The Great Debt Delusion: How Math Keeps Proving Austerity Wrong

The austerians' foundational text lost a battle with Excel -- and reality

Ignore the mistakes behind my austerity study, says Harvard professor Ken Rogoff (Reuters)

Well, it turns out history does repeat itself as farce. Especially when it comes to selling austerity.

Back in 2010, policymakers decided the economic crisis was over. Fiscal and monetary stimulus had prevented the Great Recession from turning into a great depression, and that was enough for them to pretty much declare mission accomplished. It was time for policy to go back to "normal", even though the economy had not. In other words, it was time for austerity and higher interest rates, low growth and high unemployment be damned.

There was just one small problem. It was impossible to justify tighter budgets and tighter money when the economy was still depressed. Contractionary policy is, well, contractionary, and the U.S. and Europe certainly didn't need more of that with unemployment still in the double digits.

Unless austerity isn't contractionary. Maybe cutting spending and raising taxes is good for growth!

That was the conclusion of a paper by Alberto Alesina and Silvia Ardagna that quickly became the Holy Grail of austerians everywhere. Alesina and Ardagna looked at times when advanced countries had cut their deficits in the past 40 years, and found that austerity was often expansionary, particularly when it was weighted towards spending cuts rather than tax hikes. But this wasn't quite the intellectual coup it looked like. For one, Alesina and Ardagna didn't do a very good job identifying when countries did or didn't do austerity -- their sample missed some pretty big counter-examples. For another, as Mike Konczal and Arjun Jayadev of the Roosevelt Institute showed, almost none of the countries started austerity amidst a slump, and all of them could (and did) cut interest rates to offset their deficit-cutting. In other words, Alesina and Ardagna wasn't relevant to our zero-interest rate, low-growth world.

Okay, so austerity isn't expansionary in the short-run, but maybe it's expansionary in the long-run. That was the conclusion of a paper by Carmen Reinhart and Ken Rogoff that quickly became the new Holy Grail of austerians everywhere. Reinhart and Rogoff looked at times when advanced countries have had debt of at least 90 percent of GDP going back to 1946, and found that growth tends to be much lower during these periods. Reinhart and Rogoff argued this shows that higher debt causes lower growth -- which would mean austerity causes higher growth, at least in the long run.

There are lots of problems here. Most obviously, this correlation between high debt and low growth doesn't prove much. Deficits and debt go up during a slump as tax revenue falls and safety-net spending rises. That doesn't mean debt and deficits cause the slump; the opposite. That alone is reason to discount this fiscal doomsaying. But now there might be another.

It turns out Reinhart and Rogoff might have gotten some of their data wrong. As Mike Konczal reports -- and you really need to read the link -- a new paper by Thomas Herndon, Michael Ash, and Robert Pollin of the University of Massachusetts, Amherst, identifies two big issues. First, and most amusingly, Reinhart and Rogoff seem to have made an Excel formula error. (Yes, really). As you can see below, they apparently only averaged the first 14 countries in their sample, and forgot to drag down for the last five. Oops.


This Excel error "only" reduced their growth average by 0.3 percentage points, which although not good, isn't as bad as their other mistake. That was using the wrong number for New Zealand. Now, Reinhart and Rogoff have a straightforward enough methodology. They calculate each country's average growth for each year it had debt over 90 percent of GDP, and then average those country averages. But they didn't calculate the average for New Zealand. They just used one year instead. This skewed the result more than you might think, since the average should have been 2.6 percent, and Reinhart and Rogoff used one year that had -7.6 percent growth. Double oops.

But is Reinhart and Rogoff's methodology even the right one? That's not clear. Because they just average the country averages, Reinhart and Rogoff weigh each equally, regardless of whether they had one year of bad growth or many years of decent growth. That's not necessarily wrong, but if you average all the years, rather than all the countries, you get 2.2 percent growth, versus the -0.1 percent Reinhart and Rogoff reported.

This brouhaha has been equal parts distracting and revealing. The Herndon, Ash, and Pollin paper doesn't change the big picture all that much. Growth does tend to slow down when debt is high, just not as much as Reinhart and Rogoff claimed. But whether it slows down to 2.2 percent or -0.1 percent doesn't really matter. What matters is whether it's the higher debt causing the slower growth, whatever that may be. There's still no evidence of that, and never has been.

But that doesn't mean this dispute has just been a sideshow. It should be astonishing that such an apparently shoddy paper has been as influential as it has. Now, policymakers didn't pursue austerity because of Reinhart and Rogoff, but it was the best argument they had for doing so. And they used it a lot. Think about that. A paper that couldn't even manage to enter numbers into a spreadsheet correctly became the chief intellectual justification for a horribly self-defeating economic experiment that has kept millions out of work.

We're finding out with how little wisdom the world is governed, one Excel error at a time.