lOMoARcPSD|2981718

Solutions Exercises Decision Analysis Week 2

Quantitative Business Analysis (Vrije Universiteit Amsterdam)

StuDocu is not sponsored or endorsed by any college or university

Downloaded by wesam usej (wesam22jul@Gmail.com)
 lOMoARcPSD|2981718

Solutions Exercises Decision Analysis

Exercise 1.1
A company wants to decide about investing in a new product. Market research
predicts that the probability that the product will be a success is 60%. There is a
probability of 40% that it will be a failure. Investment costs are €10,000. If the product
is a success, a profit of €80,000 is expected. If the product is a failure the company
foresees a loss of €30,000. The profit and loss do not include the investment costs.

a. Construct a decision tree to determine the decision that maximizes the expected
monetary value (EMV) for the company. Give the EMV.
b. What is the minimum probability of success of the product for which investment in
the product is still an optimal strategy if maximizing the EMV is the only criterion?

Solution of exercise 1.1

0.6
Product is success
70000
Invest 80000 70000

-10000 26000 0.4

Product is failure
-40000
1 -30000 -40000
26000

Don't invest
0
0 0
a.
The values at the terminal nodes are used for the automatic roll back calculation,
computing expected values at event nodes and maximizing (or minimizing) values at
decision nodes. These values are mentioned in the right-hand cells underneath each
branch. The maximal EMV is E 26,000.

b. Solving the inequality 7p – 4(1-p) > 0 gives p > 4/11 = 0.3636

Exercise 1.2
A country in the Middle East has placed a large order for night vision equipment at a
Dutch electronics company. However, at the moment there is an embargo on delivery
of military goods to this Middle East country. If the company decides to accept the
order, they make a profit of 10 million euro if the embargo is lifted at the time of
delivery. If the embargo is not lifted, they have a loss of 2.5 million euro. The
European Union will decide in two months time about the lifting of the embargo. The
company estimates that there is a probability of 50% that it will indeed be lifted.
Another alternative for the company is not to decide immediately, but to wait until the
European Union has taken its decision. In that case there is an probability of 20%
that the competition will get the order.

Downloaded by wesam usej (wesam22jul@Gmail.com)

 lOMoARcPSD|2981718

a. Construct a decision tree showing the decision alternatives that the company has
and the events that may take place and that influence the results of the
decision(s).
b. Associate partial cash flows to the relevant branches in the tree and determine
the payoffs at the terminal nodes (leaves) of the tree.
c. Execute a roll-back computation to determine which strategy (i.e. decision(s))
maximizes the expected profit. Give the expected profit.

Solution of exercise 1.2

The numbers in the left- most cells underneath the branches are the partial cash
flows in millions of Euro. The strategy that maximizes the expected profit is to wait for
the decision of the European Union. The corresponding expected profit is 4 million
Euro.

Please mention that this is not the only way how to represent the problem by a
decision tree. You could also draw it by exchanging event nodes 2 and 4. The lower
part of the decision tree than reads: wait and see whether the embargo is lifted or not
lifted. The latter results in payoff zero. If it is lifted there is a 20% chance that the
order goes to the competition which implies a payoff of zero. With a chance of 80%
the company gets the order and makes a profit of 10 (Milion).

Downloaded by wesam usej (wesam22jul@Gmail.com)

 lOMoARcPSD|2981718

Exercise 1.3
An oil company is considering making a bid for oil drilling off the Mexican coast. The
company has decided to bid 520 million US dollar for these rights. Experts have told
the company that they have a 60% probability of winning the job with this bid. If the
company wins the contract, it can drill on its own, it can join a joint venture, or it can
sell its rights. Should the firm drill, costs of drilling are expected to be $80 million.
Drilling expenses in the joint venture would be $45 million. The negotiation process
involved in selling the rights is expected to cost $2 million. The estimated revenues of
the three options are given in the following table. Note that costs have not been
deducted from these values.

Big Find Average Find Small Find

Drill on own $900 million $600 million $250 million
Probabilities 0.6 0.2 0.2
Joint venture $750 million $500 million $200 million
Probabilities 0.6 0.3 0.1
Sale of rights $600 million

Use a decision tree to analyze this problem and to determine the optimal strategy
(i.e. decision(s)). Give the Expected Monetary Value.

Downloaded by wesam usej (wesam22jul@Gmail.com)

 lOMoARcPSD|2981718

Solution of exercise 1.3

Costs and revenues are mentioned as partial cash flows in the left-hand cells
underneath the branches. E.g. in the case of selling the rights the revenues are
$600M - $2M = $598M. The company payed $520M for the contract, so the terminal
node value is $598M – $520M = $78M.
The decision that maximizes the expected profit is to make a bid. If the contract is
won, the optimal decision for the company is to drill on its own.
The corresponding Expected Monetary Value is $ 66 Million.

Downloaded by wesam usej (wesam22jul@Gmail.com)

 lOMoARcPSD|2981718

Exercise 1.4
An art dealer has a client who will buy the masterpiece Rain Delay for $50,000. The
dealer can buy the painting now for $40,000 (making a profit of $10,000). He can
also wait one day. The price will then go down to $30,000 but there is a probability of
1/3 that the painting will be sold to someone else. If the painting is not sold, he can
still wait another day. The price will now go down to $25,000 but there is again a
probability of 1/3 that someone else will buy the picture. If the picture is not sold the
second day, it will no longer be available.

a. Draw a decision tree for the dealers decision process.

b. Solve the tree. When should the dealer buy the painting and what is his expected
profit?
c. Would the dealer using the decision tree decide to wait two days if the price of the
painting went down to $21,000?
d. What is the maximum probability that the dealer will accept that the picture will be
sold to someone else that will make him decide to wait two days before buying the
picture? (i.e. the value of P(still available))
e. How much would the dealer want to pay to a fortune teller who would tell him on
what day the picture will be sold (given the original probabilities)? (This is called
the Expected Value of Perfect Information.)

Solution of exercise 1.4

a.

b. Buy after 1 day. Expected profit $13,400.

c. No, because (2/3)*29000 = 19333.33 < 20000

(Profit 29000 = 50000 – 21000)

d. P{still available}*25000>=20000; P{still available}>=0,8

e. The probability that the painting will be sold after the first day is 1/3, in that case is

Downloaded by wesam usej (wesam22jul@Gmail.com)

 lOMoARcPSD|2981718

must be bought immediately. This gives a contribution to the expected monetary

value of 10000*(1/3). The probability that it is sold on the second day is
(2/3)*(1/3), with a contribution to the expected monetary value of
20000*(2/3)*(1/3). The probability that it is still available after the second day is
(2/3)*(2/3), with a contribution to the expected monetary value of
25000*(2/3)*(2/3).
Here this means that with perfect information in 1/3 of all cases a profit of 10000
is made, in 2/9 of all cases a profit of 20000, and in 4/9 of all cases a profit of
25000. I.e. the expected profit with perfect information is
10000*1/3+20000*2/9+25000*4/9 = 18888.89. Without perfect information the
maximum expected profit is 13400.
The EVPI (expected value of perfect information) is defined as the maximum
improvement in the expected return if the decision maker has perfect information
about each future event. The total expected value of perfect information is:
$18,888.89 - $13,400 = $5,488,89.

Exercise 1.5
Nova Ltd is considering whether to spend € 30,000 to develop a new product. If the
development program is successful two alternative sizes of plant can be considered:

Plant Production capacity Capital cost (1000 €) Running cost

(tons) (1000 € /year)

Small 2,000 100 50

Large 5,000 200 100

The development program, which is thought to have a 70% chance of success, will
take a year to complete. After completing that program it would take one year to build
a small plant and two years to build a large plant. It is believed that one other major
company in this field is developing a competing product and the marketing
intelligence unit of Nova believe that there is a 25% probability that this company will
have their product on the market 3 years from now, a 50% probability that they will
have their product on the market 4 years from now and a 25% probability that it will
take them 5 years to get their product onto the market.

The following estimates of Nova's possible selling price have been obtained from the
marketing department:

Small Large

Output (per year) 2,000 tons 5,000 tons

Monopoly price €50 per ton €30 per ton

Competitive price €30 per ton €25 per ton

where the monopoly price applies whilst Nova is without competition and the
competitive price applies once the competing company has its product on the market.

Downloaded by wesam usej (wesam22jul@Gmail.com)

 lOMoARcPSD|2981718

a. Develop a decision tree to determine what Nova should do if the lifetime of the
Nova product is 10 years (once released onto the market). Give the optimal strategy
(i.e. decision(s)) and the corresponding Expected Monetary Value.
b. Mention a "downside" of your suggested course of action?

Solution of exercise 1.5

a. We can construct the following decision tree:

First the partial cash flows are entered from left to right:
Development costs : 30
Capital + running costs during 10 years for large plant: 1200
Idem for small plant 600.

Revenues:
Developing product takes 1 year, building large plant takes two years; i.e. after 3
years production may start. At this time competitive product is ready. This means a
production of 5000 ton at € 25 per ton during 10 years = € 1,250,000.

Downloaded by wesam usej (wesam22jul@Gmail.com)

 lOMoARcPSD|2981718

Second end node from top: competition is ready after 4 years. This means 1 year
5000 ton at € 30 per ton and 9 years 5000 ton at € 25 per ton = (5000*30 +
9*5000*25) = € 1,275,000

Third end node from top (2*5000*30 + 8*5000*25) = € 1,300,000

Fourth end node: 1*2000*50+9*2000*30 = € 640,000

Fifth end node: 2*2000*50+8*2000*30 = € 680,000
Sixth end node: 3*2000*50+7*2000*30 = € 720,000

Expected profit if product is € 26,000, provided that a small plant is built if the
development is a success.
If it is decided not to develop the product, the expected profit is 0, so the optimal
decision is to decide to develop the product.

b. The downside is, that the company risks loosing €30,000 if the project is not a
success.

Exercise 1.6
A famous writer of thrillers, Stephany Queen, has offered the motion picture company
Van Gogh the exclusive rights for her new book that will come out in one year’s time.
She asks 1 million dollars for these rights. The company has to decide whether or not
to accept the offer. The decision to make the movie will be taken after the book has
been published. The literary advisor of the motion picture company estimates that the
probability that the book will be a success is 60% and that the probability is 40% that
it will be a flop. It is expected that the movie will generate a net profit of $30 million if
the film is a hit, and that the loss will be $10 million if the film flops. If the book is a
success the probability that the film is a success is estimated to be 75%. The
probability that the film still will flop is then 25%. If the book is not a success the
probabilities for success and failure of the movie are 20% and 80% respectively.

a. Use a decision tree to determine which course of action maximizes the expected
net profit for Van Gogh.
b. Suppose the decision maker is risk averse. Which course of action will be chosen?
c. Suppose the decision maker is risk neutral. Which course of action will be chosen?
d. Suppose the decision maker is risk seeking. Which course of action will be
chosen?
e. Suppose the decision maker has the following utility function
U(x) = (x + 11)2/(40)2. Which course of action will be chosen?
f. Suppose a marketing expert can tell you whether the book will be a success. His
information is perfect. How many dollars do you want to pay to that marketing expert
(i.e the Expected Value of Perfect Information)?

Downloaded by wesam usej (wesam22jul@Gmail.com)

 lOMoARcPSD|2981718

g. Suppose the information of the marketing expert is not always correct. The
following table indicates the number of cases he is correct and the number of cases
he is not correct. How many dollars do you want to pay to that marketing expert (i.e.
the Expected Value of Sample Information)?

Book is a success Book is not a success

Prediction “Book is a 25 cases 5 cases
success”
Prediction “Book is not 5 cases 15 cases
a success”

Solution of exercise 1.6

Figure 1.6 a.

a. The strategy that maximizes the expected profit is to buy the rights and then to
make the movie if the book is a success, and not to make it if the book flops.
b. The decision maker doesn’t like risk. The optimal strategy of question a. is rather
riskful. Maybe “Don’t buy rights” is the best course of action. The answer depends on
the degree of risk aversion.
c. The same answer as in question a.

Downloaded by wesam usej (wesam22jul@Gmail.com)

 lOMoARcPSD|2981718

10

d. The same answer as in question a, because the decision-maker is not afraid of

risk. However, it depends on the degree of risk-seeking whether to make the movie
even if the book flops.
e. The pay-offs on the right side of the decision tree in figure 1.6 a. have to be
replaced by the corresponding utility function values. The function
U(x) = (x + 11)2/(40)2 provides for x=29, x=-11 and x=-1 the utility values

U(29) = (29 + 11)2/(40)2 = 1

U(-11) = (-11+ 11)2/(40)2 = 0
U(-1) = (-1 + 11)2/(40)2 = 100/1600 = 0.0625
U(0) = ( 0 + 11)2/(40)2 = 121/1600 = 0.076

Next, we place the utility values at the final nodes of the tree and perform the roll
back method. Hence we do the same as in a., however, now we are working with
utility “payoffs” instead of monetary payoffs.

Figure 1.6 e.1

The strategy that maximizes the expected utility of profit is to buy the rights and then
to make the movie if the book is a success and also if it is no success. This is a risky
strategy. Indeed, when we draw the utility function U(x)=(x+11)^2/(40)^2, as
illustrated in Figure 1.6 e.1, we get the typical convex shape of the utility of a risk
seeking decision maker.

Downloaded by wesam usej (wesam22jul@Gmail.com)

 lOMoARcPSD|2981718

11

0.9

0.8

0.7

0.6
U(x)

0.5

0.4

0.3

0.2

0.1

0
-15 -10 -5 0 5 10 15 20 25 30
x

Figure 1.6 e.2

f. We can draw another decision tree for this situation:

Figure 1.6 f.: Payoffs are in 1000 dollars.

Now, the EMV becomes $11.400M, so the EVPI = $11.400M – $11.000M = $ 0.4M.

g. In this situation we have to put the changes into a new type of decision tree, as
you can see below: (The text between quotes (“”) is the prediction of the marketing
expert.)

Downloaded by wesam usej (wesam22jul@Gmail.com)

 lOMoARcPSD|2981718

12

Figure 1.6 g.: Probabilities are in blue, EMVs in red

Now, the EMV becomes $11 M such that EVSI = $11 M – $11 M = $0 M.

Hence you don’t want to pay for the predictions of the marketing expert.
The reason is that we don’t use the information of the marketing expert – we buy the
rights anyway, whatever he or she says.

Exercise 1.7
Insurance company Hollandia has made a verbal agreement to take over the
specialized insurance company Gotham Insurance. However, just before the official
papers are to be signed, the board of directors of Gotham Insurance sell their
company against all earlier agreements to the competing company Imperia. Hollandia
does not accept this course of action and makes an official complaint at the
disciplinary board for the insurance business. The board decides that Imperia has to
pay a compensation of 20 million euro to Hollandia. Imperia refuses and offers an
amount of 5 million euro as compensation. Hollandia's lawyers analyze the situation
and advise to propose a settlement for an amount of 10 million euro before starting a
lawsuit. The lawyers expect two possible reactions by Imperia to the settlement
proposal. One possibility is that they will make a last counter offer of 7.5 million euro
(with a probability of 50 %), the other possibility is that they will refuse and wait for
the outcome of the lawsuit. If Hollandia decides to refuse all counter offers made by
Imperia, and to start a lawsuit right away, then the lawyers see three possibilities.
The court orders that the original compensation of 20 million euro must be paid; it
orders that the settlement of 10 million euro must be made, or it decides for no
compensation at all. The probabilities of these three alternatives are estimated to be
respectively 10%, 50% and 40%.

Downloaded by wesam usej (wesam22jul@Gmail.com)

 lOMoARcPSD|2981718

13

a. Which strategy maximizes the expected amount that Hollandia will receive?
b. What is the expected value of perfect information about the probability of a counter
offer of 7.5 million Euro.

Solution of exercise 1.7

a. To maximize the expected returns Hollandia must propose a settlement and
accept the counter offer if it is made. If no counter offer is made, it is assumed that
this implies the start of a lawsuit.

The Expected Monetary Value for Hollandia is 7.25 million Euro.

b. We can construct the decision tree for the situation with perfect information and
calculate the corresponding EMV. To understand the tree, assume that it starts
with someone predicting with 100% whether there will be a counter offer or not.
The chance that the prediction will be the one option or the other is 0.5.

Downloaded by wesam usej (wesam22jul@Gmail.com)

 lOMoARcPSD|2981718

14

After that, we can compare this EMV with the EMV of the initial decision tree (see
question A.). The difference is equal to zero. So, the Expected Value of Perfect
Information is 0. The interpretation for that outcome is that Hollandia can find out
herself whether there will be a counter offer - without losing the option of starting a
lawsuit. Hollandia can always propose settlement and see how Gotham reacts.

Downloaded by wesam usej (wesam22jul@Gmail.com)

 lOMoARcPSD|2981718

15

Exercise 1.8
A company wants to decide whether or not to make a proposal for a government
building project. The costs for making a project proposal are €50,000. If the proposal
is accepted the project may generate a revenue of €1,000,000 with an estimated
probability of 70%. There is also a probability of 20% that the revenues are only
€500,000 and there is even a probability of 10% that a loss of €100,000 is incurred.

a. Construct a decision tree for the described situation and determine the minimum
probability of getting the project for which a decision-maker would decide to make
the proposal? Use maximalisation of the expected monetary value as criterion.
b. Discuss the advantages or disadvantages of the expected monetary value as
criterion in a situation in which a one-time decision must be taken, versus a
situation in which the same decision process is repeated a great number of times.
c. What other criterion might be used for the selection of decision alternatives?
d. Would a decision-maker who is risk averse (or risk avoiding) make the proposal if
the probability to get the project is equal to the probability you computed in
question a.? Explain your answer.

Solution of exercise 1.8

a.
0.7
High outcome
950
1000 950

0.063291 0.2
Accepted Low outcome
450
0 740 500 450

0.1
Make proposal Loss
-150
-50 -1.9E-07 -100 -150

0.936709
Not Accepted
2 -50
0 0 -50

Don't make proposal

0
0 0

Solve p*740 – (1 – p)*50 = 0 to find p = 0.0633.

b. In the case of a one-time decision the expected monetary value is a less

realistic criterion than in the case of a process that is often repeated. The
expected monetary value is the amount you get on the long run if the same
decision process with the same probabilities is repeated a great number of
times.

Downloaded by wesam usej (wesam22jul@Gmail.com)

 lOMoARcPSD|2981718

16

c. Another criterion that may be used is the expected utility of the decision
alternatives.
d. The probability that was computed for question 1 is the probability p for which
a risk neutral decision maker will be indifferent between having €0 or playing in
a lottery in which he can win €740,000 with probability p or loose €50,000 with
probability 1-p. A risk avoiding decision maker will not accept the lottery with
this value of p.

Exercise 1.9
Brandon Appliance Corporation, a predominant producer of microwave ovens, is
considering the introduction of a new product. The new product is a microwave oven
that will defrost, cook, brown, and broil food as well as sense when the food is done.
Brandon must decide on a course of action for implementing this new product line.
An initial decision must be made whether or not to (1) market the product not at all,
(2) introduce the product in a marketing test, or (3) nationally distribute the product
from the onset. If a marketing test is conducted, Brandon must then decide whether it
wishes to abandon the product line or make it available for national distribution.
The finance department has provided some cost information and probability
estimates relating to this decision. The preliminary costs for research and
development have already been incurred and are considered irrelevant to the
marketing decision. A success nationally will increase profits by E 5.000.000, and a
failure will reduce them by E 1.000.000, while abandoning the product will not affect
profits. The market test analysis will cost Brandon an additional E 100.000.
If a market test is not performed, the probability of success in a national campaign is
estimated to be 45 percent. If the market test is performed, the probability of a
favorable test result is 60 percent. With favorable test results, the probability for
national success is estimated to be 80 percent. However, if the test results are
unfavorable, the national success probability is only 10 percent.

a. This decision problem has three initial decision alternatives:

 Do not introduce the new product;
 Use a marketing test for the model;
 Introduce the product on the national market immediately.
Construct the decision tree showing the decision alternatives that the company
has and the events that may take place and that influence the results of the
decision(s).
What is Expected Monetary Value for each of the three decision alternatives?
Which decision alternative has the highest Expected Monetary Value?
Give the optimal strategy that maximizes the Expected Monetary Value.
b. If a market test is not performed, the probability of success in a national
campaign is estimated to be 45 percent. What is de value of perfect
information (EVPI) about this probability?

Downloaded by wesam usej (wesam22jul@Gmail.com)

 lOMoARcPSD|2981718

17

Solution of exercise 1.9

a. We have the following decision tree:

0,45
Succes
5000000
Introduce 5000000 5000000

0 1700000 0,55
Failure
-1000000
-1000000 -1000000

0,8
Succes
4900000
Introduce 5000000 4900000

0 3700000 0,2
0,6 Failure
Favourable -1100000
2 -1000000 -1100000
0 3700000

Not introduce
-100000
1 0 -100000
2180000 Marketing test
0,1
-100000 2180000 Succes
4900000
Introduce 5000000 4900000

0 -500000 0,9
0,4 Failure
Not favourable -1100000
3 -1000000 -1100000
0 -100000

Not introduce
-100000
0 -100000

Not introduce
0
0 0
Introduce the product has an EMV of $ 1,700,000.
Using the marketing test has an EMV of $ 2,180,000.
Not introduce the product has an EMV of $ 0.
The decision to conduct the marketing test has the highest EMV.
Optimal strategy: Conduct market test. If test results are favourable then Introduce. If
tests results are not favourable then do not introduce.
Remark: Note that the marketing test has anEMV which is much higher than the EMV
without it. If the marketing test didn’t cost anything the difference would be $
2,280,000 - $ 1,700,000 = $ 570,000. If the price of the marketing test was
negotiable, this would be Brandon’s maximum willingness to pay for it.

b. Consider the following tree: (notice that the first three branches are the same as in
question (a))
EVPI = $2,250,000 – $2,180,000 = $70,000.

Downloaded by wesam usej (wesam22jul@Gmail.com)

 lOMoARcPSD|2981718

18

Downloaded by wesam usej (wesam22jul@Gmail.com)

 lOMoARcPSD|2981718

19

Exercise 1.10
A group of medical professionals is considering to construct a private clinic. If the
medical demand is high (i.e., there is a favorable market for the clinic), the medical
professionals could realize an annual net profit of €1.200.000. If the market is not
favorable, they could lose €400.000 annually. Of course, they don’t have to proceed
at all, in which case there is no cost. In the absence of any market data, the best the
group of medical professionals can guess is that there is a 50-50 chance the clinic
will be successful.

Questions
a. Develop a decision tree to determine which course of action maximizes the
expected money value for the group of medical specialists. Which strategy
maximizes the expected profit of the group of medical specialists? What is the
corresponding expected money value?
b. The probability that the clinic will be successful is estimated to be equal to 0,5.
However, suppose this probability is set equal to p. What is the minimum
probability of success of the clinic for which the construction of the private clinic is
still an optimal strategy (i.e. maximizing the EMV). Give a brief motivation.
c. The probability that the clinic will be successful is estimated to be equal to 0,5.
Instead of this, the group of medical professionals can get perfect information
before making his decision regarding the construction of the private clinic.
Determine how much the group of medical professionals wants to pay for this
perfect information (i.e. the Expected Value of Perfect Information, EVPI)?
d. Like earlier mentioned, the new clinic can be successful (probability 0,5).
However, the medical professionals can ask a marketing expert. This expert can
tell them more about the success or failure of the new clinic. The expert will cost
€100.000. The following table indicates the number of cases the expert is correct
and the number of cases he is not correct.
Success clinic Failure clinic
Advisor says “success clinic” 25 cases 10 cases
Advisor says “failure clinic” 5 cases 20 cases
How many dollars do you want to pay to the marketing expert (i.e. Expected
Value of Sample Information, EVSI)? Give a brief motivation.
e. Now suppose, the medical professionals will not ask a marketing expert to advice
them (i.e. the situation of question a). The group of medical professionals has
modeled its risk attitude by using the following utility function: U(-400.000)=0,
U(0)=0,01 en U(1.200.000)=1. Is the Group of medical professionals risk avers,
risk-neutral or risk-seeking? Give a brief motivation.
Which strategy maximizes the expected profit of the group of medical specialists
dealing with this risk attitude and the money values of the decisions? Give a brief
motivation.

Downloaded by wesam usej (wesam22jul@Gmail.com)

 lOMoARcPSD|2981718

20

Solution of exercise 1.10

a. The best decision is: Start Private Clinic.
EMV = € 400.000.

0.5 1.2

0.4
-0.4
0.5

b. EMVstart = p * 1.200.000 – (1-p) * 400.000

EMVnot start = 0
EMVstart > EMVnot start
p * 1.200.000 – (1-p) * 400.000 > 0
p * 1.600.000 > 400.000
p > 400.000/1.600.000
p > 0,25

At p=0,25 the group is indifferent between constructing the clinic or not, assuming
they are risk neutral (i.e. EMV is their decision criteria). For any value of p higher than
0,25 they should build the clinic.
If they are risk averse, the value of p has to be higher to make them indifferent, to
make building the clinic the best option respectively.
If they are risk seeking their best option is to build the clinic for p=0.25, as the point of
indifference is lower than 0,25.

c. EVPI = EMVwith info – EMVwithout info

EMVwithout info = € 400.000
EMVwith info = 0,5 * 1.200.000 = € 600.000
EVPI = 600.000 – 400.000 = € 200.000

1.2 1.2

0.5

Advice

0 -0.4

0.5

Downloaded by wesam usej (wesam22jul@Gmail.com)

 lOMoARcPSD|2981718

21

d. EVSI = EMVwith info – EMVwithout info

EMVwithout info = € 400.000
EM with info = (35/60) * ((25/35)*1.200.000 - (10/35)*400.000) = € 433.333
EVSI = € 433.333 - € 400.000 =€ 33.333.
The information of the expert increases the expected value of the best strategy. If the
expert predicts success, it is the best strategy to build the clinic and not to do it
otherswise. However, this increase is only =€ 33.333 . The price of the expert of
€ 100.000 is therfore too high.
25/35 1.2
0,7429
10/35
35/60 -0.4

Adv. 5/25 1.2

0
20/25 -0.4

25/60
0

e.
EUnot start = 0,01.
EUstart = 0,5 * 0 + 0,5 * 1 = 0,5
Best decision: Start Private Clinic
The Group of medical professionals is risk-seeking as the form of the utility function
is convex. To see this draw a straight line conncting the point (x,y)=(-4,U(-4))=(-4,0)
(where x is in hundredthousands) and (x,y)=(12,U(12))=(12,1). The point
(0,U(0))=(0,0.01) lies below this line which indicates convexity of the utility function.

Downloaded by wesam usej (wesam22jul@Gmail.com)