May 2007 Exam MFE Solutions

1. Answer = (B)

Let D = the quarterly dividend.

Using formula (9.2), put-call parity adjusted for deterministic dividends, we have
4.50 = 2.45 + ⎡⎣52.00 − D × e −0.01 − D × e −0.025 ⎤⎦ − ⎡⎣50 × e −0.03 ⎤⎦
= 54.45 − D × ( 0.99005 + 0.97531) − 50 × 0.970446 .
Rearranging the equation yields
D × 1.96356 = 54.45 − 4.50 − 48.5223 = 1.4277 ,
or
D = 0.73 .

2. Answer = (A)

Let p be the true probability of the stock going up. Thus,

puS + (1 – p)dS = eαh S (which is equation (11.3) on p. 347),
yielding
e αh − d
p= .
u−d
Because α = 0.1, h = 1 , u = 1.433 , and d = 0.756 , we have p = 0.52.

3. Answer = (C)
Let P denote the price of the European put option. Then,
P = 98e −0.055×½ N ( − d 2 ) − 100e −0.01×½ N ( −d1 )
by formula (12.3) with S = 100, K = 98, δ = 1% , σ = 50% , r = 5.5%, and T = ½.

Here, d1 is calculated using formula (12.2a) and is equal to 0.29755819 ≈ 0.30; d 2 is

from formula (12.2b) and is equal to −0.0559952 ≈ −0.06. From the normal cdf table,
N(0.06) = 0.5239 and N(−0.30) = 1 − 0.6179 = 0.3821.

Thus, P ≈ 98e −0.055 / 2 × 0.5239 − 100e −0.01/ 2 × 0.3821 = 11.93 ≈ 11.90.

4. Answer = (E)

For a special put option with strike price K, the payoff upon immediate exercise is
K − 50.

1
This value should be compared with P, the price of the corresponding one-period
European put option. The value of P can be determined using put-call parity:
P = Ke − r − Se −δ + C .
With S = 50, r = 4%, and δ = 8%,
P = K × e −0.04 − 50 × e −0.08 + C = 0.9608 K − 46.1558 + C .

K C P K−50
40 9.12 1.3962 −10
50 4.91 6.7942 0
60 0.71 12.2022 10
70 0.00 21.1002 20

From the table above, we see that it is not optimal to exercise any of these special put
options immediately.

5. Answer = (D)

By (12.9),
σ option = σ stock × | Ω | = 0.50 × |Ω|,
where Ω is the option elasticity.
By (12.8),
Ω = SΔ/C,
where Δ is the option delta,
Δ = e −δ T N (d1 ) (see page 383).
By (12.1),
C = Se −δ T N (d1 ) − Ke − rT N (d 2 )
= SΔ − Ke − rT N (d 2 ) .
Thus,
Ω = SΔ/C
= SΔ/[SΔ − Ke− rT N (d 2 ) ]
= 1/[1 − Ke − rT N (d 2 ) /(SΔ)]
= 1/{1 − [ Ke − rT N (d 2 ) ]/[S e −δ T N (d1 ) ]}.

We are given S = 85, K = 80, δ = 0, r = 5.5%, T = 1.

By equation (12.2a), d1 is 0.4812 ≈ 0.48; hence, N (d1 ) ≈ 0.6844.
By equation (12.2b), d2 is −0.0188 ≈ −0.02; hence, N (d 2 ) ≈ 0.4920.
With these values, we obtain
S e −δ T N (d1 ) ≈ 85×e0×0.6844 = 58.174,
Ke − rT N (d 2 ) ≈ 80×e−0.055×0.4920 = 37.2537.

2
Hence,
Ω = 1/{1 − [ Ke− rT N (d 2 ) ]/[S e −δ T N (d1 ) ]} ≈ 2.78,
and
σ option = σ stock × |Ω| ≈ 0.50 × 2.78 = 1.39 .

Remark added in April 2009: See also page 687 of McDonald (2006).

6. Answer = (C)
Because of the identity
Maximum( S1(3), S2(3) ) = Maximum( S1(3) – S2(3), 0) + S2(3),
the payoff of the claim can be decomposed as the sum of the payoff of the exchange
option in statement (v) of the problem and the price of stock 2 at time 3. In a no-
arbitrage model, the price of the claim must be equal to the sum of the exchange
option price (which is 10) and the prepaid forward price for delivery of stock 2 at
time 3 (which is e−δ2 ×3 ×S2(0)). So, the answer is
10 + e−0.1×3×200 = 158.16.

Remark: If one buys e−δ2 ×3 share of stock 2 at time 0 and re-invests all dividends,
one will have exactly one share of stock 2 at time 3.

7. Answer = (E)

By formula (24.32), the call option price is

C = P(0, T)[F N(d1) – K N(d2)],
where
T = 1,
P(0, T) = P(0, 1) = 0.9434,
F = F0, 1[P(1, 2)] = P(0, 2)/P(0, 1) = 0.8817/0.9434 = 0.934598,
K = 0.9259.
With σ = 0.05, we have
ln( F / K ) + 12 σ 2T
d1 = = 0.212011 ≈ 0.21,
σ T
d2 = d1 − σ T ≈ 0.21 – 0.05 = 0.16.
Thus,
N(d1) ≈ N(0.21) = 0.5832,
N(d2) ≈ N(0.16) = 0.5636.
Hence,
C = P(0, T)[F N(d1) – K N(d2)]
= 0.9434[0.9346×0.5832 – 0.9259×0.5636]

3
 = 0.9434×0.02322 ≈ 0.022

Remarks: (1) The footnote on page 791 points out that the call option price formula
can also be expressed as
C = P(0, 2)N(d1) – KP(0, 1)N(d2).
(2) The symbol F in the Black formula (24.32) denotes a forward price, but the same
symbol in the Black formula (12.7) denotes a futures price. There is no contradiction
because, in the Black model discussed on page 381, the interest rate is constant. It is
stated on page 146 that “if the interest rate were not random, then forward and futures
price would be the same.”
(3) Consider a forward contract, with delivery date T, for an underlying asset whose
price at time T is denoted by S(T). For t < T, the time-t prepaid forward price is
Ft ,PT [S(T)] = Et∗ [e−R(t, T) S(T)]
by risk-neutral pricing. Here, we use the notation in the last paragraph of page 783;
Et∗ means the conditional expectation with respect to the risk-neutral probability
measure given the information up to time t, and
T
R(t, T) = ∫ t
r (u )du . (24.21)
Thus, the time-t forward price is
1 1
Ft,T [S(T)] = Ft ,PT [S(T)] = Et∗ [e−R(t, T) S(T)].
P(t , T ) P(t , T )
Noting (24.20), we can rewrite this formula as
Et∗ [e − R ( t ,T ) S (T )]
Ft, T [S(T)] = .
Et∗ [e − R ( t ,T ) ]
If the short-rate, r(u), is not stochastic, then the right-hand side is
Et∗ [e − R ( t ,T ) S (T )] e − R ( t ,T ) Et∗ [ S (T )]
∗ − R ( t ,T )
= − R ( t ,T ) ∗
= Et∗ [S(T)],
E t [e ] e Et [1]
which is the formula for the time-t futures price of the underlying asset deliverable at
time T.
(4) Consider the special case S(T) = P(T, T + s). Then the time-t prepaid forward
price of the zero-coupon bond deliverable at time T is
Ft ,PT [P(T, T + s)] = Et∗ [ e−R(t, T) P(T, T + s)]
= Et∗ [ e−R(t, T) ET∗ [e−R(T, T+s)]]
= Et∗ [ e−R(t, T) e−R(T, T+s)]
= Et∗ [ e−R(t, T) − R(T, T+s)]
= Et∗ [e−R(t, T+s)]
= P(t, T + s),
where the third equality is by the law of iterated expectations. Thus, the time-t
forward price is
1 1
Ft, T[P(T, T + s)] = Ft ,PT [P(T, T + s)] = P(t, T + s),
P(t , T ) P(t , T )
which is equation (24.31).

4
8. Answer = (C)

By formulas (12.1) and (12.2a, b), with δ = 0 , the call option price is
S (0) N ( 12 σ T ) − ⎡⎣ S (0)erT ⎤⎦ e− rT N (− 12 σ T )

= S (0) ⎡⎣ N ( 12 σ T ) − N (− 12 σ T ) ⎤⎦

= S (0) ⎡⎣ 2 N ( 12 σ T ) − 1⎤⎦
where the last equality is due to the identity N(−x) = 1 – N(x).

By (20.12), the random variable ln ⎡⎣ S ( t ) ⎤⎦ is normally distributed with variance σ 2t .

Thus, statement (iii) means that σ 2 = 0.4 , and
1
2
σ T = 12 0.4 ×10 = 12 × 2 = 1 .
Therefore, the option price is
S (0) [ 2 N (1) − 1] = 100 [ 2 × 0.8413 − 1] = 68.26 .

9. Answer = (A)

This problem is a modification of the example on page 805. Note that the example is
about cap payments on a four-year loan, not a three-year loan.

An interest rate cap pays the difference between the realized interest rate in a period
and the cap rate, if the difference is positive. Observe that in this problem only ru and
ruu are higher than 7.5%.

At the “u” node, it is expected that a payment of 100×(7.704% − 7.5%) will be made
at the end of the year. Thus, the present value of the payment at the node is
100 × ( 7.704% − 7.5% )
= 0.18941 .
1 + 7.704%
At the “uu” node, it is expected that a payment of 100×(9.892% − 7.5%) will be made
at the end of the year. Thus, the present value of the payment at the node is
100 × ( 9.892% − 7.5% )
= 2.17668 .
1 + 9.892%

5
The tree below corresponds to Figure 24.5 and Figure 24.9 of McDonald (2006).

Year 0 Year 1 Year 2

9.892%
$2.17668
7.704%
$0.18941
6.000% 6.000%
$0 $0
4.673%
$0
3.639%
$0

By risk-neutral pricing, the time-0 price of the interest rate cap is

1/ 2 1/2 1/ 2
× 0.18941 + × ×2.17668
1 + 6% 1 + 6% 1 + 7.704%
= 0.08934 + 0.47649 = 0.565827 ≈ 0.57

Remark: The cap payments are not $0.18941 and $2.17668. They are
$100×(7.704% − 7.5%) to be paid one year after the u node, and
$100×(9.892% − 7.5%) to be paid one year after the uu node. One may be tempted to
put $100×(7.704% − 7.5%) at the uu node and at the ud node, and put
$100×(9.892% − 7.5%) at the uuu node and at the uud node. Unfortunately, this can
be confusing, because these cash flows are not path-independent. For example, if one
reaches the ud node via the d node, then there is no cap payment because rd is less
than 7.5%.

10. Answer = (B)

Let y = number of units of the stock you will buy,
z = number of units of the Call-II option you will buy.
If x or y turns out to be negative, this means that you sell.

Delta-neutrality means
1000 × 0.5825 = y × 1 + z × 0.7773.
Gamma-neutrality means
1000 × 0.0651 = y × 0 + z × 0.0746.

6
From the second equation (the gamma-neutral equation), we obtain
z = 65.1/0.0746 = 872.654 ≈ 872.7.
(This is sufficient to determine that (B) is the correct answer.) Substituting this in the
first equation (the delta-neutral equation) yields
y = 582.5 − 872.7 × 0.7773 = −95.8.

11. Answer = (D)

With u = 1.181 , d = 0.890 , h = 0.5 , and δ = 0 , the risk-neutral probability that the
stock price will increase at the end of a period is
e(
r −δ ) h
− d e0.05×0.5 − 0.890
p* = = = 0.465 . (10.5)
u−d 1.181 − 0.890

For the two-period model, the stock prices are

S0 = 70
Su = uS0 = 1.181× 70 = 82.67 Sd = dS0 = 0.890 × 70 = 62.30
Suu = uSu = 1.181× 82.67 = 97.63 Sud = dSu = 0.890 × 82.67 = 73.58
Sdd = dSd = 0.890 × 62.30 = 55.45

Let P0, Pu, Pd, Puu, Pud, Pdd denote the corresponding prices for the American put
option. The three prices at the option expiry date are
Puu = max(K − Suu, 0) = max(80 − 97.63, 0) = 0,
Pud = max(K − Sud, 0) = max(80 − 73.58, 0) = 6.42,
Pdd = max(K − Sdd, 0) = max(80 − 55.45, 0) = 24.55.

By the backward induction formula (10.12), the two prices at time 1 are
Pu = max(K − Su, e−rh[Puup* + Pud(1 – p*)])
= max(80 − 82.67, e−0.05/2[0×0.465 + 6.42×(1 – 0.465)])
= e−0.05/2×6.42×0.535
= 3.35,
Pd = max(K − Sd, e−rh[Pudp* + Pdd(1 – p*)])
= max(80 − 62.30, e−0.05/2[6.42×0.465 + 24.55×(1 – 0.465)])
= max(17.70, 15.72)
= 17.70.

Finally, the time-0 price of the American put option is

P0 = max(K − S0, e−rh[Pup* + Pd(1 – p*)])
= max(80 − 70, e−0.05/2[3.35×0.465 + 17.70×(1 – 0.465)])
= max(10, 10.75)
= 10.75.

7
12. Answer = (A)

Define the function f (x, t) = xe(r−r*)(T−t). Then,

G(t) = f (S(t), t).

∂ ∂2 ∂
Obviously, f ( x, t ) = e(r−r*)(T−t), 2 f ( x, t ) = 0, and f ( x, t ) = f(x, t)(r−r*)(−1).
∂x ∂x ∂t

By Itô’s Lemma, we have

dG(t) = e(r−r*)(T−t)dS(t) + 0 + f(S(t), t)(r* − r)dt
= e(r−r*)(T−t)S(t)[0.1dt + 0.4dZ(t)] + G(t)(r* − r)dt
= G(t)[0.1dt + 0.4dZ(t)] + G(t)(0.10 − 0.08)dt
= G(t)[(0.1 + 0.02)dt + 0.4dZ(t)]
= G(t)[0.12dt + 0.4dZ(t)].

13. Answer = (E)

In a Vasicek model, zero-coupon bond prices are of the form

P (t , T , r ) = A(t , T )e − B (t ,T ) r . (24.26)
Furthermore, the functions A(t , T ) and B (t , T ) are functions of T − t . Therefore, we
can rewrite formula (24.26) as
P (t , T , r ) = exp ( − [α (T − t ) + β (T − t )r ]) .

The first two pieces of data tell us:

0.9445 = e −α (2)− β (2)×0.04
0.9321 = e −α (2)− β (2)×0.05
which, by taking logarithms, are equivalent to
0.0571 = α (2) + β (2) × 0.04
0.0703 = α (2) + β (2) × 0.05

The solution of this pair of linear equations is

β (2) = 1.32
α (2) = 0.0043

The last piece of data says

0.8960 = e −α (2)− β (2) r*

Taking logarithms yields 0.1098 = α (2) + β (2)r * ,

or
r* = (0.1098 – 0.0043)/1.32 = 0.08.

8
Remark: By comparing (24.29) with (24.26), we see that the word “Vasicek” in this
problem can be changed to “CIR.”

14. Answer = (E)

This is a one-period binomial model. Let p* be the risk-neutral probability of an

increase in the stock price. (See page 321.) Then,
60 × er − 45 60 × e0.08 − 45
p* = = = 0.79988896 ≈ 0.8.
70 − 45 70 − 45

By risk-neutral pricing, the price of the straddle is

e−r[p*|70 – K| + (1 − p*)|45 – K|] = e−0.08[p*|70 – 50| + (1 − p*)|45 – 50|]
= e−0.08[p*×20 + (1 − p*)×5]
= e−0.08[15p* + 5]
≈ e−0.08[15×0.8 + 5]
= e−0.08×17 = 15.693
≈ 15.70.

15. Answer = (C)

This is a variation of Example 12.3 on page 380. Because of the discrete dividend,
we are to use the version of the Black-Scholes put option formula that is in terms of
prepaid forward prices. The prepaid forward price of the stock is
P −0.05/3
F0,1/ 2 ( S ) = 50 – 1.50e = 50 – 1.5×0.983471 = 48.5248.
We apply formula (12.2a), with S = 48.5248, K = 50, r = 0.05, δ = 0, σ = 0.3, and
T = ½, to obtain
d1 = {ln(48.5248/50) + [0.05 – 0 + (0.3)2/2]×½}/{0.3× ½ }
= {−0.02995 + 0.0475}/0.212132
= 0.082740
≈ 0.08.
(This is the same as applying the formula for d1 that follows (12.5) on page 380.)
Then,
d2 = 0.082740 – 0.212132 = −0.129392 ≈ −0.13.
It now follows from the prepaid forward price version of (12.3) that the put option
price is
50e−0.05/2N(+0.13) – 48.5248N(−0.08)
= (50×0.975351×0.5517) – (48.5248×0.4681)
= 26.9039 – 22.7145
= 4.1894
≈ 4.19.

9
Remark added in April 2009: The following sentence can be found on page 381 in
new printings of McDonald (2006). “Because the dividend is fixed, the volatility in
Example 12.3 is the volatility of the prepaid forward, rather than the volatility of the
stock.”

16. Answer = (D)

r −δ 1
Define β = − . Then, the formulas on page 403 for h1 and h2 are
σ 2
2
2r
h1 = − β + β 2 +
σ2
and
2r
h2 = − β − β 2 + .
σ2

Adding these two equations yields

h1 + h2 = −2β.
Hence, −2β = 7/9 or β = −7/18.

For r = 5% and σ = 0.3,

2
2r 7 ⎛ −7 ⎞ 2 × 0.05
h1 = − β + β 2 + = + ⎜ ⎟ + ≈ 1.51.
σ2 18 ⎝ 18 ⎠ 0.32

Alternative solution: The parameters h1 and h2 are the positive and negative roots,
respectively, of the quadratic equation
σ2 2 σ2
h + (r – δ − )h − r = 0; (*)
2 2
see the study note “Some Remarks on Derivatives Markets.” Thus,
σ2 2 σ2 σ2
h + (r – δ − )h − r = (h – h1)(h – h2).
2 2 2
Consequently,
σ2 σ2 σ2 7
r–δ− = (–h1 – h2) = ×− .
2 2 2 9
Hence, the quadratic equation (*) becomes
0.32 2 0.32 7
h + ( × − )h − 0.05 = 0,
2 2 9
the positive root of which is h1.

Remark: For a positive δ, the positive root h1 is in fact greater than 1.

Remark added in April 2009: Question 16 is not in the current syllabus of MFE/3F.

10
17. Answer = (B)

In terms of the notation in Section 14.15, K1 = 90 and K2 = 100.

By (12.1), and (12.2a, b), statement (ii) of the problem is

4 = S (0)e−δ T N (d1 ) − K 2 e − rT N (d 2 ) , (1)
where S ( 0 ) = 80,
ln( S (0) / K 2 ) + (r − δ + 12 σ 2 )T
d1 = ,
σ T
and
ln( S (0) / K 2 ) + (r − δ − 12 σ 2 )T
d2 = d1 − σ T = .
σ T
Do note that both d1 and d2 depend on K2, but not on K1.

From the last paragraph on page 383 and from statement (iii), we have
Δ = e −δ T N (d1 ) = 0.2,
and hence equation (1) becomes
4 = 80 × 0.2 − 100e − rT N (d 2 ) ,
or
e − rT N (d 2 ) = (80 × 0.2 − 4) /100 = 0.12 .

By (14.15) on page 458, the gap call option price is

S (0)e−δ T N (d1 ) − K1e − rT N (d 2 )
= 80 × 0.2 − 90 × 0.12 = 5.2.

Remark: The payoff of the gap call option is

[S(T) – K1]×I(S(T) > K2),
where I(S(T) > K2) is the indicator random variable, which takes the value 1 if
S(T) > K2 and the value 0 otherwise. Because the payoff can be expressed as
S(T)×I(S(T) > K2) – K1×I(S(T) > K2),
we can obtain the pricing formula (14.15) by showing that the time-0 price for the
time-T payoff
S(T)×I(S(T) > K2)
is
S (0)e−δ T N (d1 ) ,
and the time-0 price for the time-T payoff
I(S(T) > K2)
is
e−rT N(d2).

11
Note that both d1 and d2 are calculated using the strike price K2. We can use risk-
neutral pricing to verify these two results:
E*[e−rT S(T)×I(S(T) > K2)] = S (0)e−δ T N (d1 ) ,
which is the pricing formula for a European asset-or-nothing (or digital share) call
option, and
E*[e−rT I(S(T) > K2)] = e−rT N(d2),
which is the pricing formula for a European cash-or- nothing (or digital cash) call
option. Here, we follow the notation on pages 604 and 605 that the asterisk is used to
signify that the expectation is taken with respect to the risk-neutral probability
measure. Under the risk-neutral probability measure, the random variable
ln[S(T)/S(0)] is normally distributed with mean (r – δ – 1 σ2)T and variance σ2T.
2

The second expectation formula, which can be readily simplified as

E*[I(S(T) > K2)] = N(d2),
is particularly easy to verify: Because an indicator random variable takes the values 1
and 0 only, we have
E*[I(S(T) > K2)] = Prob*[S(T) > K2],
which is the same as
Prob*(ln[S(T)/S(0)] > ln[K2/S(0)]).
To evaluate this probability, we use a standard method, which is also described on
pages 590 and 591. We subtract the mean of ln[S(T)/S(0)] from both sides of the
inequality and then divide by the standard deviation of ln[S(T)/S(0)]. The left-hand
side of the inequality is now a standard normal random variable, Z, and the right-hand
side is
ln[ K 2 / S (0)] − (r − δ − σ 2 / 2)T ln[ S (0) / K 2 ] + (r − δ − σ 2 / 2)T
= −
σ 2T σ T
= −d2.
Thus, we have
E*[I(S(T) > K2)] = Prob*[S(T) > K2],
= Prob(Z > −d2)
= 1 − N(−d2)
= N(d2).

The first expectation formula,

E*[e−rT S(T)×I(S(T) > K2)] = S (0)e−δ T N (d1 ) ,
is harder to derive. One method is to use formula (18.29). A more elegant way is the
actuarial method of Esscher transforms, which is not part of the syllabus of any
actuarial examination. It shows that the expectation of a product,
E*[e−rT S(T)×I(S(T) > K2)],
can be factorized as a product of expectations,
E*[e−rT S(T)] × E**[I(S(T) > K2)],
where ** signifies a changed probability measure. It follows from (20.26) and
(20.14) that
E*[e−rT S(T)] = e−δT S(0).

12
To evaluate the expectation E**[I(S(T) > K2)], which is Prob**[S(T) > K2], one shows
that, under the probability measure **, the random variable ln[S(T)/S(0)] is normally
distributed with mean
(r – δ − 1 σ2)T + σ2T = (r – δ + 1 σ2)T,
2 2
2
and variance σ T. Then, with steps identical to those above, we have
E**[I(S(T) > K2)] = Prob**[S(T) > K2],
= Prob(Z > −d1)
= 1 − N(−d1)
= N(d1).

Alternative solution:
Because the payoff of the gap call option is
[S(T) – K1]×I(S(T) > K2)
= [S(T) – K2]×I(S(T) > K2) + (K2 – K1)×I(S(T) > K2),
the price of the gap call option must be equal to the sum of the price of a European
call option with the strike price K2 and the price of (K2 – K1) units of the
corresponding cash-or-nothing call option. Thus, with K1 = 90, K2 = 100, and
statement (ii), the price of the gap call option is
4 + (100 – 90)×e−rTProb*[S(T) > 100]
= 4 + 10e − rT N (d 2 ) .
On the other hand, from (ii), (iii), and (12.1), it follows that
4 = 80(0.2) − 100e − rT N (d 2 ).
Thus, e − rT N (d 2 ) = 0.12, and the price of the gap call option is
4 + 10×0.12 = 5.2.

18. Answer = (A)

In an arbitrage-free model, two assets having the same source of randomness (their
prices driven by the same Brownian motion) must have the same Sharpe ratio (which
is not necessarily constant with respect to time); see Section 20.4. With r = 4%, we
thus have
0.07 − 0.04 G − 0.04
= ,
0.12 H
or

G = 0.25H + 0.04. (1)

If f(x) is a twice-differentiable function of x, then Itô’s Lemma (page 664) simplifies
as
df(Y(t)) = f ′(Y(t))dY(t) + 1 f ″(Y(t))[dY(t)]2,
2
∂
because f (x) = 0. If f(x) = ln x, then f ′(x) = 1/x and f ″(x) = −1/x2. Hence,
∂t

13
 1 1⎛ 1 ⎞⎟
d(ln[Y(t)]) = dY(t) + ⎜ − [dY (t )]2 . (2)
Y (t ) ⎜ 2
2 ⎝ [Y (t )] ⎠⎟
We are given that
dY(t) = Y(t)[Gdt + HdZ(t)]. (3)
Thus,
[dY(t)]2 = {Y(t)[Gdt + HdZ(t)]}2 = [Y(t)]2 H2 dt, (4)
by applying the multiplication rules (20.17) on pages 658 and 659. Substituting (3)
and (4) in (2) and simplifying yields
d(ln[Y(t)]) = (G − 1 H2)dt + HdZ(t).
2
Comparing this equation with the one in (i), we have
H = σ, (5)
1
G − H2 = 0.06. (6)
2

Applying (1) and (5) to (6) yields a quadratic equation of σ,

1 σ2 − 0.25σ + 0.02 = 0,
2
whose roots can be found by using the quadratic formula or by factorizing,
1 (σ − 0.1)(σ − 0.4) = 0.
2

By condition (iii), we cannot have σ = 0.4. Thus, σ = 0.1. Substituting H = 0.1 in (1)
yields
G = 0.25×0.1 + 0.04 = 0.065.

Remark: Exercise 20.1 on page 675 is to use Itô’s Lemma to evaluate d[ln(S)].

19. Answer = (D)

The delta-gamma approximation is merely the Taylor series approximation with up to

the quadratic term. In terms of the Greek symbols, the first derivative is Δ, and the
second derivative is Γ. The approximation formula is
P(S + ε) ≈ P(S) + ε Δ + 1 ε2 Γ. (13.2 & 13.5)
2

With P(30) = 4, Δ = −0.28, Γ = 0.10, and ε = 1.50, we have

P(31.5) ≈ 4 + (1.5)(−0.28) + 1 (1.5)2(0.1)
2
= 3.6925
≈ 3.70.

14