How did one trader destroy a 233-year-old bank?

Barings was founded in 1762. It financed the Louisiana Purchase, banked the British monarchy, and survived every crisis of two centuries. In February 1995 it collapsed and was sold to ING for £1. The cause is usually written up as "derivatives". The actual cause is more specific and far more useful.

What happened

Nick Leeson ran Barings Futures Singapore. His mandate was low-risk arbitrage: the Nikkei 225 futures contract traded on both SIMEX in Singapore and the Osaka exchange, and small price differences between the two could be captured by buying on one and selling the other. Done properly this is nearly riskless, because the two positions offset.

He did not do it properly. He took directional positions — outright bets on the Nikkei's level — and concealed the losses in an unreconciled error account numbered 88888. He also sold straddles on the Nikkei: selling both a call and a put, collecting two premiums, profitable only if the index stayed in a range.

On 17 January 1995 the Kobe earthquake struck. The Nikkei fell sharply. The short straddles moved deeply against him and the concealed losses grew. Rather than close, he increased the position, buying large quantities of Nikkei futures in an attempt to support the index itself. The index did not cooperate.

Total loss: approximately £827 million — more than the entire capital of the bank. Barings was declared insolvent on 26 February 1995.

The arithmetic of the short straddle

Use rounded illustrative numbers with the index near 19,000. Sell a straddle at that strike and collect, say, 400 index points in total premium.

The bottom row is a 15% fall — large but entirely possible. The loss is more than six times the premium collected, and every further point of decline adds another point of loss. The column's only arithmetic bound is the index reaching zero — 18,600 points net of premium — which is not a limit anybody can plan around, and the call leg on the other side has no bound even in principle. This is the seller's asymmetry from Unit 2, at institutional size.

Then note the compounding decision: the response was to buy futures on top. A losing short-volatility position was funded by adding a leveraged directional one. Two independent ways to lose, stacked.

What actually caused it — and it is not futures

1. No segregation of duties. Leeson ran the trading desk and the back office that settled and reconciled his own trades. He could book a trade and then confirm it himself. This single control failure made everything else possible, and it has nothing to do with derivatives — the same arrangement would have permitted the same fraud in bonds or shares.
2. An unreconciled account nobody audited. Account 88888 was excluded from reports sent to London for years.
3. Margin funded without inquiry. Head office wired hundreds of millions to Singapore to meet margin calls, without anyone establishing what positions required them. A margin call is a question about a position; nobody asked it.
4. Incentives on reported, not verified, profit. Leeson was the star of the franchise, and reported profits were never reconciled against actual positions.

The instruments behaved exactly as designed. A short straddle paid out what a short straddle pays out; the futures marked to market daily and demanded the cash that a futures position demands. Leverage plus opacity destroyed Barings. The derivatives were the medium, not the mechanism — and that distinction is the whole point of this unit.

Try it now

1. Rebuild the straddle table for a 25% decline instead of 15%, and express the loss as a multiple of the premium collected. Then write down the worst the downside column can reach — the index at zero — and note how far past any tolerable loss it already sits. The upside column has no bound at all.
2. The index itself is below, over its full history. Navigate to January–February 1995. The earthquake struck on the 17th, but the index slid for several sessions before the cliff arrives on 23 January, a 5.6% fall in one day — the shape to notice is that the worst session came six days after the news, not with it. Measure the fall from the January level to the February low and look at the shape of it rather than the level.

3. Name the one control that would have stopped this without banning a single instrument. If your answer involves segregation of duties, you have read the case correctly.