Classic: The Correlation Trade Gone Wrong

The following was published at RealMoney on May 23rd, 2005.  It’s a little obscure, but indicative of what can happen when too much money pursues an obscure arbitrage.  If nothing else, the piece tries to explain a complex concept to those with moderate market knowledge.

Because of the market dislocations last week, I want to give a primer on the class of derivatives that really jolted the markets this week: Indexed Synthetic CDOs.  First, some definitions:

1)      CDO – Collateralized Debt Obligation.  A trust that owns bonds, loans, or credit default swaps.  The ownership in the trust is hierarchical.  There are several classes of certificates that have different interests in the trust, which get defined by which class receives losses from defaults first, second, third, etc.  The earlier that a class (or tranche) receives losses if they occur, the higher the yield a class of certificates receives.

2)      Synthetic – a term used in opposition to “cash.”  One who transacts in corporate bonds participates in the cash market.  The synthetic market in corporate credit is composed of credit default swaps.  It is called “synthetic” because it transacts in corporate credit risk without making loans to corporations.

3)      Credit default swaps – A market participant who buys default protection on a given corporation through the credit default swap market gains the right to deliver a certain amount of defaulted bonds in exchange for the par value of the bonds, when an event of default occurs for the class of bonds covered by the agreement.  In exchange for this privilege, the buyer of protection pays the seller a fixed fee for the term of the swap, which is usually five years, but can vary.

4)      Spread – That fixed fee is called the spread.  When the spread falls, the value of a credit default swap to someone who has previously sold protection becomes more valuable, in the same way that a bond price rises when its yield falls.

5)      Indexed – A third party puts together a seemingly diversified (or focused) list of companies so that investors can invest in a liquid pool of similar companies that they want exposure to, whether on a debt, synthetic, or equity basis.

Indexed Synthetic CDOs gather together the risk of debt default for a group of corporations, and parcel the risk of default out in a concentrated form to those who hold the “first loss” certificates, in exchange for a high yield.  Those who hold other certificates in the loss priority get lesser yields commensurate to the risk of taking losses.

Setting the Stage

The Indexed Synthetic CDO market rallied until March 2005.  In most cases, the more risk an investor took, the better that investor did.  The indexes were rallying.  Those willing to offer protection against the default of a wide number of corporations were willing to do so at smaller and smaller spreads.  As I stated previously on RealMoney, those spreads were too small to compensate for the possibility and severity of losses.

Also, until March of 2005, the decline in spreads was fairly uniform.  There weren’t many credits within each index that were not moving in tune with the rally.  This was significant, because it meant that results were particularly good for the “first loss” investors.  What hurts “first loss” investors are credits going into default.  If the spread on the index as a whole improves (goes lower), but a small minority of credits diverge (get wider) and then default, the “first loss” investor can get hurt, while investors with greater loss protection can still do well.

What Happened Last Week

Last week, not only did spreads rise in general, but some credits related to the auto and auto parts industries widened disproportionately.  This wouldn’t have been such a problem, except that a large number of hedge funds participated in the Indexed Synthetic CDO market doing an esoteric arbitrage trade, where the hedge funds when long the “first loss” piece, and short 2.0-2.5x the “second loss” piece.  This trade was sometimes called the “correlation trade” for reasons I will talk about in a moment.

Why do such a trade?  The lure of free money is inexorable, and the trade had been free money for a while.  So long as movements in the spreads of credits in the index remained closely correlated, the hedge would hold between the “first loss” and “second loss” pieces, and the hedged investment would earn a high riskless yield, which to a hedge fund is the holy grail; a lot of hedge fund of funds will throw money at a strategy like that.

All arbitrages boil down to buying and selling two similar securities, and attempting to profit from the price or yield spread over the anticipated time horizon of the transaction.  Arbitrages can be intelligent or foolish depending on whether the anticipated total return is large enough to compensate for the negative results if the convergence anticipated in the arbitrage does not occur.

Last week, conditions for the hedge did not hold as the credit default swap spreads on automotive-related credits rose, leading the “first loss” pieces to fall in value.  Surprisingly, the “second loss” pieces actually rose in value, as a number of players moved to close out their hedges, which put downward pressure on the prices of the “first loss” pieces, and upward pressure on the prices of the “second loss pieces.  This became self-reinforcing for a while until the close on Tuesday.  On Wednesday, hedge funds and investment banks poured fresh capital into the trade, since the risk reward ratio on the hedged trade was now more attractive, bringing the market back to a more normal state.

Effects on the equity market

This put a damper on the equity market for several reasons: first, some players feared that some of the investment banks were caught on the wrong side of the trade, or had lent to those on the wrong side of the trade.  My guess is that’s not true, but if true, it could raise systemic risk issues, which lowers equity values, as it did in 1998 during the LTCM crisis.  The risk controls at the investment banks are far superior to those at most hedge funds now, and far superior to what they were at the investment banks during LTCM.  That doesn’t mean there can’t be crises, but the preparations for a crisis are better now.  The investment banks have laid off more risks to other market participants.  The other main effect on the equity market was that yields on riskier corporate bonds rose, which usually correlates with lower stock prices.

In closing, just be aware that there are other big markets such as the credit default swap market, both in its single-credit, and indexed forms, that can have a big effect on the equity markets.  There is a lot of leverage around, and “bets gone wrong” can be big enough to knock some confidence out of the markets.  But I offer this hope: so long as the effects of the “bets gone wrong” do not affect major institutions such as investment banks, commercial banks or insurance companies, the effects on the markets should be transitory, as they were after LTCM.