Slog Music

Music, Nightlife,
and Drinks

Wednesday, April 17, 2013

How an Error in a Microsoft Excel Spreadsheet Tanked the Global Economy

Posted by on Wed, Apr 17, 2013 at 5:58 PM

The big scandal in economics this week (and yes, there are big scandals in economics) is the news that the most widely cited academic study driving austerity programs worldwide came to an erroneous conclusion due to... wait for it... an error in a Microsoft Excel spreadsheet! (Apparently a simple mistake that "all spreadsheet jockeys fear." Not Microsoft's error, a user error.)

The highly influential paper authored by economists Carmen Reinhart and Ken Rogoff warned that countries with debt-to-GDP ratios above 90 percent experience dramatically lower economic growth than countries with lower debt-to-GDP ratios. It is this alarming (but, now we know, nonexistent) 90-percent cutoff that has largely been cited as grounds for budget-slashing economic austerity in the US and abroad.

But... well... oops:

It turns out that the initial results were driven by simple computational and transcription errors. The most important of these errors was excluding four years of growth data from New Zealand in which it was above the 90 percent debt-to-GDP threshold. When these four years are added in, the average growth rate in New Zealand for its high debt years was 2.6 percent, compared to the -7.6 percent that R&R had entered in their calculation.

Since R&R country weight their data (each country's growth rate has the same weight), and there are only seven countries that cross into the high debt region, correcting this one mistake alone adds 1.5 percentage points to the average growth rate for the high debt countries. This eliminates most of the falloff in growth that R&R find from high debt levels. (HAP find several other important errors in the R&R paper, however the missing New Zealand years are the biggest part of the story.)

This is a big deal because politicians around the world have used this finding from R&R to justify austerity measures that have slowed growth and raised unemployment. In the United States many politicians have pointed to R&R's work as justification for deficit reduction even though the economy is far below full employment by any reasonable measure. In Europe, R&R's work and its derivatives have been used to justify austerity policies that have pushed the unemployment rate over 10 percent for the euro zone as a whole and above 20 percent in Greece and Spain. In other words, this is a mistake that has had enormous consequences.

... If facts mattered in economic policy debates, this should be the cause for a major reassessment of the deficit reduction policies being pursued in the United States and elsewhere.

But of course, facts don't matter.

This post has been updated since it was first published.


Comments (19) RSS

Oldest First Unregistered On Registered On Add a comment
Catherwood 1
Well, the snark is fun, I'm sure, but I've read the articles, and the paper on which the articles are based, and none of them blame the bad arithmetic on Microsoft: the software apparently did what it was told to do by the original researchers, namely, ignore data which failed to comport with their pre-conceived notions of how it should have worked. It's confirmation bias and then some. The stupid pot-shot at Microsoft is just that: stupid, and it obscures the relevant point. Way to go, Stranger journalists.
Posted by Catherwood on April 17, 2013 at 6:06 PM · Report this
SchmuckyTheCat 2
Entering an equation with a range of cells ending in 44 instead of 49 isn't a problem with Excel. It's a typo or mistake by the author. Why this headline - no clue.
Posted by SchmuckyTheCat on April 17, 2013 at 6:10 PM · Report this
@1 -- Goldy is updating this post to clarify the issue was not Microsoft's.
Posted by Christopher Frizzelle on April 17, 2013 at 6:11 PM · Report this
TomJohnsonJr 4
@1, we should bear in mind Goldy's hardly the only headline writer making that choice the last couple days on this story. It is a darn catchy way to get us to read something we might not otherwise.

My favorite view of how the "facts don't matter" angle played out among media writers and policymakers is Krugman's from a few hours ago:
Yet [they] not only grabbed hold of the alleged result, they wrote again and again as if this highly disputed claim was a known fact. Thus just a few months ago the Washington Post, attacking those who wanted to reduce the focus on deficits, wrote,

If [debt projections are] even slightly off, debt-to-GDP could keep rising — and stick dangerously near the 90 percent mark that economists regard as a threat to sustainable economic growth.

Not “some economists”, let alone “some economists who have been sharply criticized by other economists with equally good credentials”, but “economists”.

This is deciding what you want to believe, finding someone who tells you what you want to hear, and pretending that there are no other voices. It’s deeply irresponsible — and you can’t blame Reinhart-Rogoff for that mistake.
Posted by TomJohnsonJr on April 17, 2013 at 6:19 PM · Report this
It's also a bit overstated to suggest that the failed austerity policies of so many countries (including the U. S.) were the result of this one paper. Even if you believe the paper is correct (as it was written way back when) most economists feel that the policies of the U. S. and the U. K. (to name two) were stupid. Boneheaded Republicans and Tories would mention the paper, but they ignored the actual evidence presented. None of the papers suggested that we had anything to fear by the amount of debt we have. Our policies are simply the result of fear mongering by folks who have the same sort of stubborn adherence to a failed philosophy as old line communists -- ignore the evidence, this will work if we just try harder!
Posted by Ross on April 17, 2013 at 6:27 PM · Report this
Goldy 8
@1 Nothing in the body of the post implied that it was Microsoft's fault. I wrote an admittedly leading headline with a local angle, and just clarified it. I've also added a link to a post that explains the mistake from a spreadsheet jockey's perspective.

Anybody who's ever put together a complex spreadsheet in Excel knows exactly the spirit in which it was intended. We've all done what R&R did; we just didn't have lawmakers crafting economic policy based on our errors.
Posted by Goldy on April 17, 2013 at 6:28 PM · Report this
Will in Seattle 9
I have to agree with everyone that this was not Microsoft Excel's problem, but both a weighting formula that was badly chosen (one data point in NZ being weighted the same as many more in the UK) and a non-inclusion of the range for the NZ data in the first place.

Austerity doesn't work. Never has. The only austerity that works is austerity for rich people, since they tend to not spend and most of their capital flows out of the origin country elsewhere, providing little GDP to the source country.

Conclusion: Tax the Rich and hunt them down with drones when they hide assets.
Posted by Will in Seattle on April 17, 2013 at 6:35 PM · Report this
Maybe they borrowed the basic spreadsheet from the tobacco industry.
Posted by kinaidos on April 17, 2013 at 7:09 PM · Report this
Even before the spreadsheet error came to light, R-R was a highly disputed paper. They showed a correlation between low growth and high debt, but correlation doesn't prove causation, and there was good cause to believe that it was reverse causation - the low growth caused the high debt, not the other way around.

As Paul Krugman has been pointing out, the pro-austerity politicians who seized on this paper did so because it said what they wanted to hear, not because it actually established anything close to a definitive result. And now that the coding error has been revealed and the R-R result has completely collapsed... well, the pro-austerity crowd will keep on repeating the 90% thing over and over anyway, because they don't want to admit that they've been wrong about anything.
Posted by I have always been... east coaster on April 17, 2013 at 7:39 PM · Report this
rob! 12
The same warnings are being sounded about China's domestic debt (local governments and agencies which borrowed heavily from local banks to stave off the effects of the world economic crisis). It's different in the sense that there are not defaults in the manner of international borrowing—rather, as debts come due they are being routinely rolled over by their issuers, for now at least.

But because it's not international, the numbers are harder to get at. It's been considered to be at ~20% of GDP; however, Simon Rabinovitch of the Financial Times recently suggested it's more in the 50-80% range and rising toward the 90% highlighted by Reinhart & Rogoff's flawed analysis.

It's hard to see all the implications and possible consequences of that right now. Here's a link to a 5-minute audio story on All Things Considered; by tomorrow they should have added a Transcript link to the story page:…
Posted by rob! on April 17, 2013 at 9:29 PM · Report this
briantrice 13
Fine. I'll blame Microsoft Excel's design as a software designer if you all won't. Its very low conceptual ceiling makes it very difficult to audit the conceptual integrity of what people do with it. This makes it very easy to lie or mislead with Excel, whether it's to oneself or others, and hard to catch it.

Microsoft may not be legally culpable, but we don't have a legal regime that declares academic provenance to be auditable.
Posted by briantrice on April 17, 2013 at 10:06 PM · Report this
SchmuckyTheCat 14
I've been using spreadsheets since VisiCalc in the early 80s. There has never been spreadsheet software well designed to catch formula errors. Formulas are by design, every design, hidden behind the value in the cell.
Posted by SchmuckyTheCat on April 17, 2013 at 10:16 PM · Report this
I agree with @7. One bad math problem does not a global disaster make. It took a lot of assholes to accomplish the economy tank.
Posted by idaho on April 17, 2013 at 11:41 PM · Report this
rob! 16
Re: 12, the transcript is up already.
Posted by rob! on April 17, 2013 at 11:45 PM · Report this
Kinison 17
Hey Goldy, hope you dont own any Apple stock, its lost nearly 300 billion in value since November and continues to fall and dropped another 20 billion thanks to Cirrus Logic's report yesterday. While this is major news, worthy of your awesome tech knowledge, you choose to ignore that and focus only on Microsoft tabloid fluff.
Posted by Kinison on April 18, 2013 at 5:54 AM · Report this
Pope Peabrain 18
It's never been about deficit reduction anyways. This paper was the excuse they needed to protect the vast offshore wealth that's been hidden.
Posted by Pope Peabrain on April 18, 2013 at 6:06 AM · Report this
Shoulda used Matlab
Posted by wxPDX on April 18, 2013 at 8:12 AM · Report this
treacle 20
"No economy ever recovered from a downturn through austerity." - J. Stiglitz (...)
Posted by treacle on April 18, 2013 at 9:27 AM · Report this
briantrice 21
@14 I'm sorry, but there's a lot of evidence statistically against that notion:… (What We Know About Spreadsheet Errors)

The fact is, spreadsheets have an *information architecture*, and Microsoft's dominant Excel (via VisiCalc) provides the least conceptual architecture to allow people to identify the conceptual domain of a sum, say. It's really quite a simple notion that spreadsheet software often isn't just a grid, that it has some conceptual coherence. Apple's Numbers exhibits some basic design that prevents basic matrix notation errors (small, localized, finite table-sheets instead of a universal grid sheet).

Just because you're good at it, doesn't mean that it's inherently good.
Posted by briantrice on April 18, 2013 at 5:30 PM · Report this

Add a comment

Commenting on this item is available only to registered commenters.

All contents © Index Newspapers, LLC
1535 11th Ave (Third Floor), Seattle, WA 98122
Contact | Privacy Policy | Terms of Use | Takedown Policy