Linda J. Bilmes, a lecturer in public policy at Harvard University, calls her latest paper "pretty dry." That hasn't prevented it from riling high-ranking Pentagon officials -- who called her and her dean to complain about her work. When they questioned her sources of material, they ran into a bit of a problem: She did most of her research with data on federal Web sites. So what did the Pentagon do? It changed the Web sites, and now continues to trash her research.
Bilmes has become a leading expert on economic questions related to the war in Iraq, and her experience the last few weeks demonstrates how social scientists can end up in the line of political fire when their findings -- however dry -- offend government officials.
The story begins with a paper Bilmes wrote last year with Joseph E. Stiglitz, a Columbia University professor and Nobel laureate in economics. In their study,  they found that the Bush administration has seriously underestimated the economic costs of the war in Iraq. After the study was publicized, Bilmes was approached by some experts on veterans' benefits who said that one cost of the war hadn't received enough attention in their work (or from the government): the costs of caring for veterans injured in the conflict.
And that's the question that led Bilmes to prepare a 21-page study that she presented this month in Chicago at the Allied Social Sciences Association meeting. The presentation of "Soldiers Returning From Iraq and Afghanistan: The Long-Term Costs of Providing Veterans Medical Care and Disability Benefits"  went off without controversy and might have escaped Pentagon notice. But Bilmes also published an op-ed version of her findings in the Los Angeles Times.  The Pentagon did notice that piece.
The central argument of the new Bilmes paper is that so many soldiers are being injured that the costs of caring for them over their lifetimes is likely to be $350 billion, or up to twice that, depending on how long the war lasts. The high cost is the result of huge advances in military medicine that have greatly reduced the chances that a soldier injured in Iraq will die. As a result, the ratio of injuries to deaths -- 16:1 by her estimate -- is higher than in any other war in U.S. history. (By comparison, in Vietnam the ratio was 2.8:1 and in World War II the ratio was 1.6:1.)
Bilmes uses a series of calculations based on the types of care those injured will require over their lifetimes to offer various scenarios for the costs of the care, and she also argues that the current veterans' health-care system is not ready for the influx of injured or the associated costs. She offers suggestions for streamlining the process of getting injured veterans the benefits they have earned. And while both her studies and the op-ed are critical of the Bush administration's response (or lack thereof) to the veterans' health needs, the tone is academic, not polemic.
What set off the Pentagon was Bilmes' estimate for the current number of injured of 50,500. William Winkenwender Jr., assistant secretary of defense for health affairs, called the Los Angeles Times, Bilmes, and David T. Ellwood -- dean of Harvard's John F. Kennedy School of Government -- to complain that the real figure is less than half that -- just over 22,000. When Bilmes was asked where she got her data, she pointed out that it came from the Department of Veterans Affairs, which in turn gets its data from the Pentagon.
The Pentagon investigated further and found that the VA "misunderstood" the Pentagon's reports, according to Cynthia Smith, a Department of Defense spokeswoman. She acknowledged that the VA had been using numbers consistent with what Bilmes reported, but said that once the Pentagon explained "the error," the Veterans Affairs department changed its Web site so its injury numbers are consistent with those of the Pentagon.
Why the misunderstanding and the "error"? The original figures from Veterans Affairs were for "non-mortal" injuries. But that doesn't include only those who are shot at in combat. That includes people who get sick, people who are in accidents and so forth -- a group of people that is as large as those injured in combat. The Pentagon doesn't want those people counted.
Bilmes points out that a soldier in an accident in Iraq is as entitled to health care as a soldier who is shot. And she points out that she wrote an economic analysis looking at the question of how much all of this care was going to cost. Leaving out half of those injured would have resulted in seriously flawed numbers -- when the whole point of her work in this area is to help people figure out how much money will be needed for the U.S. to meet obligations it has made to its soldiers.
Smith, the Pentagon spokeswoman, does not dispute Bilmes on the point that soldiers are entitled to health care regardless of how they are injured. "They are all cared for," she said. So if Bilmes is correct that she's counting injured veterans who are entitled to health care and the source for her data is the U.S. government (before the Pentagon had the public data changed), why is Smith issuing statements saying that Bilmes is engaging in "gross distortion," as she said in an e-mail? And why is a top Pentagon official calling Harvard suggesting that numbers are erroneous when they are just not the numbers the Pentagon wants out?
When pressed that Bilmes was just using a more inclusive definition of injured, and not making any mistake or distortion, Smith suggested that she hadn't revealed that she wasn't using the only definition of injured possible, asking twice: "Does she say that in her study?"
Actually Bilmes says exactly that. And she does so in her paper's first footnote, in the introduction of her study. In that footnote, she writes that the Pentagon tracks injuries in several ways, and that one count it uses covers only those wounded with bullets, shrapnel, etc. And Bilmes writes explicitly that this would result in a ratio of injured to killed of 8:1 (although she notes that even this ratio would be larger than that of any previous U.S. war).
To Bilmes, what's infuriating is that the Pentagon is saying she is wrong on points of fact when they aren't dealing with what she actually wrote. "I have no problem with them calling me or anyone to talk about my paper," she said. "But what I think is inappropriate is that they seem to be responding without having read my paper."
While the Pentagon was entitled to its view of which figure of injured it wanted to focus on, she said, it was also unfair for it to object to others' arguing that other numbers were worth examining. "I think it is inappropriate for the Pentagon to put pressure on a junior faculty member doing a piece of scholarly research on the somewhat data-driven subject on disability costs just because they may want to have a different number in public."
Bilmes said that she thought the military should look at the issue from the perspective of making sure that soldiers get the help they may need in the years ahead. While she worked in the Clinton administration (rising to the level of assistant secretary of commerce), she said she viewed this project and her interest in these issues as "nonpartisan." She stressed that she wasn't writing to criticize the Iraq war, but was doing economic analysis -- analysis designed to help those in the armed forces.
Harvard has been supportive, Bilmes said. The university is featuring her research on its home page. 
While the Pentagon continues to attack, others in Washington are praising the paper.
U.S. Sen. Barack Obama, the Illinois Democrat exploring a run for the presidency, was briefed on the findings and cited them in introducing legislation last week to demand better accounting by the military of casualties in Iraq and improvements in the health benefits offered.
"The Pentagon and VA need to come clean on the true costs of the Iraq war on our troops," Obama said in a statement. "It doesn't make a difference whether you were hit by enemy fire, or injured because your vehicle crashed, or got sick because of serving in a war zone. The effects on the soldiers and their families are the same. And the impact in terms of the current fighting force and future demands on the VA are also the same."
And the veterans groups that asked Bilmes to explore the issue are also backing her.
Stephen L. Robinson, a Gulf War vet who is director of government relations for Veterans for America (formerly the Vietnam Veterans of America Foundation), called her findings "incredibly valuable." The Iraq conflict is unusual, he said, in that so many soldiers are being injured out of combat. There are all kinds of reasons for that, he said, but it's a reality.
"We think it's reprehensible that the Pentagon would call Linda Bilmes and bitch her out over the phone and put pressure on the school" about this, Robinson said. "I want to ask the Department of Defense, 'Why aren't you doing reports on why there are so many injured and on their care?' It's ridiculous to be ignoring all of those [non-combat] injuries. Why are they attacking this research?"