The Norwegian Sovereign Wealth Fund's $92 Million Excel Error 49
FT Alphaville: Last year, Norway's $1.5tn sovereign wealth fund revealed that it had lost NKr980mn, roughly $92mn, on an error relating to how it calculated its mandated benchmark. Here's what Norges Bank Investment Management said at the time: "In February this year, a calculation error was discovered in the composition of the index we're measured against. This error led to a marginal overweight in US fixed income relative to global fixed income. When this was discovered, we immediately set about correcting it, but because the fund is so large, the return was 0.7 basis points. Due to this our previously reported positive relative return of NOK 118 billion was adjusted down to NOK 117 billion."
It is a good example of how even tiny operational mistakes can have mammoth-sized consequences in nominal terms when you manage one of the world's biggest pools of capital. Sometimes a mistake can even lead to a windfall -- such as in 2021, when NBIM apparently made NKr582mn by accidentally accumulating an outsized position in a rising stock. But the 2023 index snafu is by far the biggest the fund has registered, almost twice as large as the cumulative operational-accidental losses it suffered from 2010-20. Alphaville was intrigued. What exactly went wrong? Well, in a recently-released anthropological report commissioned to investigate its own culture, NBIM seems to have inadvertently revealed just how minuscule the mistake was.
Here's an NBIM employee called "Simon" recounting the debacle to the report's author, Tone Danielsen. Alphaville's emphasis below: "Last year (spring 2022) we had an off-site. One of our workshops was on 'Mistakes and how to deal with them.' We wrote post-it notes, classifying them into different categories from harmless to no-goes. One of my post-it notes, I remember it vividly, read: Miscalculation of the Ministry of Finance benchmark. I placed it in the category unforgivable.
When I wrote that note, I honestly couldn't even dare to think about the consequences. And less than a year later, I did exactly that. My worst nightmare. It was a manual mistake. My mistake. I used the wrong date, December 1st instead of November 1st which is clearly stated in our mandate. The mistake was not revealed until months later, by the Ministry of Finance. They reported back that the numbers did not add up. I did all the numbers once more, and the cause of the mistake was identified. I immediately reported to Patrick [Global Head] and Dag [Chief]. I openly express that this was my mistake, and mine alone. I felt miserable and was ready to take the consequences -- whatever they might be."
It is a good example of how even tiny operational mistakes can have mammoth-sized consequences in nominal terms when you manage one of the world's biggest pools of capital. Sometimes a mistake can even lead to a windfall -- such as in 2021, when NBIM apparently made NKr582mn by accidentally accumulating an outsized position in a rising stock. But the 2023 index snafu is by far the biggest the fund has registered, almost twice as large as the cumulative operational-accidental losses it suffered from 2010-20. Alphaville was intrigued. What exactly went wrong? Well, in a recently-released anthropological report commissioned to investigate its own culture, NBIM seems to have inadvertently revealed just how minuscule the mistake was.
Here's an NBIM employee called "Simon" recounting the debacle to the report's author, Tone Danielsen. Alphaville's emphasis below: "Last year (spring 2022) we had an off-site. One of our workshops was on 'Mistakes and how to deal with them.' We wrote post-it notes, classifying them into different categories from harmless to no-goes. One of my post-it notes, I remember it vividly, read: Miscalculation of the Ministry of Finance benchmark. I placed it in the category unforgivable.
When I wrote that note, I honestly couldn't even dare to think about the consequences. And less than a year later, I did exactly that. My worst nightmare. It was a manual mistake. My mistake. I used the wrong date, December 1st instead of November 1st which is clearly stated in our mandate. The mistake was not revealed until months later, by the Ministry of Finance. They reported back that the numbers did not add up. I did all the numbers once more, and the cause of the mistake was identified. I immediately reported to Patrick [Global Head] and Dag [Chief]. I openly express that this was my mistake, and mine alone. I felt miserable and was ready to take the consequences -- whatever they might be."
Re: (Score:2)
And building a proper system might well have cost ten times that.
Re: (Score:2)
Independent parallel programming might have caught the error. Just have two individuals create independent implementations of the same calculations. Neither can see the other's work. A third party views the outputs and checks that they match. If not, they have to find the error. I don't think that would cost 10x the amount.
There's no such thing as 100% accuracy. Errors always happen. It's just a matter of how much error one is willing to accept, and if the stakes are high enough, you do a cost-benefi
Re:Hmm (Score:4, Insightful)
And building a proper system might well have cost ten times that.
The error didn't really matter. It caused them to invest slightly more in one fund than another, and the overweight fund didn't earn quite as much.
But that was due to mostly random chance. The error could have just as easily resulted in a higher return, and back in 2021 a similar error did exactly that, resulting in an inadvertent 582 million Nkr windfall.
Re: (Score:2)
Re: (Score:1)
OK so maybe they could automate that whole report and take the human out of it. But converting what an Excel spreadsheet does to "code" in some organizations might also be "premature optimization".
For example, next year they might need to change the way some calculations are done and the resulting reports. Due to changes in taxes, laws, etc.
With Excel, the current perso
Re: (Score:2)
I would have thought something in the process would autofill the start date. This sounds like it's an annual thing, so building in "start in November" seems doable.
Of course, it's also the kind of thing that will be considered "nobody's gonna mess that up without noticing" until something like this happens...
Re: (Score:2)
You want a pair of human eyes, because code writer isn't going to assume responsibility for errors. Because brain behind those eyes can be trained and vetted by you.
This is something most nerds have real problems grasping. When it comes to software, responsibility problem is real, and the reason why pretty much all software EULAs try to shed any semblance of responsibility for what software does as much as they can get away with, and in most cases, more than that.
Re: (Score:2)
I would have thought there a better systems and tools for managing $1.5tn than Excel. To cheap to build a proper system cost them $92 million.
Even more surprising, not only were the guys responsible not fired, but their boss sent them an email telling them not to sweat it, because shit happens [gearrice.com]:
Neither Simon nor Patrick du Plessis were fired. Nicolai Tangen, CEO of the fund, He opted instead to send them a reassuring email.:
These things happen! We run complex operations and what surprises me most is that historically we have had very few if any incidents like this. You are both super professionals and have contributed significantly to the success of NBIM.
Don’t let this ruin your weekend. Onward and upward!
Re: (Score:2)
Fire you?! I just spent 92 million US dollars training you to never, ever, EVER do that again!
Re:Hmm (Score:5, Insightful)
Even more surprising, not only were the guys responsible not fired,
The only thing surprising is that you find it surprising. An organisation that fires an employee for fat fingering a mistake rather than acknowledging the shortcomings of process and practices that could lead to a single mistake going unnoticed would be unimaginably toxic. This isn't the 1930s anymore.
Re: (Score:2)
In 1930s, people who knew and were good enough at math to do what spreadsheet software does today required a lot of training. In part because all math was done by hand, errors like these were a norm and priced into the service.
You didn't fire your accounting professional because he miscalculated somewhere, because whoever you replace him with will do same kinds of errors about as often in a good case, and more in a bad one. Because every time you get a new accountant in pre-spreadsheets era, you need time t
Re: (Score:2)
Of course you did. Blaming the person was something that was done a lot in the old days, and in some parts of the world still done now. Professional or otherwise.
Pretending that simply having spreadsheets means accountants need to be any less professional and trained to use the systems and processes in place at your company today compared to the 30s is silly. That is one of the few things which hasn't changed.
Re: (Score:2)
"Some parts of the world", certainly.
Parts of the world that developed into wealthy nations? The general rule for those is the opposite. It's one of the main cultural reasons why some nations develop and some don't. Culture of personal responsibility at every level and leadership culture that nurtures it by allowing people to correct their mistakes and learn from them.
Re: (Score:2)
You're confusing US that made itself the wealthiest nation with US of today.
Strong men lead to good times. Good times lead to weak men.
Re: (Score:2)
they would 100% fire someone (and have in the past) in my company for doing something like this and my company has total garbage spreadsheets which we aren't allowed to fix because..."sarbanes oxley"...
Re: (Score:2)
Re: (Score:2)
they would 100% fire someone (and have in the past) in my company for doing something like this and my company has total garbage spreadsheets which we aren't allowed to fix because..."sarbanes oxley"...
That's because you work for a toxic company that places responsibility on individuals rather than processes and procedures. It sounds like an absolute shithole to work for.
Re: (Score:2)
honestly, compared with other similar jobs I've had, it's not so bad generally really
I mostly know what's important and make an effort to mitigate the risk associated with errors
Re: (Score:2)
People always think you can build "something better" especially when they think it'll only cost a few million. You sound like the type of person managing government system upgrades or migrations, the kind that go 4x overbudget and then fail to meet the original design use case. No you won't be building anything better for under $92million. Not if you realistically address all requirements of your excel users.
The reality is, as much as you don't like the idea of a Microsoft product being used, Excel absolute
Re: (Score:2)
It was a manual mistake. My mistake. I used the wrong date, December 1st instead of November 1st
Re: (Score:2)
Only fools (MBAs) use Excel.
Real professionals use a database where you don't make stupid mistakes like this.
Re: (Score:2)
Only fools (MBAs) use Excel.
Real professionals use a database where you don't make stupid mistakes like this.
You're claiming that no one has ever entered an incorrect date into a database? I find that hard to believe.
Since I've found incorrect data in databases before, I find your blind faith in them foolish. Which I guess means that by your own criteria you are eligible to use Excel.
Re: Hmm (Score:2)
Databases have error checking
Context (Score:2)
This is like someone who makes $150,000 a year making a mistake that costs them $9.
Re: (Score:2)
Re: (Score:2)
Re:Context (Score:5, Interesting)
Or did it cause them to put a little more of their fund than they intended into a stock that undeperformed a little? That could be called a loss in hindsight, but was just as likely to be an accidental gain.
Re: (Score:2)
My question as well. The FT article teases several times with "what really happened?" and answers the question with a human interest anecdote.
It sounds to me like somebody made an oops and as a result a report over-estimated their profits, which they had to revise. Somebody, presumably trawling for tidbits through this anthropological report, wrote a story about it and a bunch of weird websites and the Financial Times carried it.
Re: (Score:2)
Norwegian Milton Waddams is still complaining (Score:2)
about his drink at his retirement resort.
Systems mistakes don't have a single cause (Score:5, Insightful)
I openly express that this was my mistake, and mine alone.
This is broken thinking. Processes never go wrong due to a single mistake. If you think there was only one mistake, then the second mistake was not designing the process to have a check for that category of mistake. If this computation was so important, surely there should have at least been a second set of eyes on it. I guess we're reading this article because there was a second pass on the computation, it just happened months too late.
Re: (Score:1)
This is correct thinking. Where you assume responsibility for your mistakes, rather than blame "the system".
There's a reason why nations where people are culturally trained to behave like this are rich and peaceful, whereas nations where people are trained to blame "the system, the process, etc" as you suggest are universally much worse.
Re: (Score:3)
No, the GP is correct. Psychologically no one attaches as much importance to their own mistakes as to those of others, so your "culture of assuming your personal responsibility" cannot be the sole cause of the system working better; if it were, a culture of "uncovering the mistakes of colleagues and making them pay for them" would work even better.
What really improves a system is "making sure that mistakes are discovered and corrected, no matter who makes them". Owning the errors in your work may be a part
Re: (Score:2)
You're utterly confused. My guess you're anglo fed on the current woke culture and "personal responsibility is whiteness, and whiteness is evil and must be exterminated", which causes you to regurgitate the following woke talking points uncritically. Let me explain:
>cannot be the sole cause
"You're wrong because it's not the only factor"
This is reductio ad absurdum. Main factor is rarely if ever the only one. "All/Not all" is one of the primary ways woke theorists use to confuse normal people. Fight this
Re: (Score:2)
So not Excel? (Score:2)
Amateurs at work (Score:2)
How pathetic. Save a million or two per year to have actually competent experts on staff, lose a lot more. It is time to stop tolerating incompetent management.
It doesn't matter (Score:2)
It's not as if balance accounts are physical magnitudes. Accounting is a human tool used by humans to make decisions, a number that summarizes a million inputs into a comprehensible codified simulation.
In a system as big and complex as a country's finances, what matters most for making decisions is the changes in the inputs, that causes the value to be updated (cybernetics 101); so, the actual exact value only matters to the few companies or projects on the margin, that would be placed in a different class
Inconsequential (Score:2)
For certain values of "unforgiveable" (Score:2)
Your category. Not necessarily everyone's category.
I once killed a man, by forgetting an important part of a procedure. As soon as the emergency was over (the ambulance took the guy away, the paramedic still working on the corpse ; he was a DOA), I realised my mistake. I have never forgiven myself for not living up to my own standards. But everyone who has been fully trained for such situations, to whom I've described the events, has told me to forgive myself. Stil