International Fixed Income
Topic 6B:LTCM
1
(From last time)
B. Peso Crisis
Bythe end of 1993, the MP traded at 3.1/$ and was
overvalued by PPP measures (3.5-4.1). [Managed float].
» Mexico’s international reserves surged towards 30bn by year-end.
» In March, these reserves dropped 10bn when presidential candidate
Colosio was assassinated. Mexico’s reserves were almost completely
depleted through November 1994.
» December 1, the new president, Zadillo, took office and shortly after
devalued the peso, letting it float freely.
2
Mexico’s International Reserves
30
25
20
15 Reserves
10
5
0
Ja M M J Se No Ja M M J Se No Ja M
n- ar- ay- ul- p- v- n- ar- ay- ul- p- v- n- ar-
93 93 93 93 93 93 94 94 94 94 94 94 95 95
3
0
1
2
3
4
5
6
7
8
10/11/1994
10/25/1994
11/8/1994
11/22/1994
12/6/1994
12/20/1994
1/3/1995
1/17/1995
1/31/1995
2/14/1995
Mexican Peso
2/28/1995
3/14/1995
3/28/1995
4
PESO
0
1
2
3
4
5
6
7
8
10/11/1994
10/25/1994
11/8/1994
11/22/1994
12/6/1994
12/20/1994
1/3/1995
1/17/1995
1/31/1995
2/14/1995
2/28/1995
Mexican Peso
3/14/1995
3/28/1995
0
10
20
30
40
50
60
70
5
PESO
Brady
Cetes & Currency
90 9
80 8
70 7
60 6
Cetes-91
50 5
Cetes-182
40 4
MP/$
30 3
20 2
10 1
0 0
5
5
4
5
5
95
95
96
l-9
-9
-9
-9
-9
p-
n-
n-
ov
ov
ar
ay
Ju
Ja
Ja
Se
N
N
M
M
6
C. Russian Default
By June 1998, the stock market was at a two-year low; and
the gov’t failed to collect enough funds from T-bill auctions
to repay outstanding debt.
August was a bad month
» 3-mth moratorium on debt payments
» 50% devaluation of ruble
» halting of FX trading
» debt rescheduling that would lead to unprecedented losses
September
» Liberal reformers resigned
» Russian stock market dropped over 45%
7
Ruble
1400 35
1200 30
1000 25
800 20 Russ. T-bills
600 15 Ruble/$
400 10
200 5
0 0
1/1/1998 1/1/1999 1/1/2000
8
0
10
20
30
40
50
60
70
80
4/3/1996
10/3/1996
4/3/1997
10/3/1997
4/3/1998
10/3/1998
4/3/1999
10/3/1999
4/3/2000
Russia, $-denominated 3%, 2003
9
Russia Bond
Outline for LTCM lecture
Chronology of events
Review one way of creating leverage in fixed income
markets
Examples
» Off-the-run versus on-the-run Trreasuries
» $ swaps versus Treasuries
» International government bonds
» International credit spreads
Summary
10
Chronology
After being forced to leave Salomon in 1991, in wake of the Treasury auction
rigging scandal, John Meriwether founded Long Term Capital Management in
late 1993.
There was an exodus from Salomon during this period, giving LTCM
(sometimes called Salomon North) a “dream team” of traders/researchers in the
fixed income area.
LTCM’s “original” strategy was to take advantage of spreads between liquid
and illiquid instruments in similar markets, and bank on a convergence of these
securities - a so-called date with destiny. These strategies are highly levered
because they are viewed as relatively riskless in the long run.
In their first (shortened) year (1994), LTCM earned 20%, with staggering
returns of 43% and 41% (in 1995&1996). In 1997, their returns dropped to
19%, partly because they were getting to big to get the same levered returns.
With amassed capital of $7 billion, they returned 2.7 billion to investors,
leaving only the original investors and themselves. The size of their positions,
however, were maintained.
11
Chronology continued…
Their trade positions had also ventured out of their usual “convergence” trades
into other areas, such as risk arbitrage, credit spread bets, emerging market debt,
corporate bond spreads, etc…
Also, in 1998, the partners bought an option from UBS, which in essence was
structured such that UBS got some equity but mostly provided a loan (which
gave them temporary significant equity in LTCM, which would get bought out
by the partners later). The motivation on LTCM’s part was tax driven.
In early summer,
» LTCM reported its first losses to investors, primarily in the MBS area.
» On July 17th, Salomon closed down its fixed income arbitrage group (to get out of
that type of business); this induced a price pressure effect which led to a further drop
of around 10%.
» On August 17, Russia defaulted on its debt, with an immediate flight to “quality”, i.e.,
to liquid instruments, the opposite of LTCM’s trades.
» On August 21, swap spreads moved 20 basis points (a normal move would be around
1-2bps). LTCM lost $550 million.
» On Sept. 2, Merriwether sent a letter stating the fund had lost $2.5 billion over the
year (or 52% of its value), with August producing $1.8 billion in losses.
» Soon after that date, LTCM began to look for an infusion of capital, including
meetings with Soros, Robertson, Buffet, Goldman Sachs, etc...
12
Chronology continued….
» In the third week of September, Bear Stearns (LTCM’s clearing agent) asked for
$500 mm in collateral to continue clearing trades. On Friday, September 18th, with
market conditions deteriorating, there was discussion about LTCM’s solvency.
» On Sunday, Sept. 20th,the NY Fed looked at LTCM’s portfolio and found that LTCM
had over $50 billion of long-short positions in securities, notional amounts of futures
contracts of $500 billion, notional swaps of $750 billion, and options over $150
billion. It was clear that LTCM failure would affect many counterparties.
» On Monday, Sept. 21st, they lost another $500 mm, half of it on short positions in
long-term equity options. LTCM’s capital base was now only $600mm. In the
evening, a consortium of banks and investment banks met to discuss what to do.
» On Tuesday, Sept. 22nd, Warren Buffet offered to buy the portfolio for $250mm,
recapitalize it with $3 billion, with no management role for Meriwether. They had
until 12:30pm to decide whether or not to do it. The offer was declined.
» That evening, an alternative plan was devised, which required 13 banks to put in
about $250 mm each, and then to oversee the management of the portfolio. There
would be no fire sale of assets. The existing investors, including the LTCM partners,
would maintain a 10% ownership, or roughly $400 mm.
13
Chronology continued…
In the first two weeks after the bail-out, LTCM continued to lose money. There
were even rumors, false as it turns out, that the consortium was in trouble. By
mid-November, markets had clamed down (e.g., Brazil did not default, the
Asian crisis was diminishing, etc…), and the trades began to pay off.
By June, 1999, the fund was up 14.1%, net of fees, from the bailout date.
Monies were gradually being paid back to the original investors and about $1
billion was returned to the consortium.
By December 1999, the fund was “officially” closed, with the consortium
getting their money back, and Meriwether announcing a new fund, Relative
Value, keeping just a few of the original Salomon Partners.
Who really lost ex-post?
» The LTCM partners lost $1.3 billion, most of their stake in the fund.
» UBS lost $690 million, most of the value of their equity and the loan they had put in
to the fund.
» Bank of Italy, Sumitomo Bank, and Dresdner Bank lost around $100mm.
» A number of investors lost money in the $25-50mm range, including Merill Lynch’s
deferred compensation plan.
14
The Repo Market
The repo market (along with bank debt) is the primary way
institutions borrow/lend in fixed income markets. (we will
discuss this market in depth in a few weeks).
Simultaneously borrow $X from the dealer to purchase
securities, and then lend the same securities as collateral to
the dealer.
Note that the haircut (i.e., margin) represents the dealer's
protection against the collateral losing value.
Institution repays the dealer $X plus interest in return for
the security. Institution bears market risk (credit risk aside).
15
Example 1: Off- versus On-the-Run
Treasuries
Consider the 6.125 11/15/2027 off-the-run Treasury versus
the 5.5 08/15/28 on-the-run Treasury.
Characteristics of Bonds on 10/05/98
1
1 0
/7/
1 8
5 /8
/0 5
2 2
0
2
2 1
i
e
l
d
Y 4
.
8 4
5 .
2
7
u
r
a
t
i
o
n
D 1
4
.
5 5
213
4
.
C
o
n
v
e
x
i
t
y 1 3
7 4
3 7
16
The Strategy
Thebasic idea of the strategy is to buy off-the-runs and
short the on-the-runs.
» Since their cash flows should be pretty similar, they should have
similar yields (because they are governed by the same discount
rates). In fact, since the yield curve is upward sloping, the higher
coupon bond (i.e., the off-the-run) should have a slightly lower yield
as it relatively emphasizes earlier cash flows which are governed by
lower rates.
This was one of the strategies of Long Term Capital
Management, and, in fact, had been one for many years.
The coupon effect aside, the figure on the next page graphs
the spread between the on-the-run and the closest off-the-
run long-term treasury bond.
17
Off- versus On-the-Run Spreads
a is
sn
i
st
BoP
6
1
p
r
e
Sd
a
4
1
2
1
0
1
8
6
4
2
0
/
1/
52 /
91 /
6 6 7 7 8 8 99 1
32 /
71 /
02 /
4 7/1/
205
/
2 /
2 6/
6 77 88 8 9 9
2 /
0 3/ 73 /
1/ 42
11 /8
a
de
t
18
Explanation
a 5 basis point spread was not unusual. The
Historically,
coupon effect aside, and assuming that any duration is
hedged out in a portfolio context, this means that on every
$1 mm invested, Long Term Capital was earning $5,000 for
a .5% annualized return.
» Now assume Long term Capital took a $1 billion dollar position on
all the off-the-run/on-the-run trades combined, but with only $10
mm of capital (i.e., 1% haircut).
» With $10mm capital, they expect to earn $5mm for a 50%
annualized return.
What actually happened?
» The spreads widened 12-14 basis points, perhaps a great long-term
trade, but giving a short-term mark-to-market loss of about 1% (after
some calculations) of the $1 billion --- in other words, the entire
$10mm capital invested. 19
What happened after the bailout?
a is
sn
i
st
BoP
0
2
p
r
e
Sd
a
5
1
0
1
5
0
/ 6 / 7 / 8 / 9 0 16
19 20
5 17 2
6 / 7 / 8/ 9 2 5 1 1
2 3 14 / 1 / 0 /
7 /0 2
2
1 1 /1 9
/ 72
22 / 811 9
6 0 / 8/ 2 1 11
3
6 / 7 / 8 / 9/
7 18 0 / 1
34 /1 /
1 1
92 6
a
t
de
20
Example 2: swap spreads
A plain vanilla interest rate swap is one in which you
exchange fixed cash flows for floating-rate cash flows
(often benchmarked against LIBOR (London Interbank
Offer Rate, appropriate for A-rated London banks on their
dollar deposits).
As we will learn about swaps later in the course, just realize
for now that a swap has very little credit risk; so the fixed
rate from a swap will reflect the par rate on Treasuries plus
the curve of the LIBOR spread above Treasuries.
The figure graphs the swap spreads over the period.
21
Swap Spreads
a i
sn
i
st
BoPs
/
1
84
/
2
84
0
0
1 /
0
15
0
0
8
0
6
0
4
0
2
0
y
r
2 y
r
3 5
y 7
r yr 0 3
y 0
1r r
y
Mtt
u
ai
ry
22
The Strategy
One of Long-Term Capital's strategies was to be long swaps
and short treasuries --- the thinking being that swaps
spreads were going to tighten.
Over the weekend of August 21st-23rd, Russia essentially
defaulted on its debt. The swap spread widened about 20
basis points during this period, which saddled Long Term
Capital with substantial losses (as their positions were
highly levered).
23
What happened after the bailout?
a i
s n
i
st
BoPs
/
1
84
/
2
84
0
0
1 /
0
15
0
1
/
18
1
0
8
0
6
0
4
0
2
0
y
r
2 3
y
r y
r
5 7
yr 0 3
y
1r y
r
0
Mtt
u
ai
ry
24
Example 3: European convergence
trade
With approach of European Monetary Union, one popular
trade was to expect foreign gov't bonds to trade at similar
yields, as their discount rates would eventually converge.
One popular one was to long Italian governments and short
German government bonds, as Italian gov'ts were offering
higher yields.
25
Italian vs. German gov’t bonds
rS e 88
m ro4
u
e ru n 9
c
T t t s //
1 S cu
m e1 8
r u o5
er r n/0
/9
T t t s0
a
re
t t
a
re
Grn
ema Grn
e
m a
a
Ii
ta
ln ai
l
t n
Ia
5
.
5 55
.
5 5
.
4
5 5
.
4
4 4
.
5
3 .
5
3
3 3
y r r r
r3 70
1 yy y3 1yr y
y ry r
r3 7 0
3
y r r
r50
2 yy1 2 yy
yr r
r510
mtt
u
ai
ry mty
u
ai
t
r
26
The Strategy
Rather than converging, the government bonds actually
diverged in yields, leaving the highly levered Long Term
Capital saddled with losses. In essence, their invested
capital on these trades has disappeared.
27
What happened after the bailout?
Te8
es
r u9
m8
S1
t o
r n
u1
c /
t 1
r/
a
t
e
r
e
r
Ga
n
m
t
a
l
i
a
In
.
5
5
5
.
5
4
4
.
5
3
3
y
rrrrr rr
2 5 1
y
1 3 7 3
y y 0
y 0yy
mt
u
r
i
t
ay
28
Example 4: European credit spreads
As you’ll see later in the course, swap spreads represent the
spread between future LIBOR (bank offer rates) and future
T-bills. During the 1996-1997 period, UK swap spreads
began to rise relative to German swap spreads. In other
words, UK swaps (relative to their treasuries) were cheap
compared to German swaps (relative to their treasuries).
LTCM went long UK swap spreads and short German swap
spreads, betting they would eventually converge.
29
Swap Spreads: UK & Germany
1.4
1.2
1
0.8
UK spd
0.6
Ger. Spd
0.4
0.2
0
1/8/1996
-0.2 1/8/1997 1/8/1998 1/8/1999 1/8/2000
30
0
1
0.2
0.4
0.6
0.8
-0.4
-0.2
1/8/1996
5/8/1996
9/8/1996
1/8/1997
5/8/1997
9/8/1997
1/8/1998
5/8/1998
9/8/1998
1/8/1999
5/8/1999
9/8/1999
1/8/2000
Swap Spreads: UK & Germany
Diff.
31
Summary
Infact, we could have described relative trades in the
Commercial Mortgage-Backed Securities market, the
straight MBS market, the spread between corporate bonds
and governments in various markets, Danish MBSs, risk
arbitrage trades in the equity market, volatility trades
(especially in the French stock market), etc... Nearly every
one of the trades moved against Long Term Capital in the
summer of 1998, with extreme moves in late August &
September. As we have seen, leverage, while usually
increasing the returns, can wipe out the capital quickly.
32