« Playing Nice | Main | Untroubled by Inequality »

The Cost of War

26 Oct 2007 07:37 am

Ever since I read Kastellec & Leoni's "Using Graphs Instead of Tables to Improve the Presentation of Empirical Results in Political Science" I've been slightly annoyed with the frequent presentation of data in table format. What's more, I think the audience expects a multimedia component from today's Professional Blogger, so I like a graph with bright colors. Hence, I took Table 1 from Peter Orzag's "Estimated Costs of U.S. Operations in Iraq and Afghanistan and of Other Activities Related to the War on Terrorism" and made this thing below which may or may not be superior to the original table (unfortunately, I'm not much of a graph maker, my new dream is to find some kind of crack graph-making intern):

cboestimatesiraq.png

Under scenario one, the number of deployed troops is reduced to 30,000 by 2010. Under scenario two, the number of deployed troops is reduced to 75,000 by 2013. One takeaway here is to keep in mind that the next time you hear someone say the country can't afford to spend an additional $50 billion a year on something (mass transit, day care, schools, whatever) that this is a bit less than the low cost scenario for future military ventures. As my old boss Bob Kuttner likes to say, we have a way of finding out that we can afford the things that our political leaders decide it's important to afford.

On the other hand, puzzling through this I'm left baffled as to why House Budget Committee John Spratt asked for these particular things to be estimated. Running Iraq and Afghanistan together seems to blunt whatever political point it is one was trying to make, and doing so in the context of a discussion of potential troop withdrawals actually leaves it unclear what policy option we're considering. Are our 75,000 troops supposed to be in Iraq or in Afghanistan? It may not make a large financial difference which country you deploy them to, but obviously those are two very different policies. Consequently, the main value of this exercise may be the portion near the end where Orzag compares his methodology to that of the famous Stiglitz & Bilmes paper and argues convincingly that they mishandled the question of the long-term treatment costs for Iraq-related brain injuries.

Share This

Comments (17)

"I like a graph with bright colors."

As do we all.

-----

"It may not make a large financial difference which country you deploy them to, but obviously those are two very different policies."

Yup. But Spratt's looking at budgeting, not foreign policy.

From his bailiwick, it doesn't matter if the troops are stationed in Iraq or Antarctica. It just matters how much is being spent outside the normal budget every year.

Also, I'll mentioned that I'm shocked how small a percentage of overall spending is taken up by Veterans Benefits.

That can't possibly be referring to total Veterans Benefits spent over the course of Iraq veterans' lifetimes, can it?

That can't possibly be referring to total Veterans Benefits spent over the course of Iraq veterans' lifetimes, can it?

It probably is. If it looks small, that's because it's dwarfed by the massive cost of conducting a major military operation. Fighting a war really is mind-blowingly expensive.

I'm not sure if Matt's conversion of the table into a bar chart was intended to point out the inherent silliness of the CBO report, but that's the effect it had on me. Both the bar chart and the original table have a "trying-to-justify-a-conjecture" quality-- similar to a graph with only one data point.

Matt,

As a political scientist, I am glad to see political bloggers using this work. You should check out some of Ed Tufte's books on the visual display of quantitative data, including, naturally, "The Visual Display of Quantitative Data," and "Beautiful Data," which I appreciate just for the title.

Rather, the title is "Beautiful Evidence." Oops.

"It probably is. If it looks small, that's because it's dwarfed by the massive cost of conducting a major military operation. Fighting a war really is mind-blowingly expensive."

I know American style combat eats up the dollars like an SUV eats up petrol, but I've really got to assume those Veterans benefits are only for the next couple of years, not over the course of Iraq veterans' lifetimes.

There's pensions and a lot of healthcare that gets spent over a lifetime, if I understand things correctly. And the fact that the "so far" period has a lower Veterans' benefits percentage than the future periods would seem to support my contention...

I'm not much of a graph maker

Try Elements of Graph Design, by Stephen M. Kosslyn, who's a well-known vision scientist, or Elements of Graphing Data by William S. Cleveland, who'se a well-known statistician. Both are very good. (Tufte has some nice ideas, good insights, and great examples of bad graphs, but a few of his recommendations aren't the best.)

I've really got to assume those Veterans benefits are only for the next couple of years, not over the course of Iraq veterans' lifetimes.

N.B., dude: each tick on that y-axis represents 200 billion dollars. Eyeballing it, it looks like Scenarios One and Two allow for about ten billion dollars in veteran's benefits--which, for a group of 100,000 vets over a lifetime, seems about right.

"each tick on that y-axis represents 200 billion dollars"

Sure. But eyeballing those graphs seems to indicate that Veterans benefits are only 1% - 2% of total expenditures, which seems light to me unless the accounting system is only calculating benefits paid currently.

And again, I do think the way that Veterans benefits take a lower percentage in past expenditures than in the future scenarios seems to back up my interpretation that this is only accounting for Veterans benefits spent currently.

(Also, I love that we're both too lazy to download and read the PDF to see who's right.)

"my new dream is to find some kind of crack graph-making intern"

would you settle for a crack-making graph intern?

Follow the link and read the report. The text makes it clear that the line item for veterans' benefits constitutes expenditures during the time period stated, not during the lifetimes of the veterans in question.

Thank you, NCProsecutor.

-----

It would be interesting to see the war cost numbers with future Veterans benefits calculated in.

Those estamates seem way low. It says we "only" spent $600 billion so far. I would have to research it, but I believe we've spent way more than that. Bush just requested $200 billion alone.

But then that chart seems crazy too. How can "scenario one" be less than "so far." Are scenario one and two added to so far? So you are showing past and future expendatures on one chart? If so, that is NOT a good chart. If you are showing past and future there needs to be a line separating them with the past bar below and the future bars above

Here's something else that seems off: based on eyeballing alone, it looks as if the scenario 1 spends $550 billion over the next three years (~$180 billion a year) while scenario 2 spends $1,050 billion spread over six years ($175 billion a year).

So keeping more troops in theater over a longer period results in less money spent per year? Shit, if we sent 500,000 troops to occupy Iraq indefinitely, maybe the whole thing would pay for itself.

I also recommend looking over some of Tufte's work, although I guarantee you will be further annoyed at 95% of the graphs and tables you come across.
My pet peeves: the terrible colors that are the 'defaults' in Excel as well as the gray background.

The real point to take away from this is that US military operations are just pure theft from the taxpaying population.

It's pure wealth transfer from the entire population to the tiniest portion of the rich - those who run defense-related industries. Well, they might not be the "tiniest" portion anymore, given the size of the military-industrial complex, especially if you add in the finance and oil components.

But still, this is pure, utter theft.

Especially when you realize that the US military could defend the population of the US and its real national interests with a military one tenth the size and one tenth the cost - if it was properly designed.

You can easily turn off the gray background that's the default in Excel. Doing that makes the graph more vivid.


Comments closed November 09, 2007.

Copyright © 2008 by The Atlantic Monthly Group. All rights reserved.