How a Spreadsheet Error Hid the Size of a Trade
In 2012, JPMorgan’s Chief Investment Office built a large position in credit derivatives—contracts whose value depends on the perceived risk of companies or groups of companies failing to repay debt. The position was intended to manage risk, but it grew so large that the bank’s own buying and selling could move prices in the market.
To track the danger, the bank used a model called Value at Risk, or VaR. It estimates how much a portfolio might lose over a set period under ordinary market conditions. VaR is not a forecast or a guarantee; it is a signal managers can use to judge whether a position is growing too risky.
The Mistake Changed the Risk Signal
A spreadsheet used in calculating the portfolio’s VaR contained a manual copy-and-paste error. That error distorted the calculation, making the portfolio appear less risky than it was. The model’s output therefore gave managers a weaker warning about the size of the exposure.
The spreadsheet did not place trades by itself. People made the trading decisions. But risk estimates help shape how closely a position is watched and whether managers believe it is within acceptable limits. When the estimate is wrong, a large position can look manageable on paper even as its real-world exposure grows.
Why the Loss Kept Growing
As market conditions turned against the trades, JPMorgan tried to reduce the position. But the portfolio was so large that unwinding it was difficult: selling or hedging in a market can affect the prices needed to complete the next trade. The effort to get out therefore became part of the problem, and losses mounted to more than $6 billion.
The episode showed how a tiny data-handling slip can matter far beyond one cell. A spreadsheet can feed a formal risk system, and that system can influence decisions across a bank. The consequential failure was not simply that someone pasted incorrectly; it was that the resulting number was trusted without catching how badly it misrepresented the risk.
