On 26 February 1995, Barings — Britain's oldest merchant bank, financier of the Napoleonic Wars, banker to the Queen — was sold to the Dutch bank ING for \pounds 1. One pound. Not because nobody wanted its business, its staff or its two-hundred-and-thirty-three-year reputation, but because a single trader in a Singapore back office had run up losses of roughly \pounds 827 million — more than twice the bank's entire capital — and someone had to inherit the hole.

That trader, Nick Leeson, was twenty-eight years old. He was not a genius who out-smarted the market with some devious secret strategy. His actual trades were, in hindsight, fairly ordinary directional and volatility bets on Japanese markets — the kind a stress test exists precisely to catch before they become catastrophic. What made Barings unique was not the trade; it was that nobody with the authority to stop him ever looked. This lesson is a case study in operational risk — the risk that a firm's own people, processes and controls fail — as distinct from the market risk the rest of this course has spent so much time measuring.

The job Leeson was never supposed to have

In 1992, Barings sent Leeson to Singapore to run its operation on SIMEX, the Singapore International Monetary Exchange, trading futures and options on the Nikkei 225 (the main Japanese stock index) and Japanese government bonds. He was good at the job, and Barings gave him a second one on top of it: he was put in charge of the back office too — the settlements team that confirms trades, matches cash, and reconciles what the front-office traders say happened against what actually happened in the market.

This is the single sentence that explains Barings' collapse: Nick Leeson was both the trader placing the bets and the back-office manager checking whether those bets were sound. In a properly run trading operation this is unthinkable — the two functions are kept apart by design, a discipline called segregation of duties, precisely so that the person with an incentive to hide a bad trade is never the same person auditing it. Barings skipped this because Leeson's Singapore desk was small, profitable, and far from head office in London — and because, for a while, he kept reporting extraordinary profits.

With both hats on, Leeson opened a error account with the number 88888 — nominally to record small, harmless trade-settlement discrepancies. In practice, it became a hiding place. Every losing trade he made could be booked into 88888 instead of onto his real trading book, kept off the P&L reports London actually looked at, and the losses simply accumulated there, unseen, for over two years.

What he was actually trading

Two overlapping strategies fed account 88888. First, Leeson sold large numbers of short straddles on the Nikkei 225 — simultaneously selling call options and put options at the same strike. A short straddle collects premium up front and pays off handsomely if the index stays roughly still; it loses badly if the index makes a big move in either direction, because one leg or the other finishes deep in the money. It is, in effect, a bet that volatility will stay low.

Second, as his concealed losses grew, Leeson tried to trade his way back to break-even by taking increasingly large directional long positions in Nikkei 225 futures — betting the index would rise. By January 1995 he held futures contracts worth an estimated \$7 billion notional, a position size wildly beyond anything his official mandate (low-risk arbitrage between SIMEX and the Osaka exchange) allowed — and paid for it by telling London the margin calls funded a client's own large, safe arbitrage trades. Nobody in London reconciled that story against the exchange's own records, because the man who would have had to raise the alarm was Leeson himself.

Both bets needed the same thing to be true: a quiet, range-bound Japanese market. On 17 January 1995, that assumption broke catastrophically.

The Kobe earthquake and the unwind

The Great Hanshin earthquake struck the Japanese city of Kobe before dawn on 17 January 1995, killing more than 6,000 people and devastating one of the country's major industrial ports. Japanese markets reacted the way markets react to sudden, large, genuinely unpredictable bad news: the Nikkei 225 fell sharply and volatility spiked, dropping from around 19,000 to below 17,000 within about a week.

For Leeson this was the worst possible outcome twice over. The short straddles lost money on the volatility spike itself, and the long futures positions — bought precisely to try to recoup earlier losses — lost money on the falling index. Rather than close the position and report the disaster, Leeson doubled down again, buying still more Nikkei futures on the theory that a rebound would bail everything out at once. It didn't come in time. By late February the concealed losses in account 88888 had reached roughly \pounds 827 million, comfortably exceeding Barings' entire shareholder capital of around \pounds 350 million. Leeson fled Singapore, was arrested in Frankfurt days later, extradited, and eventually served four years in a Singapore prison. Barings, unable to meet its obligations, was declared insolvent and sold to ING for a token \pounds 1 — ING took on all of Barings' assets and liabilities (and the hole in them) in exchange.

Why this matters for a risk manager

Everything else in this course — Value at Risk, the Greeks, stress testing — measures market risk: how much a portfolio's value might move. Barings is the canonical lesson in operational risk: the risk that a firm's own people and processes fail in ways that have nothing to do with where the market goes. No VaR model would have "priced" Leeson's fraud, because VaR assumes the reported positions are the real positions. The control that would have caught him wasn't a better model — it was a boring, unglamorous rule: the person who takes a risk must never be the person who checks it. Modern banks now enforce segregation of duties, independent risk reporting lines, and position limits precisely because of the Barings case, which is still taught in every bank's compliance training three decades later.