« Star-Ledger 03-23-07 / Be About Peace Day | Home | Rep. Woolsey Bill HR 508 to Withdraw from Iraq 1-27-07 »

Cost of Iraq War Studies by Dan O'Flaherty

Iraq War Cost Estimates: A brief survey
by Dan O'Flaherty, Columbia University; Treasurer South Mountain Peace Action

March 13, 2007

Studies differ in their sophistication. The least sophisticated just look at authorized budgetary expenditures of the federal government. The intermediate level looks at budgetary expenditures that have not been authorized yet, but almost certainly will have to be. The final level looks at both budgetary expenditures and other costs borne by American residents, and to be borne by American residents. These studies are the most sophisticated but also the most uncertain, since they require their authors to construct projections both of what the world will look like over the next 15 years (tough enough) and what it would have looked like had the war not started (even tougher).

A. Budgetary studies

The definitive study of the budgetary impact of the Iraq war is by the Congressional Research Service. This is updated as of September 2006, before the 'surge':

The most interesting finding there is that budgetary costs had risen by mid-2006 from about $5 billion a month to almost $10 billion a month by 2006. The Boston Globe did a summary story of this finding:

This has some impact on the older studies of more comprehensive costs.

The National Priorities Project uses the same accounting basis -- Congressional appropriations to develop its ticker:

This is at $456 billion today. This site lets you disaggregate the costs to the state of New Jersey (around $18.6 billion today) and to South Orange ($57 million) and Maplewood ($76 million). I discuss this disaggregation below; these figures seem low to me.

An interpretation of these figures is available from Eric Margolis at the Hall Institute website

He calls Iraq "The Second Most Expensive War in American History."

B. Intermediate study

Linda Bilmes looks at the projected budgetary cost of caring for wounded veterans. She points out that the wound/death ratio in Iraq at around 16/1 is the highest of any war, almost 10 times the comparable ratio for World War 2. The study is intermediate in that it looks only at budgetary cost, but does not restrict itself to explicit appropriations that have already occurred. These are appropriations that will have to be made (of course, Congress could refuse to care for wounded veterans, but this is highly unlikely and most people would think morally reprehensible). She finds a long-term cost in the $350 billion to $700 billion range. This paper was completed last year, when there was a chance that the Baker commission recommendations would be accepted; because they have not been the bottom of this range is not possible now. The paper is available at:

Notice that this item alone, which is not included in the Congressional appropriations discussed above, is likely greater than the entire directly appropriated budget cost to date.

C. Comprehensive studies

There are two studies that attempt to estimate the full costs to Americans of the Iraq war, not just the budgetary cost or the budgetary appropriations. Necessarily, these look to the future, since the war is not over yet, and costs continue to be incurred.

The original study by Joe Stiglitz and Linda Bilmes was released at the American Economic Association meetings in January 2006. Their estimate was $1.2 to $2.0 trillion. This included not only budgetary expenditures and veterans costs, but the investments needed to return the military to its pre-war preparedness, the costs of a share of higher oil prices, and macroeconomic sluggishness. They treat only a small fraction of the oil price increase as due to the war, but that fraction causes severe terms of trade losses for the American people as a whole. The study and much related work is available at the website for the Economists for Peace and Security:

The study itself is at

A different study was done by Scott Wallsten and Katrina Kosec under the auspices of the American Enterprise Institute. This study was released shortly after Bilmes-Stiglitz.,filter.all/pub_detail.asp

This study does not include macro and oil costs, and reaches a bottom line of $1.2 trillion. It notes that the war precluded the need for containing Saddam, and estimates that the long-term costs of doing that would be about $400 billion. Note that this figure assumes that Saddam, in the absence of US intervention, would live eternally.

AEI and Brookings have a web-page where you can make your own assumptions about the future course of the war and see what it implies in the Kosec-Wallsten model:

Bilmes and Stiglitz produced a rejoinder to Kosec-Wallsten and updated their analysis in October 2006:

The bottom line in this paper is $2.3 trillion, a little more than the original paper. This paper includes only $127 billion for veterans, and so if Bilmes were to revisit it today she would probably add another $300-400 billion.

Two quick discussions of these numbers are available from David Leonhardt in the New York Times, “What $1.2 Trillion Can Buy”

and Kevin Hassett, an economist with the American Enterprise Institute:,filter.all/pub_detail.asp

John Foster argues that even these figures are seriously understated because they do not take into account losses suffered by Iraqis. The argument is not humanitarian; it is simply that the US will probably eventually have to spend a lot of money cleaning up the damage done there, as in the Marshall plan.

D. Effect on South Orange and Maplewood

Disaggregating these numbers is also difficult.

The most explicit attempt to do so is the National Priorities Project. As noted, they deal only in appropriated budgetary cost. They implicitly assume that this spending is incremental—that in the absence of the war, taxes would be decreased across the board and the deficit and all other programs would remain the same. Thus they rely on detailed existing studies of existing revenue sources to attribute shares to each state. That’s how they got the $18.6 billion for New Jersey.
For allocation within state, they do not explicitly state what they do, but say only that local shares are based on population and median income. Thus they attribute 0.41% of New Jersey costs to Maplewood and 0.31% to South Orange.
One way to check this allocation is to look at 1999 income in the 2000 census. That yields

Shares of New Jersey Aggregate Household Income, 1999

Maplewood South Orange
Total 0.38% 0.30%

For households l.t. $200k 0.33 0.21
For households g.t. $200k 0.61 0.65

Thus the NPP allocation is approximately proportional to aggregate household income. But much of the federal tax system is progressive.

The largest piece of the federal tax system is the personal income tax. Much of the personal income tax is raised from households with income (AGI) over $200,000. In 2000, 45.7% of income tax collections came from households above $200,000 and in 2003, 42.0% did (Statistical Abstract 2007, table 474). If we apply these proportions to the South Orange and Maplewood shares of New Jersey aggregate income, then we would expect 0.46% of New Jersey federal income taxes to come from Maplewood in 2000, 0.47% in 2003. Similarly, we get 0.41% for South Orange in 2000 and 0.39 for 2003.

That would bring March 13 estimates of expenditure to date to $85-87 million for Maplewood and $72-76 million for South Orange.

Of course, much is wrong with these estimates. The personal income tax may not be the marginal source of revenue (although some other revenue sources like the estate tax and possibly corporate income taxes are at least as progressive as the income tax), and the counterfactual may have been more other expenditures, not less revenue. To the extent that the cost of the budgetary cost of the war will be borne in the future, if the marginal revenue source of the federal government is debt, the South Orange-Maplewood share may be less, because the populations of these towns are not growing as fast as the US population. But the tax system that raises these funds in the future may be more progressive than the 2003 tax system.

As far as the comprehensive estimates go, on a national level these run in the neighborhood of 3-5 current war-to-date budgetary expenditures. But it would probably not be appropriate to multiply these Maplewood-South Orange estimates by 3-5 to get comprehensive costs. This is because most of the non-budgetary costs will not be borne in proportion to income tax payments. One would expect the South Orange-Maplewood shares of these costs to be considerably lower (maybe 2/3 to ½).

sitedesign by Darby Communications