I was thinking about thinking about this:
Dynamic Scoring, by Andy Harless: Suppose that, at the beginning of the fiscal year, Congress appropriates $100 billion extra for infrastructure projects. At the end of the fiscal year, how much higher will the deficit turn out to be, compared to what it otherwise would have been?
The obvious answer, and the one that usually seems to be implicitly assumed by the media and the pundits, is $100 billion. But if you think about it carefully, it should become obvious that the obvious answer is the wrong answer.
The government is going to use most of that money to hire people and to buy things. Many of the people it will hire are people who were previously unemployed. Many are leaving other jobs which will subsequently be filled by people who were unemployed. These previously unemployed people, who may have been collecting benefits, will now be paying taxes. Those taxes will reduce the deficit, as will the reduced benefit payments. Moreover, for the businesses from which the government purchases, their profits will rise, and they will pay additional taxes on those additional profits. And they may expand and hire new people, or retain people that would otherwise have been laid off. And (if you believe in a multiplier effect), all the newly employed people, as well as the owners of the businesses, will spend more money, thus providing more profits and more employment for others, who will also pay taxes and stop collecting benefits. And so on. The ultimate effect of the original expenditure on the budget deficit will be considerably smaller than $100 billion.
This is called dynamic scoring. ... In the past, dynamic scoring has met with a lot of skepticism – and with good reason. Under normal economic conditions (by which I mean those that prevailed from 1953 through 2007), it's not clear that budget changes have any significant indirect effects on revenues and expenditures. Supply-siders claimed that the incentive effect of tax cuts would increase incentives for economic activity and thereby result in increased revenues. (I mean, "increased" relative to the static estimate of the revenue loss, not increased relative to what would happen without the tax cut. The latter idea had a lot of play in the popular press, but it was seldom taken very seriously by economists.) Keynesians (the old-fashioned kind) claimed that tax cuts and expenditure increases would increase demand and thereby result in increased revenues (again, relative to the static estimate). But...
Mainstream economic analysis said they were both wrong. Many economists think there are major supply-side benefits to more efficient taxation, but most such economists think those are primarily long-run benefits (faster growth over a span of time) rather than benefits that would significantly affect revenues in the short run. The Keynesian argument would make sense if monetary policy were passive, but in fact, the Fed has its own goals, and its goals don't necessarily change in response to fiscal policy. And of course the Fed takes fiscal policy into account when deciding how to accomplish those goals. So if a tax cut or an expenditure increase were expected to create, say, a million extra jobs, then, under normal economic conditions, the Fed would simply raise interest rates enough (according to its best estimate) to destroy a million jobs. (If the Fed didn't think the demand for those million jobs would be potentially inflationary, then it would already have tried to create them.)
But today's economic conditions are not normal. The Fed, like most everyone else, is expecting the recession to be a severe one, a potentially deflationary one, but the Fed is running out of options for how to deal with it. Contrary to what happens under normal conditions, the Fed will make no attempt to offset the effects of fiscal policy; indeed, it will enthusiastically welcome the help. The old-fashioned Keynesians, whose advice about dynamic scoring was (properly, in my opinion) considered wrong or irrelevant for so long, can now dust off their computers and start giving meaningful dynamic estimates of the effects of budget changes. ...
But there has never been a consensus beforehand about how dynamic scoring should be done, or even about the direction of the effects. In the past, the only conservative approach was to use static scoring – to ignore any indirect effects that budget changes might have on the ultimate deficit.
There is still no consensus about the details. But today one can hardly doubt that the indirect effects of stimulus policies on the budget will partly (if not entirely) offset their direct effects, or that the indirect effects will be large enough to be important. In today’s environment, static scoring is not just conservative, it's fundamentally unreasonable.
The details have to be negotiated... But next time you think that an $800 billion stimulus plan will add $800 billion to the national debt, think again.
I was thinking more along the lines of the traditional supply-side argument, i.e. the dynamic effects from the higher growth rate we'd have with improved infrastructure, but as noted above, these "are primarily long-run benefits."
Update: Paul Krugman:
Bang for the buck (wonkish): Mark Thoma says he was thinking about thinking about this; I was actually thinking about it. Anyway, it’s true: the cost of an effective fiscal stimulus, in terms of adding to the government’s debt, can (and should) be much less than the headline cost.
Consider an increase in government spending; assume that the interest rate is fixed (a good assumption right now, because interest rates are up against the zero lower bound). Then textbook analysis says that if the stimulus is dG, the increase in GDP is 1/(1 - c(1-t)) where c is the marginal propensity to consume out of income and t is the marginal tax rate. Suppose c is 0.5 and t is 1/3; then the multiplier is 1.5, which is more or less the conventional wisdom right now.
But if $100 billion in spending raises GDP by $150 billion, and the marginal tax rate is 1/3, $50 billion of the spending comes back in additional revenue. So bang for the buck — increase in GDP per dollar of added debt — is 3, not 1.5. Since the main concern about stimulus is that it will add to government debt, it’s this bang for the buck measure, rather than the multiplier, that’s relevant. And 3 sounds a lot better than 1.5.
Take this a bit further: $150 billion is about 1 percent of GDP, which Romer and Bernstein say means a million jobs; so this says $50,000 per job, which is a much better number than the critics have been throwing around (plus many more workers with full-time rather than part-time jobs).
Bang for the buck also heightens the contrast between effective and ineffective stimulus policies. Stay with c = 0.5, t = 1/3, and look at the effects of a tax cut; the multiplier is 0.75, half that for public investment, but bang for the buck is 1, only 1/3 that for investment.
So thinking about how stimulus comes back via revenues is important.