Jeffrey LewisA Potemkin Defense: Integrated Flight Tests Hate Our Freedoms

Arms Control has obtained the full text of the Final Report of the Missile Defense Agency’s Independent Review Team (IRT), first reported by Bradley Graham in the Washington Post.

Graham, however, missed the scoop—the IRT perversely recommends reducing future flight tests because … this is so priceless … more failures might undermine the system’s meager deterrent value:

Successful test intercepts will send a strong message to adversaries of the U.S., who may be dissuaded by the effectiveness of the system from investing further in ballistic missile forces and/or be deterred from attacking the U.S., our deployed forces, our allies, and friends. Therefore, successful flight testing is a strategic issue as well as a key to continued successful system development.

The IRT recommends five changes, the cumulative effect of which will be to make testing less likely:

1. Establish a More Rigorous Flight Readiness Certification Process
2. Strengthen Systems Engineering
3. Perform additional ground-based qualification testing as a requirement for flight testing
4. Hold contractor functional organizations accountable for supporting prime contract management
5. Assure that the GMD program is executable

The general effect of these recommendations is to create a presumption against conducting any scheduled flight test—what the IRT calls “Prove why should fly”. For good measure, the IRT recommends making the next integrated flight test a “non-intercept” test.

Imagine that: First, MDA rushes a defense that won’t defend to meet a deadline that just happens to coincide with a Presidential election. Then, MDA scales way back on necessary testing, lest the bad guys figure out the damn thing doesn’t work.

Brace yourself, it gets worse. Kim Jong-Il probably knows the system is dog; the dunce who concerns me is George W. Bush.

Bush told a Pennsylvania audience that potential adversaries know “You fire, we’re going to shoot it down”—a disturbing remark that suggests Administration officials may have an exaggerated notion of effectiveness of the system.

Undersecretary E.C. Aldridge similarly told Congress that he believed the system would “would perform with 90 percent effectiveness against a missile launched by North Korea.” Senator Carl Levin (D-MI) nearly choked before suggesting Aldridge have another look at the classified number. Apparently, however, MDA is sticking to the Aldridge line—several officials told the Post that the classifed estimate was “greater than 80 percent.”

In Technical Realities: An Analysis of the 2004 Deployment of a U.S. National Missile Defense System, Lisbeth Gronlund, David Wright, George Lewis and Phil Coyle explain that Aldridge and some MDA officials are making personal judgements because there is no testing data that indicates otherwise:

… there is essentially no information from the flight-test program on which to base an estimate of the system’s defensive capability. This is clearly understood by the same Pentagon officials who have made highly optimistic claims about the system effectiveness, as indicated by a discussion between Senator Levin and then Under Secretary of Defense for Acquisition Aldridge at a March 2003 hearing. After Aldridge stated that the initial Block 2004 system would be 90 percent effective against a North Korean missile, Levin countered that the classified numbers did not support his claim. Aldridge then expressed surprise that there even was a classified number “because we don’t know yet, until we get into the testing process …

Gronlund et al conclude that “claims about the performance of a weapon system must be based on the results of a rigorous test program, not the personal beliefs of any individual in the absence of such data.”

What’s so bad about false confidence?

“Leading military and political leaders to believe they have options that are not in fact realistic,” Gronlund et al argue, “can be dangerous and, at the least, contribute to bad decision making”:

If a country were preparing to launch missiles at the United States, beliefs about the effectiveness of the U.S. missile defense system could also affect a decision about whether to use precision-guided weapons to try to destroy the missiles on the ground in advance of their potential launch.

Similarly, if administration officials believed that the GMD system could reliably intercept ballistic missiles launched by North Korea, they might be less motivated to pursue diplomatic means to address the North Korean missile program.

It is not difficult to find examples in which the perceptions of high-level policy makers differed starkly from the technical assessment of experts who were more familiar with the details of a situation. A striking example is the explosion of the space shuttle Challenger in 1986. It is clear in retrospect that the technical experts who understood the space shuttle in detail knew that the
unusually cold temperatures on the night of the launch represented a significant risk if the launch proceeded. But this was not understood by the high-level officials who made the decision to launch, and the result was disastrous.

Seriously, either test the damn thing or spend the money on Special Forces.


  1. Allen Thomson (History)

    I note that the report notes:

    System reliability design is based on multiple interceptors per engagement

    Yeah. Nov 1996 – Aug 1997 I was working for a contractor that was supporting BMDO’s countermeasures office, which provided a fair insight into the NMD of the time. NMD was essentially the same as though bigger than today’s GMD.

    The baseline firing doctrine of NMD was four interceptors per “credible target”—things which the defense, at the time of interceptor launch commit, couldn’t be sure weren’t real warheads.

    In the best case, an ICBM would release one and only one thing that might be a warhead, in which case NMD would launch four interceptors its way to ensure an adequately high probability of kill.

    If the ICBM released more than one object (N) and discrimination couldn’t be done before the decision to launch had to be made, 4*N interceptors would be fired. Objects that might be warheads were called “credible objects,” or “credible targets.”

    In some cases, for some ICBM launch point/target combinations, the intercept time-lines were so tight that the interceptors had to be launched before the number of objects associated with individual ICBMs could be determined, let alone discriminated. In which case
    the defense would fall back on intelligence estimates of “anticipated credible targets” associated with a given ICBM type and immediately launch a corresponding number of interceptors.

    This quickly led to a need for large numbers of interceptors. If the attack involved, say, five more or less simultaneous ICBM launches and each ICBM carried one warhead plus one good-enough decoy, the defense would fire 5*2*4 = 40 interceptors. And then have to provide them with guidance updates as they flew out to their intercept volumes.

    Even if the doctrine were to be reduced to two interceptors per credible target, the number of interceptors needed got big pretty fast.

    A way to try to partly get around this, of course, is to launch multiple kill vehicles on each interceptor, which MDA seems to have started working on under the aptly named “Multiple Kill Vehicle” program.

  2. Gordon Mitchell (History)

    The placebo rationale for missile defense has a long lineage –

    * 1983. Lt. Gen. Daniel O. Graham justified SDI on the grounds that even if technically unsound, it could “severely tax, perhaps to the point of disruption, the already strained Soviet technological and industrial resources.”

    * 1983. Edward Teller suggested that even if SDI’s only benefit would be to increase Soviet defense expenditures, “we would have accomplished something.”

    * 1991. Rep. George Atkins [R-NY], defending Patriot’s Gulf War record, said “the primary benefit of Patriot is a political and psychological one, which I would say militarily is like a placebo.”

    * 1991. Rep. Frank Horton (R-NY) argued, “Thanks to widespread public perception of Patriot effectiveness, Israeli defense officials were apparently better equipped to resist public pressure to intervene in Iraq.”

    Jeffrey adeptly pointed out one policy upshot of a government commitment to psychological perception control via missile defense test design and schedule manipulation – self-deception. Other direct entailments include systematic suppression of contrary scientific data and demonization of those who might denude the placebo efficacy of missile defense programs by telling the truth. For more on the 2001 roots of this logic in the NMD program:

  3. Earl Kirkman (History)


    Quck, somebody get a list together for Bolton to wave around at this next set of hearings.

  4. Nick Schwellenbach (History)

    “Then, MDA scales way back on necessary testing, lest the bad guys figure out the damn thing doesn’t work.”

    It seems fairly obvious that the bad guys will figure out it doesn’t work, if we’re too scared to even test it.

  5. Muskrat (History)

    Can I claim a priority on the creation of the phrase “Cuttlefish syndrome” to describe an administration confused by its own ink? (

    In your hearts, you know we’re one optimistic briefing away from Bush declaring that the system has ALREADY shot down several North Korean missiles.

  6. Stephen (History)

    I may be misunderstanding the terms out of ignorance, but I wonder if, on page 17 of the report, the authors are unfavourably contrasting the current ‘spiral development’ process with a ‘closed loop’ design feedback process that would see problems fully resolved at lower levels before testing would move on to higher levels.

    I ask because on page 18, the report mentions 1990s ‘acquisition reform’ that led to a reduction in specifications and standards before going on to suggest that GMD, in the light of recent failures ‘should adopt an expanded set of specifications and standards related to the design and development of the Ground-based Midcourse Defense hardware and software.’

    I realize I may be misreading it, but I wonder if the report is suggesting that the supposed ‘reform’ hasn’t worked with respect to GMD?

    Also: is there any timeline for an official response to the report’s recommendations? The Washington Post story seems to leave the report on Rear Admiral Kathleen Paige’s desk for now.

  7. Johann (History)

    Re. SF and Iraqi Scud (non) use, Annex A, Volume II of the ISG’s final report tells us that Iraq didnt fire any Scuds because it didnt have any.

    And because I enjoy stirring up trouble I must also point out that the annex’s missile by missile break out it isnt clear how many if any were destroyed by SF/TACAIR action in either of the Coaltion- Saddam scraps. This can only make the skepticism over claims in the ‘Gulf War Airpower Survey’ seem even more well founded.