CHAPTER 30

Convexity, Timing, and Quanto Adjustments

Practice Questions

30.1
Suppose first that the correlation between the underlying asset price and interest rates is
negative and we have a long forward contract. When interest rates increase, there will be
a tendency for the asset price to decrease. The increase in interest rates means that an
investor would like a positive payoff to be early and a negative one to be late. The
negative correlation means that a negative payoff is more likely.
When interest rates decrease, there will be a tendency for the asset price to increase. The
decrease in interest rates means that an investor would like a positive payoff to be late and
a negative one to be early. The negative correlation means that a positive payoff is more
likely.
This argument shows that a negative correlation works in the investor’s favor. Similarly,
a positive correlation works against the investor’s interests.

30.2
(a) A convexity adjustment is necessary for the swap rate.
(b) No convexity or timing adjustments are necessary.

30.3
There are two differences. The discounting is done over a 1.0-year period instead of over
a 1.25-year period. Also a convexity adjustment to the forward rate is necessary. From
equation (30.2), the convexity adjustment is:
007 2  022  025 1
 000005
1  025  007
or about half a basis point.
In the formula for the caplet, we set Fk  007005 instead of 007 . This means that
d1  05642 and d2  07642 . The caplet price becomes
025 10e006510[007005N (05642)  008N (07642)]  0.0531

30.4
The convexity adjustment discussed in Section 30.1 leads to the instrument being worth
an amount slightly different from zero. Define G ( y ) as the value as seen in five years of
a two-year bond with a coupon of 10% as a function of its yield.
01 11
G( y)  
1  y (1  y ) 2

01 22
G( y )   
(1  y ) (1  y )3
2
 02 66
G( y )  
(1  y ) (1  y ) 4
3

It follows that G(01)  17355 and G(01)  46582 and the convexity adjustment that
must be made for the two-year swap- rate is
46582
05  012  022  5   000268
17355
We can therefore value the instrument on the assumption that the swap rate will be
10.268% in five years. The value of the instrument is
0268
 0167
115
or $0.167.

30.5
In this case, we have to make a timing adjustment as well as a convexity adjustment to the
forward swap rate. For (a), equation (30.4) shows that the timing adjustment involves
multiplying the swap rate by
 08  020  020  01 5 
exp     09856
 1  01
so that it becomes 10268  09856  10120 . The value of the instrument is
0120
 0068
116
or $0.068.
For (b), equation (30.4) shows that the timing adjustment involves multiplying the swap
rate by
 095  02  02  01 2  5 
exp     09660
 1  01
so that it becomes 10268  0966  9919 . The value of the instrument is now
0081
  0042
117
or –$0.042.

30.6
(a) The process for y is
dy   y dt   y y dz
The forward bond price is G ( y ) . From Itô’s lemma, its process is
1
d [G ( y )]  [G ( y ) y  G ( y ) y2 y 2 ] dt  G ( y ) y y dz
2

(b) Since the expected growth rate of G ( y ) is zero

(c) Assuming as an approximation that y always equals its initial value of y0 , this
 shows that the growth rate of y is
1 G ( y0 ) 2
  y y0
2 G ( y0 )
The variable y starts at y0 and ends as yT . The convexity adjustment to y0
when we are calculating the expected value of yT in a world that is defined by a
numeraire equal to a zero-coupon bond maturing at time T is approximately y0T
times this or
1 G( y0 ) 2 2
  y y0 T
2 G( y0 )
This is consistent with equation (30.1).

30.7
(a) In the traditional risk-neutral world, the process followed by S is
dS  (r  q)S dt   S S dz
where r is the instantaneous risk-free rate. The market price of dz -risk is zero.
(b) In the traditional risk-neutral world for currency B, the process is
dS  (r  q  QS S Q )S dt   S S dz
where Q is the exchange rate (units of A per unit of B),  Q is the volatility of Q
and QS is the coefficient of correlation between Q and S . The market price of dz -
risk is QS Q .
(c) In a world that is defined by a numeraire equal to a zero-coupon bond in currency A
maturing at time T,
dS  (r  q   S P )S dt   S S dz
where  P is the bond price volatility. The market price of dz -risk is  P
(d) In a world that is defined by a numeraire equal to a zero-coupon bond in currency B
maturing at time T,
dS  (r  q   S P  FS S F )S dt   S S dz
where F is the forward exchange rate,  F is the volatility of F (units of A per unit
of B, and  FS is the correlation between F and S . The market price of dz -risk is
 P  FS F .

30.8
Define:
P(t,T): Price in yen at time t of a bond paying 1 yen at time T
ET(.): Expectation in world that is defined by numeraire P(t T )
F: Dollar forward price of gold for a contract maturing at time T
F0: Value of F at time zero
F: Volatility of F
G: Forward exchange rate (dollars per yen)
G: Volatility of G
We assume that ST is lognormal. We can work in a world that is defined by numeraire
P(t T ) to get the value of the call as
P(0 T )[ ET (ST ) N (d1 )  N (d2 )]
 where
ln[ ET ( ST )  K ]   F2 T  2
d1 
F T

ln[ ET ( ST )  K ]   F2 T  2
d2 
F T
The expected gold price in a world that is defined by a numeraire equal to a zero-coupon
dollar bond maturing at time T is F0 . It follows from equation (30.6) that
ET (ST )  F0 (1   F GT )
Hence the option price, measured in yen, is
P(0 T )[ F0 (1   F GT ) N (d1 )  KN (d2 )]
where
ln[ F0 (1   F  GT )  K ]   F2 T  2
d1 
F T

ln[ F0 (1   F  GT )  K ]   F2 T  2
d2 
F T

30.9
(a) The value of the option can be calculated by setting S 0  400 , K  400 , r  006 ,
q  003 ,   02 , and T  2 . With 100 time steps, the value (in Canadian dollars) is
52.92.
(b) The growth rate of the index using the CDN numeraire is 006  003 or 3%. When
we switch to the USD numeraire, we increase the growth rate of the index by
04  02  006 or 048 % per year to 3.48%. The option can therefore be calculated
using DerivaGem with S 0  400 , K  400 , r  004 , q  004  00348  00052 ,
  02 , and T  2 . With 100 time steps, DerivaGem gives the value as 57.51.

30.10
(a) We require the expected value of the Nikkei index in a dollar risk-neutral world. In a
yen risk-neutral world, the expected value of the index is
20 000e(002001)2  20 40403 . In a dollar risk-neutral world, the analysis in Section
30.3 shows that this becomes
20 40403e030200122  20 69997
The value of the instrument is therefore,
20 69997e0042  1910848

(b) An amount SQ yen is invested in the Nikkei. Its value in yen changes to
 S 
SQ 1  
 S 
In dollars this is worth
1  S  S
SQ
Q  Q
 where Q is the increase in Q . When terms of order two and higher are ignored, the
dollar value becomes
S (1  S  S  Q  Q)
The gain on the Nikkei position is therefore S  S Q  Q
When SQ yen are shorted the gain in dollars is
1 1 
SQ   
 Q Q  Q 
This equals S Q  Q when terms of order two and higher are ignored. The gain on
the whole position is therefore S as required.

(c) In this case, the investor invests $20,000 in the Nikkei. The investor converts the
funds to yen and buys 100 times the index. The index rises to 20,050 so that the
investment becomes worth 2,005,000 yen or
2 005 000
 2011033
997
dollars. The investor therefore gains $110.33. The investor also shorts 2,000,000 yen.
The value of the yen changes from $0.0100 to $0.01003. The investor therefore loses
000003  2 000 000  60 dollars on the short position. The net gain is 50.33 dollars.
This is close to the required gain of $50.

(d) Suppose that the value of the instrument is V . When the index changes by S yen
the value of the instrument changes by
V
S
S
dollars. We can calculate V  S . Part (b) of this question shows how to manufacture
an instrument that changes by S dollars. This enables us to delta-hedge our
exposure to the index.

30.11
To calculate the convexity adjustment for the five-year rate, define the price of a five year
bond, as a function of its yield as
G ( y )  e 5 y

G( y)  5e5 y

G( y)  25e5 y
The convexity adjustment is
05  0082  0252  4  5  0004
Similarly, for the two year rate the convexity adjustment is
05  0082  0252  4  2  00016
We can therefore value the derivative by assuming that the five year rate is 8.4% and the
two-year rate is 8.16%. The value of the derivative is
024e0084  0174
If the payoff occurs in five years rather than four years, it is necessary to make a timing
adjustment. From equation (30.4) this involves multiplying the forward rate by
  1 025  025  008  4 1 
exp     098165
 108
The value of the derivative is
0.24×0.98165e-0.08×5=0.158.

30.12
(a) In this case, we must make a convexity adjustment to the forward swap rate.
Define
6
4 100
G( y)   
i 1 (1  y  2) (1  y  2) 6
i

so that
6
2i 300
G( y )   i 1

i 1 (1  y  2) (1  y  2) 7

6
i (i  1) 1050
G( y )   i2

i 1 (1  y  2) (1  y  2)8
G(008)  26211 and G(008)  85329 so that the convexity adjustment is
1 85329
 0082  0182 10   000338
2 26211
The adjusted forward swap rate is 008  000338  008338 and the value of the
derivative in millions of dollars is
008338  100
 4.617
10320

(b) When the swap rate is applied to a yen principal, we must make a quanto
adjustment in addition to the convexity adjustment. From Section 30.3, this
involves multiplying the forward swap rate by e02501201810  09474 . (Note that
the correlation is the correlation between the dollar per yen exchange rate and the
swap rate. It is therefore 025 rather than 025 .) The value of the derivative in
millions of yen is
008338  09474 100
 6.474
10120