KABUL, Afghanistan -- U.S. officials, when looking to quickly illustrate progress in Afghanistan, have in the past few months highlighted a 7 percent drop last year in what they call "enemy-initiated attacks." Fewer attacks, the reasoning went, meant Afghans were safer, and that the Taliban was weaker.
The problem: There was no decline; the numbers were wrong.
The U.S.-led NATO coalition said Tuesday that it had discovered a clerical error in its reporting, and that the number of enemy-initiated attacks -- defined as attacks with guns, mortars, rockets or improvised explosive devices -- remained constant from 2011 to 2012.
Although the mistake may be embarrassing, it is not likely to greatly change perspectives about how the war is going. That is in part because, outside of official circles, few analysts have seen the violence statistic as good news. In fact, the same measure, looked at over a wider sample of years, actually depicts a dramatic growth in violence since 2009, when U.S. commanders first began inching toward a counterinsurgency strategy that focused on reducing violence, rather than solely battling militants.
Previous coalition reports, which use bar graphs to plot the level of enemy-initiated attacks each month instead of specific figures, showed about 2,000 attacks in July 2009, for instance. That was before the Obama administration dispatched tens of thousands of fresh troops in a bid to stabilize Afghanistan. A year later, with all of the U.S. surge forces in Afghanistan, the number of monthly attacks had roughly doubled, to about 4,000.
Three years later, in July 2012, enemy-initiated attacks stood just above 3,000 -- a decline, to be sure, but still far higher than the number of attacks before the surge and the strategic shift toward protecting Afghans.
The coalition has sought to overcome the problematic numbers by saying 80 percent of the enemy-initiated attacks have occurred in areas where fewer than 20 percent of Afghanistan's 30 million people live. But with no public comparative data for years past, it is unclear whether that represents a change or just a long-standing feature of the war.
On Tuesday, the coalition said the mistake in the 2012 numbers resulted from the Afghan security forces' omission of a few months' worth of data on enemy-initiated attacks. The error was inadvertent, it said, and was discovered during what was described as a "quality-control check."
Once the Afghan numbers were added, it became apparent that the statistics the coalition released were incorrect, and the report that included them was pulled from the coalition website.
"This was a record-keeping error that we recognized and have now corrected," coalition spokeswoman Erin Stattel said.
The incorrect statistics were first reported by The Associated Press, which noticed that the report containing them had been taken off the coalition website and inquired about the reason. The coalition then said it was revising the figures.
The revised figures are being audited and will soon be reposted, Ms. Stattel said.