Long-Term
Capital Management (LTCM)
Monday,
September 06, 2006
2:55
PM
·
Lessons
·
Analysis
The LTCM fiasco is full of lessons
about:
1. Model risk
2. Unexpected correlation or the
breakdown of historical correlations
3. The need for stress-testing
4. The value of disclosure and
transparency
5. The danger of over-generous
extension of trading credit
6. The woes of investing in star
quality
7. And investing too little in
game theory.
The latter because LTCM’s partners
were playing a game up to hilt.
John Meriwether, who founded
Long-Term Capital Partners in 1993, had been head of fixed income trading at
Salomon Brothers. Even when forced to leave Salomon in 1991, in the wake of the
firm’s treasury auction rigging scandal (another marker buoy), Meriwether
continued to command huge loyalty from a team of highly cerebral relative-value
fixed income traders, and considerable respect from the street. Teamed up with
a handful of these traders, two Nobel laureates, Robert Merton and Myron
Scholes, and former regulator David Mullins, Meriwether and LTCM had more
credibility than the average broker/dealer on Wall Street.
It was a game, in that LTCM was
unregulated, free to operate in any market, without capital charges and only
light reporting requirements to the US Securities & Exchange Commission
(SEC). It traded on its good name with many respectable counterparties as if it
was a member of the same club. That meant an ability to put on interest rate
swaps at the market rate for no initial margin - an essential part of its strategy.
It meant being able to borrow 100% of the value of any top-grade collateral, and
with that cash to buy more securities and post them as collateral for further borrowing:
in theory it could leverage itself to infinity. In LTCM’s first two full years of
operation it produced 43% and 41% return on equity and had amassed an investment
capital of $7 billion.
Meriwether was renowned as a
relative-value trader. Relative value means (in theory) taking little outright
market risk, since a long position in one instrument is offset by a short
position in a similar instrument or its derivative. It means betting on small
price differences which are likely to converge over time as the arbitrage is spotted
by the rest of the market and eroded. Trades typical of early LTCM were, for example,
to buy Italian government bonds and sell German Bund futures; to buy theoretically
underpriced off-the-run US treasury bonds (because they are less liquid) and go
short on-the-run (more liquid) treasuries. It played the same arbitrage in the interest-rate
swap market, betting that the spread between swap rates and the most liquid
treasury bonds would narrow. It played long-dated callable Bunds against Dm swaptions.
It was one of the biggest players on the world’s futures exchanges, not only in
debt but also equity products.
To make 40% return on capital,
however, leverage had to be applied. In theory, market risk isn’t increased by
stepping up volume, provided you stick to liquid instruments and don’t get so
big that you yourself become the market.
Some of the big macro hedge funds
had encountered this problem and reduced their size by giving money back to
their investors. When, in the last quarter of 1997 LTCM returned $2.7 billion
to investors, it was assumed to be for the same reason: a prudent reduction in
its positions relative to the market.
But it seems the positions weren’t
reduced relative to the capital reduction, so the leverage increased. Moreover,
other risks had been added to the equation. LTCM played the credit spread
between mortgage-backed securities (including Danish mortgages) or double-A
corporate bonds and the government bond markets. Then it ventured into equity
trades. It sold equity index options, taking big premium in 1997. It took
speculative positions in takeover stocks, according to press reports. One such was
Tellabs whose share price fell over 40% when it failed to take over Ciena, says
one account. A filing with the SEC for June 30 1998 showed that LTCM had equity
stakes in 77 companies, worth $541 million. It also got into emerging markets, including
Some of LTCM’s biggest competitors,
the investment banks, had been clamouring to buy into the fund. Meriwether
applied a formula which brought in new investment, as well as providing him and
his partners with a virtual put option on the performance of the fund. During
1997, under this formula [see separate section below, titled UBS Fiasco], UBS
put in $800 million in the form of a loan and $266 million in straight equity.
Credit Suisse Financial Products put in a $100 million loan and $33 million in
equity. Other loans may have been secured in this way, but they haven’t been
made public. Investors in LTCM were pledged to keep in their money for at least
two years.
LTCM entered 1998 with its capital
reduced to $4.8 billion.
A New York Sunday Times article
says the big trouble for LTCM started on July 17 when Salomon Smith Barney
announced it was liquidating its dollar interest arbitrage positions: "For
the rest of the that month, the fund dropped about 10% because Salomon Brothers
was selling all the things that Long-Term owned." [The article was written
by Michael Lewis, former Salomon bond trader and author of Liar’s Poker. Lewis
visited his former colleagues at LTCM after the crisis and describes some of
the trades on the firm’s books]
On August 17,1998
Most of LTCM’s bets had been
variations on the same theme, convergence between liquid treasuries and more
complex instruments that commanded a credit or liquidity premium. Unfortunately
convergence turned into dramatic divergence.
LTCM’s counterparties, marking
their LTCM exposure to market at least once a day, began to call for more
collateral to cover the divergence. On one single day, August 21, the LTCM
portfolio lost $550 million, writes Lewis. Meriwether and his team, still
convinced of the logic behind their trades, believed all they needed was more capital
to see them through a distorted market.
Perhaps they were right. But
several factors were against LTCM.
1. Who could predict the
time-frame within which rates would converge again?
2. Counterparties had lost
confidence in themselves and LTCM.
3. Many counterparties had put on
the same convergence trades, some of them as disciples of LTCM.
4. Some counterparties saw an
opportunity to trade against LTCM’s known or imagined positions.
In these circumstances, leverage
is not welcome. LTCM was being forced to liquidate to meet margin calls.
The company, which was providing
annual returns of almost 40% up to this point, experienced a
Flight-to-Liquidity. This prompted a bail-out of $3.625 bn by the banks,
organized by the Federal Reserve Bank of
The total losses were found to be
$4.6 billion. The losses in the major investment categories were (ordered by
magnitude):
* $1.6 bn in swaps
* $1.3 bn in equity volatility
* $430 mn in Russian and other emerging markets
* $371 mn in directional trades in developed countries
* $215 mn in yield curve arbitrage
* $203 mn in S&P 500 stocks
* $100 mn in junk bond arbitrage
* no substantial losses in merger arbitrage
Long Term Capital was audited by
Pricewaterhouse LLP. The lead partner on the engagement was John Reville
(Pricewaterhouse LLP -
The significance of the events
surrounding the collapse of Long-Term Capital Management and its subsequent
loss of $4.4 billion should not be underestimated. The breakdowns at LTCM
included an overexposure to leverage, sovereign, model, liquidity, and
volatility risk. In addition, the firm lacked the diverse revenue streams of
the Wall Street investment banks of which it liked to compare itself. The
unraveling of events that led to the firm’s demise include Russia’s
announcement on August 17, 1998 that it was “restructuring” its debt, or lengthening
the terms of the payout on short-term bonds. In actuality this comprised a
“default event” and the markets, with a newly acquired suspicion of all
sovereign instruments, witnessed a mass unwinding of credit risk positions. The
strategists at Long-Term Capital (including two Nobel prize winners, Robert
Merton and Myron Scholes) remained convinced that their mathematical models
would hold up under the stress, and that the markets would behave as they had done
in the past. LTCM had predicted that the markets could only go down by a
certain percentage before they would correct themselves within an assumed time
frame. This did not happen. On one day alone, August 21, 1998, the firm lost $550,000,000.
Half of that money was lost in a single trade: a short position in five-year
equity options. The fund was on the verge of liquidation in mid-September, and
the fear was that this would have caused a large chain reaction through very significant
market disruption as major broker-dealers moved to cover derivative trades with
LTCM. There was an additional fear that the amount of leverage on LTCM’s books
was unknown.
Marketwide Repricing of all Risk
The profits from LTCM's trading strategies were
generally not correlated with each other and thus normally LTCM's highly
leveraged portfolio benefitted from diversification. However, the general
flight to liquidity in the late summer of 1998 led to a marketwide repricing of
all risk and these positions then did all move in the same direction. As the
correlation of LTCM's positions increased, the diversified aspect of LTCM's
portfolio vanished and large losses to its equity value occurred. Thus the
primary lesson of 1998 and the collapse of LTCM for Value at Risk (VaR) users
is not a liquidity one, but more fundamentally that the underlying covariance
matrix used in VaR analysis is not static but changes over time.
Market Can
Stay Irrational Longer Than You Can Stay Solvent
In the end, the basic idea of LTCM
was correct, in that the values of government bonds did eventually converge
after the company was wiped out. Nonetheless, the incident confirms an insight
often (though perhaps apocryphally) attributed to the economist John Maynard
Keynes, who is said to have warned investors that although markets do tend
toward rational positions in the long run, "the market can stay irrational
longer than you can stay solvent."
The fall of LTCM is an important
example of the principle that arbitrage is not riskless. This undermines the
claim of efficient market theorists that markets must converge instantaneously
to efficient prices because of the action of rational investors who will
immediately take advantage of pricing anomalies in markets. Markets are not
perfect devices — they consist of people entering into contracts based on
expectations. When unexpected events depress confidence, investors panic and
reason goes out the window. When this took place in 1929, the stock market
crash that followed led to the Great Depression.
Excessive Leverage
At the time of LTCM's downfall, it had an implied
balance-sheet leverage ratio of more than 25-to-1 (assets of $125+ billion over
equity capital of $4.8 billion). As market conditions worsened, LTCM's
size and leverage, combined with the sheer number of trades it had on its
books, contributed to a serious deterioration in the liquidity of many markets
as LTCM and countless other market participants sought simultaneously to unwind
losing positions.
Stress Testing
Some securities firms did not adequately stress test
their exposures to hedge funds, leading them to underestimate their level of
risk exposure; Of course, during the third quarter of 1998, statistical
measurements of potential exposure became less relevant as market volatility
increased beyond the historical levels incorporated into the risk models. In
other words, many firms did not realize that they were in a “100-year flood” scenario,
and as a result, they did not have adequate safeguards built into their models
to alert them to this fact. And they often did not factor concentration and
liquidity risks into their risk assumptions.
1. LTCM's risk management
Despite the presence of Nobel
laureates closely identified with option theory it seems LTCM relied too much
on theoretical market-risk models and not enough on stress-testing, gap risk
and liquidity risk. There was an assumption that the portfolio was sufficiently
diversified across world markets to produce low correlation. But in most
markets LTCM was replicating basically the same credit spread trade. In August and
September 1998 credit spreads widened in practically every market at the same time.
LTCM risk managers kidded
themselves that the resultant net position of LTCM’s derivatives transactions
bore no relations to the billions of dollars of notional underlying
instruments. Each of those instruments and its derivative has a market price
which can shift independently, each is subject to liquidity risk. LTCM sources
apparently complain that the market started trading against its known
positions. That seems like special pleading. Meriwether et al must have been in
the markets long enough to know they are merciless, and to have been just as merciless
themselves. "All they that take the sword shall perish with the
sword."
2. Risk management by LTCM counterparties
Practically the whole street had a
blind spot when it came to LTCM. They forgot the useful discipline of charging
non-bank counterparties initial margin on swap and repo transactions.
Collectively they were responsible for allowing LTCM to build up layer upon
layer of swap and repo positions. They believed that the first-class collateral
they held was sufficient to mitigate their loss if LTCM disappeared. It may
have been over time, but their margin calls to top up deteriorating positions
simply pushed LTCM further towards the brink. Their credit assessment of LTCM
didn’t include a global view of its leverage and its relationship with other
counterparties. A working group on highly leveraged institutions set up by the
Basle Committee on Banking Supervision reported its findings in January, 1999
drawing many lessons from the LTCM case. It criticized the banks for building
up such exposures to such an opaque institution. They had placed a "heavy
reliance on collateralization of direct mark-to-market exposures" the
report said. "This in turn made it possible for banks to compromise other
critical elements of effective credit risk management, including upfront due
diligence, exposure measurement methodologies, the limit setting process, and
ongoing monitoring of counterparty exposure, especially concentrations and
leverage."
The working group also noted that
banks’ "covenants with LTCM did not require the posting of, or increase
in, initial margin as the risk profile of the counterparty changed, for
instance as leverage increased". Another report in June, 1999 by the
Counterparty Risk Management Policy Group, a group of 12 leading investment
banks, suggested many ways in which information-sharing and transparency could
be improved. It noted the importance of measuring liquidity risk, and improving
market conventions and market practices, such as charging initial margin.
Securities firms often lent to and
traded with hedge funds without having a comprehensive view of their
creditworthiness, particularly their off-balance sheet positions.
3. Supervision
Supervisors themselves showed a certain blinkered view
when it came to banks’ and securities firms’ relationships with hedge funds,
and a huge fund like LTCM in particular. The US Securities & Exchange
Commission (SEC) appears to assess the risk run by individual broker dealers,
without having enough regard for what is happening in the sector as a whole, or
in the firms’ unregulated subsidiaries.
The sad truth revealed by this
testimony is that the SEC and the NYSE were concerned only with the risk ratios
of their registered firms and were ignorant and unconcerned, as were the firms
themselves, about the market’s aggregate exposure to LTCM. Bank of England
experts note the absence of any covenant between LTCM and its counterparties
that would have obliged LTCM to disclose its overall gearing.
______________________________ Menu ______________________________
Executive
Summary - Market Inefficiencies -
Risk Premia - Illiquidity
- Eventful Periods - Properties of Returns - Performance Measurements - Risk Management - Portfolio Context -
Conditional Performance
Evaluation