Home Mail MemberMap Chat (0) Wallpapers
Go Back   ChiefsPlanet > The Ed & Dave Lounge > D.C.

Reply
 
Thread Tools Display Modes
Old 04-17-2013, 06:18 AM  
Loneiguana Loneiguana is offline
Veteran
 
Loneiguana's Avatar
 
Join Date: Nov 2012
Location: Springfield
Casino cash: $5915
That's not a mundane detail, Michael

Influential Reinhart-Rogoff Pro-Austerity Research Riddled With Errors: Study

Influential research by U.S. economists Carmen Reinhart and Ken Rogoff, touted by policymakers pushing government austerity in the United States and Europe, is riddled with errors, a bombshell new academic study claims.

The findings may not have much impact on the debate over government debt, and it probably won't cause those who have spent the past several decades panicking over government debt to stop their panicking. But it seriously erodes the intellectual underpinnings of the pro-austerity argument -- and makes the damage done by austerity in Europe and the U.S. in recent years all the more poignant.

"This is a mistake that has had enormous consequences," wrote Dean Baker of the Center for Economic and Policy Research. "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."

The new paper, by Thomas Herndon, Michael Ash, and Robert Pollin of the University of Massachusetts, Amherst, set out to reconstruct the findings of an influential 2010 paper by Reinhart and Rogoff, called "Growth In A Time Of Debt." Reinhart and Rogoff, of the University of Maryland and Harvard, respectively, claimed that economic growth slowed fairly dramatically for countries whose public debt crossed a threshold of 90 percent of gross domestic product.

The problem is that other economists have been unable to recreate Reinhart and Rogoff's findings. Herndon, Ash and Pollin now say they were able to do so -- but only by leaving out big, important pieces of data.


Using the same spreadsheet that Reinhart and Rogoff used for their research, Herndon, Ash and Pollin found that "Growth In A Time Of Debt" was built around a handful of significant errors. Correcting for those errors changes the findings dramatically: Average GDP growth for high-debt countries jumps from negative 0.1 percent to 2.2 percent.

The most important error appears to be a failure to include years of data that showed Australia, Canada and New Zealand enjoying high economic growth and high debt at the same time. Including all the years of data boosts New Zealand's average economic growth rate under high debt to 2.58 percent, from negative 7.6 percent. Given the small amount of data used in Reinhart and Rogoff's study, this has a huge impact on the overall findings.


Another error seems to be a simple failure to use an Excel spreadsheet correctly, as highlighted by economist Mike Konczal at the Roosevelt Institute's Next New Deal blog. In building a formula to calculate average economic growth rates, Reinhart and Rogoff appeared to leave off several lines of data in their spreadsheet.

"We literally just received this draft comment, and will review it in due course," Reinhart and Rogoff wrote in a long and detailed emailed statement Tuesday afternoon. "Nevertheless, the weight of the evidence to date -- including this latest comment -- seems entirely consistent with our original interpretation of the data." (Read the whole statement here.)

Update: At about 2:00 a.m. ET on Wednesday, Reinhart and Rogoff emailed a fuller statement defending their work, the full text of which you can read here. They admit to the spreadsheet error, but defend their weighting of data and say they have corrected for some of their omissions in subsequent work. They stand by the gist of their conclusion, that higher debt is associated with lower growth.

Even before the errors cited in the new study came to light, many economists doubted Reinhart and Rogoff's conclusion that high debt causes low growth, given the glaring chicken-and-egg problem at the heart of the research. Did these countries have slow growth because they had high debt, or did they have high debt because they had slow growth?

(Reinhart and Rogoff noted in their Tuesday statement that they have been careful not to claim that high debt causes slow growth, but rather that it has an "association" with slow growth.)

Beyond that, Baker notes, there were lots of other reasons to question Reinhart and Rogoff, including the fact that their gloomy conclusions about debt relied heavily on slow U.S. economic growth immediately after World War II. At the time, the U.S. was deep in war debt and dismantling its war machine. That relatively brief state of affairs was quickly followed by arguably the greatest economic boom in history.

Despite these questions, Reinhart and Rogoff's 90-percent threshold has been discussed ad nauseum in the press and used frequently to justify austerity measures in the U.S. and Europe, as detailed by Quartz's Tim Fernholz. The 2012 version of the pro-austerity budget plan of Rep. Paul Ryan (R-Wis.) cites Reinhart and Rogoff by name, and specifically refers to the 90-percent threshold.

Washington's constant state of panic over government debt and budget deficits has contributed to severe cutbacks in government spending that have slowed economic growth and helped keep unemployment high. The situation has been even worse in Europe, particularly in troubled nations like Greece, where austerity has been enforced as a bailout condition, only to result in slower growth and higher debt burdens.

Still, that evidence has been in sight all along, yet the pro-austerity crowd has kept up its drumbeat of deficit panic. This new research probably won't change that.

"There are other threshold papers out there," The Economist's Ryan Avent tweeted. "And people hate to change their minds."



http://www.huffingtonpost.com/2013/0...n_3094015.html

http://www.slate.com/blogs/moneybox/...ad_coding.html

Posts: 3,884
Loneiguana would the whole thing.Loneiguana would the whole thing.Loneiguana would the whole thing.Loneiguana would the whole thing.Loneiguana would the whole thing.Loneiguana would the whole thing.Loneiguana would the whole thing.Loneiguana would the whole thing.Loneiguana would the whole thing.Loneiguana would the whole thing.Loneiguana would the whole thing.
  Reply With Quote
Old 04-17-2013, 06:15 PM   #2
DanT DanT is online now
Veteran
 

Join Date: Sep 2000
Location: Davis, CA
Casino cash: $5670
Very interesting. I've seen mistakes like that get made in Excel all the time. When I get Excel data from colleagues, I usually will import it into SAS, Stata or R and redo the analysis there. But even there bugs can result and lead to wrong results.


Reinhart and Rogoff were good to make their data available to the graduate student who found these mistakes. I also appreciate that they acknowledge that they mistakenly excluded five countries and I would assume that was a completely innocent mistake. The weighting issue is a serious one. I would prefer the country-year weighted average that the UMass study used ( http://www.peri.umass.edu/236/hash/3...blication/566/), not the country weighted average that Reinhart and Rogoff used, but I understand that there are advantages and disadvantages to either approach. I do think that the UMass scholars are correct, though, in describing the Reinhart and Rogoff approach as unconventional. If you are trying to figure out the average batting average for a baseball club, most folks would weight the individual player's batting average by the number of at bats from that player. It would be kind of weird to just take the simple average of all of the individual player averages, which weights pitchers and backup players exactly the same as regular starters.

The revised number from the UMass team is more than 2 percentage points higher in magnitude than Reinhart and Rogoff's number and puts it about 1 percentage point away from the average annual GDP growth seen in countries with lower levels of debt. That 1 percentage point gap may not, in fact, even be statistically significant. It's a helluva lot closer to the annual GDP growth in the lower-debt countries than it is to the original number that Reinhart and Rogoff report. I disagree with Reinhart and Rogoff's comments this week that the revised analysis still supports the claim that average GDP growth rates tend to be lower when debt levels are relatively high. They don't have convincing evidence for that claim and their flawed numbers were the major ones that had been used as evidence by policymakers favoring austerity. More fundamentally, Reinhart and Rokoff's claim in only about a correlation, not causality. They acknowledge as much but many people try to argue that there's a causal link between high debt levels and lower GDP growth, despite having no convincing empirical evidence of causality.


Also, the fact that studies are often not reproducible is a very serious problem that is starting to draw a lot of attention. A characteristic of good science is that it be reproducible, but there are lots of examples where studies are not reproducible, either because of innocent problems or outright fraud. In another posting today, I tried to reproduce a number in a study and was unable to do so ( http://chiefsplanet.com/BB/showpost....6&postcount=80 ) . The irreproducibility can result from incorrect usage and interpretation of statistical methods, but it can also result from bad data management practices or sloppy programming. http://www.peri.umass.edu/236/hash/3...6388b1/publica
tion/566/

Last edited by DanT; 04-17-2013 at 06:50 PM..
Posts: 3,064
DanT must have mowed badgirl's lawn.DanT must have mowed badgirl's lawn.DanT must have mowed badgirl's lawn.DanT must have mowed badgirl's lawn.DanT must have mowed badgirl's lawn.DanT must have mowed badgirl's lawn.DanT must have mowed badgirl's lawn.DanT must have mowed badgirl's lawn.DanT must have mowed badgirl's lawn.DanT must have mowed badgirl's lawn.DanT must have mowed badgirl's lawn.
  Reply With Quote
Old 04-17-2013, 06:33 PM   #3
DanT DanT is online now
Veteran
 

Join Date: Sep 2000
Location: Davis, CA
Casino cash: $5670
Some useful links:

To find the UMass study, go here:

http://www.peri.umass.edu/236/hash/3...blication/566/

To find a response by Reinhart-Rogoff, go here:
http://blogs.wsj.com/economics/2013/...e-to-critique/
Posts: 3,064
DanT must have mowed badgirl's lawn.DanT must have mowed badgirl's lawn.DanT must have mowed badgirl's lawn.DanT must have mowed badgirl's lawn.DanT must have mowed badgirl's lawn.DanT must have mowed badgirl's lawn.DanT must have mowed badgirl's lawn.DanT must have mowed badgirl's lawn.DanT must have mowed badgirl's lawn.DanT must have mowed badgirl's lawn.DanT must have mowed badgirl's lawn.
  Reply With Quote
Old 04-17-2013, 07:17 PM   #4
FD FD is offline
Veteran
 
FD's Avatar
 

Join Date: Dec 2006
Location: Los Angeles
Casino cash: $5778
Their result was irrelevent yesterday and its irrelevent today, the fact that it was wrong doesn't change anything. The fact that some proponents of austerity ever cited it shows how weak their case is.
__________________
Homer: [looking at watch] Two hours? Why'd they build this ghost town so far away?
Lisa: Because they discovered gold over there!
Homer: It's because they're stupid, that's why. That's why everybody does everything.
Posts: 3,553
FD 's adopt a chief was Sabby PiscitelliFD 's adopt a chief was Sabby PiscitelliFD 's adopt a chief was Sabby PiscitelliFD 's adopt a chief was Sabby PiscitelliFD 's adopt a chief was Sabby PiscitelliFD 's adopt a chief was Sabby PiscitelliFD 's adopt a chief was Sabby PiscitelliFD 's adopt a chief was Sabby PiscitelliFD 's adopt a chief was Sabby PiscitelliFD 's adopt a chief was Sabby PiscitelliFD 's adopt a chief was Sabby Piscitelli
  Reply With Quote
Old 04-18-2013, 05:57 AM   #5
Loneiguana Loneiguana is offline
Veteran
 
Loneiguana's Avatar
 

Join Date: Nov 2012
Location: Springfield
Casino cash: $5915
Quote:
Originally Posted by DanT View Post
Very interesting. I've seen mistakes like that get made in Excel all the time. When I get Excel data from colleagues, I usually will import it into SAS, Stata or R and redo the analysis there. But even there bugs can result and lead to wrong results.


Reinhart and Rogoff were good to make their data available to the graduate student who found these mistakes. I also appreciate that they acknowledge that they mistakenly excluded five countries and I would assume that was a completely innocent mistake. The weighting issue is a serious one. I would prefer the country-year weighted average that the UMass study used ( http://www.peri.umass.edu/236/hash/3...blication/566/), not the country weighted average that Reinhart and Rogoff used, but I understand that there are advantages and disadvantages to either approach. I do think that the UMass scholars are correct, though, in describing the Reinhart and Rogoff approach as unconventional. If you are trying to figure out the average batting average for a baseball club, most folks would weight the individual player's batting average by the number of at bats from that player. It would be kind of weird to just take the simple average of all of the individual player averages, which weights pitchers and backup players exactly the same as regular starters.

The revised number from the UMass team is more than 2 percentage points higher in magnitude than Reinhart and Rogoff's number and puts it about 1 percentage point away from the average annual GDP growth seen in countries with lower levels of debt. That 1 percentage point gap may not, in fact, even be statistically significant. It's a helluva lot closer to the annual GDP growth in the lower-debt countries than it is to the original number that Reinhart and Rogoff report. I disagree with Reinhart and Rogoff's comments this week that the revised analysis still supports the claim that average GDP growth rates tend to be lower when debt levels are relatively high. They don't have convincing evidence for that claim and their flawed numbers were the major ones that had been used as evidence by policymakers favoring austerity. More fundamentally, Reinhart and Rokoff's claim in only about a correlation, not causality. They acknowledge as much but many people try to argue that there's a causal link between high debt levels and lower GDP growth, despite having no convincing empirical evidence of causality.


Also, the fact that studies are often not reproducible is a very serious problem that is starting to draw a lot of attention. A characteristic of good science is that it be reproducible, but there are lots of examples where studies are not reproducible, either because of innocent problems or outright fraud. In another posting today, I tried to reproduce a number in a study and was unable to do so ( http://chiefsplanet.com/BB/showpost....6&postcount=80 ) . The irreproducibility can result from incorrect usage and interpretation of statistical methods, but it can also result from bad data management practices or sloppy programming. http://www.peri.umass.edu/236/hash/3...6388b1/publica
tion/566/
Very well said, thank you for the comment.
Posts: 3,884
Loneiguana would the whole thing.Loneiguana would the whole thing.Loneiguana would the whole thing.Loneiguana would the whole thing.Loneiguana would the whole thing.Loneiguana would the whole thing.Loneiguana would the whole thing.Loneiguana would the whole thing.Loneiguana would the whole thing.Loneiguana would the whole thing.Loneiguana would the whole thing.
  Reply With Quote
Old 04-18-2013, 05:58 AM   #6
Loneiguana Loneiguana is offline
Veteran
 
Loneiguana's Avatar
 

Join Date: Nov 2012
Location: Springfield
Casino cash: $5915
Quote:
Originally Posted by FD View Post
Their result was irrelevent yesterday and its irrelevent today, the fact that it was wrong doesn't change anything. The fact that some proponents of austerity ever cited it shows how weak their case is.

Like Paul Ryan and his budget.
Posts: 3,884
Loneiguana would the whole thing.Loneiguana would the whole thing.Loneiguana would the whole thing.Loneiguana would the whole thing.Loneiguana would the whole thing.Loneiguana would the whole thing.Loneiguana would the whole thing.Loneiguana would the whole thing.Loneiguana would the whole thing.Loneiguana would the whole thing.Loneiguana would the whole thing.
  Reply With Quote
Reply

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is On

Forum Jump




All times are GMT -6. The time now is 03:54 PM.


Powered by vBulletin® Version 3.8.0
Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.