Understanding the Subprime Crisis: A Narrative, Part Three

(Part One.  Part Two.  9 am – promoted by ek hornbeck)

Part Three: The Rise of Long-Term Capital Management and the Superportfolio

This is an extremely compressed version of the story of Long-Term Capital Management, perhaps the most written about corporate failure besides Enron of the last two decades.  If you wish to learn more, I highly recommend Lowenstein’s book When Genius Failed, linked below.

John Meriwether launched the limited partnership of Long-Term Capital Management in 1994.  Limited partnerships are the actual name of the entities commonly known as “hedge funds”.  The name hedge fund is in fact a misnomer; they originally developed that name because the funds were designed to “hedge” against losses by being more conservative than mutual funds, but have developed into the opposite.  The appeal of such funds is that limitations on the number of partners and the overall wealth of those allowed to join (no more than 99 people or entities of a total value of over $1 million, or 500 people or entities worth over $5 million – with those worth less entirely excluded) are coupled with a nearly total lack of government regulation.  Mutual funds are forced to disclose their portfolios and to maintain certain levels of diversification and leverage; limited partnerships are not.

Joining Meriwether as the partners of LTCM were former Salomon arbitrage group members Larry Hilibrand, Eric Rosenfeld, Victor Haghani, Greg Hawkins, and three very notable additions: economists Robert Merton and Myron Scholes, and David Mullins, who was the number-two at the Federal Reserve under Alan Greenspan and had previously been considered his heir apparent.  With such a roster, LTCM launched with capital of $1.4 billion, the largest such launch in financial history.  It had such disparate investors as the national bank of Italy and the President of Merril Lynch, and had an elegant and innovative structure, with the company that employed the partners and traders being a Delaware-registered management services company employed by a Cayman Islands partnership (the fund itself) financed by six international dummy corporations from whom investors in different nations would buy their shares.

The first hugely successful trade conducted by LTCM had its roots both in the past of Salomon and the future crisis that affects markets today.  Of the myriad of forms of collectivized mortgage obligations invented by Lewis Ranieri at Salomon, the most popular by far was the division of CMOs into interest-only bonds, representing only the interest payments of mortgage-holders, and principal-only bonds, representing only payments of the principal of mortgage loans.  The interest-only bonds, called IOs, had higher rates of return but more volatility, as early repayments would reduce the interest return, while principal-only bonds were the opposite.  In the early 1990s, the Fed had lowered interest rates, leading to many homeowners refinancing their mortgages, driving down the value of IOs relative to POs.  Then, interest rates were increased, shifting the balance again.  In 1994, as LTCM began trading, the Federal Reserve, in one of its last actions as Mullins was leaving the Fed for LTCM, lowered interest rates again.  Strangely, however, this did not lead homeowners to rush to refinance right away again, leading LTCMs partners believing that IOs were considerably undervalued relative to POs.  

LTCM therefore decided to buy IOs while shorting IOs.  To “balance” the trade, limiting the risk, they also purchased a number of Treasury bonds equal to the amount of IOs they purchased, reasoning that since Treasurys increase in value in proportion to the rise in interest rates, while the opposite is true of mortgages, this would insulate them in case of the rise or fall of all CMOs.  The purpose of this is important: what LTCM wanted to do was to make the absolute value of their PO and IO holdings irrelevant and make the only factor whether the difference, or spread, between their values increased or decreased.

LTCM did something else to maximize their advantage in this trade.  They didn’t actually buy the POs or the Treasurys at all.  They simply signed contracts with banks saying they would pay for the banks to purchase the bonds for them, a practice known as “repo financing” in investment banking.  Through the use of repo financing, LTCM leveraged its trades by astronomical proportions; in fact, this was the key to their success.

LTCM shortly began posting remarkable profits.  In 1995, fueled in particular by the IO trade, showed returns of 59% before fees, and 43% after.  During its first two years, LTCM earned $1.6 billion.  But hidden in those numbers were some facts which ought to have caused concern, and impact directly on the problems of today.  At the end of 1995, LTCM’s equity capital (its cash on hand) was $3.6 billion.  But LTCM’s assets (the various bonds, stocks, etc. that they owned) totalled $102 billion.  It had leveraged its capital 28 to 1.  Doing the math, we can find that its actual return on total capital was merely 2.45% – the gaudy number of 59% was due entirely from the willingness of LTCM to magnify its exposure on each trade through massive leverage.

Few things demonstrate better both the appeal and the risk of arbitrage.  Any intelligent person will recognize that people will generally find themselves believing something incorrect in groups.  These people acting as a group will create market inefficiencies, or situations where the value of something is either too high or too low.  However, these inefficiencies will be very small and hard to find.  But what if a group of really smart experts with supercomputers spent their time finding these small inefficiencies, and then found a way of turning the nickels of over and undervalued assets into millions?  Why, it would be a sure way of creating money.  Of course, the problem is that the very existence of the inefficiencies is the risk; by leveraging yourself so highly to turn the nickel an asset is mispriced by into millions, if the inefficiency worsens, say by another nickel, you will lose the same massive amount.  And the underlying irrationality which made the nickel profit possible in the first place is the same irrationality which could worsen, meaning the lever for your destruction is already in place.

Why didn’t LTCM worry about this?  Because of an economic theory advanced by Merton: continuous-time finance.  In simple terms, continous-time finance suggests that the price of an asset, say a share of Google, does not go from $100 to $700 in a leap, but rather from $100 to $100.01 to $100.02 and so on until it reaches $700.  The amount of time in each transition may be a fraction of a second, but this is how the theory claims it works.  Merton was an evangelical proponent of this, and he had many admirers, inside the academic world and outside of it.  However, the teachers of Merton and Scholes were not so convinced.  Paul Samuelson, Merton’s professor at MIT, thought it made for wonderful theory, but felt that the market failure of information inequality swamped such movements in real life, forcing prices to move in fits and spurts.  Eugene Fama, who had been Scholes’ thesis advisor, did a series of studies showing that the bell curve of option prices showed a far higher instance of events more than five standard deviations from the mean than in normal projection (a phenomenon known as “long tails”), which assailed the heart of the Black-Scholes model’s applicability to real-world finance.

But with 59% returns, these contrary views found few listeners within LTCM, or on Wall Street.  LTCM had no trouble raising another billion in capital, and other similar funds along with in-house projects at every major investment bank quickly tried to follow in LTCM’s footsteps.  LTCM was notoriously secretive, to the extent that its partners once bought back the rights of photos of themselves from Buisnessweek to ensure they could not be reprinted, but it was hard to keep the nature of their trades entirely secret.  The academic work which was its basis was, of course, widely available.  And despite a fanatical effort to use different banks for each leg of their trades, some aspects of their trades invariably became known.

The result was that competitors flooded LTCM’s market.  This meant several things: first and most visibly, arbitrage opportunities became harder to come by.  LTCM found that its opportunities became rarer, and that spreads closed far more rapidly.  They ended up taking two major actions in response; in 1997, they returned the capital and profits of many of their investors, saving the bulk of the fund’s equity for their own stakes (which all of them relentlessly pumped into the fund, keeping almost nothing for themselves), and they began to make much riskier trades in arenas where they had much less expertise, such as stocks.

But the deeper problem was that in effect, all of Wall Street was holding a portfolio similar or identical to LTCM’s.  If LTCM made money, Wall Street would make money.  But if other firms lost money, those losses would spill over to LTCM, and vice versa.  This phenomenon, of the entire industry having a portfolio of idenctical or similar assets, has been named the “superportfolio”.  And in the summer of 1998, two things happened which were outside of LTCM’s control which had disasterous consequences.

In 1997, Salomon Brothers, still limping from the massive fines imposed by the SEC and lacking the wealth created by Ranieri’s mortgage group and Meriwether’s arbitrage group, was bought by the notoriously risk-adverse Travelers Corporation.  With the arbitrage desk showing a loss, and having been middling since the departure of Meriwether, chairman Sandy Weill chose to shut down the arbitrage group, and ordered Salomon to liquidate its holdings in early summer 1998.  Because Salomon liquidated its holdings with speed rather than losses being its concern, they sold nearly all their holdings at a loss.  But because nearly every other bank and hedge fund held similar assets, the value of their holdings was also driven down by the liquidation.  As the liquidation caused those firms to lose money, more of them were forced to liquidate their holdings to meet margin calls (done at the end of business each day), which drove the prices down further, and so on.

Meanwhile, in August of 1998, Russia defaulted on its ruble-denominated debt and ordered a three month ban on Russian banks complying with forward contracts in foreign exchange (which many funds had used to hedge against the risk of Russian default).  This had little direct effect on LTCM, which was not heavily invested in Russia.  But other members of the superportfolio were badly exposed.  Credit Suisse lost $1.3 billion.  A fund called High-Risk Opportunities was forced into bankruptcy, while owing huge sums to Bankers Trust, Credit Suisse, and particularly Lehman Brothers, whose exposure was so great that rumors began that they too were headed to bankruptcy.  As with the sell-off of Salomon’s arbitrage holdings, these losses forced more and more members of the superportfolio to sell their assets at current prices.

On this stage, LTCM was poised to lose money in numbers larger and faster than any such loss had ever happened before, requiring the largest bailout in American history.

This post includes information from several sources not available online.  The major sources are: