WOHLSTETTER, SOVIET STRATEGIC FORCES, AND NATIONAL INTELLIGENCE ESTIMATES

• Document Type: CREST
• Collection: General CIA Records
• Document Number (FOIA) /ESDN (CREST): CIA-RDP83M00171R001600010001-9
• Release Decision: RIPPUB
• Original Classification: S
• Document Page Count: 8
• Document Creation Date: December 16, 2016
• Document Release Date: March 29, 2005
• Sequence Number: 1
• Content Type: PAPER

Approved For Release 2005/04/18 : CIA-RDP83M00171R001600010001-9
SECRET
Soviet weaponry-
,overestimated or underestimated?
WOHLSTETTER, SOVIET STRATEGIC FORCES, AND
NATIONAL INTELLIGENCE ESTIMATES
25X1
The summer 1974 issue of Foreign Policy carried an article by Professor Albert
Wohlstetter titled "Is There a Strategic Arms Race?" In the article Wohlstetter
took up the question of myths and realities in the "arms race." He sought to
demonstrate that much of the public debate over the arms race has been driven
by myths, among others the myth of overestimation-that is, the widespread
belief that the Pentagon systematically overestimates the strength of Soviet
strategic forces. Using the Defense posture statements as his basis, Wohlstetter
showed that in fact during the mid-1960s the tendency was to underestimate
in such things as ICBMs, SLBMs, and bombers (although earlier, in the
"missile gap" era, the tendency had of course been quite the contrary).
Since the article was published, I have examined National Intelligence
Estimates going back to 1960 to see whether or not the same charge could be
directed at the official judgments of the corporate body whose task it is to
communicate the views of the U.S. intelligence community, the U.S. Intelligence
Board, as set forth in the NIEs. I also checked estimative history as regards
anticipated qualitative improvements in weapon systems and their predicted
operational dates.
I am satisfied on the basis of my research that Dr. Wohlstetter is essentially
correct in the case of ICBMs during the 1960s. There are some minor differences
between what the intelligence community said and the data Wohlstetter used,
but not enough to make a case against Wohlstetter's findings.
Take some examples of how the USIB tended to underestimate:
NIE 11-8-63 forecast a spread of 370 to 670 launchers for mid-1969.
The actual count in mid-'69 was 858.
NIE 11-8-64 was even more extreme. It forecast 410 to 700 launchers
for mid-'70. The actual count was 1,292.
The worst of the lot was the NIE issued in 1966. It forecast a spread
of 800 to 1,120 for mid '72. The actual count was 1,527. With the exception
of the initial year, the actual count exceeded the projected annual spreads
over the entire period covered by the estimate. (See Chart 1.)
In reading the past NIEs, I hoped to find some clear rationale for the
repeated underestimation. In many years the community expressed views essen-
tially along the lines that the Soviets would not deploy as many ICBMs as the
U.S. for fear of touching off a new round of deployment in the U.S., or that
they would be content with a retaliatory force somewhat smaller than the U.S.
force. In 1967, the rationale given for estimates such as 11-8-66 was that the
SECRET 1
Approve For Release 2005/04/18 : CIA-RDP83M00171R001600010001-9
 Approved For Release 2005/04/18 : CIA-RDP83M00171R001600010001-9
SECRET NIE Performance
ICBM Launchers: NIE 11-8-66 v. Fact
1600 r-
300 1.
Mid-66 Mid-67 68
76
SECRET
2
Approved For Release 2005/04/18: CIA-RDP83M00171R00?T0001-9
 Approved For Release 2005/04/18 : CIA-RDP83M00171R001600010001-9
NIE Performance SECRET
Soviets saw political and psychological advantages in having an ICBM force
roughly the same size as that of the U.S., and that was said to be the goal
of their deployment program. In other years, however, no clear rationale was
presented for the projected numbers.
In 1968, the text of the NIE established a lower side of the range for the
future at 1,360 launchers (the number operational and under construction at
that time). Because of several factors that could influence the size of the force,
however, it did not estimate the maximum size it might reach.*
I thought, when embarking on this venture, that it would be found that
we had credited the Soviets with the ability to make rapid strides in weapons
technology (leading to MIRVs, high accuracies, etc.,) and that consequent
improvements in quality would permit Moscow to limit the quantity of weapons
deployed. But this was not the case. Our judgments on when the Soviets were
likely to introduce certain qualitative improvements into their systems fitted
pretty well with what has actually happened.
The "Missile Gap"
I suspect, but obviously cannot document the fact, that part of the reason
for the repeated underestimation of the growth of Soviet ICBM forces was a
subconscious (or maybe even conscious) overreaction by the intelligence com-
munity to the gross overestimation of Soviet ICBM growth during the days of
the "missile gap." The intelligence community took quite a public flailing for
that error of judgment.
The "missile gap" era began in August 1957 when the Soviets carried out
the first test firing of an ICBM. That firing and subsequent ones served to
convince a large segment of the U.S. intelligence community, as well as sizable
elements of Congress and the Department of Defense, that the Soviets were
preparing to embark on an ICBM deployment program involving large numbers
of missiles. From the late 1950s until September 1961, the tocsin was repeatedly
sounded that the Soviets were outpacing the United States in ICBM production
and deployment. Several statements made by Khrushchev during those years,
both public and private, seemed to be encouraging such thoughts. The "missile
gap" was much discussed during the Presidential campaign of 1960, and the
NIE for that year serves in part to tell why. (See Chart 2.)
NIE 11-8-60, dated 1 August 1960, contained three numerical estimates of
Soviet ICBM strength for mid-1963. The Air Force estimated 700, the CIA 400,
and the Army and Navy 200. State and the J-2 of the Joint Staff stated that they
thought the number would be somewhere between the CIA number and the Air
Force number. By mid-1963 the actual number deployed was less than 100.
In NIE 11-8-61, dated 7 June 1961, opinion was again well divided. CIA
estimated that by 1964 there would be 200 to 400 ICBMs deployed; State INR's
spread was 300 to 500; Army and Navy liked 150 to 300; and the Air Force
projected 850. By mid-1964 the number actually deployed was 191. The Air
*State, DIA, and the military services all took footnote to this omission. They considered
1,800 launchers to be the upper limit.
Ap eTd For Release 2005/04/18: CIA-RDP83M00171R001600010001-93
 Approved For Release 2005/04/18 : CIA-RDP83M00171R001600010001-9
SECRET NIE Performance
ICBM Launchers: NIE 11-8-60 v. Fact
800 r-
State & J-2:
Somewhere
between
A and B
4 Approved For Release 2005/04/18 : CIA-RDP83M00171R001 ODO1-9
 Approved For Release 2005/04/18 : CIA-RDP83M00171R001600010001-9
NIE Performance SECRET
Force, bolder than the rest of the community, also projected ahead to 1966,
estimating that by that time the Soviets would have 1,450 in the field. By 1966
the actual number deployed was 250. (See Chart 3).
Data collected during the late summer of 1961 showed the community how
wrong its estimates had been, and a Memorandum to Holders of NIE 11-8-61
was issued in September of that year. All agreed that there were probably 10
to 15 ICBMs deployed at that time (there were actually only 4) and that by
mid-1963 the spread would he some 75 to 125; the number actually deployed
by mid-1963 was 91, close to the center of the spread.
The Cuban Impact
I also feel that part of the reason for repeated underestimating was a lack of
appreciation on the part of the intelligence community of how bitter Khru-
shchev, and probably others in the Soviet hierarchy at the time, felt about the
"facing down" they experienced as a result of the Cuban missile crisis in 1962.
At that time the Soviets had less than 50 ICBM launchers operational, and we
knew it. The Soviets knew they were dealing from a position of weakness and
probably at least suspected that we were aware of their lack of ICBM strength.
The chronology of the growth of their ICBM force fits very neatly with a deci-
sion that might have been taken shortly after the pullout from Cuba to expand
their ICBM force at a rapid rate and probably to a size never originally intended.
Deployment of the SS-9 and SS-1.1 really started to take off four years after the
crisis.
Heavy Bombers
Wohlstetter is again correct in the case of heavy bombers. From 1960 through
1971, the NIEs always phased out the Bison and Bear bombers at a faster rate
than actually occurred. There has, in fact, been no reduction in the heavy
bomber force for the past six years. I still think that the estimative judgments
were logical, albeit erroneous. Why the Soviets would go to the expense of
retaining such a small fleet of obsolete heavy bombers defies well-reasoned
explanation.
Submarine-Launched Ballistic Missiles
In the case of Soviet submarine-launched ballistic missiles, Wohlstetter is
somewhat wrong in his assertion that the tendency was to underestimate. The NIE
history on those systems is mixed. From 1961 through 1963, the NIEs over-
estimated. The spread of 160 to 250 projected for 1970 in the 1964 NIE was
exceeded, but only in that one year. (See Chart 4.) (It is only fair to point out
that the only NIE Wohlstetter used for these allegations about numbers of
SLBMs was the 1964 projection, and that isn't cricket.)
The subsequent 1965 estimate erred slightly on the high side in its projection
for mid-1966, showing a spread of SLBMs for that year of 122 to 137. The actual
count was 108. The projection for mid-1967 held at 122 to 137 and was pretty
good but for the wrong reason; the unpredicted advent of the first Y-class in
1967 boosted the actual count to 124, within the NIE spread. In projecting beyond
Approve For Release 2005/04/18 : CIA-RDP83M00171R001600010001-95
 Approved For Release 2005/04/18 : CIA-RDP83M00171R001600010001-9
SECRET NIE Performance
ICBM Launchers: NIE 11-8-61 v. Fact
1500 T--
1400 F
1300 }-
1000 =-
900 ~--
i
800 L
700
600 F-
500
400 1
f
1 Army & Navy
I
y
tv ICIA 1
6 Approved For Release 2005/04/18 : CIA-RDP83M00171R00101-9
Air Force
 Appmee M r elease 2005/04/18 : CIA-RDP83M00171R001600010001-9
SECRET
SLBM Tubes: NIE 11-8-64 v. Fact
300 r-
100 Oct
1964
69
1967, the NIE considered only the ballistic missile submarines, and did not
specify the numbers of tubes per boat. In doing so, however, the estimate
was not too bad. It projected some 50 such submarines, including perhaps 7 of a
new class, for mid-1970. The actual number was 47, including 13 of the new
Y-class.
Our projection in 1966 for mid-1971 fell well below the actual count for
that year, principally because we had not yet established a production rate for
the Y-class, and the Soviets had 21 of that class operational by 1971, as opposed
to an estimate of 10. The NIE in 1967 also fell short in its projection for mid-69
and for mid-72 for the same reason. The 1968 estimate was quite accurate in
its prediction, as was the one in 1969. The 1970 NIE's record for the years
'71, '72, '73, and '74 was under, under, over, over-but the margin of error in each
instance was 20% or less.
The main reason for the overestimates in SLBM strength in the early 1960s
was the fact that the size of the ballistic missile submarine force remained static
from 1962 through 1966. The natural tendency in estimating when a new weapon
Approved For Release 2005/04/18 : CIA-RDP83MOO171ROO1600010001-9
 SECRET proved For Release 2005/04/18: CIA-RDP83M0~1~71~20?1600010001-9
er ormance
system is seen coming in is to project a continuing growth in that system and
to anticipate the development of new, improved models. In retrospect, it is now
clear that the Soviets opted to hold their SLBM force at a modest level of three-
tubed G- and H-class submarines, equipped with short range missiles, until
something more nearly approaching the U.S. Polaris system could become avail-
able. And it did-the Y-class, fitted with 16 tubes.
In sum, the USIB repeatedly erred after the mid-1960s in two out of three
categories. No one can claim this as a triumph, and let us hope that the short-
comings prove instructive in the future. As Wohistetter has observed, however:
"Predicting the size and exact mixture of a potential adversary's weapon deploy-
ments several years hence is a hard line of work. It is intrinsically uncertain,
reversible by the adversary himself between the time of prediction and the actual
deployment."
8 Approved For Release 2005/04/18: CIA-RDP83M00171R0011;UDD'10001-9