Forces are finally converging for a genuine debate on President Bush's missile-defense program. The Republican-controlled Congress is looking for ways to cut $9 billion from the military budget (which, at $420 billion, is getting unmanageable even for hawkish tastes). It's becoming painfully clear that rogues and terrorists are more likely to attack us with planes and trains than with nuclear missiles. And a recent series of technical studies—bolstered on Thursday by a high-profile Senate hearing—has dramatized just how difficult, if not impossible, this project is going to be.
Bush's budget for next year includes $10.7 billion for missile defense—over twice as much money as for any other single weapons system. This summer, he's planning to start deploying the first components of an MD system—six anti-missile missiles in Alaska, four in California, and as many as 20 more, in locations not yet chosen, the following year.
Yet, except by sheer luck, these interceptors will not be able to shoot down enemy missiles. Or, to put it more precisely, Bush is starting to deploy very expensive weapons without the slightest bit of evidence that they have any chance of working.
In the past six years of flight tests, here is what the Pentagon's missile-defense agency has demonstrated: A missile can hit another missile in mid-air as long as a) the operators know exactly where the target missile has come from and where it's going; b) the target missile is flying at a slower-than-normal speed; c) it's transmitting a special beam that exaggerates its radar signature, thus making it easier to track; d) only one target missile has been launched; and e) the "attack" happens in daylight.
Beyond that, the program's managers know nothing—in part because they have never run a test that goes beyond this heavily scripted (it would not be too strong to call it "rigged") scenario.
It's as if some kid were to hit a baseball thrown by a pitching machine straight down the middle at 30 mph and, on the basis of that feat, claimed he could hit whatever Mark Prior might throw him from a real mound, pitch after pitch after pitch, without fail.
There is, in other words, a vast distance between the Pentagon's current level of testing and the level that would need to be done before anyone could begin to claim that a missile-defense system might shoot down real enemy missiles in a real nuclear attack.
The latest annual report by Thomas Christie, the Pentagon's director of operational testing and evaluation, reveals just how incalculably vast this distance is. (The report was published with no fanfare at the end of last year and has appeared on private Web sites—but not the Pentagon's—in the past two weeks.)
Christie's bottom line is that we're rushing into this thing blind. Assessments of the system's capabilities are based primarily on "modeling and simulations" or on canned tests of "components and sub-systems," not on "operational tests of a mature, integrated system." Nothing can be reliably inferred from these data, because we don't know enough about the actual system that might be built and, therefore, don't know whether it bears any resemblance to the simulations. Or, as Christie puts it: "Due to the immature nature of the systems they emulate, models and simulations cannot be adequately validated at this time."
Step back and look at what a missile-defense system would involve. Broadly speaking, it would be a meshing of six separate operations: 1) an early warning radar, which would detect a missile launch; 2) satellite-based sensors that would distinguish missiles from deliberate decoys and random space clutter; 3) X-band radar that would track the missiles and control the firing of "kill vehicles" (anti-missile missiles that would shoot down enemy missiles); 4) the kill vehicles themselves; 5) booster rockets to launch the kill vehicles; and 6) the automated command-control-communications network that would connect all the above into a seamless system.