American pessimism.

¹⁵ See Schelling 1960 (199–201).

¹⁶ Throughout the process, it is worth noting that the Soviet Union was exhausted too: It had suffered roughly 20 times more casualties than the United States. Moscow correspondingly had no desire to turn the Berlin Blockade into the Berlin War (Harrington 2012:77–78). However, for the purposes of the commitment problem described, Soviet war exhaustion had little impact on the strategic interaction. Proliferating acts as a fait accompli to the opposing state. It is up to the opposing state to launch preventive war to stop it, which Washington was unwilling to do at the time.

¹⁷ American leaders also expressed moral concerns about how such a poorly targeted preventive war would undoubtedly result in a high number of civilian casualties, though others thought not engaging was immoral (Bhite and Hamel 2005:375). See Silverstone 2007 (51–75) for an overview of the normative concerns in Washington.

¹⁸ To wit, the calls for preventive war continued in Washington even after the Soviet Union started producing nuclear bombs (Buhite and Hamel 1990:376–381), though the overall consensus was the West was only ready to endure the number of causalities a proxy war could create.

¹⁹ In reality, Soviet commanders had tactical nuclear weapons—capable of striking Florida—available without needing launch codes from Moscow (Allison 2012:11). However, as a matter of establishing willingness to fight, Kennedy’s beliefs trump strategic realities.

²⁰ The estimated risk of nuclear war was substantially smaller, though (Gaddis 1997:269).

²¹ One might notice an apparent disconnect between the model and this case study here. The model says that if the costs of war are decreasing over time, a commitment problem can result. However, “costs” in this context refer to a combination of the physical toll and state resolve. Thus, even though war would have been more expensive, American “costs” in the crisis bargaining sense would have decreased due to resolve’s interaction.

²² Of course, a surgical preventive strike was possible in Cuba because of U-2 aerial photography. The United States did not have this luxury in the late 1940s to stop the Soviet Union from first acquiring nuclear weapons. Still, Kennedy knew that a strike on Cuban soil would inevitably kill Soviet troops and consequently spark a greater conflict with the Soviet Union. He was nonetheless willing to run this risk.

²³ One might then wonder why the United States did not intervene before the People’s Republic of China’s first test in 1964. Policymakers in Washington briefly considered it a possibility, especially as the Republic of China sought assistance. Compared to the Soviet Union, however, the United States’ differences with China were small. Regardless of war exhaustion, this raised the functional cost of intervention to unacceptably high levels; indeed, “Chinese nuclear capabilities would not [have posed] a major threat to U.S. interests, much less change the balance of power in East Asia” (Burr and Richelson 2000:56). As a result, Washington passed on the conflict and made peace with Beijing almost two decades before making peace with Moscow.

²⁴ With a population of around 140 million at the time, this means that roughly 1 in every 400 Soviets were involved in the project.