Reinhart and Rogoff couldn’t justify austerity before it was debunked

Last week, The Century Foundation hosted a Twitter chat forum in which Mike Konczal of the Roosevelt Institute, TCF fellow Mark Thoma and I discussed the Reinhart and Rogoff kerfuffle and its implications for the policy debate over austerity (here’s the Storified “transcript”). Nearly every facet of this incident has been thoroughly covered in the blogosphere—see Mike’s post for a summary of the Herndon, Ash, and Pollin (2013) paper debunking R&R; Arindrajit Dube’s post on reverse causation; my colleague Josh Bivens’ post on R&R’s response to reverse causation criticism; Paul Krugman on R&R’s obfuscating rebuttal; and Dean Baker’s post on R&R’s purported role in the policy debate.

But what’s gone entirely missing, as far as I can tell, and what I struggled to explain in sub-140-character increments, is that R&R’s reported finding—that “median growth rates for countries with public debt over 90 percent of GDP are roughly one percent lower than otherwise; average (mean) growth rates are several percent lower [and slightly negative]”—couldn’t justify austerity even before it was debunked.

Back in early 2010, pundits and policymakers immediately seized on R&R’s now-invalidated results to justify austerity policies, so the paper’s methodological debunking has been correctly interpreted as a major defeat for the austerity movement. But the Beltway interpretation of R&R was based on a false premise from the get-go. Robert Samuelson’s predictably unhelpful addition to the R&R debate—his half-hearted defense of R&R’s “minor mistakes” is scattered with objectively inaccurate revisionist history—perfectly encapsulates this widely propagated false dichotomy: “It’s ‘austerity’ versus ‘stimulus.’ If debt exceeding 90 percent of GDP is hazardous, then the case for austerity seems stronger.” (Emphasis added.)

Rephrased, it seems to Samuelson that austerity should reduce the debt ratio. But public policy is better guided by evidence than gut feelings.

Causality, data and methodological questions aside, the policy relevant takeaway from R&R’s reported debt “tipping point” for growth was that policymakers should prevent debt ratios from rising to 90 percent. So for R&R to be an economic argument for austerity, austerity would have to reduce debt ratios—but austerity has produced the opposite effect throughout Europe. Holland and Portes (2012) concluded that European austerity not only had a larger-than-expected adverse effect on growth but has perversely raised debt ratios: “In both the UK and the Euro Area as a whole, the result of coordinated fiscal consolidation is a rise in the debt-GDP ratio of approximately 5 percentage points.” Of the twelve counties analyzed, Ireland was the only one in which fiscal consolidation over 2011-2013 was not found to be entirely counterproductive with respect to debt ratios.

The debt ratio is simply government debt as a share of GDP, so it rises if: a) GDP falls cet. paribus; b) nominal debt rises cet. paribus; or c) nominal debt rises relatively faster than GDP growth. But when GDP falls, the debt ratio increases through two mechanisms: a decreased denominator and also increased nominal debt through fiscal feedback effects. So a dollar of austerity will reduce GDP by the policy’s associated fiscal multiplier, and a fraction of that austerity-induced reduction in GDP will increase the cyclical portion of the budget deficit by depressing tax receipts and increasing spending on automatic stabilizers (e.g., unemployment benefits). For the United States, every dollar the economy moves away from potential output adds roughly $0.37 to budget deficits—and this effect is leveraged by the policy’s associated fiscal multiplier.

The kicker is that fiscal multipliers are currently elevated because monetary policy cannot be employed to cushion fiscal retrenchment (advanced economies’ central banks have short-term policy rates stuck at or near the zero lower bound of nominal interest rates), and the interest rate “crowd-out” is currently blocked (at full employment, fiscal retrenchment would be partially cushioned by increased private investment as decreased public borrowing lowered interest rates). Blanchard and Leign (2013) (PDF) document that policymakers and fiscal authorities greatly underestimated fiscal multipliers early in the downturn, and identified best estimates (PDF) of the government spending multiplier ranging from 0.9 to 1.7.

For policy savings to lower the U.S. debt ratio from current levels, the associated fiscal multiplier must be under roughly 0.9—below Blanchard and Leigh’s range of estimates—suggesting that any government spending cuts will increase the near-term debt ratio. We estimate that sequestration spending cuts being implemented will push the U.S. debt ratio higher in 2013.

Essentially, austerity policies in Europe and the United States are trading weaker growth and larger cyclical budget deficits for smaller structural budget deficits, to the net effect of worsening relative near-term fiscal positions. R&R’s reported results never provided any economic cover for this trade.

Holland and Portes, let alone a cursory look at the U.K.’s experience with austerity, should have been a nail in the coffin for using R&R as an economic argument justify austerity. Then again, Krugman declared that R&R “has been completely discredited” in July 2010 after Josh and former EPI colleague John Irons (2010) identified serious theoretical and empirical flaws, particularly reverse causation with the contemporaneous correlation between growth and debt.

But R&R’s debt “tipping point” played out as an expedient fig leaf for far too many pundits and policymakers interested in austerity because it seemed like the responsible thing—or worse, the pretense of painful choices seeming necessary advanced ulterior policy agendas—economics be damned. Depressingly, austerity policies are all too likely to continue being implemented on both sides of the Atlantic without any perceived economic justification whatsoever, just as austerity has been escalated without any empirical economic justification for some time.