Lecture 4

Forwards and Futures on Currencies

HKUST, FINA3203, G. Panayotov

 Examples
U.S. dollar (USD): world’s dominant currency
▶ large proportion of the reserves of central banks (and major financial institutions)
▶ predominantly used for international trade and financial transactions
▶ a safe-heaven currency (park your funds there in times of trouble, similar to gold)
▶ however, the above does not mean it is risk-free (sometimes it fluctuates quite a lot)!

Euro (EUR): the other major reserve currency, but not competitor (yet)
▶ can move against USD
▶ other major: Japanese yen (JPY), British pound (GBP), Swiss franc (CHF), ...

Australian dollar (AUD): commodity currency

▶ countries that depend heavily on the export of certain raw materials
▶ other examples: Canadian dollar (CAD), New Zealand dollar (NZD) ...

Hong Kong dollar (HKD): pegged currency

▶ (almost) fixed exchange rate against USD
▶ other examples: Saudi riyal (SAR), Qatar riyal (QAR)
HKUST, FINA3203, G. Panayotov
 Exchange rate quotes

How rates are quoted is extremely important! Mistakes can lead to disasters (or at
least losing your job)!!

XXX/YYY typically means how much of the quote currency (YYY) is needed to buy
one unit of the base currency (XXX).

But never use certain quote unless you are perfectly sure what it means!

Well, you will not make a mistake with JPY (say 104 JPY per 1 USD).

But how about USD/CHF=1.01 and CHF/USD=0.9901?

For further details, see ”Lehman FX [Link]” or ”Merrill Lynch FX primer

[Link]” posted on Canvas.
HKUST, FINA3203, G. Panayotov
 Hong Kong vs. London
A U.S. asset management firm needs to invest $1 million in government bonds for
one year. It has two choices: Hong Kong and London

The one-year interest rates in the three economies are all r = 2.96% (continuously
compounded)

Today, the exchange rates (quoted as US dollars per one unit of foreign currency) are:

HKD/USD = S0HKD = 0.129

GBP/USD = S0GBP = 1.429

Investors know that HKD has constant (pegged) exchange rate to the USD:
S0HKD = STHKD = 0.129

Let’s also assume that investors expect GBP to increase to STGBP = 1.658 or decrease
to STGBP = 1.20 with equal probability

HKUST, FINA3203, G. Panayotov

 Hong Kong vs. London (cont’d)

Table 1: US investor’s payoff at time T = 1

Scenarios Average

GBP appreciates GBP depreciates no change in GBP

GBP/USD exchange rate STGBP 1.658 1.200 1.429

HKD/USD exchange rate STHKD 0.129 0.129 0.129

Payoff if investing in UK bonds $1.1951 mil. $0.8649 mil. $1.03 mil.

Payoff if investing in HK bonds $1.03 mil. $1.03 mil. $1.03 mil.

1.658
1.1951 mil. = × 1 mil. × e 0.0296
1.429
1.200
0.8649 mil. = × 1 mil. × e 0.0296
1.429
0.129
1.03 mil. = × 1 mil. × e 0.0296
0.129
Question: is the USD here the quote currency or the base currency?
HKUST, FINA3203, G. Panayotov
 How can London help to address the FX risk?
Solution: Currency forward and futures intermediated by London’s financial
institutions can help international investors to eliminate the exchange rate risk

Forward and futures contracts:

▶ US investors short GBP forwards (exchange GBP for USD)

▶ British firms go long GBP forwards (exchange USD for GBP)

▶ They are counterparties to the dealers (financial institutions in London)

The currency forward contract:

▶ zero payment (value) at inception t = 0

▶ the long side will purchase GBP at maturity T = 1 by paying the forward price of
F0,T = $1.429 for each £1 (1.429 USD for one GBP)

HKUST, FINA3203, G. Panayotov

 How can London help to address the FX risk? (cont’d)
Table 2: US Investors’ Payoff at Time T = 1

Scenarios Average

GBP appreciates GBP depreciates GBP unchanges

Spot GBP/USD exchange rate STGBP 1.658 1.200 1.429

Spot HKD/USD exchange rate STHKD 0.129 0.129 0.129

Payoff if investing in UK bonds $1.1951 mil. $0.8649 mil. $1.03 mil.

Payoff if shorting 1/1.429 × e 0.0296 mil. forward −$0.1651 mil. $0.1651 mil. $0 mil.

Total payoff $1.03 mil. $1.03 mil. $1.03 mil.

Payoff if investing in HK bonds $1.03 mil. $1.03 mil. $1.03 mil.

Total Payoff $1.03 mil. $1.03 mil. $1.03 mil.

−0.1651 mil. = 1/1.429 × e 0.0296 mil. × (F0,T − 1.658)

0.1651 mil. = 1/1.429 × e 0.0296 mil. × (F0,T − 1.200)
HKUST, FINA3203, G. Panayotov
 The forward exchange rate

Assume the quote currency is USD. The continuously-compounded US interest rate is

r , and the foreign rate is r ∗ .

r ∗ can be viewed as the dividend yield generated from holding the foreign currency.

Then we know from Lecture 2:

∗
F0,T = S0 e (r −r )T

S0 is the spot exchange rate, i.e. how many US dollars we pay to buy one unit of
foreign currency now

F0,T is the forward exchange rate.

HKUST, FINA3203, G. Panayotov

 Quiz questions

Let’s assume that investors believe that the price of the underlying will increase over
the next month. How does that impact the one-month forward price?

Now let’s assume that you are the only person in the world who knows that the US
dollar will appreciate against the foreign currency in the future. Assume also that the
exchange rate is quoted as US dollars per one unit of foreign currency (for example,
1.429 USD per one GBP). How do you use a forward to make a profit?

Will your answer above change if the exchange rate was quoted as units of foreign
currency per one US dollar (for example, 0.7 GBP per one USD)?

HKUST, FINA3203, G. Panayotov

 Squeeze history

The first bear squeeze was operated by the Russian government in 1894, when
considerable speculative short-selling of rubles had developed on the Berlin Bourse

The Czar’s Minister of Finance, Serguei Witte, ordered his agents in Berlin to buy
large amounts of ruble notes for forward delivery

Then Witte forbade suddenly the export of ruble notes from Russia.

It was physically impossible for all short-sellers of rubles to honour their forward
contracts. Russian official forward purchases had amounted to many times the
number of ruble notes outside Russia. The ruble short-sellers were at Witte’s mercy

Witte finally authorised finally the export of extra rubles, but at an artificially high
price

HKUST, FINA3203, G. Panayotov

 Covered interest rate parity (CIP)

∗
Again, the forward rate is F0,T = S0 e (r −r )T
. This is also called the CIP condition

If this equation does not hold, then there is an arbitrage opportunity:

∗
▶ If F0,T > S0 e (r −r )T , you short the forward and borrow U.S. dollars, convert them into
the foreign currency at the spot and buy the underlying (i.e. invest in foreign bonds).
∗
▶ If F0,T < S0 e (r −r )T
, do the opposite. (Can you specify exactly this opposite?)

But sometimes, arbitrage may fail and CIP may be violated

An interesting development was observed during and after the financial crisis in 2008.

HKUST, FINA3203, G. Panayotov

 Calculations for the previous slide

Again, we quote how many US dollars (USD) are needed to buy one unit of foreign
currency (FCU)

Value at time 0 Payoffs at time T

∗
If F0,T > S0 × e(r−r )T :

Short forward 0 F0,T − ST

∗ ∗
Borrow S0 e −r T USD, convert into 0 ST × e (−r +r ∗ )×T − S0 × e (r −r
∗
)×T
∗
e −r T FCU, and invest that at r ∗
∗
Total 0 F0,T − S0 × e (r −r )T (> 0)

∗
If F0,T < S0 × e(r−r )T :

Long forward 0 ST − F0,T

∗ ∗ ∗
Borrow e −r T FCU, convert into 0 S0 × e (r −r )×T − ST × e (−r +r ∗ )×T
∗
S0 × e −r T USD, and invest that at r
∗
Total 0 S0 × e (r −r )T − F0,T (> 0)

HKUST, FINA3203, G. Panayotov

 CIP violations: EUR/USD
Panel A. Pre−Crisis
2
1 year Fwd
1.5 6 month Fwd

1
Percentage

0.5

−0.5

−1
12/30/1998 12/30/1999 12/29/2000 12/31/2001 12/31/2002 12/31/2003 12/30/2004 12/30/2005 1/1/2007

Panel B. Financial Crisis

2
1 year Fwd
1.5 6 month Fwd

1
Percentage

0.5

−0.5

−1
6/1/2007 10/2/2007 1/31/2008 6/2/2008 10/1/2008 1/30/2009 6/2/2009 10/1/2009 2/1/2010 6/2/2010

The graphs show by how much (in percent) did the actual forward price exceed the
theoretical one from our formula

HKUST, FINA3203, G. Panayotov

 CIP violations: GBP/USD
Panel A. Pre−Crisis
4
1 year Fwd
6 month Fwd
2
Percentage

−2

−4
6/15/1990 6/16/1992 6/16/1994 6/17/1996 6/17/1998 6/16/2000 6/18/2002 6/17/2004 6/19/2006

Panel B. Financial Crisis

4
1 year Fwd
3 6 month Fwd
2
Percentage

−1

−2

−3
6/1/2007 10/2/2007 1/31/2008 6/2/2008 10/1/2008 1/30/2009 6/2/2009 10/1/2009 2/1/2010 6/2/2010

HKUST, FINA3203, G. Panayotov

 Why did CIP fail in 2008-2009?
In this case, an arbitrage trade required the following (exactly as on slide 10):
(a) short forward;
(b) borrow dollars;

(c) convert them into Euro;

(d) invest in Euro (i.e. buy Euro bonds)

But point (b) failed during the financial crisis, as dollars became scarce. Because it
was valuable to hold dollars during a financial crisis for liquidity management (like
sometimes it is convenient to hold physically a commodity)
▶ US dollars (typically in the form of US Treasuries) are the main collateral accepted for
short-term borrowing transactions.

▶ Excessive risk aversion during a crisis means that everyone wants to hold US dollars
(the safe asset) so they can use them in such borrowing.

▶ Therefore, it was difficult to find lenders of USD at the time!

So, US Treasuries have “convenience yield” when everyone needs USD as cash collateral

HKUST, FINA3203, G. Panayotov

 More generally, why may arbitrage fail?
Lack of borrowing capacity
▶ Arbitrage trades require long-short strategies and leverage to increase the return.

▶ During market disruptions, banks are excessively risk averse. Borrowing is difficult.

Funding risk
▶ An arbitrage strategy must be held until maturity T to payoff.

▶ During market disruptions, losses are more likely. More frequent margin calls.

▶ Difficult to execute arbitrage strategies due to limited funding capacity.

Fear of withdrawals
▶ Investors withdraw money from hedge funds when they incur losses.

▶ This is more likely to happen during market disruptions and forces the fund to close
positions before arbitrage pays off.

HKUST, FINA3203, G. Panayotov

 Uncovered interest rate parity (UIP)
Assume that the forward exchange rate equals the spot rate that the market expects
today for time T : F0,T = E0 [ST ]

On slide 12 in L2 we said that such an assumption will not hold for a stock, due to a
risk premium. However, we can argue here that an exchange rate always has two
sides, and what is risk for one side is benefit for the other, hence FX is different.

∗
Because CIP (F0,T = S0 e (r −r )T
) typically holds in the data, from the above we get:
∗
E0 [ST ] = S0 e (r −r )T
(UIP)

▶ take logs on both sides of UIP and rearrange, so: ln E0 [ST ] − ln S0 = (r − r ∗ )T

▶ say r < r ∗ . Then UIP says that a low-interest rate currency is expected to appreciate!!
▶ if we have monthly data (T=1/12), we can test UIP in the data using a regression:

ln St+1 − ln St = α + β(rt − rt∗ )T + εt+1

If UIP holds, the estimate of β should be equal to 1. However, a large literature has
found that often this estimate is much less, and can even be negative!! Therefore,
UIP and the assumption made at the top of this slide are suspicious (puzzle).
HKUST, FINA3203, G. Panayotov
 Currency carry trade
The trade is simple: borrow in low-rate currencies and lend in high-rate currencies.

Let USD be the home currency and euro the foreign currency. If the US interest rate
r$ is lower than the euro interest rate re , then borrow in USD and invest in a euro
deposit, so we gain from the interest rate difference.

From the regression on the previous slide we also expect to gain from the change in
exchange rate, because (i) E0 [ST ] = S0 e β(r$ −re )T and (ii) β < 0 and r$ − re < 0.
Therefore, E0 [ST ] > S0 (the euro is expected to appreciate).

Don’t forget that we need to convert back into USD at T to calculate return

So, our return has two sources: extra interest earnings and appreciation of the euro.
Note that this is excess return, because we have done the trade using borrowed
money, so we have already paid the risk-free US rate.

Recall: excess return can only be earned for bearing risk (Finance 101). But what is
the risk in the carry trade? And for whom? These are tough questions – people have
tried to answer them for 40 years, and are still trying...
HKUST, FINA3203, G. Panayotov
The carry trade from US and euro perspectives (∗ not for exam)
The standard way to do carry trade is with forwards. Assume maturity is T = 1 year,
r$ = 2% and re = 3%. The spot exchange rate today (S0 ) is 1 USD per 1 euro, and
the spot exchange rate at time T (ST ) will be 1.02 USD per 1 euro.

The US investor buys one euro forward. The forward price is F0,T = S0 × e (r$ −re )T
which is 0.99 USD per 1 euro notional. She is long euro against USD.

Her payoff at T is ST − F0,T = 1.02 − 0.99 = 0.03 USD. This is a 3% gain if she
invested a USD amount equivalent to the euro notional at the spot exchange rate
today (”full collateralization”) – in our example S0 = 1, so she will need to invest 1
USD. In practice she may just need to post collateral, but let’s ignore leverage.

The European investor doing the same trade sells one USD forward and is short USD
Eur
against euro. The forward price for her is F0,T = 1/0.99 = 1.0101 euro per 1 USD
Eur
notional (we switched the quote!). Also, ST = 1/1.02 = 0.9804 euro per 1 USD.
Eur
Her payoff at T is F0,T − STEur = 1.0101 − 0.9804 = 0.0297 euro. With full
collateralization this is very close to a 3% gain.

Therefore, the US and European investors get the same return from the carry trade
HKUST, FINA3203, G. Panayotov
 Borrowing/lending vs. a forward (∗ not for exam)
Slide 16 says that the carry trade borrows in low-rate currencies and lends in
high-rate currencies. And slide 17 says that the US investor goes long one euro
forward to do the trade. Is that correct?

Let’s see that a currency forward is pretty much like borrowing in one currency,
investing in the other, and at maturity converting back to the original currency:

Let’s say that S0 is the current exchange rate (S0 units of the home currency (USD)
per one unit of the foreign currency (EUR)), ST is the exchange rate at maturity T ,
∗
and the forward price for that maturity today is F = S0 ∗ exp (r −r )T , as in our
definition from slide 8.

Then the payoff of a long forward at maturity, in the home currency, is:
∗
ST – F = ST − S0 ∗ exp (r −r )T .
∗
On the other hand, if we borrow today S0 ∗ exp −r T units of the home currency at
rate r , convert to the foreign currency, invest that at r ∗ , and at time T convert back
∗
to the home currency, our payoff is ST ∗ exp −r ∗T ∗ exp r ∗T –S0 ∗ exp (r −r )T = ST – F ,
exactly as above. Also, in both cases we have a zero cashflow at time 0 (today).

Therefore, the two strategies are equivalent.

HKUST, FINA3203, G. Panayotov
 Case study: Chinese Copper Carries

After 2010, clear temptation to do CNY-USD carry trade: interest rate differential
4-5% and CNY appreciating!

Problem: Capital controls. That is, banks in China (onshore) are not allowed to lend
you USD just so that you can do your carry trade
HKUST, FINA3203, G. Panayotov
 Chinese Copper Carries (cont’d)
So, Chinese firms become creative. One play - using a Letter of Credit (LC)
an LC is a most standard (and allowed) way to finance imports: a bank agrees to
ensure the payment of an import after the importer provides necessary documents.
Key: importers in China are allowed to pay for the LC with some delay (say 6 months)
Then the trade is: buy copper using LC, sell it to the market, invest the proceeds in a
higher yielding CNY deposit. At the payment date for the LC liquidate the deposit
and use it to pay. Rinse and repeat.

HKUST, FINA3203, G. Panayotov

 Chinese Copper Carries (cont’d)
However, every time you do the carry trade above you need to import and sell locally
more and more physical copper (but even the Chinese market has physical limits)

So, how about using the same copper to finance multiple carry trades?

Solution: use warrants of bonded copper. Now instead of moving physical copper,
only the warrants change hands

HKUST, FINA3203, G. Panayotov

 Chinese Copper Carries (cont’d)
Offshore trader A sells warrant of bonded copper to onshore party B. This is copper
in a bonded warehouse (e.g., the free-trade zone at Shanghai’s Waigaoqiao port),
which is still considered ”offshore” - it is exempted from VAT payment, and has not
cleared customs yet. So, B imports copper from A, and A is paid with a USD Letter
of Credit (LC), issued by onshore bank D.

Party B sells the copper to its offshore subsidiary C by sending to C the warrant
documentation (and not the physical copper, which stays in the bonded warehouse).
C pays B cash in USD or CNH (i.e., the offshore CNY). Then B goes to bank D and
converts the USD or CNH into onshore CNY, with which it can do a CNY-USD carry
trade, or can invest in anything else.

Offshore subsidiary C sells the warrant back to A (again, no move in physical copper
which stays in bonded warehouse ”off-shore”), and A pays C in cash with USD or
CNH cash, with some discount for the service.

The above three steps are repeated as many times as possible (before the regulator
intervenes). In this way one obtains multiple LC’s on the same copper warrant. The
same tonne of bonded copper may have been used 10 times or more within a year, a
main limitation being the time to do the paperwork.

HKUST, FINA3203, G. Panayotov

 Variations on the theme
Besides copper, others commodities with high value-to-density ratios have been used
for the same purpose, e.g. gold, nickel, soybeans, palm oil, rubber, zinc, aluminum...

Instead of using ”apparent” imports and LC’s as above, Chinese firms also did
”apparent” exports, possibly on an even larger scale

Exporter could ship their goods in and out of the bonded area frequently, with these
each time being counted as exports. Domestic bank doesn’t mind, because it earns
fees from the guarantee letters, currency conversion and the WMP business.
HKUST, FINA3203, G. Panayotov
What was the likely scale of these trades?

HKUST, FINA3203, G. Panayotov