Toomre Capital Markets LLC

Real-Time Capital Markets -- Analytics, Visualization, Event Processing, and Intelligence

Amaranth Advisors Down to $3.4 Billion with Energy Sale

The New York Times in an article entitled Hedge Fund Sheds Assets in Energy written by Jenny Anderson has more details on the sale of Amaranth Advisors’ energy assets, which was highlighted yesterday afternoon in this previous TCM post. This article confirms that Amaranth sold its entire book of energy trades to J.P. Morgan and Citadel Investments Group. According to a letter sent to the fund group’s investors,

The sale leaves Amaranth, once a well-regarded $9.25 billion fund, with $3.4 billion, down 55 percent from the beginning of the year and down 65 percent for September. (TCM emphasis added)

Without knowing any specific about the particulars of the Amaranth Advisors situation (other than what has been reported publicly), one can categorically point to a complete failure of leading-edge risk management techniques. How the heck does such a reputable hedge fund group with multiple strategies get so off-sides that it can lose 65% in approximately three weeks? As an outsider, one can be stupefied by the Amaranth turn of events; imagine being an investor or even worse a fiduciary with millions of dollars invested with what some might now regard as investment cowboys. This NYT article continues with more information about the shell-shocked investors:

More than half of the [investors] are funds of hedge funds that saw Amaranth as a blue-chip name, stocked with star traders and an impressive track record. Amaranth’s future will depend, in large part, on how quickly investors decide to redeem their funds.

Investor anger is directed not only at the fund’s loss of more than 50 percent of its value in just a week, but the fact that Amaranth — after having a bad month in May, with funds falling 9.5 percent to 10.5 percent in value — apparently reassured investors that it would better manage its risk in the future.

Still, the May losses alarmed some investors, even as the fund continued to show returns of more than 29 percent for the year. Nervous investors who called or turned up at Amaranth’s expansive offices in Greenwich, Conn., trying to determine if the traders were taking too much risk, got a simple message: No.

“After May there was a lot of concern and they spent a lot of time explaining and demonstrating that if the month had ended three days later, losses would be a lot less,” said one investor who spoke on the condition that he not be identified because he is trying to get his money back. “They explained why it wouldn’t happen again.

It has been reported elsewhere that the Amaranth funds started 2006 with assets under management of approximately $7.4 billion. Assuming that reports are correct that Brian Hunter, Amaranth’s head energy trader, had already made $2 billion in profits by the end of April 2006, and then lost approximately $1 billion in May 2006, one has to wonder about how much volatility there was in Amaranth’s daily profit/loss statement.

Value-At-Risk (“VAR”) is one measurement technique that risk managers use to estimate the risk inherent in a portfolio of investment positions. Typically, prospective price movements are simulated based on some period of historic price history and a distribution of prospective results are calculated; then, one estimates the amount of potential losses at various confidence intervals such as 90%, 95%, 99% and even 99.5%.

Assuming the typical interval of 99% and the observed portfolio volatility, the VAR numbers at Amaranth must have been amazing, even before the most recent events. In round numbers, it took Amaranth approximately 80 trading days to make $2 billion and only 20 trading days to lose $1 billion. Without knowledge of the specifics of the Amaranth positions, it is difficult to estimate the distribution of daily P&L results. However, simple averages give some idea of the volatility of the fund; going up it was making approximately $25 million per day and down approximately $50 million per day. The key question is/was what amount volatility should be assumed in assessing possible tail events (i.e. chance of ruin). Clearly, someone overlooked or ignored these important results that already were evident as of May 2006.