In a liquidity trap, it is very low relative to its benefits--the deadweight loss from the extra future taxation needed to amortize the extra debt is the cost, and the benefits are large and immediate.
The true fiscal cost of stimulus : As I get ready for the CAP and EPI events, I’ve been thinking more about the issue of crowding in. (See also Mark Thoma.) And I’m coming more and more to the conclusion that the public debate over fiscal stimulus, which views it as an agonizing tradeoff between possible benefits now and certain costs later, is wildly off base. Just to be clear, we’re talking about fiscal stimulus in a liquidity trap — that is, under conditions in which conventional monetary policy has lost traction, in which the Fed would set interest rates much lower if it could. Under more normal conditions the conventional view of stimulus is more or less right. But we’re in liquidity-trap conditions now, and will be for a long time if official projections are at all right. So what does that imply?
First of all, as I and others have pointed out, fiscal expansion does not crowd out private investment — on the contrary, there’s crowding in, because a stronger economy leads to more investment. So fiscal expansion increases future potential, rather than reducing it. And yes, there’s some evidence to that effect beyond the procyclical behavior of investment.... Still, it does burden the government with higher debt, requiring higher taxes or some other sacrifice in the future. Or does it? Well, probably — but not nearly as much as generally assumed.
Here’s why: first, in the short run fiscal expansion leads to higher GDP, which leads to higher revenues, which offset a significant fraction of the initial outlay. A billion dollars in stimulus probably leads to only $600 million or a bit more in additional debt. But that’s not the whole story. Crowding in raises future GDP — which raises future tax revenues... [and] offsets... some of the burden of debt service....
And that, in turn, means that penny-pinching on stimulus is deeply, destructively foolish.