I have not written here on the New START treaty, in part because everything that can be said has been said, well, almost everything…see below. The treaty is in no way revolutionary. I don’t think Reagan would bat an eyelash at it. Yet, while there is widespread bipartisan support for the treaty, including almost all the leading defense specialists from former Republican administrations, there is also some opposition to the treaty, with the Heritage Foundation having taken it on as a cause. Some of the critiques are truly bizarre, such as the treaty does not address Russian tactical nuclear weapons or North Korea. (On that last point, would one of the critics please sketch out how we would have included North Korea in the negotiation?) Of course, no past arms control treaty has ever covered every type of weapon and if New START is not ratified then any chance of negotiating limits on tactical nuclear weapons is off the table completely. (The treaty does not cure world hunger either, another good cause.)
The one issue that opponents consistently latch onto is the supposed limits on missile defense. There is language in the preamble drawing attention to the connection between offensive and defensive missiles and in the text there is a limit on converting offensive missile launchers to be able to launch defensive missiles. Administration spokesmen have addressed these criticisms by saying the preamble language is not binding. I find it very strange that advocates of missile defense would like to argue that there is no connection between offensive and defensive missiles. Of course there is a connection between the two of them. Isn’t one supposed to shoot down the other? Isn’t that a connection? It is like arguing there is no connection between ships and torpedoes. (I think the connection is actually quite weak because defensive missiles probably cannot shoot much down, but that is a different story.) Simply saying that doesn’t seem to change much.
In the treaty proper, there is language that prohibits putting defensive missiles in offensive missile launchers. The administration has responded to criticism of this provision by saying it does not restrict us from doing anything we actually want to do. It is like agreeing to a restriction to not base our missiles on horse-drawn wagons. We don’t want to do that anyway, so why worry?
What I have not seen is an explanation of how, if the restriction is something we don’t care about, it got into the treaty in the first place. I have searched, perhaps it is out there, but I have not found anything. And I have not asked the negotiators directly either, so what follows is my speculation based on work long ago.
First of all, there was, at one point, at least some interest in doing this. During the START negotiations and the heyday of the Star Wars program, I was doing analytical and technical work at the Institute for Defense Analyses and the idea of mixing offensive and defensive missiles in silos did come up. I was doing mathematical modeling of the effectiveness of various defensive systems and, in particular, how defenses affected first strike stability (that is, the incentive for either side to strike first in a crisis).
Either side will want to have high confidence that, if they were planning on striking first, the attack would be successful and that is a challenge. Most of the missile silos will be destroyed with one warhead but if only some need to be hit with a second warhead and the attacker doesn’t know which ones, then all of the silos have to have a second warhead targeted on them. (This assumes that the attacker cannot do damage evaluation during the attack.) So failing in just some cases makes a big jump in the number of warheads the attacker needs. Defense of the missile silos exacerbates this problem.
It turned out that a system that had missile launchers that could hold either offensive or defensive missiles had some interesting mathematical properties that made a first strike very difficult. If the Soviets tried for a first strike and the defense system could predict which warheads were heading for which silos, then the defender could give up on some silos entirely, simply abandon them, and focus all the defenses on the rest. Say that the defensive system could get two warheads each that were heading to half of the silos. To be certain to overcome the defenses and still get two warheads on each silo, the Soviets then have to target four warheads on each silo. That means that the undefended silos get multiply whacked by four warheads but it makes no difference to the defender, once the missile inside is destroyed, it’s destroyed, destroying it tens times over doesn’t gain the attacker anything.
Now, if the defensive missiles are in identical silos and the attacker can’t tell which are which then he must attack the defensive silos too because they might have offensive missiles in them. Of course, by the time the attacking warheads arrive, the defensive silos will be empty because the defensive missiles would have long since been launched and the attacker is completely wasting warheads on empty silos. (This is what we called a “warhead sponge,” some useless target that just soaked up warheads.) If the defensive missiles were in, say, half the silos and the Soviets had to target four warheads on each silo, that means they would have to target a total of eight warheads for each silo that actually contains an offensive missile. It is getting hard to carry out a first strike! One can get into a great deal of detail, like whether the defensive missiles are launched in a single wave or some are held in reserve to specifically target incoming warheads missed in the first attack, but the basic idea is unchanged. I have no idea whether any such schemes were suggested to the Soviets in negotiations but they, no doubt, would have thought of something similar.
What is important to see here is that this trick only works if the defensive system is geared toward defending the ICBM fleet from attack. It doesn’t work if the attacker is going for, say, cities. Then the offensive missiles go for their targets and the defensive missiles go for the offensive missiles. The United States has a quite plausible first strike capability against Russia but Russia has no first strike capability against the United States, if for no other reason than the U.S. has so many warheads on invulnerable submarines. (Russia also has submarines but they rarely put to sea.) So the Russians might have an interest in this trick to defend their missiles silos but it is of little interest to the Americans. The restriction is, I suspect, entirely irrelevant but if it were significant at all, it favors the U.S. at the expense of Russia.
So why did this provision get into the treaty? It seems pretty obvious to me: it’s the counting rules. We say previous treaties limited warheads to such-and-such a number but they didn’t really. There was no technical way to count warheads. So there was a limit, not on warheads, but on delivery vehicles—that is, missiles—and each delivery vehicle was counted as though it had a certain number of warheads, regardless of how many it actually had. This number was based on the maximum number of warheads that the missile had ever been tested with. So it is perfectly possible that no missile actually had as many warheads as it is counted as. The “treaty limited” number of warheads was the number of delivery vehicles multiplied by the defined number of warheads per delivery vehicle, not the actual number. Well, not really…since we couldn’t count the number of delivery vehicles either, we counted the number of launchers, that is, the missile silos and assumed one missile per silo. So when we talked about limiting warheads, what we were really doing was limiting a surrogate for a surrogate for warheads, we were really limiting concrete-lined holes in the ground and using a simple mathematical formula to convert that into warhead numbers. New START is different because it limits launchers, missiles, and warheads. But for the launcher limit to make sense and to simplify the counting rules, it has to be clear-cut what can and cannot be in each silo. If we mixed offensive and defensive missiles in identical silos then even more frequent and more intrusive inspections would be required to sort it all out. I suspect the motivation for the restraint was to clarify the meaning of silo counting rules. This is nothing more than the kind of restraint that, if two different missiles could be stored in the same kind of canister, then the canister will count as the more dangerous of the two missiles or all bombers that are nuclear capable are counted as nuclear capable whether they are intended to carry nuclear weapons or not. We would have the same effect if we allowed the Russians to store grain in some of their silos and missiles in others. It just complicates things.
The scheme looks attractive if the goal is to protect missiles from a first strike but it has a major flaw: The construction of missile silos is very obvious from satellites. Construction of missiles themselves, less so. So, if the side with the mixed offense-defense array of silos wanted to, they could secretly build up missiles and quickly switch the defensive missile for offensive missiles, switching over from keeping a defended second-strike ICBM force to having an all offensive first strike force. This potential would worry the other side.
The rule applies to submarine-launched as well as land-based missiles. Mitt Romney, having not embarrassed himself enough with his Washington Post op-ed has come back with a piece in the National Review to try again. In regard to the prohibition on launcher conversion, he writes: “Under its terms, there could be an average of four or more SLBM tubes on each of our strategic submarines that no longer contain ballistic missiles but may not be converted for defensive interceptors, and so are empty.” So now he is suggesting we have mixed loads of offensive and defensive missiles on our submarines. However would the enemy game that? First of all, let’s put aside that submarines remain stealthy by remaining hidden beneath the sea and sticking antennae up is not the best way to remain hidden in the middle of the ocean so they are not in constant communication with the rest of the world. Thus, being an incommunicado submarine is a disadvantage when launching defensive missiles that must be launched within a one to three minute time window to hit their targets. Moreover, mixed loads invite attack on the submarine. The enemy will simply launch a couple of missiles over the area where the submarine is lying perfectly hidden and impossible to find. The submarine then reveals its location to any radar or sonar within a thousand miles as soon as it launches its first defensive interceptor. And where is the next enemy nuclear bomb headed? To the submarine, of course. No submariner would want to have mixed loads of offensive and defensive missiles. Governor Romney, though lack of understanding of nuclear strategy and tactics, has again made an issue of not being able to do something that no serious military planner wants to do.