Krugman:

Think of the government budget as involving tradeoffs similar to
those an individual household makes. On one side, there are all kinds
of things the government could be doing, from dropping freedom bombs to
providing children with dental care; think of each of these things as
involving a certain marginal benefit per additional dollar spent, with
the marginal benefit declining in the total amount spent on each
concern…

What the government should do,
in this case, is set all the marginals equal: the marginal benefit of
an additional dollar spent on bombs, dental work, national parks, soup
kitchens, etc, should all be equal, and this common marginal benefit
should equal the marginal cost of raising an additional dollar of
revenue.

Now suppose a disaster strikes. What this does is raise
the marginal benefit of spending on disaster relief. The appropriate
response is to move all the marginals to get them in line: spend less
on everything else, and also raise more in taxes.

Landsburg:

In a radical departure from his previous expressions of dissillusionment, Paul Krugman has implicitly declared in his latest blog post
that we are now living under the best of all policy regimes…

Since Krugman has carelessly neglected to spell out an important
detail of his argument, let me fill in the gap for him: The Ricardian
conclusion does not come from thin air; instead it follows logically
from certain premises, key among which is that you’re starting from an ideal policy regime.

[…]

Unless
you believe that everything is perfect to begin with, the Ricardian
argument fails, leaving you with no reason to believe that the cost of
new spending should be spread widely. Instead, you should start by
cutting back on your least wise activities.

One of my favorite interpretive principles is, “He said what he meant.”  If we apply this principle to Krugman, then Landsburg’s got him dead-to-rights.  Krugman is being Panglossian.  It is bizarre for a guy who thinks that Satan controls half the U.S. government to think that the status quo is remotely optimal.  And when the status quo is far from optimal, economics tells you to cut your worse programs first.

I have just one point to add.  Even if we did start from an optimal state, microeconomics still wouldn’t tell us to cut a little of everything.  Why not?  Fixed costs.  Every proposed cut sparks a political battle.  Each of these battles is costly.  In order to cut a little from a million different programs, you’d have to fight a million costly battles.  Under the circumstances, standard, Panglossian micro tells us to pick fewer, bigger battles: To cut some programs a lot, and leave the rest alone.