THE CENTRAL INTELLIGENCE AGENCY is marking its 50th anniversary, and it has a new director, George J. Tenet. Last week, Tenet announced his staff, a crew of veterans. The agency -- much embattled, and with good reason -- needs all the help it can get. It should take a particularly hard look at its intelligence estimates.
Those estimates come in a few basic models: straight, erroneous, and cooked to order. You cook an estimate by making assumptions and ignoring contrary evidence so as to reach a desired conclusion. That is why I can say, given the evidence available, that the National Intelligence Estimate for 1995 was probably cooked -- though plain incompetence cannot be ruled out. That estimate concluded that no other country than the "declared powers" -- Britain, France, Russia, and China -- could pose a ballistic-missile threat to the United States for the next 15 years.
The Clinton administration declared a state of emergency in 1994 over the threat of nuclear proliferation, but even so it is dead set against shielding the country from ballistic missiles. The administration points to the 1972 ABM Treaty, despite its total violation by the Soviet Union for three decades -- the Soviets deployed two generations of nationwide ballisticmissile defenses with some 10,000 to 12,000 missiles. When the last Congress, in December 1995, tried to mandate deployment of a minimal ABM system to protect against rogue attacks, President Clinton issued a veto. The 1995 estimate was just what the administration needed to justify its policy: No threat, no need to deploy anti-missile defenses, just hedge your bets with R&D and wait and see whether a threat emerges.
The estimate's assumptions were exceedingly fragile, according to the General Accounting Office. They were as follows: First, the Missile Technology Control Regime (MTCR) would sharply limit the transfer of technologies. Second, no country would sell an ICBM to another. Third, those countries already possessing the means to build ICBMs would not do so. Fourth, the development of an ICBM would require five years of flight-testing.
The MTCR was created in 1987 in an attempt to curb worldwide missile proliferation. When the Soviet empire collapsed, Russia, the former republics, and other nations were invited to join the MTCR in an attempt to curb worldwide missile proliferation. It is not a proper treaty -- just a collection of declarations by governments. And one of those governments, Russia's, does not exactly inspire confidence. Estimates of illegal exports from Russia since 1991 run as high as $ 200 billion. Up to one half of the Russian economy is controlled by "firms" created by the old KGB in the last days of the empire and staffed by them now. The new kleptocrats employ numerous bodyguards and small private armies. Many (probably most) of them are engaged in activities such as prostitution, weapons-smuggling, drug- smuggling, and money-laundering.
Legal economic activity, meanwhile, has declined drastically at the expense, not only of ordinary citizens, but also of scientists and engineers, who can market their skills to rogue nations. A number of Russian missile and nuclear experts are known to be working in India, Iran, Iraq, and Libya. The Russian Federal Security Service admits that it cannot keep track of them all. The United States almost certainly does not know how many such Russians have already emigrated or anything about their specific skills and destinations. Put together (a) what the Chinese and North Koreans are selling, (b) a few hungry Russian scientists and engineers, and (c) the money and political will of bellicose enemies, and the 1995 estimate quickly goes up in smoke.
The Soviet Union had large stocks of ballistic missiles of all types that were neither accounted for nor destroyed under the START and INF treaties. The Russians have admitted to inheriting a stockpile of 45,000 to 60,000 nuclear warheads, but the Soviet government produced enough plutonium and enriched uranium to make many more than that. As of 1993, some 30 nuclear warheads had been reported missing from Russian stockpiles. At least 17 pounds of outgoing uranium have been seized by Russian and East European authorities. No one can say how much is smuggled out.
How about the 1995 estimate's claim that the United States can rest easy about ballistic missiles for the next 15 years or so? Under current rules, Russia and Ukraine may export long-range ballistic missiles as "space boosters," which can deliver nuclear weapons to full ICBM ranges. A few skilled people with the money to shop the market would not require 15 years to hit a major city with such missiles.
As for the five years of flight-testing, only 27 months elapsed between the initial launch and the operational status of the first Soviet ICBM. Subsequent models generally required less than two years of flight-testing. It took the Soviets only five years to go from the SCUD missile (with a range of 150 miles) to their first ICBM, and seven years to attain their first ICBM with storable liquid fuel -- and that was 40 years ago. As William R. Graham, an ex-science adviser to President Reagan, has pointed out before Congress, most of these technologies are now taught in graduate schools. SCUD and longer-range missiles are for sale on the open market. A few skilled people working for a determined, nefarious regime can make a critical difference in the time required to assimilate and extrapolate the essential technologies into longer-range systems.
Members of the House National Security Committee -- including its chairman, Floyd Spence -- found the intelligence estimate unconvincing and perceived it to be at odds with previous estimates. In response, the administration selected a panel, headed by former CIA director Robert Gates, to undertake a review. While stoutly denying that the estimate was cooked, the panel's report paints a picture of haste and incompetence at least as damning as the earlier GAO report.
Says the Gates report, "An Estimate with conclusions which may be unwelcome to a policy requester -- or which alters previous judgements -- provides unusually comprehensive analysis, clearly states the reasons for any change in previous judgements, explores alternative scenarios, and is candid about uncertainties and shortcomings in evidence." It goes on to say that an unauthorized Russian missile launch "is a remote possibility," but one that "would appear to be technically feasible" -- a truly alarming prospect.
But in its zeal to validate the estimate, the Gates panel commits a few sins of its own. Thus, "In the days of the Soviet Union, strategic force estimates for years tended to avoid questions of doctrine and purpose, in no small part because there were no clear answers, and the issues were so violently disputed." But plenty has been known about such "doctrine and purpose" since at least 1960 and the revelations of Col. Oleg Penkovsky, plus Marshal V. Sokolovsky's 1962 book Soviet Military Strategy.
The CIA possesses ample evidence that the agency simply refuses to accept. For example, proof that the Soviets violated the ABM Treaty by deploying two generations of defenses has been languishing in CIA files for years. Former Soviet ambassador to the United States Anatoly Dobrynin published some of this documentation in 1995. But the Gates panel said not a word about it. Soviet secrets were almost invariably safe from the CIA -- and from academics and the media -- if they were placed on newsstands and in bookstores.
Was the CIA's behavior all the result of human frailty, or did it have something to do with the agency's own characterization of its corporate culture -- "We may not always be right, but we are never wrong"? Alternatively, has the CIA operated for decades with a hidden agenda? The Gates panel further contended left that "policy makers can have high confidence" that foreign development of missiles and weapons of mass destruction "will be reported promptly." In fact, the CIA has a long history of failing to report major developments promptly or, worse, of misleading policymakers on the basics, as in Iraq.
According to U.N. inspectors and Iraqi defectors, the first test of an Iraqi nuclear device was scheduled for sometime between three and 18 months after the beginning of the Gulf War in 1991. The full extent of Iraq's chemical- and biological-weapons programs may not be known even now. We never knew how many missiles Iraq had, nor were we able to find a single mobile SCUD launcher before it launched, and all that in a country not noted for its tropical-rain-forest cover.
On balance, the weight of the evidence indicates that 1995's National Intelligence Estimate was adroitly cooked to order. Assumptions were set up to lead to a desired conclusion. The CIA's own pieties concerning such estimates were violated. The job was rushed. It was an inexcusable performance by an increasingly inexcusable agency.
William T. Lee served both the Central Intelligence Agency and the Defense Intelligence Agency, and now writes and lectures on cold war history.