3 ms·

Because Reinhart and Rogoff was such a massive deal.

by ycitm 13y ago

kenjackson 13y ago
But the R&R error had almost nothing to do with Excel. The same mistake could have been made with R or pen and paper. They were careless in a way I wouldn't expect a high schooler to be careless. Not in some way that Excel did something unexpected. It would be like blaming C# because you did foreach over collection A, rather than collection B.
- ycitm 13y ago
  You're right, of course. The problem in the R&R case is not with Excel itself, but that it's being used with a lack of rigour. If it were being done in C# or R I would expect unit tests and so on. I'm not saying that would make such a mistake impossible, but programmers have processes and tools for a reason. In the OPs case the problem actually is with excel though.
tfigment 13y ago
Basically this alone may have significantly change the course of several European nations for the next decade because of policy made on the that paper. NPR's Planet Money team just released a podcast [1] that goes into the basics of that. Last year my company was having supply chain issues. We have complicated 3 tier supply chain for a critical component with odd shipping restrictions and mix of batch and continuous process and other items. We don't control the the vendors but pay on yield and know what material enters and leaves each supplier. We desperately needed insight into why we had delivery issues as the suppliers were not very forth coming understandably. I was tasked with writing an app that let analysts run dozens of scenarios and give allow them to tweak all aspects of our models to gain insight into yield and schedule. I could have written a python script or Java or C# app (since I dont have SCM or MES software) or I could write an Excel spreadsheet. In 2 weeks, I wrote a spreadsheet that modeled the basics including complicated recycling and exposed everything to analysts. We debottlenecked several logistics issues with that and kept the business running and that spreadsheet still is updated daily. I can guarantee writing an engine+ui+reports in any language and allow the level of flexibility Excel provides would have taken me 6 months or more. The SCM or MES software would ultimately serve all of companies better but who has $200K-$1M+ to spend on these things for every issue when $200 + 2 weeks can get you 80% of the way. [1] http://www.npr.org/blogs/money/2013/04/19/177999020/episode-357-how-much-should-we-trust-economics http://www.npr.org/blogs/money/2013/04/19/177999020/episode-...
- joe_the_user 13y ago
  Not really. The R&R paper was criticized from the start but it suited a mindset that was happy to use what ever was convenient. Many hack economic studies that suit the dominant agenda are published each month. By happenstance, I know that Kenneth Rogoff is pretty much a professional liar, having made a survey of his predictions before and after the bubble (I found a pre-bubble interview with him deriding a doomsayer and a post-bubble interview with him being a doomsayer and expressing anger at the people, such his pre-bubble incarnation, who said everything was great). But I'm sure there are many, many Kenneth Rogoffs in the economics field, since much of it involves validating existing policy. So the R&R paper probably influenced nothing, was just icing on a cake that was already baked. The discovery of the paper's errors, on the other hand, is more of an outre event. Unlike your average mediocre product, it just happened that this paper had these error that were so bad they couldn't be waved away. Well, these hacks are done, to be replaced by other hacks. The error discovery may force a short backtracking on austerity idiocy but I wouldn't count on it.

Because Reinhart and Rogoff was such a massive deal.

by ycitm 13y ago

Because Reinhart and Rogoff was such a massive deal.

But the R&R error had almost nothing to do with Excel. The same mistake could have been made with R or pen and paper. They were careless in a way I wouldn't expect a high schooler to be careless. Not in some way that Excel did something unexpected. It would be like blaming C# because you did foreach over collection A, rather than collection B.

ycitm 13y ago
You're right, of course. The problem in the R&R case is not with Excel itself, but that it's being used with a lack of rigour. If it were being done in C# or R I would expect unit tests and so on. I'm not saying that would make such a mistake impossible, but programmers have processes and tools for a reason. In the OPs case the problem actually is with excel though.

tfigment 13y ago

Basically this alone may have significantly change the course of several European nations for the next decade because of policy made on the that paper. NPR's Planet Money team just released a podcast [1] that goes into the basics of that. Last year my company was having supply chain issues. We have complicated 3 tier supply chain for a critical component with odd shipping restrictions and mix of batch and continuous process and other items. We don't control the the vendors but pay on yield and know what material enters and leaves each supplier. We desperately needed insight into why we had delivery issues as the suppliers were not very forth coming understandably. I was tasked with writing an app that let analysts run dozens of scenarios and give allow them to tweak all aspects of our models to gain insight into yield and schedule. I could have written a python script or Java or C# app (since I dont have SCM or MES software) or I could write an Excel spreadsheet. In 2 weeks, I wrote a spreadsheet that modeled the basics including complicated recycling and exposed everything to analysts. We debottlenecked several logistics issues with that and kept the business running and that spreadsheet still is updated daily. I can guarantee writing an engine+ui+reports in any language and allow the level of flexibility Excel provides would have taken me 6 months or more. The SCM or MES software would ultimately serve all of companies better but who has $200K-$1M+ to spend on these things for every issue when $200 + 2 weeks can get you 80% of the way. [1] http://www.npr.org/blogs/money/2013/04/19/177999020/episode-357-how-much-should-we-trust-economics http://www.npr.org/blogs/money/2013/04/19/177999020/episode-...

joe_the_user 13y ago
Not really. The R&R paper was criticized from the start but it suited a mindset that was happy to use what ever was convenient. Many hack economic studies that suit the dominant agenda are published each month. By happenstance, I know that Kenneth Rogoff is pretty much a professional liar, having made a survey of his predictions before and after the bubble (I found a pre-bubble interview with him deriding a doomsayer and a post-bubble interview with him being a doomsayer and expressing anger at the people, such his pre-bubble incarnation, who said everything was great). But I'm sure there are many, many Kenneth Rogoffs in the economics field, since much of it involves validating existing policy. So the R&R paper probably influenced nothing, was just icing on a cake that was already baked. The discovery of the paper's errors, on the other hand, is more of an outre event. Unlike your average mediocre product, it just happened that this paper had these error that were so bad they couldn't be waved away. Well, these hacks are done, to be replaced by other hacks. The error discovery may force a short backtracking on austerity idiocy but I wouldn't count on it.