Amaranth Lost Six Billion Dollars Betting on Natural Gas Spreads
A multi strategy hedge fund collapsed in 2006 after a single trader accumulated an enormous position in natural gas futures. The fund was diversified in name only.
The Fund
Amaranth Advisors operated as a multi strategy hedge fund, meaning it ran several distinct approaches with capital allocated across them. The premise of that structure is diversification, since different strategies should not fail simultaneously.
In practice, the energy trading book grew to dominate both the risk and the returns. When it failed in September 2006, losses in the region of six billion dollars forced the fund to wind down within weeks.
The Trade
The positions were spread trades in natural gas futures, betting on the price relationship between contracts for different delivery months rather than on the outright direction of gas prices.
Natural gas has pronounced seasonality because of heating demand, and winter contracts typically trade at a premium to shoulder season contracts. Positions were built expecting that relationship to widen.
Spread trades are often described as lower risk than outright positions, because both legs move together and much of the directional exposure cancels. That framing encourages larger position sizes, which is exactly where the danger enters.
A spread trade has less risk per unit, which invites more units. The reduced risk is frequently spent rather than kept.
The Size Problem
The positions reportedly represented a very large share of open interest in the relevant contracts. At that scale the fund was not participating in the market so much as constituting it.
Two consequences follow. Other participants can identify the position and trade against it, knowing that an exit must eventually occur. And the position cannot be reduced without moving prices adversely, so the theoretical ability to cut losses does not exist in practice.
The preceding year had featured hurricane disruption to Gulf production, which had produced exactly the price behaviour the strategy anticipated. When the following season brought no comparable disruption and inventories were ample, the spread moved decisively the other way.
Why the Risk Systems Did Not Prevent It
Standard risk measurement relies substantially on value at risk, which estimates potential loss over a horizon at a confidence level, calibrated on historical price behaviour.
That approach has a specific blind spot. It measures how much the position could lose if prices move as they have historically, and it says nothing about whether the position can actually be closed. A model can report an acceptable figure for a position that would take weeks to exit and would move the market throughout.
Liquidity adjusted measures and position limits relative to open interest address this, and they are precisely the controls that were inadequate here.
The Lesson About Diversification
The fund had multiple strategies and one exposure that mattered. Diversification measured by the number of activities is meaningless if capital and risk concentrate in one of them.
The practical test is contribution to variance. Which book explains most of the fund's return volatility? If one answer dominates, the fund is that book plus some decoration, regardless of the organisational chart.
The Bottom Line
Amaranth held a spread position too large for the market it traded in, inside a fund that was diversified on paper. Measure concentration by contribution to risk, and treat any position large relative to open interest as unexitable.