This publication is now archived.
The aim of President Bush’s “New Way Forward,” announced in January 2007, was to improve security in the Iraqi capital and foster an environment for national reconciliation. In announcing the deployment of thirty thousand additional troops to Iraq, Bush predicted, “daily life will improve, Iraqis will gain confidence in their leaders, and the government will have the breathing space it needs to make progress in other critical areas.” Opinion sharply divides over whether this has been achieved, and some claim that Pentagon statistical methodology fails to reflect the real levels of violence in the country. A trio of independent reports offers a grim assessment of political gains, but conflicting statistics on civilian deaths shroud the debate over Iraq’s future, and U.S. involvement in it. The U.S. military insists progress on the ground can be measured, but some experts accuse the Pentagon of “cherry-picking” positive data.
Military Numbers on Violence
Gen. David H. Petraeus, commander of U.S. forces in Iraq, told Congress on September 10 and 11 that civilian casualties in Iraq—including those killed and wounded—dropped dramatically during the surge of troops. According to numbers compiled by the U.S. military and cited by the general before Congress, Iraqi civilian deaths linked to violence declined by over 45 percent between December 2006 and August 2007, from three thousand deaths monthly to around 1,500 a month. The drop was even greater—70 percent—in Baghdad, where monthly deaths fell from approximately 2,220 to around five hundred. Overall, ethno-sectarian violence declined by about 55 percent; the drop was 80 percent in Baghdad. Car bombings and suicide attacks were also down by nearly 50 percent between March 2007 and August 2007.
Security gains have been most dramatic in Anbar province, the general said, where Sunni tribes have begun to cooperate with coalition forces to root out al-Qaeda insurgents. Petraeus said attacks in Anbar dropped substantially between October 2006 and August 2007, from roughly 1,350 per month to “a bit over two hundred.” Not all provinces fared as well; two north of Baghdad have remained highly volatile. “The trend in Ninevah province, for example, has been much more up and down, until a recent decline, and the same is true in Salah ad Din province,” the general said.
To be sure, the pace of civilian killings in Iraq remains high. Michael E. O’Hanlon, a senior fellow at the Brookings Institution who has closely followed the war’s evolving statistics, says average daily attacks on Iraqi civilians and coalition forces have dropped by 25 percent since August 2006. But attacks still hover at around 120 a month. Brookings’ Iraq Index, which culls data from UN, Pentagon, and other sources, estimates that as many as 2,800 civilians may have been killed in July 2007, down from 3,500 in November 2006—a 20 percent decline but still higher than the Pentagon’s numbers. The Brookings’ data also shows the Iraqi death rate cited by Petraeus for 2007 remains well above levels in 2004, 2005, and early 2006.
Nonmilitary Sources of Data
Independent reports published before Gen. Petraeus’ appearance on Capitol Hill paint a far bleaker picture than the military’s assessment. The Government Accountability Office (GAO) concluded in August 2007 the average number of daily attacks against “coalition, Iraqi security forces, and civilians” remained relatively static—at about 170—between October 2006 and July 2007, with a slight drop between June and July. An August 2007 National Intelligence Estimate (NIE) made a similar conclusion, without offering statistics. A third report on the readiness of Iraqi Security Forces concludes the country’s soldiers and police largely are incapable of protecting its citizens.
Nongovernmental analysis of civilian attacks and casualties has been even more critical; it also contradicts military numbers. An Associated Press analysis of civilian casualty statistics, published September 1, 2007, concluded “at least 1,809 civilians” were killed in August 2007, the second-highest monthly level in 2007. The AP, which based its analysis on Iraqi police reports, said 27,564 Iraqi civilians have been killed since April 2005, when the news service began collecting data. Iraq Body Count, a British group that monitors civilian deaths, has counted more than 71,000 civilian deaths from violence since 2003. The Economist says the civilian death toll “almost certainly exceeds 100,000.”
The Military’s Collection Process
The Pentagon has come a long way since 2002 when Gen. Tommy Franks, the former commander of forces in Afghanistan, famously declared, “We don’t do body counts.” Today the military does, indeed, count bodies. The question is how well.
Violence assessments are prepared by Multi-National Forces Iraq (MNF-I) with data culled from Iraqi and coalition incident reports. Information is continually updated as information from the field is verified. Army Col. Michael J. Meese, an advisor to Petraeus, says the military expanded its ability to crunch numbers in 2006 by adding more analysts and specialists, who “literally spend all their time doing daily reports, weekly reports, (and) monthly reports.”
A senior military officer who has served in Iraq says that data is constantly under review. “The death toll in the recent truck bombings against the Yezidi villages in northwest Iraq, for example, when verified by our Special Forces teams after the dust literally settled, was a good bit lower than original accounts reported, and we thus adjusted that report downward,” the officer wrote in a letter to CFR Senior Fellow Max Boot, which he posted on Commentary’s blog. The officer said MNF-I data is considered “the most accurate” measure of violence in Iraq by the Central Intelligence Agency and Defense Intelligence Agency, which helped draft the August 2007 NIE.
Discrepancies in the Numbers
The clearest rift between military and nonmilitary assessments surfaced with the GAO report. According to Col. Meese, the GAO did not include all data from July and August in its report, a period of dramatic improvements on the ground in Iraq. “When I say ‘changed substantially,’ attacks and security incidents are down significantly,” Meese says. “I think eight out of the last eleven weeks, attacks have been dropping. They are now at the lowest level since June 2006.” In addition, military experts say some statistics from Iraqi sources are unreliable or double counted; they have dismissed the GAO’s counting methodology as flawed.
By contrast, the GAO says it is uncomfortable with MNF-I formulas for measuring sectarian versus nonsectarian violence. “Part of the problem that we had in reaching a conclusion about sectarian violence is there are multiple sources showing different levels of violence with different trends,” GAO Comptroller General David Walker told lawmakers in September 2007. “Let’s just say that there are several different sources within the administration on violence, and those sources do not agree.” He said his agency’s analysis remains classified and did not elaborate.
Either way, CFR’s Boot says the Pentagon has been consistent in its reporting, making trend analysis possible. “It’s very hard to have accurate statistics in the context of Iraq. But what I’m confident in is the military reporting has remained consistent,” Boot says. “Statistics give you trends; there’s no question that the trends are positive in the last six months. I don’t know how you can dispute that.”
Experts see a multitude of reasons to question the military’s assessments. For one, analysts inside and outside the military accuse the Pentagon of “cherry-picking” information by selectively redefining measures of violence. One senior intelligence official tells the Washington Post: “If a bullet went through the back of the head, it’s sectarian. If it went through the front, it’s criminal. Depending on which numbers you pick you get a different outcome.” Petraeus refutes the claim. Additionally, an MNF-I spokesman told the Post the U.S. military does not keep accurate statistics of Shiite-on-Shiite and Sunni-on-Sunni violence, despite a conclusion within the intelligence community that “factional rivalries” are a leading contributor to instability in Iraq, particularly in Basra.
Experts also say Baghdad has become more segregated as a result of the surge, which is likely contributing to the reduction in attacks. Tens of thousands of Sunnis have fled, and those that remain are walled in by massive (NYT) U.S.-made blast walls. Some sources directly disputes U.S. commanders’ contention that civilian deaths have decreased in Baghdad. An unnamed Iraqi Ministry of Interior official told McClatchy Newspapers that the number of Iraqis killed by violence soared in August 2007—to 2,890—nearly double the number cited by Petraeus. The actual numbers may be even higher. The Iraq Study Group report noted in 2006 violence is significantly underreported.
Rand Beers, a counterterrorism adviser to both Bush presidents and President Clinton, says he believes Gen. Petraeus has selectively focused on trends beginning in December 2006 to enhance the perception of military progress. Beers, who heads the Washington-based National Security Network, also suspects the military of inflating pre-December 2006 violence numbers to paint its surge efforts in a more positive light. Lawrence J. Korb, a former senior Pentagon official in the Reagan administration, is equally skeptical: “It’s a selective use of statistics to buttress his case,” Korb says.
Wherever the truth lies, some experts find reason for optimism in the statistical warfare. James Dobbins, a former Bush administration envoy to Afghanistan and an authority on post-conflict zones at the RAND Corporation, says at least the United States is measuring. “We’re finally counting the right thing, which is how many Iraqis are getting killed. In a counterinsurgency this is the fundamental metric,” Dobbins says. “Now, are counting them properly? As the saying goes, there are lies, damn lies, and statistics.”