The Equation Series
Menu

Part III: Building a System That Works

Chapter 15

Arbitrage and Relative Value

On 2 September 1998 John Meriwether sent a letter to the investors of Long-Term Capital Management. It told them that the fund had lost 44% in August (Federal Reserve History). The fund was four and a half years old. Meriwether had founded it in Greenwich, Connecticut in February 1994 with the core of the bond-arbitrage group he had run at Salomon Brothers until August 1991, when he resigned after one of his traders, Paul Mozer, was found to have submitted false bids at Treasury auctions (companion account). He had added Myron Scholes and Robert Merton, who would share the Nobel prize in economics in October 1997, and David Mullins, who left the vice-chairmanship of the Federal Reserve Board to join. The fund earned about 43% in 1995 and 41% in 1996 (Federal Reserve History, gross or net not stated); the President's Working Group on Financial Markets, which reported on the fund in April 1999, called those years "about 40 percent" net of fees.

The method was the Salomon desk's. Buy a cheap security, sell an expensive, nearly identical one, and wait for the gap to close. Short the newest 30-year Treasury and buy the slightly older one, because the new bond's liquidity premium fades as it ages. Bet that swap spreads in the US, Britain and Europe would revert to normal. Buy Royal Dutch and sell Shell, two lines of the same oil company that traded at different prices. From 1996 the fund also sold long-dated equity-index options, merger arbitrage and emerging-market debt. At the end of 1997, after returning about $2.7 billion to outside investors without cutting its positions, it had $4.8 billion of capital on 1 January 1998, balance-sheet assets of $129 billion, leverage of 28 to 1 and $1.3 trillion of notional off-balance-sheet derivatives (PWG). It dealt with about 75 repo counterparties and about 50 derivatives counterparties, many of which asked for little or no initial margin because of who the partners were.

Losses began in May. Capital was $4.1 billion at 31 July (PWG). Then Russia devalued and defaulted on its domestic debt in mid-August, and investors everywhere sold whatever was illiquid, unpopular or risky and bought whatever was not. Nearly every LTCM position was long the first kind and short the second. Unrelated spreads in different countries widened together because the same investors were fleeing the same kind of risk. Capital fell to $2.3 billion at 31 August, while assets were still over $125 billion (PWG). The leverage the PWG measured on 1 January capital was "more than 25 to 1"; on the capital that was left, it was about 54 to 1 (companion arithmetic). Banks running copies of LTCM's trades sold too; Salomon Smith Barney had already closed its US bond-arbitrage desk in July (reported; from memory in the companion, verify).

The preview ends here. The full chapter is in the book.

In this chapter

  1. 15.1Same claim, two prices
  2. 15.2What the catalogue finds
  3. 15.3The signature trade that did not pay
  4. 15.4A worked example: right at maturity, wrong on the path
  5. 15.5Who survives the path
  6. 15.6Beyond the United States
  7. 15.7How an expert reads an arbitrage