Excel-lent Mistakes: A Hall of Shame

This Spreadsheet Could Ruin Your Life (And Definitely Your Bonus)
You think it’s just Excel—until it’s not. In this episode of Overleveraged, Overconfident, we unpack five real-world spreadsheet fails that triggered $24 million losses, $6 billion meltdowns, government investigations, and even global austerity policies. From Barclays hiding rows in a bankruptcy deal to Fannie Mae's $1.2 billion “oops,” and JP Morgan’s $6B model built by one guy in London with a PhD and no backup plan, it’s a masterclass in how bad inputs = career-ending outputs. If you've ever copy-pasted a formula and prayed no one noticed…this one's for you.

Come for the Excel horror stories, stay for the schadenfreude.

This episode was researched and written by Cherise Lloyd. Some music was created by using UDO. You can find show notes, transcripts, and sources at www.overleveragedoverconfident.com. Follow us on Instagram at Overleveraged Overconfident. Support the show on Patreon and help us keep deep-diving into those spectacular financial fails.. Overleveraged Overconfident is part of Seven Seven Spider, which specializes in deep dive research and financial storytelling. All research uses publicly available sources, so no confidential or regulatory information, just my keen sense of spotting nonsense.

Transcript

Cherise 2 0:02

A quick warning before we get started. Today's episode contains mild swearing and detailed descriptions of spreadsheet induced panic. If you've ever had to explain a VLOOKUP error to a managing director, or worse a judge, this may be triggering. You've been warned. imagine this. It's 11:37 PM You're a junior associate at a white shoe law firm. Or a senior vice president at a major bank or an accountant working on a bankruptcy, filing a risk model, You open the Excel file, 1000 rows, 24,000 cells. You're tired, hungry, and running on courtroom deadlines and sheer caffeine. You resize the rows so it looks nicer for a PDF, or you copy and paste in the wrong columns, or you add up the numbers instead of average them, and then you hit save. Glad to be done and finally have dinner. You pack up and head home. You have no idea. You've just cost the firm millions of dollars by incorrectly adding the risk on a trade or forgetting one formula. That's right. A single click buried in a spreadsheet. Just detonated a landmine, ready to hear how you can end your career. Cause congressional hearings and bring down the entire management team, we'll polish up your resume as we pay homage to the most powerful, most dangerous software in corporate history. Microsoft Excel.

anchor 1:38

The telecom giant that once employed over thousands of workers, has officially filed for Chapter 11 bankruptcy protection. Company cited mounting debt and declining revenues as primary factors likely see their investments wiped out entirely while bond holders may recover only premium investors. This marks the end of what was once considered a blue chip. Look at the timeline. They were still issuing rosy guidance to investors. Full sense of this, was this primarily a case? Of being over leveraged or was this more about overconfidence? I would say overleveraged. It feels like the whole market is overconfident, neither. It's a perfect hedge. Just hang on. Welcome to Over Leveraged Overconfident, the podcast where spectacular risk meets more spectacular ego. This is where the hedge fund tightens. The FinTech darlings and the CDO Cowboys swagger into risks like they own the term sheet. But all of these stories will end with subpoenas, sell offs, and sob stories. Hi, I am Cherise. I've spent the last 15 years with a front row seat to some of the most incredible financial meltdowns from my time at the Federal Reserve to the area of optimistic leverage finance. And I've seen some very smart people do some very dumb things with extremely large sums of money. This is where we will look at investors who talk their book, bankers Who Believe the model, and speculators who go home empty handed. These are the stories where conviction replaces confirmation, and everyone believes in fast money. Since the risk of failure may seem extremely unlike. Even if it's highly reoccurring.

Cherise 2 3:25

First launched in 1985. Xcel has become the duck tape of global capitalism, loved by analysts, feared by auditors, and overused by literally everyone. It's the place where people model billion dollar deals and somehow also plan their kids' birthday parties. For some, it's a playground color coded, conditionally formatted full of macros and promise for others. It's a mind feel of invisible formulas, broken links, and the ever haunting reference error. While Excel was never built to be a legal document, a compliance system, or a trading engine that hasn't stopped anyone from trying today, we're counting down the top five Excel blunders spreadsheet fails so epic that they torch stock prices, triggered lawsuits and cost companies hundreds of millions of dollars, all because someone hit the wrong key. Forgot a minus sign. Or my personal favorite hit a row they meant to delete. These are the stories that remind us. Excel may just be sales and formulas, but in the wrong hands. It's a financial weapon of mass destruction. Number one, TransAlta cut and paste disaster.

anchor 4:51

We depend on transmission providers. By partnering with them, we make electricity accessible for consumers through the electrical grid. It's a pretty complex process, which requires lots of talented people from operators and maintenance people to engineers, legal, finance, and corporate support to ensure TransAlta associates have the resources, tools, and knowledge they need to operate effectively. In a nutshell, we're a big company with a bunch of moving parts all working together to generate and transmit power using a variety of energy sources in a manner that's safe, reliable, and environmentally responsible. That's what TransAlta does.

Cherise 2 5:29

Let's talk about copy and paste, the beautiful little shortcut that saves all of us from retyping our email addresses for the millionth time and how it became a$24 million mistake. Because apparently when you're dealing with power transmission contracts with tens of millions of dollars, the same level of care you use to copy a meme into your group chat is perfectly fine. This is the Trance Alta story, a major Canadian power generator that managed to turn excel's most basic function into their most expensive lesson. The best part, the CEO had to explain to investors that they had basically fat fingered their way into a financial loss transAlta isn't some scrappy startup run by addicted 20 somethings. This is a major Canadian utility company think big. They generate power, they trade contracts and they hedge risks like grownups or so we thought part of their business is power transmission, passing power through lines, something you have to bid on. In May, 2003, they're working on the US power transmission contracts. Standard stuff. For a utility their size. They use an Excel spreadsheet to manage and submit the bids. You bid on contracts, you hedge the risk and you make sure the numbers are right because millions of dollars are at stake. Some unnamed, permanently scarred analyst is sitting there with a spreadsheet full of data, rows and rows of numbers, all carefully calculated, all representing real money and real contracts. They copy the data, they paste it, sort it, and prepare their final submission. Somewhere between the copying and the pasting and the sorting, something went wrong. What CEO Steve Snyder would later describe as literally a cut and paste they did not detect. The analysts failed to align the rows correctly, just pasted them in one or two columns over. As a result, higher price bids got assigned to lower demand transmission paths. Oddly enough, TransAlta ended up winning bids at higher prices than they attended. And once submitted, it was final. There were no do-overs, no mulligans. This was the big boy world of energy trading, and someone had just made a very expensive mistake. The result, a$24 million loss because someone hit control plus V at the wrong time, in the wrong place with the wrong data. Steve Snyder, the CEO had to eat crow on a conference call with investors. It must have felt like telling everyone they were playing amateur hour. The Canadian press certainly thought, so. The TransAlta disaster is an example of something profound about operational risk. It's not always the sophisticated stuff that gets you. Sometimes it's the mundane, the more routine it feels, the less attention you pay to it, but it's still a process that can go wrong. TransAlta had all the systems and processes you would expect for a major utility, but then when it came to the final step, the human checking the human work, they relied on the same quality control. You used to copy a recipe from a website. Number two, Barclays bankers buy more than they wanted. Before we dive into this one, a quick apology to all attorneys listening. You may experience secondhand anxiety. As someone who has sent their fair share of frantic midnight emails during the great financial crisis, this story hits a little too close to home. Let's set the scene. It's September, 2008. Lehman Brothers has just filed for bankruptcy, the fourth largest investment bank in America, and everyone is scrambling like it's Black Friday at Target, except instead of discounted TVs, they're fighting over the scraps of 150 8-year-old investment bank.

9:59

Lehman Brothers is going bankrupt. And financial markets from Asia to Europe are doing their utmost to prevent Monday from turning from dark to black employees of America's fourth largest investment banks saw the writing on the wall late Sunday after talks to pull them back from the abyss collapsed. British firm, Barclays Bank said no thanks while Bank of America decided instead to help another struggling bank, Merrill Lynch in a merger deal worth over$50 billion.

Cherise 2 10:29

Enter Barkley's capital. Ambitious, aggressive, and not exactly known for subtlety, the bank sees an opportunity in Lehman's best bits. They had excellent timing and deep pockets. Barclays had actually sought to acquire all of Lehman Brothers in the days before the bankruptcy. But UK regulators overwhelmed by their own financial crisis, blocked the deal. Now, they were part of what's called an asset scramble, a frantic environment where financial institutions scramble to acquire the viable assets from a distressed firm. They moved quickly. By September 18th, just three days post meltdown, they were ready to buy. They wanted the trading desks, Lehman's, Midtown, Manhattan office, and the human capital, the non-toxic parts. They made an offer for only the icing at a fire sale price of 1.35 billion. The catch. Lehman's bankruptcy wasn't just messy. It was astronomically complex, which increased the likelihood of things going wrong. We're talking about more than 900,000 derivatives contracts with the notional value of more than$35 trillion. So Barclays did what bankers do and created an Excel spreadsheet. This critical spreadsheet was intended to serve as the definitive listing of all the contracts they were interested in acquiring from Lehman Brothers, because apparently even billion dollar banks use the same software you use to track your fantasy football league. This spreadsheet included all the contracts. The winners and the losers. So it was huge. Thousands of rows of data and 24,000 cells making any manual verification process incredibly difficult. But here's where someone at Barclays made a choice that would haunt them forever. Instead of deleting the rows on the spreadsheet that contained losing contracts they didn't want, they just hid them. They also added a little column with an N in the cell to mean no or not critical. Let me repeat that for the people in the back. They hid the rose in Excel in what would become a legally binding document. To be clear, it was Barclay's bankers that sent this file to their lawyers at clearly Gottlieb at 7:50 PM on September 18th. The court deadline midnight same day. They provided no explanation of the hidden rose, the annotations with the letter N, or even the number of excluded contracts. The legal team had four hours and 10 minutes to review format and submit a document with over a thousand rows. That would determine which contracts Barclays would be legally obligated to acquire. Just a personal aside, I worked in finance and my Excel skills stop at the occasional pivot table. My attorney friends think V lookup is a Supreme Court case. The pitfalls of Excel that the first year banking analysts get drilled on are totally foreign territory to most legal eagles. So it makes sense that the associate at clearly gottlieb a third year lawyer who probably thought they were just dealing with formatting the bread and butter of the legal profession decides to globally resize all the rows to make the document easier to read when printed. Or in this case, converted to PDF and boom. Just like that, the hidden Rose became visible. And once that PDF hit the bankruptcy court system at 11:37 PM game over, the error wasn't discovered until October 1st, two weeks after the deal had been approved. Barclays tried to fight it. Of course, they filed legal motions, scheduled hearings, did everything short of claiming their dog ate their homework. But the court basically said, sorry, not sorry. You submitted it. We approved it, deal with it. The company was forced to swallow the losses of all 179 contracts. The exact amount is classified. But sources describe it as in the millions, number three, the 1.2 billion Excel error, the toppled Fanny May. This is How one Excel formula became the loose thread that unraveled a$10.6 billion accounting fraud at Fannie Mae because nothing says financial integrity, like a spreadsheet error that crashes your stock by 97%. Fannie Mae Short for Federal National Mortgage Association was a United States government sponsored entity. Created in 1938 during the Great Depression to expand access to affordable housing. Its main job by mortgages from banks and lenders, bundle them into mortgage-backed securities and sell them to investors, injecting liquidity into the housing market. Fannie Mae doesn't make home loans directly. Instead, it keeps the money flowing to lenders so that they can issue more mortgages. It operated in a public, private hybrid zone, publicly traded, but created and assumed to be backed by the federal government. It was the best of all worlds and the envy of civil servants in dc. The staff and management made market money with government contract security, and they loved it, spending millions on lobbyists to protect their golden business model. This irked some members of Congress

16:36

that's generally the kind of talk the financial community loves to hear Mr. Brains and reason enough for there to be growing talk. Uh, that if Treasury Secretary Rubin, for example, were to step down as is perennially always argued, you are the perfect man to take his place. How do you feel about that? Uh, well, my desire is for Bob Rubin to stay treasury secretary for life. What if he dozens? Uh, if he doesn't, I'm sure the president will find, uh, many qualified people, but I'm very happy. You seem the top, you keep seeming to top that list. Well, I'm, I topped the one list I was interested in, which is the list of people to be CEO of Fannie Mae. And, uh, that's where I'm gonna be in. So if they asked, you'd say no. Uh, what I'm saying is that I will continue as, as CEO of Fannie Mae. Alright. Franklin Rain

Cherise 2 17:20

This caused a little tension a few years into re's controversial tenor as CEO. things took a turn. It's 2003, only two years since Enron imploded and became a huge sinkhole fraud created by interesting accounting and shell game tactics to ensure management got big payouts and people were still pretty pissed about it. This all kicks off with an unknown financial modeler at Fannie Mae working in Excel. Probably someone making decent money who may have even felt good about their Excel skills. It's the end of the quarter and their job is to work out the pricing on loans purchased by Fannie Mae for reporting under a new accounting rule that had just been passed. Thanks to Enron, they were working on what the company called complicated calculations required in the implementation of FAS 1 49, which is accountant speak four. This is about to get messy. The oops, the modeler wrote the wrong formula. Not a typo, not a rounding error. A fundamentally flawed piece of logic that becomes not just a whistleblower, but a full on air horn. This wasn't just any calculation. This was tied to unrealized gains. Those phantom profits that look great on paper until reality comes a knocking the error, a$1.2 billion overstatement. To put that in perspective, that's enough money to buy the entire GDP of small countries. The point is, it's a lot of money. Since it was discovered in an internal review it seemed the answer was simple. Come clean. Just say it was a simple mistake. Rains and the company tried to brush it off as, Hey, this is the first time we're new at this. No big deal. We'll just issue a restatement of earnings and move along. If this was a normal company with normal problems, the story may have ended there. Company makes error, fixes error. Everyone moves on. But this was Fannie Mae. We're talking about a government-sponsored enterprise with the financial complexity of a Rubik's cube designed by statistical mathematicians. Also in the past year, Fannie Mae's, sister firm and chief competitor, Freddie Mac. Had to make some embarrassing accounting admissions, and everyone was on edge. Fannie Mae had recently been giving everyone the warm fuzzies by saying they definitely absolutely crossed their hearts and hoped to die or not using creative accounting. And Pinky Swear did not have the same earning smoothing problems that had surfaced at their competitor. So the market didn't roll with this. The reaction was, let's call it Biblical Fannie Mae's. Stock price went from$73 to$2 and 25 cents faster than you could say, career limiting move. That's a 97% drop. Folks, I've seen crypto crashes that were gentler. Members of Congress were beginning to wonder if senior management was up to their job. Their job was to account for mortgages, and now it seemed like they weren't even able to do that. Competently representative Richard Baker saw the writing on the wall immediately, his response to the spreadsheet error. Here we go again. First Freddy, now Fanny, he wasn't talking about a technical glitch. He saw a pattern. The spreadsheet error was like lifting the curtain. How exactly was the firm calculating expenses and income using a manual Excel spreadsheet so that regulators got curious and started looking? Enter O-F-H-E-O, the Office of Federal Housing Enterprise Oversight. They took one look at this honest mistake and thought, you know what? Let's dig a little deeper. What they found makes that Excel error look like a rounding issue. From 1998 to 2004, Fannie Mae had been running what can only be described as double entry accounting if that meant two sets of books. They were manipulating earnings with the creativity of a Hollywood screenwriter and the moral flexibility of a used car salesman. The techniques they used read like the Greatest Hits album of corporate fraud. First, they had cookie jar reserves, an approach you could think of as financial meal prep. When times are good, you save some of the earnings for later. When times are bad you dip into your stash. Except this wasn't budgeting wisdom. This was accounting manipulation, designed to make their earnings look smoother than a freshly waxed floor. Then there was the creative interpretation of hedge accounting rules. They basically treated accounting standards like the terms of service agreements. You know, those things you click agree to without reading, because who has time for that? And my favorite, the most extraordinary measure that they took to avoid recording losses on declining assets, they literally ignored all the accounting fallout of the last few years and couldn't see that what worked before would not work. Now, here's what makes that story spectacular. The motivation. It's not unique. It wasn't some abstract accounting philosophy or misguided attempts to help shareholders. This was cold, hard executive bonuses. When the dominoes started falling, they fell fast. O-F-H-E-O Launched a full investigation in September, 2004, finding additional widespread accounting errors that had made the original spreadsheet mistake look like a typo. Franklin Rains retired in 2004 under what can only be described as a mushroom cloud of suspicion. The SEC concluded in 2006 that Fannie Mae had engaged in financial fraud involving multiple violations of accounting principles from 1998 to 2004, with the goal of falsely portraying stable earnings growth and maximizing executive bonuses. The final tally.$10.6 billion in overstated profits. Hundreds of millions of ill-gotten bonuses and a$1 billion just to clean up the mess. After eight years of litigation and 101 civil charges, the executives were cleared. In 2012, the court said there hadn't been enough evidence to prove intent. The Excel error did not cause the fraud, but it certainly kicked off the investigation that revealed it. It was that loose thread that when pulled, unraveled years of systemic deception. Without that mistake, who knows how long this party would've continued. Number four, late nights and manual entries. How JP Morgan lost 6 billion. This is about JP Morgan Chase and how they managed to lose$6 billion by essentially playing financial jingga with a blindfold on. It is the story of what happens when brilliant people convince themselves that changing the math changes reality, and you can guess how that turns out. So picture this, it's December, 2011. Basel three banking regulations are breathing down. JP Morgan's neck, like that friend who always wants to split the dinner bill equally when they order a lobster. Basel three controls how assets are rated for risk and how much capital a bank should have. The bank's chief investment office, let's call them CIO, has a problem. They have this massive synthetic credit book, and it's risky. It's exceeding the limits of what the bank can and should hold, making regulators and even the risk department nervous. Most people would think, okay, time to sell some stuff and take our lumps, but that would be boring. The CIO looked at the situation and decided that everyone was being dumb and saw no reason to be nervous. This was all perfectly fine. It was just how they measured the risk, their brilliant solution. Don't change the risk. Just change how you measure it. Enter our protagonist, pat Hagan senior Quant in London. Smart guy PhD, smarter than you and me in math. And here's the thing about being smart. Sometimes it means you think you can rationalize away reality. Pat's boss came to him with what was basically a mission and possible scenario. Hey Pat, we need you to build a new risk model. Oh, and we need it yesterday. Also it needs to reduce the level of risk on the stuff we already bought and don't wanna sell. But don't worry, we'll tell everyone. It's just for accuracy. So Pat gets to work with no background on modeling synthetic credit. This man whose bonus depends on the model working exactly the way management wants, is left completely alone to build it and explain it. No oversight, no automation, no backup. Just Pat and some spreadsheets and the capital position of a United States Depository Institution. He comes up with a model that does what they need. By July 27th, 2012, they flip the switch on Pat's new model, overnight, literally their risk drops by 50%.$132 million in value at risk becomes 66 million just like that, without a single position changing, without selling anything. The Bank's model Review Group took a look at this setup and basically said, well, it's completely insane and it has to be updated manually, which, come on, people are fallible. Since it's managed manually, there's a limit as to how many transactions can be inputted and in these markets, you're only going to update it at the end of the day. Ridiculous. But then they shrugged and said, eh, close enough. The model group did make Pat and his boss promise to automate the process since it was crazy. The model would only be updated every 24 hours by one person, but that takes money and time. And apparently everyone felt that Pat had plenty of both. So it turned out that every night, pat was manually entering trading data into multiple spreadsheets. He was burning the midnight oil, probably surviving on coffee, and that kind of desperation that comes from knowing your entire career depends on getting it right. A Phd Was now in charge of data entry and no one was interested in changing that setup. Pat's model should have sparked some initial caution, but they did what any rational person would do. When they think that they're suddenly become half as risky, they double down. Well, these guys actually triple down. They went from 51 billion in exposure to 157 billion by the end of March. The magical manual Excel spreadsheet allowed the CIO to stay within risk compliance limits, but get big enough to actually move the market. This attracted media attention and earned the French Trader in London, the nickname the London Whale, a hundred million dollars a year employee in the CIO department, which was supposed to reduce the firm's risk. Meanwhile, pat was manually updating his spreadsheet every night with even more positions and trades three times more than when he started this charade. At this point, he's probably wondering how his life became a Financial Times horror story. Those spreadsheets over time became corrupted and riddled with errors. There was one particular beautiful mistake where instead of dividing by the average, the firm divides by the sum. This actually means something when you're talking about billions. For all of Pat's work and late nights, it was still shoddy and obviously wrong. He might be smart, but sometimes he just wanted to go home. The fact that the model was untethered to the reality of the risk, the CIO was taking became clear on April 12th, 2011 when the CIO lost$400 million. This little fact was missing from the model. You know the model that's supposed to tell you how much you. Can lose, can't even matter how much you did lose. So 400 million just gone like it fell behind the couch cushions of global finance. You'd think this would be a Houston. We have a problem moment, but no, they went, that's weird and just kept going. Finally, in May, 2012. Pat using that PhD. Finally realized his model was fundamentally broken. The guy who built it looked down at his own work and basically said, Hey, wait a minute. This is completely wrong. On May 10th, JP Morgan had to publicly admit that their risk model was something as accurate as a Magic eight ball. Jamie Diamond had allowed himself to be recorded, referring to the whole thing as a tempest in a teacup. And that's when they thought it was only 400 million. They switched back to the old model and suddenly their risk numbers shot right back up. However, this only lasted a moment. In September, 2012, they changed the model again, which lowered their risk to only 20% because apparently the lesson they learned from nearly destroying their bank wasn't. Maybe we should measure risk more accurately, but maybe we should be more subtle about fudging the numbers. When the CIO's positions were taken outta their control and the losses counted, the bank had lost over$6 billion. A bank backed by the FDIC Congress was, shall we say, concerned. Here's what Senator Carl Levin had to say about poor Pat and his work and question why a manual spreadsheet was used for a billion dollar position.

32:09

For the CIO's$350 million billion dollar portfolio, including the synthetic credit portfolio, was being run manually using error prone spreadsheets with operational flaws. Mr. Wyland, did you know that Mr. Hagen was doing nightly manual data entry and sometimes staying up into the wee hours of the night to get it done? Were you aware of that fact? I wasn't aware of the details of the, of the manual work he was doing. I did know there were spreadsheets involved, and I did know that Mr. Hagan often stayed late at night. Mr. Bacon. Um, the OCC told us that the operational problems, the spreadsheets, the lack of an automated database, the calculation and formula error, which lowered these VAR results were shocking and absolutely unacceptable. Do you agree? I do.

Cherise 2 32:49

The lesson here, sometimes the simplest explanation is the right one. If your risk drops by 50% overnight, you're probably not getting smarter about risk management. You're probably just getting dumber about risk measurement. Excel spreadsheets are fallible and people get tired. number five, even smart people can screw up excel. The austerity debacle

anchor 33:17

The economic theory underpinning austerity policies being followed by governments worldwide may be flawed. That's the allegation made in a study by the University of Massachusetts. It claims to have found coding errors on the Excel spreadsheet used by the academics who produce the theory, which could invalidate their conclusions.

Cherise 2 33:39

On January 4th, 2010, the Marriott Hotel in Atlanta was buzzing with the annual meeting of the American Economic Association. Among the many presentations, one paper stood out, grabbing the attention of policy and economists alike growth in the time of debt. Authored by the highly respected Professor Carmen Reinhardt and the former IMF Chief Economist Ken Rogoff their stark findings resonated deeply amid the global economic crisis. Economic growth they found dramatically slows when a country's debt exceeds 90% of its gross domestic product. This was merely an academic exercise. It offered a potential roadmap for nations struggling with massive debt. This paper became the only thing politicians seemed to read from cover to cover policymakers, eager for solutions, started citing its findings as gospel. Back at the University of Massachusetts Amherst, a graduate student named Thomas Herdon thought that the results looked too clean, the correlation to clear cut. Luckily, his professors had given his class a critical assignment. Pick an economics paper and try to replicate the results. It's a standard exercise for aspiring researchers designed to hone their analytical skills. With his curiosity peaked. Thomas chose growth in the time of debt. This was 2010 during the global recession, and at the heart of the debate was the crucial question, should governments increase debt to stimulate growth and climb out of the slump? Or should they aggressively cut spending and raise taxes to get the public debt under control. Reinhart and Rogoff's paper provided powerful ammunition for the latter influential figures like EU commissioners and the uS Republican politician, Paul Ryan both referenced the 90% debt to GDP limit to back their austerity strategies. As US politicians debated stimulus measures, europe was creaking under the weight of forced austerity, and the UK's new coalition government was promising deep cuts. Meanwhile, Thomas Hardin's homework assignment wasn't going well. No matter how hard he tried, he couldn't replicate their findings. He thought he had made a gross error. He was, after all, only a student and his professors, including Michael Ash, were initially certain he was making mistake, given that these were imminent Harvard professors. Ash even said, come on Tom. This isn't too hard. You've just gotta sort this out. But Thomas reviewed their work over and over without any clarity. By the end of the summer, his professors were beginning to doubt the fault was with their student. They encouraged Thomas to continue the project by reaching out to the Harvard professors himself. After a few emails, Reinhardt and Roff provided Thomas with the actual working spreadsheet they had used. When he opened it, Thomas was well versed in the numbers and he could see right away something was off. He even called his girlfriend over to double check, but he was right. While the spreadsheet contained data for 20 countries, the Harvard professors had accidentally only included 15 of the 20 countries in key calculations of average GDP growth in countries with high public debt, Australia, Austria, Belgium, Canada, and Denmark. Were simply missing an oops moment on a grand scale. The error became more apparent after a quick formula check, but the spreadsheet error wasn't only the issue. Thomas and his professors found incomplete data and questionable weighting that made the results seem unequal. The data emission made a substantial difference, and the weighing of the data was a methodological choice that was controversial, but also significantly impacted the results. On April 15th, Thomas and his professors published their findings as a draft working paper. Their conclusion, high levels of debt are still correlated with lower growth, but the most dramatic results of the original paper vanished. The relationship between high debt and lower growth is much gentler, and there are many exceptions to that rule. The stark black and white findings were actually more shades of gray. In response, Reinhardt and Rogo showed extreme grace that you often don't see in a statement to the BB, C. They expressed gratitude to heron for his careful attention and for pointing out the important correction. The error was sobering. However, they did not believe this regrettable slip significantly affected the paper central message or their subsequent work. Accidents can happen, and scientific progress often comes from identifying previous mistakes. But this paper provided intellectual rationalization that affected how people think about the world. And when it affects politicians, it eventually affects how the world works. While the error kept nuance out of the debate, it's questionable that it would've changed anything. No one was going to argue that Greece should spend more.

anchor 39:27

The observation and the subsequent conclusion that countries must cut. Public spending has meant hardship for millions. Policy makers around the world have imposed austerity measures in a desperate bid to ensure a growing economy. However, analysts don't expect governments to discard the theory as the saying. You shouldn't spend more than you earn. Still rings true.

Cherise 2 39:51

And so there you have it, five little spreadsheet errors and a few billion dollars later, and we have learned absolutely nothing from JP Morgan's whale trading on the wrong formula to TransAlta's$24 million. Cut and paste catastrophe. To Barkley's legally binding itself to toxic Lehman contracts with a hidden row. Each of these scripts had one thing in common. Excel wasn't the problem. The people were. It's just a tool, a glorified calculator with delusions of grandeur. But when it's used to price trades, submit bids, build financial models, and submit court documents, it becomes something more. A single point of failure wrapped in grid lines and good intentions. So sure it's flexible, it's fast, it's familiar, but it's also where typos become legal liabilities where one minus sign or the lack of one can swing billions, where the people who built the model are always conveniently no longer with the firm, and yet we keep using it because it's there and because it works. Until it doesn't. So if you're listening to this while staring in an open Excel spreadsheet, ask yourself, did you check that formula? Did you delete that row or just hide it? Did you send the right version to legal? Because no one gets fired for using Excel until you do Sweet dreams. This episode was researched and written by Cherise Lloyd. Music was created by using UDO. You can find show notes, transcripts and sources@overleveragedoverconfident.com. Follow us on Instagram at over leveraged overconfident. Support the show on Patreon and help us keep deep diving into those spectacular financial fails. Speaking, which if you have a favorite corporate disaster you'd like to hear about, reach out and connect with us@sevensevenspider.com. Over leveraged overconfident is part of Seven seven Spider specializing in deep dive research and financial storytelling. All research uses publicly available sources, so no confidential or regulatory information, just my keen sense of spotting nonsense. Until next time as we continue tracking the inevitable journey from hubris to bankruptcy.