Postrel correctly points out that old-fashioned Keynesian theory would peg the national cost of the war as being less than the Budget impact (now put by the Bush Administration at $75 billion), because some of the resources used otherwise would be unemployed. However, as she also points out, today's economists tend to minimize the Keynesian argument.
'Jane Galt" describes an article by Jamie Galbraith that, among other things, adds together the Budget cost of the war and the "opportunity cost" of doing something else, such as expanding health care spending. This is double-counting. The Budget cost is the opportunity cost.
For Budget purposes, I believe that it is important to estimate the cost of the war. However, I do not see the value added in trying to evaluate the war on the basis of financial costs and benefits. In that respect, discussing the financial cost of war inappropriately trivializes what is at stake. The costs and benefits of the war are almost entirely intangible (or difficult to translate into financial terms), and they depend on the long-term outcome. This is one instance where economic analysis takes a back seat, in my opinion.