TCS Daily

To Nuke, Or Not to Nuke?

By James D. Miller - October 21, 2002 12:00 AM

In war, against threats or only in retaliation, when should the U.S. strike with atomics? America wants to deter other nations from acquiring atomic weapons and to prevent U.S. troops from being attacked by mass-killing devices. Unfortunately, these two objectives conflict.

If Saddam uses chemical weapons against U.S. troops, should the U.S. respond atomically? We certainly don't need to go nuclear to crush even a chemically-enhanced Iraqi army, so the goal of our response strategy should be to deter other adversaries. Answering a chemical attack with atomics would certainly deter countries like Iran from striking us with chemical weapons. Unfortunately, it would also create incentives for future foes to acquire and use nuclear weapons, for if they know they would pay the atomic penalty for chemical attacks, why shouldn't they go all the way once they decide to cross the chemical threshold?

Our deterrence dilemma parallels the legal problem of marginal deterrence. Should a criminal who has killed once fear murdering again? If we impose the death penalty on all captured murderers, multiple murderers face no additional marginal punishment. Indeed, the higher the punishment for killing once, the lower the extra marginal punishment that can be imposed on multiple murderers. To deter multiple murderers we therefore need to set a low punishment for one murder to allow for significant enhanced punishments for those who kill repeatedly. Of course, the lower the punishment for single murders the greater their occurrence. We must therefore decide if it's more important to deter those considering killing one or many.

To determine our optimal response to chemical weapons we need to estimate the relative damage chemical and atomic weapons could cause our troops. I imagine we will always be vulnerable to atomic assaults, but I suspect we either already have or could develop protective gear that effectively shields our troops from chemical weapons. If so, we should respond only conventionally to chemical attacks. If, however, fear of chemical attacks would seriously limit our future military ability, we should use nuclear weapons to strike at foes who hit us with chemical weapons, even though such a nuclear response would limit our ability to marginally deter atomic attacks.

We could mitigate our marginal deterrence dilemma by having different levels of atomic response. We could perhaps answer a chemical attack with limited tactical nuclear strikes and engage in full-scale strategic nuclear assaults only against those who hit us with atomic weapons. Still, the higher the punishment we impose on those who use chemical weapons, the lower the additional punishment we can inflict on those who strike us with atomics.

Similar considerations manifest when deciding whether to execute a first-strike atomic attack. Imagine that after deposing Saddam Bush moves to defang North Korea, the nuclear-armed member of the Axis of Evil. Should we consider a preemptive atomic assault? Doing so would provide the maximum deterrence against countries considering acquiring atomic weapons, for it would signal to rogue states that merely possessing atomic weapons makes them likely to face nuclear attack. Of course, under such a response doctrine, once a rogue state acquires atomic weapons they would have greater incentives to use them before losing their weapons to a U.S. nuclear strike. Consequently, when determining when we should strike with nuclear weapons, the U.S. must decide what is more important: to deter their use or acquisition.

James D. Miller is an assistant professor of economics at Smith College.

TCS Daily Archives