The counting of Iraqi civilian casualties is a hotly
politicized topic in Washington these days. Depending
on one's vantage point, the numbers are either sloping
upward or downward. Generals in Iraq say the numbers
of deaths are down. Anti-war Democrats say they are
up, or at least, aren't budging much from their
late-2006 highs. It also depends on which data is used
and how it's interpreted (average monthly killed
versus average daily killed). Employing the latter,
the Pentagon's most recent quarterly report [PDF]
refutes notions that the surge has resulted in a drop
in violence, given that civilian casualties daily were
lower in the few months preceding the surge than they
were afterward. But interpreted differently, as
General David Petraeus did before Congress in
September, there has been a 45 percent drop in total
fatalities against civilians, including noted
reductions in sectarian bloodshed, IEDs, and "high
profile" attacks. So which version of events or data
is correct?
Here's my take on the topic: First, it is extremely
difficult to gauge with any accuracy how many
civilians are killed in Iraq, so any empirical
findings should be greeted with skepticism. That is
because census data is poor; many deaths go
unreported; and there is often double-counting because
it is oftentimes difficult to determine a civilian
from a combatant. Morgue, hospital or government
statistics are also notoriously unreliable. And any
statistical analysis by outside researchers generally
relies on cluster sampling, which poses problematic
issues with sample sizes because of Iraq's shifting
demographics.
Second, even if casualties are down overall this year,
they are still dangerously high and way above
2004-2005 levels, not to mention that whatever number
is reached does not include those tortured or
kidnapped. Plus, many Iraqi Muslims simply bury their
dead and do not go through an undertaker or hospital.
Third, trend lines are often short-lived and futile to
predict. Proclaiming civilian casualties are sloping
downward is akin to saying the insurgency is in its
last throes. Plus, these trend lines obscure the fact
that any temporary decline in casualties can be
explained because of the ethnic cleansing of once
heterogeneous Iraqi cities and neighborhoods. Of
course, an Iraq comprising just Sunni Arabs could
result in zero sectarian casualties but a homogeneous
Iraq is presumably not the end goal.
Finally, the fault for all this confusion lies with
the U.S. military, which decided early on it was not
worth counting Iraqi civilian casualties. Even the
term given to their tragic loss at the hands of U.S.
forces -- "collateral damage" -- smacks of bureaucratic
hubris. (A chilling segment on 60 Minutes reports that
30 civilians killed was the magic number Pentagon
officials could live with when targeting a
"high-value" terrorist in Afghanistan; anything higher
requires approval from the defense secretary or
president.)
The Pentagon should keep closer tabs on Iraqi (and
Afghan) civilian deaths, especially when the United
States is at fault. Only the American military has the
presence on the ground and resources to catalog
reliable statistics. It won't, for obvious PR reasons,
but it has in the past. After military operations in
Somalia and Kosovo in the 1990s, for instance, the
Center for Disease Control and Prevention conducted
body count estimates of local civilians. They should
do the same for Iraq and Afghanistan.