More on Madison's Nightmare

by Chuck Spinney

I received the following email from Robert Jordan Prescott, who asked

"How does our defense budget compare with the defense budgets of other nations?"

I put together the following graphic comparing US defense budget spending vs. alleged "threats and adversaries" (per Bacevich's Washington Rules, an excellent book also getting at "Madison's Nightmare")

It is a good question, and I have taken the liberty of reproducing Prescott's graphic below:
600 wide.jpg

Mr. Prescott's question is useful, for at least two reasons:

First, he adds useful additional information by comparing U.S. speeding to the cumulative total of defense spending by our potential adversaries.  This is a more comprehensive picture than I presented, and I thank him for peeling one more layer off the onion.

Second, and more importantly, at least in my opinion, he inadvertently illustrates a problem of how best to handle the inevitable data uncertainties when trying to make such a comparison.  These uncertainties are always present and an analyst ignores them at his peril -- but they can be accounted for.  Mr. Prescott's information illustrates how I try to handle this problem.

Mr. Prescott's graphic is based on the Military Balance, 2010, published by the UK-based International Institute for Strategic Studies, IISS - an excellent source.  Readers may recall that I extrapolated my estimate of China's spending from data produced by the Stockholm International Peace Research Institute (SIPRI), also an authoritative source.  The difference between my estimate for China and the IISS estimate for China illustrates how I try to deal with uncertainties that are always present in this kind of data.  The rule of thumb is to construct an analysis in such a way that a reasonable exploration of the effects of the uncertainties in its data tends to make the conclusion stronger and show that the analyst is not cherry picking data to make his point -- quite the contrary in fact.

The IISS estimate of Chinese spending is about $70 billion whereas mine, which was based on a generous extrapolation of SIPRI data, was about $125 billion. Also, note the IISS data in Prescott's graphic  for US spending ($690 billion in 2010) does not appear to include the quite necessary component of the nuclear weapons program in defense spending that is funded by the Department of Energy. (This can be seen by examining the BA data in Wheeler's table for 2010 in my original posting or the outlay data for 2010 in the President's budget at the link referenced before Wheeler's table was introduced).

So, if one takes these differences and uncertainties into account, the lower bound of the multiple I used to make my arguments probably understated the point I was trying to make.

Chuck Spinney retired from the Defense Department in 2003 after 33 years service (bio) and now lives with his wife and dog on a sailboat in the Mediterranean.