Welcome to DU! The truly grassroots left-of-center political community where regular people, not algorithms, drive the discussions and set the standards. Join the community: Create a free account Support DU (and get rid of ads!): Become a Star Member Latest Breaking News General Discussion The DU Lounge All Forums Issue Forums Culture Forums Alliance Forums Region Forums Support Forums Help & Search

jazzimov

(1,456 posts)
Sun Apr 21, 2013, 05:02 PM Apr 2013

Austerity Excel error - was it on purpose?

The most influential "research" that Austerity proponents most commonly quote (Reinhart-Rogoff) has recently been found to contain a rookie Excel error. Under further scrutiny, more of the data has been found to have been cherry-picked.

But what about the spreadsheet itself? From what I've read (which may be incorrect, granted) Reinhart-Rogoff did not allow expert economists to see the spreadsheet upon which their report was based. This makes me suspicious. However, they did let a "rookie" economist access to the original spreadsheet. This rookie found the rookie spreadsheet error - when Reinhart-Rogoff did their averages, they left out 5 rows in their equation - which changed the entire results, and essentially destroyed their conclusions.

If I were Reinhart-Rogoff, I would have double- and triple-checked my work before I published. Granted, I would have focused on more "intellectual" aspects rather than simple Excel equations - but I would have checked those as well. The thing about Excel is that it is so user-friendly that all you have to do is click on the formula cell and it will OUTLINE the cells involved, in color-coded outlines!

So, it's not really a matter of reading the formula, but simply clicking on the formula cell and then looking at the spreadsheet and VOILA! the error would be frickin' OBVIOUS!

Further, this begs the question "why were the game-changing examples congregated at the bottom of the rows, so that they could so easily be excluded"?

so WHY was a "rookie" given access to find such a rookie mistake? Especially when it was denied to those considered more "expert" in the field? Perhaps they thought he would be over-whelmed by the data? Perhaps they thought he could be easily discredited - although such an obvious rookie error could hardly be discredited.

Perhaps they wanted to be caught. After all, they had whatever funding they wanted - although surely they knew that after the mistake was found their reputations would be ruined and the funding would dry up.

But does it really matter? The fact is that Austerity has been discredited - both in Models and in Reality.

27 replies = new reply since forum marked as read
Highlight: NoneDon't highlight anything 5 newestHighlight 5 most recent replies
Austerity Excel error - was it on purpose? (Original Post) jazzimov Apr 2013 OP
I use Excel all the time daleo Apr 2013 #1
Agreed. And certainly not when sending out the results to the world. JanMichael Apr 2013 #10
Summary more than an excel error FogerRox Apr 2013 #26
It's got to reek of incompetence, if not conspiracy as well Demeter Apr 2013 #2
They probably thought he was a loyal rookie Warpy Apr 2013 #3
There wasn't that much data to crunch - it's an embarrassing error, to be honest. reformist2 Apr 2013 #4
Measure twice, cut once. dipsydoodle Apr 2013 #5
The error was letting anyone see the spreadsheet jmowreader Apr 2013 #6
Precisely . . . markpkessinger Apr 2013 #23
Yes, it's a mistake Lithos Apr 2013 #7
Exactly. I often ask for a "second set of eyes" jazzimov Apr 2013 #8
No, their error was meainingless, the study is still valid, nothing changes ... Scuba Apr 2013 #9
heehee truebluegreen Apr 2013 #16
If It Were Just the Range Error, On the Road Apr 2013 #11
The first discovered "error" Ruby the Liberal Apr 2013 #13
comment @ DK summarizes FogerRox Apr 2013 #25
Excellent summary Ruby the Liberal Apr 2013 #27
"They" didn't publish this, grad students did Ruby the Liberal Apr 2013 #12
Oh, yes, I am loving Dr. Krugman just for being there and doing what he does... CTyankee Apr 2013 #17
I just can't wait for the finger pointing to begin Ruby the Liberal Apr 2013 #18
if it were not a mistake would they just have given the guy their raw data La Lioness Priyanka Apr 2013 #14
I'd say it was on purpose. truebluegreen Apr 2013 #15
it wa$ a huge mi$take. nt killbotfactory Apr 2013 #19
Once of the things I learned about writing journals and lab reports when I was in college Victor_c3 Apr 2013 #20
The weighting of the bins was much more important than the missing rows Recursion Apr 2013 #21
Has anybody found an error yet that did not support Reinhart/Rogoff's point? pa28 Apr 2013 #22
Of course it was on purpose. FogerRox Apr 2013 #24

JanMichael

(24,881 posts)
10. Agreed. And certainly not when sending out the results to the world.
Sun Apr 21, 2013, 06:57 PM
Apr 2013

I always back check my own calculations and when it involves serious issues I send he results to people I know that are statistical gods compared to me. That way a simple stupid error never gets loose.

That is why I think they are either bald faced liars OR ignorant fools and tools that should send their degrees back in shame to their alma maters as they are obvious frauds.

FogerRox

(13,211 posts)
26. Summary more than an excel error
Mon Apr 22, 2013, 02:29 PM
Apr 2013

The kicker with R&R is that they didn't use bad (0+ / 0-)

data. Their data was fine for their purposes. But it's what they did to the data that made them wrong.
They combined results from countries that had fiat currencies with countries that do not control their own money supply.
They used an arbitrary and bizarre weighting system that weighted small countries that confirmed theri conclusions much more heavily than large countries which didn't.
They ignored the "correlation doesn't equal causality" caveat and worse, they assumed an illogical causal chain:
High debt to GDP ratios => economic recession rather than the more logical economic recession => greater debt.
THEN they made the infamous "Excel coding" which just happened to leave out data that would have affected their conclusion.

by dallasdunlap on Mon Apr 22, 2013 at 10:40:37 AM EDT

 

Demeter

(85,373 posts)
2. It's got to reek of incompetence, if not conspiracy as well
Sun Apr 21, 2013, 05:06 PM
Apr 2013

But it does explain a lot...the belief in fairies, etc., which dominates modern economic "theory" and ideology.

Warpy

(111,222 posts)
3. They probably thought he was a loyal rookie
Sun Apr 21, 2013, 05:16 PM
Apr 2013

and yes, it smacks of being entirely on purpose since a lot of their other data appears to have been cherry picked.

I am delighted he wasn't as loyal as they thought he was.

jmowreader

(50,546 posts)
6. The error was letting anyone see the spreadsheet
Sun Apr 21, 2013, 05:29 PM
Apr 2013

You don't need a spreadsheet to prove austerity doesn't work; all you need do is look at history. There is not one case in history where applying austerity measures didn't reduce the economy. If you are looking at too energetic of an economy, short periods of austerity can throttle it back to rational levels. Applying austerity to a contracted economy makes it worse.

Now, austerity has a purpose for the rich. If you are a developer who wants to build homes in a national park, you might want to get the congressman you own to throw on some austerity. When austerity then ruins the economy, suggest the government could make money by selling federal land.(This is happening in the West.)

Real life case study: Greece. Its problems have three origins. First is near-universal corruption. Second is near-universal tax evasion.(When people refer to tax evasion as the national sport you know you're screwed.) And the last is austerity. Corruption and evasion got them to the point where they tried austerity...and when it didn't work they tried it again... The ultimate cure for Greece's woes would be to get rid of austerity, deport every politician in the country to an island (think "Escape From New York&quot and replace them with Harvard poli sci students, and ship over every Greek-speaking IRS auditor we have. But they won't do that.

markpkessinger

(8,392 posts)
23. Precisely . . .
Sun Apr 21, 2013, 11:28 PM
Apr 2013

. . . based on the economic history of our own and many other countries, this tudy should have been met with considerable skepticism at the outset. I mean, we had a considerably higher debt-to-GDP ratio at the end of WWII, but it surely doesn't appear to have stood in the way of a huge economic expansion in the pot-war era. And under NO circumstances should governments be basing economic policy decisions upon the results of a single, unreplicated study. My hypothesis is that the 0.1% accepted its findings, no questions asked, because it fit rather nicely with their own agenda of keeping their taxes obscenely low.

Lithos

(26,403 posts)
7. Yes, it's a mistake
Sun Apr 21, 2013, 05:48 PM
Apr 2013

One born out by people seeing what they want to see. It is harder to see fault in things you want to see.

However, says a lot in their professionalism and academic credentials that they did not open up their model for full peer review prior to publication.

L-

jazzimov

(1,456 posts)
8. Exactly. I often ask for a "second set of eyes"
Sun Apr 21, 2013, 06:41 PM
Apr 2013

because I know from experience that if you overlook an error once, you'll tend to overlook it every time.

Which is why peer review is so important.

Unless you're trying to hide something.

 

Scuba

(53,475 posts)
9. No, their error was meainingless, the study is still valid, nothing changes ...
Sun Apr 21, 2013, 06:51 PM
Apr 2013

... and it was Acorn that rigged the election for Obama, the Koch brothers are philanthropists (they gave millions to the arts!!!), Paul Ryan is a "numbers wonk" and your "proof" will be lost in a sea of bought-and-paid-for media pundits crying that only austerity can save us!!!

The fact this flawed study was used by so many to steal so much without any peer review is prima facie evidence of the power of the oligarchs.

On the Road

(20,783 posts)
11. If It Were Just the Range Error,
Sun Apr 21, 2013, 06:57 PM
Apr 2013

there would be a possibility of a genuine mistake. This is one of the most common errors with the Sum function in Excel.

What convinces me it was deliberate is the funky method of averaging years, which weights a long stable period in a particular "band" as equivalent to a single year in that band. This is not something anyone with a grasp of high school math would think is an accurate way of averagaing. And it just *happens* to reinforce the authors' point.

The errors in the Reinhart-Rogoff study appear to me to be a result of attempting to arrive at a number at all costs. It happens -- which is why some type of peer review is so valuable.

Ruby the Liberal

(26,219 posts)
13. The first discovered "error"
Sun Apr 21, 2013, 07:39 PM
Apr 2013

specifically excluded nations who did not suffer economic downturns during times of high debt/budget deficits.

Which was the whole POINT of their study. It was justification for austerity - and has been cited as proof-positive, A-grade, un-challengeable evidence for neo-cons for eons.

And. They. Lied.

FogerRox

(13,211 posts)
25. comment @ DK summarizes
Mon Apr 22, 2013, 02:28 PM
Apr 2013

The kicker with R&R is that they didn't use bad (0+ / 0-)

data. Their data was fine for their purposes. But it's what they did to the data that made them wrong.
They combined results from countries that had fiat currencies with countries that do not control their own money supply.
They used an arbitrary and bizarre weighting system that weighted small countries that confirmed theri conclusions much more heavily than large countries which didn't.
They ignored the "correlation doesn't equal causality" caveat and worse, they assumed an illogical causal chain:
High debt to GDP ratios => economic recession rather than the more logical economic recession => greater debt.
THEN they made the infamous "Excel coding" which just happened to leave out data that would have affected their conclusion.

by dallasdunlap on Mon Apr 22, 2013 at 10:40:37 AM EDT

Ruby the Liberal

(26,219 posts)
12. "They" didn't publish this, grad students did
Sun Apr 21, 2013, 07:36 PM
Apr 2013

who likely didn't feel empowered to double check the work of the high and mighty - they just typed up what they were told to type up.

I am loving that Dr Krugman is having such a field day with this.

CTyankee

(63,900 posts)
17. Oh, yes, I am loving Dr. Krugman just for being there and doing what he does...
Sun Apr 21, 2013, 09:25 PM
Apr 2013

bless him...glad we have him in our time of need...

Ruby the Liberal

(26,219 posts)
18. I just can't wait for the finger pointing to begin
Sun Apr 21, 2013, 09:34 PM
Apr 2013

on whose "job" it was to peer review this disaster.

Vindication.

 

truebluegreen

(9,033 posts)
15. I'd say it was on purpose.
Sun Apr 21, 2013, 08:01 PM
Apr 2013

These cretins don't give a fig about the truth, they only care about their own ideology. At best, they made some boneheaded errors but disregarded them because the end result fit their pre-conceived ideas.

Victor_c3

(3,557 posts)
20. Once of the things I learned about writing journals and lab reports when I was in college
Sun Apr 21, 2013, 09:42 PM
Apr 2013

is make your data match your results!

You know what you are supposed to get and, unless you want to do it all over again or get a crappy grade, fudge the numbers you got from your chemistry lab to make it work out like it is supposed to!

So yup, it was done on purpose.

Recursion

(56,582 posts)
21. The weighting of the bins was much more important than the missing rows
Sun Apr 21, 2013, 10:35 PM
Apr 2013

I could see the rows being an actual error (possibly), but the way they weighted bins, so that 20 years of high growth/high debt in the UK is balanced by 1 year of low growth/high debt in NZ, was absolutely deliberate.

pa28

(6,145 posts)
22. Has anybody found an error yet that did not support Reinhart/Rogoff's point?
Sun Apr 21, 2013, 11:18 PM
Apr 2013

I can't think of a better example (or a more effective one) of trading on credentials to manufacture agreement for your own conclusion.

This whole episode really helps explain the unbelievable smugness of decision makers and pundits who all seemed to agree that austerity was the only serious answer. Anybody who happened to disagree was considered to be something beyond just marginal. I think the best word would be ignorant.

Ignorant because they did not have the "numbers" or the "facts" and any alternative solution they might have was not even worthy of discussion.

We had a purposeful deception here and moving forward just remember there are dozens of lavishly funded think tanks out there who are perfectly willing to parse and bend data in the same way to support the same flawed conclusions.

Latest Discussions»General Discussion»Austerity Excel error - w...