Post #371,169
2/10/13 1:01:56 PM
|
![New](/static/images/lrpd.gif)
Kwak: How Excel blew up the London Whale.
http://baselinescena...ortance-of-excel/
The issue is described in the appendix to JPMorganÂs internal investigative task forceÂs report. To summarize: JPMorganÂs Chief Investment Office needed a new value-at-risk (VaR) model for the synthetic credit portfolio (the one that blew up) and assigned a quantitative whiz (Âa London-based quantitative expert, mathematician and model developer who previously worked at a company that built analytical models) to create it. The new model Âoperated through a series of Excel spreadsheets, which had to be completed manually, by a process of copying and pasting data from one spreadsheet to another. The internal Model Review Group identified this problem as well as a few others, but approved the model, while saying that it should be automated and another significant flaw should be fixed.** After the London Whale trade blew up, the Model Review Group discovered that the model had not been automated and found several other errors. Most spectacularly,
ÂAfter subtracting the old rate from the new rate, the spreadsheet divided by their sum instead of their average, as the modeler had intended. This error likely had the effect of muting volatility by a factor of two and of lowering the VaR . . .Â
I write periodically about the perils of bad software in the business world in general and the financial industry in particular, by which I usually mean back-end enterprise software that is poorly designed, insufficiently tested, and dangerously error-prone. But this is something different.
<boggle>
(via Economist's View - http://economistsvie...r-02-10-2013.html )
Cheers,
Scott.
|
Post #371,170
2/10/13 1:22:03 PM
|
![New](/static/images/lrpd.gif)
Not surprised ... and yes, I worked at a bank
--
Drew
|
Post #371,901
2/27/13 4:33:34 PM
|
![New](/static/images/lrpd.gif)
No boggle at all
I've spent years automating exactly the type of work described.
I sit next to a person who does their job, keying all day, from a variety of sources.
They have their own mental business rules that have to be documented. All those cut&pastes of rows of spreadsheet columns, and the occasional cell overwrite killing you happened all the time.
They hated me, know I will be automating the task they are doing, so they always exaggerated the difficulty, hoping I will give up.
No dice.
So they do it very quickly, skipping over and redoing for mistakes that I couldn't catch, then abuse me for asking them to explain what just happened. I was just too slow or stupid to handle it, and obviously when they complained to their boss I'd stop distracting them from their job.
I have sat through this performance with DOZENS of people through my career.
Most of the time I give up on THEM. I document what they do from and input/output blackbox perspective, and replace it all from scratch.
I ALWAYS found errors in current implementation of business rules, either in that they produced obvious erroneous results, or they didn't follow the spec. If I was to tell the customer (always working on a customer system), then we'd have to refund a shitload of money. So we didn't.
But if I changed the system to be "correct" in my version of it, then the next month's production run would vary too much from the previous, setup red flags, and then be investigated by the customer.
So replacement systems would be bug-for-bug compliant until I could slide the fixes in without being noticed.
|
Post #371,912
2/27/13 8:30:23 PM
|
![New](/static/images/lrpd.gif)
The scale is the boggle.
What boggled me was the fact that an office of JP Morgan was throwing around hundreds of millions or billions of dollars based on simple risk calculations in a spreadsheet that was not checked for accuracy.
It shows that management didn't care how Iksil made his money as long as he did so. IOW, management wasn't doing its job but was sitting back and collecting all the rewards.
Yeah, that's not really surprising. But still...
Cheers,
Scott.
|
Post #371,918
2/27/13 9:13:40 PM
|
![New](/static/images/lrpd.gif)
nah, largest phone company in the US
tracks all of its equipment on excel spreadsheets by lowest cost contractors who are not really skilled and don't really care. Now guess what software is building derivative tranches? Betcha that shit is in there somewhere
Any opinions expressed by me are mine alone, posted from my home computer, on my own time as a free American and do not reflect the opinions of any person or company that I have had professional relations with in the past 57 years. meep
|
Post #371,930
2/28/13 6:56:42 AM
|
![New](/static/images/lrpd.gif)
Pocket change
|
Post #371,949
2/28/13 10:20:15 AM
|
![New](/static/images/lrpd.gif)
Got out just ahead of that
I was working for a company that did mortgage paperwork. We picked up Ameriquest's appraisal business, more than doubling our income. (I think it was more like tripling. It was a lot.)
I left the company soon after. Now they're not doing so well. C'est la vie.
--
Drew
|
Post #371,961
2/28/13 1:24:56 PM
2/28/13 1:38:47 PM
|
![New](/static/images/lrpd.gif)
I was tossed out about 2 years before the peak
The 1st year the windows centric project manager tried to replace Oracle/Perl/dozens of linux box parallel processing and my hand picked project successor that I had been training and then working with for years, and gave it to a Windows buddy who attempted to rewrite it for a SQL Server/VB environment with a variety of reporting tools tacked on.
It had taken me 6 months to sell/plan it, once sold: 3 months to create the infrastructure (buy big solaris/little linux boxes, Gbit/FC/large disk), while coding dozens of variations of possible production methods using Perl and Oracle and SyncSort and (whatever), created a full month's worth of processing and then optimized it down to a week, threw a chair at my boss when he told me to work SMARTER (thanks towelboy), and then babysat it for 2 month's of productions, then trained my mentee and moved onto my next project, occasionally being called in for design/code meeting for addons.
It ran for 3 YEARS and was the largest amount of profit that company had ever seen, at the cost of 20% of a perl/unix programmer and 50% of a clerical assistant/production person. Paid for the lights and rent and a lot of celebrations (of which I was not part of).
It took him about a year before he turned it on. In that process he alienated the current staff on purpose, most left and/or were fired, and there were multiple failed monthly productions, in which the company had to write checks to ameriquest in the $300,000 range.
After he claimed it was all my system's fault, and then turned his on, they lost the account within a couple of months and destroyed the database division of the company.
Poof.
![Expand](/static/images/plus.png)
Edited by crazy
Feb. 28, 2013, 01:38:47 PM EST
|
Post #373,899
4/16/13 9:34:23 PM
4/16/13 9:50:51 PM
|
![New](/static/images/lrpd.gif)
Excel error blows up the world.
http://www.motherjon...or-destroys-world
From Mike Konczal, summarizing a new study that says Carmen Reinhart and Ken Rogoff made a coding error in a famous paper claiming that economic growth slows down in countries with debt levels above 90 percent of GDP:
[...]
PK has more - http://krugman.blogs...ing-error-batman/ and http://krugman.blogs...rogoff-continued/
<sigh>
:-(
[edit:] Ryan McCarthy at Felix's blog has a roundup - http://blogs.reuters...ties-beef-rogoff/
Cheers,
Scott.
|
Post #373,911
4/17/13 10:07:46 AM
|
![New](/static/images/lrpd.gif)
interesting
you have 90% of your financial income garnished by debt, so by working hard and servicing that debt while paying living expenses and saving what is left you are much better off than a guy who only has 50% debt, paying living expenses and saving what is left. Good to know
Any opinions expressed by me are mine alone, posted from my home computer, on my own time as a free American and do not reflect the opinions of any person or company that I have had professional relations with in the past 58 years. meep
|
Post #373,920
4/17/13 12:28:05 PM
|
![New](/static/images/lrpd.gif)
Meh.
The national debt is like a 50 year mortgage at a low interest rate. It's the (low) cost of having a modern society.
Nobody is going to "call" the debt and demand that it all be paid at once.
That's also why people don't become bankrupt when they take out a 30 year mortgage that is several times their annual salary.
HTH.
Cheers,
Scott.
|
Post #373,922
4/17/13 12:31:05 PM
|
![New](/static/images/lrpd.gif)
they do go bankrupt when the debt service is 90% of their
salary
Any opinions expressed by me are mine alone, posted from my home computer, on my own time as a free American and do not reflect the opinions of any person or company that I have had professional relations with in the past 58 years. meep
|
Post #373,923
4/17/13 1:16:39 PM
|
![New](/static/images/lrpd.gif)
And where has that happened?
|
Post #373,924
4/17/13 1:28:31 PM
|
![New](/static/images/lrpd.gif)
Re: And where has that happened?
Any opinions expressed by me are mine alone, posted from my home computer, on my own time as a free American and do not reflect the opinions of any person or company that I have had professional relations with in the past 58 years. meep
|
Post #373,926
4/17/13 3:42:37 PM
|
![New](/static/images/lrpd.gif)
US cities control own currencies?
|
Post #373,928
4/17/13 3:53:29 PM
|
![New](/static/images/lrpd.gif)
Of course... in "Bocksley-nation"
--
greg@gregfolkert.net
PGP key 1024D/B524687C 2003-08-05
Fingerprint: E1D3 E3D7 5850 957E FED0 2B3A ED66 6971 B524 687C
|
Post #373,930
4/17/13 3:59:52 PM
|
![New](/static/images/lrpd.gif)
boxleystan
It's gotta be one of the 'stans.
|
Post #373,932
4/17/13 4:23:33 PM
|
![New](/static/images/lrpd.gif)
My favorite stan is musial.
|
Post #373,943
4/17/13 5:57:10 PM
|
![New](/static/images/lrpd.gif)
Snrk.
Regards, -scott Welcome to Rivendell, Mr. Anderson.
|
Post #373,944
4/17/13 6:05:16 PM
|
![New](/static/images/lrpd.gif)
hemenahemanahemana
10 points if you dont need der google
Any opinions expressed by me are mine alone, posted from my home computer, on my own time as a free American and do not reflect the opinions of any person or company that I have had professional relations with in the past 58 years. meep
|
Post #373,934
4/17/13 4:33:48 PM
|
![New](/static/images/lrpd.gif)
weimar ring a bell?
Any opinions expressed by me are mine alone, posted from my home computer, on my own time as a free American and do not reflect the opinions of any person or company that I have had professional relations with in the past 58 years. meep
|
Post #373,936
4/17/13 4:56:30 PM
|
![New](/static/images/lrpd.gif)
Let me know when you have a $1,000,000 bill.
|
Post #373,938
4/17/13 5:09:34 PM
|
![New](/static/images/lrpd.gif)
have to wait till I get home where I have a color printer
Any opinions expressed by me are mine alone, posted from my home computer, on my own time as a free American and do not reflect the opinions of any person or company that I have had professional relations with in the past 58 years. meep
|
Post #373,939
4/17/13 5:23:08 PM
|
![New](/static/images/lrpd.gif)
they have them in kentucky
Any opinions expressed by me are mine alone, posted from my home computer, on my own time as a free American and do not reflect the opinions of any person or company that I have had professional relations with in the past 58 years. meep
|
Post #373,948
4/17/13 7:45:33 PM
|
![New](/static/images/lrpd.gif)
Heh. :-)
|
Post #373,950
4/17/13 9:28:44 PM
|
![New](/static/images/lrpd.gif)
NC
|
Post #373,951
4/18/13 12:05:32 AM
|
![New](/static/images/lrpd.gif)
Felix: Chart of the day - Reverse Causality Edition
http://blogs.reuters...ausality-edition/
ThatÂs what youÂre seeing in the charts. Both of them have the same axes: GDP growth on the y-axis, and debt/GDP on the x-axis. Both of them plot the correlations in the dataset, with the dark line being the signal and the dotted lines showing the 95% confidence interval. And just as in the main dataset, the correlations are much clearer at low levels of debt/GDP than they are at higher levels.
But the two charts are different, all the same, especially at levels of debt/GDP above that 90% level. If you look at the left-hand chart, it shows that it really doesnÂt matter how much debt you have: youÂre likely to average about 3% GDP growth a year over the next three years. On the other hand, if you look at the right-hand chart, it shows that the more debt you have, youÂre significantly more likely to have experienced low growth in the past three years.
In other words, the causation here seems about as clear as causal analysis can ever be: low growth causes high debt, rather than high debt causing low growth. Indeed, once you get past 90% of GDP, your debt load doesnÂt seem to have any significant effect on future growth at all!
IOW, R&R got the causality wrong, and ignored the strongest part of their evidence.
Cheers,
Scott.
|
Post #373,952
4/18/13 7:17:39 AM
|
![New](/static/images/lrpd.gif)
It won't matter
Math challenged people will move on to a new piece of paper to pray to and believe in, because they know what is RIGHT, and they KNOW someone will give them their "evidence", sooner or later.
|
Post #373,955
4/18/13 9:16:36 AM
4/18/13 9:19:11 AM
|
![New](/static/images/lrpd.gif)
Re: It won't matter
Just like I seen nut-bags already collude the Boston Marathon Bomb and the Waco Texas Ammonium Nitrate factory explosion as being related. As in:
Oh NOES! The terrierists has gotten us twice in the past few days! WE NEED ATTACK THEM! EFF OFF <INSERT RACIAL OR COUNTRY SLUR>!!1!1!11! YOU ARE DEAD!!!!1!!1!
And those Boston Bombing is related to the Waco Texas Explosion, much the same way the way that Alaska's Temperature gradient matches the Amazon River's water flow through sediment carry. (As in *not*)
You know... meh.
--
greg@gregfolkert.net
PGP key 1024D/B524687C 2003-08-05
Fingerprint: E1D3 E3D7 5850 957E FED0 2B3A ED66 6971 B524 687C
![Expand](/static/images/plus.png)
Edited by folkert
April 18, 2013, 09:19:11 AM EDT
|
Post #373,956
4/18/13 9:55:30 AM
|
![New](/static/images/lrpd.gif)
Uh huh...
A fertilizer factory blows up in a state where half the deep thinkers are schnauzers and they need terrorists? Maybe they were thinking of terrierists?
|
Post #373,961
4/18/13 10:48:09 AM
|
![New](/static/images/lrpd.gif)
haha...
You said Deep Thinkers in Texas...
--
greg@gregfolkert.net
PGP key 1024D/B524687C 2003-08-05
Fingerprint: E1D3 E3D7 5850 957E FED0 2B3A ED66 6971 B524 687C
|
Post #374,002
4/18/13 9:20:10 PM
|
![New](/static/images/lrpd.gif)
Finally: The explanation is found.
|
Post #374,004
4/19/13 12:56:11 AM
|
![New](/static/images/lrpd.gif)
And something I can understand! :)
Alex
|
Post #374,018
4/19/13 10:46:57 AM
|
![New](/static/images/lrpd.gif)
Beautiful comment
--
Drew
|
Post #374,125
4/22/13 2:57:45 PM
4/22/13 3:10:51 PM
|
![New](/static/images/lrpd.gif)
..which leads to Another:
link, a ways down
http://www.ssa.gov/p...ngpapers/wp31.pdf
An indication of JUST HOW mathematically-removed-from Reality IS: virtually every "formula" within the pseudo-science loosely called, Economics.
[Aside/not to speak ill of the departed/maybe some reason why ... I never could get Beep to answer for me, "Just What. Is. It. that ... 'conservatives' want to conserve??]
This paper is a litmus for the 'fact' we see in all related bloviations:
The Sheer Quantity of words.. words.. employed merely to demonstrate the defects within a previous [Language Murder?]
makes the case for placing Econ firmly in the realm of astrology--but without the necessary assist of Owl entrails [for checksum-error calcs. maybe?]
Ed: PS, on another page.. the LRPD always achieves concision VS verbiage ...
It's like reading nadsat, only without the clarity.
--Thou Sayest, GRR
![Expand](/static/images/plus.png)
Edited by Ashton
April 22, 2013, 03:10:51 PM EDT
|
Post #374,061
4/20/13 6:20:31 AM
|
![New](/static/images/lrpd.gif)
Here's Thomas Herndon's paper, and the data files.
|
Post #374,152
4/23/13 11:40:57 AM
|
![New](/static/images/lrpd.gif)
it isn't an Excel problem
it was human error.
Satan (impatiently) to Newcomer: The trouble with you Chicago people is, that you think you are the best people down here; whereas you are merely the most numerous.
- - - Mark Twain ÂPuddÂnhead WilsonÂs New Calendar, 1897
|
Post #374,166
4/23/13 6:36:46 PM
|
![New](/static/images/lrpd.gif)
I think the point is
that Excel contributes strongly to human error.
|
Post #374,168
4/23/13 6:49:54 PM
|
![New](/static/images/lrpd.gif)
Because it is taken for Golden...
even when is a Golden colored turd.
--
greg@gregfolkert.net
PGP key 1024D/B524687C 2003-08-05
Fingerprint: E1D3 E3D7 5850 957E FED0 2B3A ED66 6971 B524 687C
|
Post #374,185
4/24/13 11:16:04 AM
|
![New](/static/images/lrpd.gif)
Any software can give you a polished turd
if you don't double-check the results.
Satan (impatiently) to Newcomer: The trouble with you Chicago people is, that you think you are the best people down here; whereas you are merely the most numerous.
- - - Mark Twain ÂPuddÂnhead WilsonÂs New Calendar, 1897
|