Yet more proof that Krugman really, really, really needs to get a blog:

Some people, bizarrely, think that I don't know that spending plays a role in the deficit. Well, duh. If you look at my book Fuzzy Math, p. 75, you'll see a table I took from Auerbach and Gale. It shows that if you replace the unrealistic assumption of zero growth in real discretionary spending with the more reasonable assumption of constant spending per capita, the projected 10-year surplus falls almost $500 billion. If you use the even more reasonable assumption that discretionary spending remains constant as a share of GDP, the projection falls more than $600 billion more. So going from zero real growth in discretionary spending to keeping such spending constant as a share of GDP - which is the implicit assumption in my back-of-the-envelope calculation above - subtracts more than $1.1 trillion from the budget projection. That's still well short of the $1.7 trillion in direct and indirect costs from the Bush tax cut (close to $2 trillion if you ignore the nonsense about expiring tax cuts in 2011), but it's substantial.

Naturally, Krugman concludes that this, plus the cuts, means the possibility of "deficits forever", even factoring in the output gap. And the common answer of "just cut spending regardless?" He covers that too:

Smaller government is a great catchphrase, until you actually start cutting things like mine safety and nutritional aid for poor children. Apparently some people even think that, as Martha Stewart would say, it's a good thing to trick veterans into not getting health care.

Then again, considering the yeoman work that Max and J. Bradford DeLong have been doing lately, maybe that would be a little too much.