[Page 695]

155. Letter From the Chairman of the President's Foreign Intelligence Advisory Board (Anderson) to President Ford 1

Dear Mr. President:

The National Intelligence Estimates should be among the most important documents issued by the intelligence community. They are the natural backdrop to guide the Department of Defense in formulating force levels and R&D programs, and should serve Congress in their authorization and appropriation hearings. Certain of them also serve as the foundations from which to derive arms limitation negotiating positions. Underlying each of these objectives is the presumption that the NIE will substantially influence the thought processes of key Government decision-makers regarding Soviet military capabilities.

In our view, NIE 11–3/8–74 (“Soviet Forces for Intercontinental Conflict Through 1985”)2 is seriously misleading in the presentation of a number of key judgments and in projecting a sense of complacency unsupported by the facts; as a consequence, it is deficient for the purposes it should serve.

This NIE assesses that for the next ten years it is extremely unlikely that the Soviets will conclude they could launch an attack which would prevent devastating U. S. retaliation. This judgment is presented confidently, with the force of fact, although the cumulative evidence on which it is based is conflicting, often flimsy, and in certain cases does not exist:

—With respect to Soviet ICBM accuracy and the survivability of the U. S. MINUTEMAN force, the data is inconclusive and has been very differently interpreted by the experts. A number of uncertainties which have puzzled analysts for six years have been accommodated in the NIE by averaging the worst and best cases when the data could readily support either interpretation;

—with regard to Soviet antisubmarine capabilities, it assumes our POLARIS/POSEIDON submarines will remain invulnerable through 1985; yet about three months before issuance of the NIE, we observed Soviet experiments in submarine detection and trailing which are not yet understood by the U.S. intelligence community and which give very serious pause to this optimistic judgment;

[Page 696]

—with regard to Soviet capabilities against our bombers, it ignores the vulnerability of SAC bases to cruise and ballistic missiles from submarines operating off U. S. shores, vulnerability of the aircraft to mid-course intercept, and does not take adequate account of emerging data which may indicate an improving Soviet low altitude air defense capability;

—the NIE gives the appearance of a net assessment and thus the added weight of “operational” consideration, when in substance it is not. (For example, it assumes without detailed examination the survivability of the U. S. command and control apparatus, and accepts optimistic and unproven data regarding U. S. silo hardness.)

These general criticisms may be best illustrated by a brief review of available evidence which contrasts with NIE judgments in three critical areas: Soviet ICBM accuracy, POLARIS vulnerability, and U. S. bomber penetrability.

Soviet ICBM Accuracy

The hard data on both the presently deployed Soviet ICBM force and the new Soviet ICBMs does not allow any confident, precise determination of accuracy. The SS–9 accelerometer data collected over the last ten years can be interpreted as to give either a relatively good accuracy or the rather poor accuracy stated in the usual assessments. The choice of the poor number has been made by a set of judgments which has been questioned by informed and resaonable analysts. These analysts can support the view that the data indicates a significantly better accuracy than that assessed by the community.

Concerning the new Soviet ICBMs—the SS–18 and SS–19—there is no hard information indicating the basic guidance and control accuracy of these vehicles, but only the telemetry information that both the re-entry vehicle (RV) deployment and RV reentry behavior of the systems are malfunctioning in a fashion very similar to the malfunctions exhibited in the early flight tests of the U.S. POSEIDON and MINUTEMAN III MIRV'd systems. One would expect, as in the case of the U.S. systems, that these difficulties would be worked out in the next few years, probably before the systems are deployed in large numbers.

Under these circumstances the systems could be quite accurate in the near future with circular error of probabilities (CEPs) comparable to that of MINUTEMAN II (about 1/6 of a nautical mile). There appears to be no hard infromation which would negate the possibility of the systems being even somewhat better than MINUTEMAN.

In the case of the smaller payload SS–17, there is some information concerning the quality of the inertial equipment. If this data is interpreted in a straightforward manner, one would conclude that the system's accuracy is rather poor (about 1/3 of a nautical mile). How[Page 697]ever, the telemetry from early flights of this system showed what was apparently a bias in the data, perhaps deliberately inserted by Soviet technicians to assist them in separating out their information, which puts the whole question of the straightforward interpretation in doubt.

The above is in contrast with the unusually confident position of the NIE that the accuracies of the new Soviet ICBMs lie between 1/4 and 1/2 of a nautical mile and that improvements below 1/6 of a nautical mile would require a new generation of ICBMs. The difference between the NIE assessment and the possible greater accuracy suggested herein is equivalent to an almost 10-fold increase in explosive effect on target.

SLBM Survivability

The NIE asserts that there should be little worry as to the survivability of the SLBM force now or in the next 10 years. This conclusion is based partially upon U. S. superiority in “classical” ASW techniques, and partially on judgments that nonconventional techniques are unlikely to be highly successful. It is known, however, that the Soviets are conducting experiments in detecting and trailing their nuclear submarines, using nonconventional techniques; these techniques are not understood by the U.S. technical community. Photographs of some Soviet equipment, and data from subsequent experiments suggest that the Soviets have the ability to detect a submarine at substantial distances. It now appears that the Soviets are practicing the integration of this technique into ASW activities involving warships. In addition, there appear to be one or more other nonconventional systems being tested in submarines on which we have no data regarding operational capabilities.

The Soviets are pursuing at least twenty different ASW programs in a very aggressive fashion. It is very possible that this technologi-cal area will yield capabilities not yet realized by the U. S. R&D community.

Since we cannot plan on always getting solid intelligence, it may be a very long time before we are able to determine the nature of these new threats. Under these circumstances, it is imprudent to make judgmental conclusions that minimize the potential for a technological breakthrough for the next ten years and thus future Soviet capabilities in this vital area.

Bomber Penetration

The conclusion that the Soviet air defenses today are relatively ineffective against the planned U.S. low-altitude bomber strikes is based on a large amount of intelligence information which suggests two deficiencies. First, it is assumed that the most heavily deployed Soviet surface-to-air missile ( SAM), the SA–2, which carries the burden of defense against low altitude penetrators, primarily carries a high-explosive [Page 698](non-nuclear) warhead; second, that the ground-controlled intercept (GCI) sustem which must direct the aircraft interceptors to their targets is relatively inaccurate against low-flying aircraft.

While both of these conclusions may be justified by information collected in the past, data is beginning to emerge which suggests a potential for marked change within the period of the estimate. Specifically, Soviet homeland air defense practice altered significantly about 1972. Prior to that time, fewer than 3% of the target aircraft were at altitudes below 500 meters. After that time the percentage began to rise; it is now about 30% and seems to be growing. This might relate to the growing ability of the SA–2 to cope with low altitude threats through the use of a nuclear warhead option—an option which recent intelligence indicates has spread to perhaps one-half of the SA–2 deployments.

The above change may also reflect an improving capability against low altitude penetrators in a number of other areas where there are intelligence gaps, such as: (1) improved GCI vectoring accuracy through better site-to-site and site-to-aircraft data links; (2) employment of the mobile low-altitude SAMs of Ground Army Forces; (3) tactics such as radiation homing on interceptor aircraft and SAMs, and a nuclear warhead on the SA–2, to negate or degrade U.S. electronic countermeasures; (4) emergence of at least a partial look-down-shoot-down capability on the MIG–23, which is now operational with Ground Army Forces.

For the longer term, many Soviet activities seen at their R&D facilities are not fully understood. A pole-mounted, mobile radar has been observed which could extend the low altitude coverage of existing SAMs or could form the basis for a new SAM system. A high-performance SAM and SAM radar is being tested, probably for the ground forces, but which could have dual capability for homeland defense. A variety of other types of air defense radars, some elevated, are undergoing unknown tests.

Taken as a whole, the uncertainties inherent in a comprehensive assessment of Soviet air defense capabilities do not support the NIE view that “. . . it is unlikely that the Soviets will be able to cope with sophisticated low altitude attacks during the next 10 years.”

Having identified what we believe to be serious deficiencies in this NIE, there follows a series of observations examining the nature of the problems and some suggestions for their resolution.

Observations on the Intelligence Estimating Process

The root cause of the problems experienced both by the intelligence community and the users of intelligence is the lack of factual evidence and the difficulty of forecasting ten years into the future. Because [Page 699]of the importance attached to some intelligence subjects, there is an understandable desire to fill some of the intelligence holes with judgments. These judgments can then gain an acceptance approaching fact, and can then lead both the intelligence community and the users of intelligence into a single viewpoint which rejects alternatives, and can persist too long. Only when some surprise arises, totally contrary to the intelligence trend, is the pattern broken and another “review” ordered of the intelligence effort.

When decisions must be made, they are almost always based on incomplete information. When they involve intelligence information, the decision-maker should wish to know not only the facts but also the best judgments of the intelligence community and have some feeling for the uncertainties connected with these conclusions, including other possible situations consistent with the data. These uncertainties should lead the decision-maker to consider whether he should hedge his bets or to be prepared for possible reverses connected with failures of actions (or inactions) based on these assessments.

This is not an easy process; no one knows how to weigh judgmental uncertainties. For this reason we look upon the process of attempting to analyze and communicate uncertainties in the area of national assessment as a process with which we must continue to experiment, trying various modes in an attempt to find a more satisfactory procedure. This leads us to the following suggestions:

Suggestions for Resolving Observed Deficiencies

I. Those aspects of intelligence which are considered critical by key decision-makers should be subject to separate and competitive anal-yses and such alternate views as are developed should be presented to the President and other users. In our view, this suggestion deserves the highest priority for consideration and implementation.

II. To avoid the tendency of decision-makers to force the intelligence community to come up with positions when the data is too meager, the following suggestion may be helpful. The user should formulate his alternative choices of action in such a way as to permit the intelligence community to marshal its evidence around each alternative. Thus, the community would be asked to make its best case that we face a serious problem, and its best case that we do not.

The purpose of this suggestion is to try to maintain an awareness of the limitations in the intelligence information. In addition, it stimulates the user to provide important feedback to the intelligence community on his interests and problems which, in turn, can motivate the intelligence community to provide a more complete and useful product to the user.

[Page 700]

III. NIE 11–3/8–74 has the tendency to phrase the estimate as a net assessment, that is, to include an assessment of U. S. capabilities in the face of the threat in question. We suggest that the National Security Council adopt a three-step process. The first step is the generation of a purely intelligence document, the NIE, which carefully avoids the impression that a net assessment has been performed. The second step would involve a genuine net assessment, requiring participation by both the intelligence community and other agencies (Department of Defense, State, etc.), under the aegis of the NSC. The third and final step would involve a thorough critique of the net assessment document for the NSC by an entity which is enabled to function with an appropriate degree of independence.


We believe that the policy-maker would be better served by an NIE which clearly identifies that which is fact and that which is judgment, and which identifies the intelligence gaps prevailing at various stages in the analytic process. The product would also be more useful if the decision-makers provided more specific guidance regarding the relevant, contemporary issues with which they are confronted, and on the most useful format for presentation of the intelligence.

Finally, Mr. President, we recommend that you direct the NSC to implement these suggestions, insofar as possible, with respect to formulation of this year’s NIE on Soviet Strategic Forces which is now in progress and, as appropriate, to the national intelligence estimating process.


George W. Anderson, Jr. Admiral, USN (Ret.)
  1. Source: Ford Library, National Security Adviser, NSC Program Analysis Staff Files, Box 25, Subject Series, PFIAB/NIEs, 1975–76 (5). Top Secret; [Handling restriction not declassified]. Anderson handed this letter to Ford during the President's August 8 meeting with PFIAB, the record of which is printed as Document 154.
  2. Document 149.