The Great Debate over whether we should cut taxes or increase government spending to stimulate the economy has sent many of us noneconomists digging into the evidence economists have gathered on the topic. Most of it is, as Nate Silver put it after reading Christina and David Romer’s big paper (pdf) on the macroeconomic effects of postwar tax changes, “the sort of thing that will make your head spin.” But the even more dramatic conclusion I have come to after reading through the Romer-Romer paper and the most-cited recent work on the macroeconomic impact of changes in government spending, by UC-San Diego’s Valerie Ramey (pdf), is that they don’t even begin to resolve the debate.
This issue, basically, is that in a complex economy measuring cause and effect is really hard. In response to this difficulty, both Ramey and the Romers (sadly, they haven’t started a band together, at least not as far as I know) have devised ways to separate signals from noise. And in both cases, this has meant excluding the sort of tax cuts and spending increases being contemplated today.
The Romers did this by compiling an exhaustive narrative of tax changes over the past six decades, and separating them into two main categories: endogenous (that is, motivated by a desire to stimulate a faltering economy or pay for planned increases in spending) and exogenous (“any tax change not motivated by a desire to return output growth to normal”). The effects of endogenous tax changes are especially hard to disentangle from other economic factors, their reasoning went, so they focused on exogenous tax changes and found a pretty substantial—as much as 3-to-1—negative impact on GDP from tax hikes and positive impact from tax cuts. The one exception: Tax increases explicitly intended to reduce a budget deficit didn’t have much of a negative effect. The Romers also found a 2.1-to-1 impact for tax changes intended to counter a slowdown, but expressed very little confidence in that number—for the reasons expressed above and also because there just haven’t been many such tax cuts in recent decades.
So the Romers identified a bigger-than-generally-assumed economic impact from tax changes, but didn’t have much to say about precisely the kind of tax change being contemplated now, and didn’t have anything at all to say about the relative efficacy of tax changes vs. spending changes. (Brad DeLong, who has worked down the hall from them at UC Berkeley for years, has something to say on what he thinks their stance is.)
Which brings us to Ramey, who in trying to identify government “spending shocks” big enough to have a measurable impact on GDP ended up focusing on sudden increases in defense spending through the years. The impact on GDP was unimpressive: 1.4-to-1. But remember, that’s just defense spending. Why only focus on defense spending? Because, Ramey writes, “most nondefense spending is state and local spending on potentially productive activities such as education.” For the purpose of getting a cleaner measure of the impact of government spending, she has excluded much of the kind of spending being contemplated today. And as with the Romers, there was no attempt to compare the impact of government spending with tax cuts—and the two studies seem different enough that lining up the Romer’s 3-to-1 next to Ramey’s 1.4-to-1, as Greg Mankiw did in the New York Times on Sunday, is really comparing apples to, I dunno, raspberries.
The reality here is that there is simply no clear answer in the data—especially since we only have really good data for the post-World War II era and the U.S. has not experienced a recession quite like this one in the post-World War II era. Which forces us to fall back on theory, and not just economic theory. If you believe that government is capable of spending money intelligently—or at least more intelligently than the currently paralyzed private sector—then more spending is the ticket. If you don’t, you’re going to lean toward tax cuts or no action at all.
Update pneogy points in the comments to a new Paul Krugman post in which he makes the case for spending vs. tax cuts. It still suffers from the limitation that it’s based on theory, not any kind of conclusive empirical evidence, but Krugman adds an interesting twist:
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.
I should add that I lean toward the idea that some big part of the stimulus plan ought to be spending. But I’m basing that on theory and ideology and hunches, not hard evidence. And I also think it’s hard for government to quickly come up with smart ways to spend $800 billion. That’s the assessment of Obama’s economic team, too, which is why they’re including tax cuts in their stimulus plan.