Rob Farley argues, contra Ed Morrissey, against deploying a partially effective>ballistic missile defense (BMD) system:

Ed, let me explain something to you, slowly and carefully. Missile defense, at least when conceived as a response to the threat of nuclear attack on the United States, needs to be “complete and perfect.” Otherwise it’s useless. There are virtually no foreign policy goals that a President will consider worthwhile if there’s a 5% risk that the destruction of American cities will result. 80% doesn’t cut it; 95% doesn’t, and probably not even 99%. This is not a new objection to missile defense; analysts have understood that defense against nuclear armed ballistic missiles needs to be 100% for quite some time, which is why so many intelligent people have rejected the possibility that a missile defense shield could provide useful protection for the United States. Now, it’s fair to say that the same logic does not apply to conventional ballistic missile attacks on either cities or military targets; in those cases, an 85% effective missile shield is useful. But for preventing Minneapolis from disappearing under a nuclear mushroom cloud, not so much.

Rob’s wrong. There are a number of arguments in favor of a less-than-perfect BMD system.

• Accidental Launch. If a small number of nuclear missiles launch as a result of malfunction or malfeasance, then I can imagine wanting a partially effective defense system. A small number of incoming nukes present the “best case” scenario for such a system providing adequate protection–we wouldn’t be dealing with an enormous number of warheads and decoys. And even one less hit might amount to millions of fewer casualties.

• Ensuring First-Strike Dominance. Why do the Russians and Chinese dislike BMD despite the likelihood that they would be able to get enough missiles through to make war very difficult for the US to contemplate? Because they worry about a US preventive or preemptive strike. While a partially effective system would almost certainly be useless against a Russian first strike, it might prove sufficient to deal with whatever the Russians had left after a US counterforce strike. While Russia and China might be able to take effective countermeasures, such as further enhancing the survivability of their arsenals and deploying more missiles, consider a “new” nuclear power, like Iran, Pakistan, or India. The US already enjoys overwhelming nuclear superiority against such states, so even a mediocre BMD system might be just enough of a safety net to allow the US to contemplate a first strike in the event of a crisis or conventional hostilities.

• Enhancing Force Projection. Even if the US isn’t contemplating a first strike against a new nuclear power, US policymakers would certainly prefer to minimize the ability of such states to deter US coercive diplomacy, or even US intervention, by threatening to use nuclear weapons against the US homeland or against US troops. Even if a new nuclear power could make a reasonable bet that a few of their warheads might get through a BMD or THAAD system, the combination of a such systems and US retaliatory capability might reduce the credibility of their threat to cross the nuclear threshold.

Some even make a more cynical argument: given that a country like Iran knows that, in return for the loss of a division or a city, the US could turn it to glass, it follows that they cannot make a credible nuclear threat against the US. But such a threat might be enough to preclude the public from supporting, for instance, a US intervention in the Middle East. In this case, a President might find it useful to invoke the protection of a BMD system–even knowing it probably wouldn’t stop everything–in order to reassure key constituencies and “allow” the threat of US nuclear retaliation to prevent a conventional war from escalating.

Of course, deploying a BMD system isn’t costless. It might trigger a nuclear arms race with the Chinese and the Russians, enhance crisis instability and the risks of preemptive strikes, and so forth. Even if we accept these arguments in favor of a partially effective system, therefore, we could–and probably should–oppose its deployment.

But I still think Rob’s wrong to conclude that “it’s nonsense all the way down.”