The Agenda

NRO’s domestic-policy blog, by Reihan Salam.

Relitigating ARRA



One argument you hear very frequently these days — Matt Yglesias has made it with characteristic effectiveness and wit – is that conservatives are hypocrites because they embraced Keynesian policies in 2001 while rejecting them in the face of the current economic crisis:

Once upon a time an asset bubble burst, but there was little leverage involved and the ensuing downturn was relatively mild. The federal reserve had room to run in terms of cutting interest rates, and the previous ten years’ worth of fiscal policy had seen a series of measures, some bipartisan (1990 & 1997) and some partisan (1993) to improve the country’s budget situation. But the newly inaugurated young president argued that the country needed to enact a large discretionary fiscal stimulus program to combat the downturn, even thought his would shatter the fragile consensus that had guided improvements in the fiscal posture. Oddly, this stimulus program would be phased in and out over a ten-year time horizon. Even odder, the nominally temporary nature of the stimulus was clearly a fraud—everybody understood that the key authors of the stimulus in fact intended the policy change to be permanent in nature.

I refer, of course, to the Economic Growth Tax Relief Reconciliation Act of 2001, a.k.a. “the Bush tax cuts,” which IIRC were roundly applauded by most of the right-of-center economists who can today be found assuming a debt-averse and stimulus-skeptical posture.

There are a few small problems with this analysis: (a) the Economic Growth Tax Relief Reconciliation Act of 2001 was deeply flawed and (b) the deficits that resulted were far smaller than what we’re dealing with now. Indeed, one could argue that the Bush tax cuts and the Fed’s monetary expansion were the kind of extreme macroeconomic policy swings that paved the way for the economic crisis we’re struggling with now. This doesn’t strike me as a strong case for yet another round of extreme macroeconomic policy swings.

More recently, Matt cites a Brad DeLong post in which he plugs in alternative multipliers cited in a new National Affairs essay by Greg Mankiw to conclude that by Mankiw’s (supposed) reckoning, the ARRA should have been wildly successful.

The $787 billion ARRA was about 2/3 spending increases and 1/3 tax cuts. Jared Bernstein and Christie Romer estimated that with a spending multiplier of 1.57 and a tax multiplier of 0.99 that the ARRA would boost production over its lifetime by $787 x 2/3 x 1.57 + $787 x 1/3 x 0.99 = $1,083 billion.

Now comes Greg Mankiw to say that they overestimated the impact of the ARRA because their multipliers were wrong: that he prefers a tax multiplier of 3 (from Romer and Romer) and a spending multiplier of 1.4 (from Valerie Ramey) and that as a result the right estimate was that the ARRA would boost production over its lifetime by $787 x 2/3 x 1.4 + $787 x 1/3 x 3 = $1,521 billion.

Having read Mankiw’s essay fairly carefully, I can’t see anywhere where he says that “he prefers a tax multiplier of 3 (from Romer and Romer) and a spending multiplier of 1.4 (from Valerie Ramey).” Rather, I thought that he was introducing readers to recent findings in the scholarly literature concerning tax multipliers and spending multipliers.

Indeed, Mankiw begins his essay by underlining the uncertainty underlying these calculations. How does one go from the following sentence

There appears to be a growing body of evidence, then, suggesting that taxes may be a better tool for fiscal stimulus than conventional models have indicated.

to DeLong’s post? One hypothesis is that DeLong, a terrifically talented writer, knows that very few of his readers will actually read Mankiw’s essay, and he believes that his “side of the argument” will profit from making Mankiw look foolish. 

Moreover, Mankiw suggests that not all stimulus measures, including tax-focused stimulus measures, aren’t created equal. Indeed, he says this explicitly:

Of course, not all tax cuts or credits are created equal, just as not all direct government spending is.

One straightforward implication of that sentence is that tax multipliers will vary across different kinds of tax cuts and tax credits, just as different forms of direct government spending will have different multipliers. This is yet another reason why DeLong’s decision to apply the same multiplier to all of the unfunded temporary tax cuts and tax credits embedded in ARRA is a head-scratcher if we assume that he is arguing in good faith.

All that said, I think Matt makes a good point when he writes the following:

Meanwhile, I think that endlessly relitigating ARRA tends to detract from the main problem facing right-of-center and left-of-center proponents of fiscal stabilization alike, namely that the political mechanism of enacting discretionary stimulus through congress is extremely clumsy and leads to bad policy outcomes. The most helpful thing to do before the next downturn would be to establish in advance a reasonably streamlined way of preventing state and local government from hiking taxes and cutting services in the middle of a recession.

There is something tiresome about relitigating the ARRA, just as there is something tiresome about endlessly relitigating the Bush tax cuts. I would prefer that we all recognized that ARRA was very poorly designed, not unlike the Bush tax cuts, and craft new legislation designed to fund capital investment on a more sustainable basis and, as Matt recommends, establish a more streamlined set of automatic stabilizers. The trouble is that people who sincerely believe that ARRA was the best imaginable outcome given the political circumstances facing the Obama White House are, in my view, pretty wrongheaded about the next steps we should take. Hence the ongoing debate. 


Subscribe to National Review