Discover insights behind a staggering $81 trillion transaction error by one of the world’s largest financial institutions. How prepared are you for such operational failures?
In April 2024, Citigroup committed an astronomical blunder after trying to freshly deposit only 280 dollars to the escrow account of one of its customers located in Brazil but because of a human error in a system error, the customer received an incredible 81 trillion dollars into his account. The mistake was in the form of a backup system seldom utilized resulting in the automatic input of 15 zeros as default. The two employees missed the error, and, luckily, a third one felt there was an anomaly after around 90 minutes. The money was simply charged back in few hours and never came out of the bank.
Though no financial damage was done, it happened that the incident, which was classified as a so-called near miss, was reported to the Federal Reserve and the Office of the Comptroller of the Currency (OCC). It constitutes one of 10 $1 billion or greater so-called near miss in 2024, a minute decrease as compared to the previous year reported by Citigroup. These incidents do not ask to be reported formally to regulatory bodies yet they are very uncommon to the entire banking sector.
The accident follows as Citigroup is still under the regulatory spotlight due to operational and data controls infirmity. CEO Jane Fraser is in the process of overseeing a significant upheaval program, pushing to modernise the bank risk and compliance systems, but the incident highlights how legacy systems and manual workarounds are still extremely problematic. Already, Citigroup has suffered a history of operational errors, such as the mistaken $900 million payment to Revlon creditors last year, its flash crash caused by a trader error earlier this year, that have already resulted in regulatory fines and ousted leaders.
The $81 trillion miscue despite the reforms sheds light on how one press of a key throws doubts on even the largest of financial institutions in the world.