This commentary appeared in Financial Times on February 4, 2004.
There are three possible reasons for the apparent failure of the US
and British intelligence establishments to assess accurately the state
of Iraq's weapons of mass destruction: mendacity, incompetence or systemic
failure. So far, no evidence has emerged of knowing falsification or culpable
errors of judgment, suggesting that in the end we may be left with the
This apparent intelligence failure may have stemmed from the natural
tendency of intelligence analysts to accentuate the downside of any risk,
compounded by a preference among policy makers in this instance to accentuate
the particular risk that Saddam Hussein might possess WMD.
Intelligence agencies operate under permanent pressure never to miss
any impending disaster. After Pearl Harbor, after the attacks of September
11 2001 and after a number of unheralded disasters in between, large-scale
inquiries have been mounted into the intelligence failures that allowed
the US to be taken unawares. Comparable scrutiny has rarely applied to
the many exaggerated appreciations and unduly pessimistic analyses issued
by western intelligence agencies over these same decades. Indeed, this
may prove the first time in history that a false warning received the
level of scrutiny normally reserved for an unheralded disaster.
This is not to say that intelligence agencies always get it wrong. On
the contrary, they are right more often than not. But there is always
a margin of error in any intelligence assessment, sometimes small, sometimes
quite large. These inevitable errors become a source of national controversy
when they underestimate a threat, not when they overestimate one.
The intelligence community's natural tendency to stress the downside
of any risk is in most instances offset by the policy maker's penchant
for emphasising the upside. Government leaders do not, as a rule, like
being told that their current policies will not work - or, worse yet,
that they are courting disaster. The dialogue between intelligence analyst
and policy maker is therefore normally one between congenital pessimists
and optimists. This debate produces, in most cases, a more balanced perception
than either side would come to if left on its own.
However, in cases where policy makers have their own reasons to emphasise
the downside risk, their dialogue with intelligence analysts tends to
accentuate rather than counteract the latter's natural caution.
Intelligence analysts are used to receiving push-back from policy makers
who do not like what they are being told. The US and British intelligence
communities are sufficiently insulated from overt political pressures
to hold their own in such circumstances. In the instance of Iraqi WMD,
however, the dialogue between intelligence and policy communities was
not one between pessimists and optimists, but between pessimists and greater
pessimists. The result was an almost inevitable emphasis on the downside
This process of mutual negative re- inforcement also occurred throughout
several decades of western Sovietology, producing consistently exaggerated
appraisals of Soviet economic prowess and political cohesion. Here, too,
the natural caution of intelligence communities was complemented, rather
than counteracted, by policy makers' need to sustain public support for
large defence budgets.
Faulty intelligence appraisals did not lead the US into war against the
Soviet Union, however. Deterrence and containment were then key components
of the American strategy. The US was not going to attack unless the Soviet
Union attacked first, no matter how grave a threat it was perceived to
A policy of pre-emption sets a higher threshold for intelligence. It
is one thing to expand one's defences on the basis of inconclusive evidence;
it is quite another to attack a foreign nation on that same basis.
It is quite possible that both the US and British inquiries will, in
the end, find no evidence of dishonesty or incompetence with which to
explain their assessments of Iraqi WMD. Neither is it obvious how our
intelligence systems can be altered to eliminate institutional bias in
There is always a danger that inquiries of this sort will be driven by
the need to lay blame and direct retribution. In this case, however, the
proper conclusions may not be to fix our intelligence process but rather
to alter the manner in which intelligence assessments are received and
regarded, to establish greater accountability for false warnings, to educate
the public and legislators to the inherent limitations of intelligence
and, ultimately, to raise the bar for pre-emptive attacks in the future.
The writer is a former US assistant secretary of state and special adviser
to the president; he is currently director for international security
and defence policy wit h the Rand Corporation.
Explore All Topics »