Assignment 2: Decision Analysis

Due: 23:59 October 6 (Sunday), via Moodle

Problem 1 (True or False)

Answer whether each statement is True or False. If your answer is False, explain why the
statement is False.
1) A decision-making environment in which the decision-maker doesn’t have
information about the probabilities of various outcomes is “decision-making under
risk.”

Answer: FALSE. It is the decision-making under uncertainty.

2) The expected value with perfect information always equals the sum of the best
expected monetary value and the expected value of perfect information.

Answer: TRUE.

3) The alternative with the maximum expected monetary value is also the one with the
maximum expected opportunity loss.

Answer: FALSE. The alternative with the maximum expected monetary value is the
one with the minimum expected opportunity loss.

4) Risk seekers have diminishing marginal utilities from obtaining one more dollar.

Answer: FALSE. It is a risk-avoider.

5) In a decision tree, decision nodes are represented by a circle.

Answer: FALSE. It is represented by a square/box/rectangle.

6) The efficiency of sample information is defined as the expected value of sample

information divided by the expected value with perfect information.

Answer: FALSE. It is the EVSI/EVPI, not EVSI/EVwPI.

Problem 2 (Two Funds)
A financial advisor has recommended two mutual funds for consideration: Fund A and Fund
B. The return that will be achieved by each of these funds depends on the state of the
economy, which can be categorized as strong, average, or weak. A payoff table has been
constructed to illustrate this situation:
INVESTMENT STRONG AVERAGE WEAK
ECONOMY ECONOMY ECONOMY
Fund A $8,000 $2,000 -$6,000
Fund B $3,000 $1,000 0
Probability 0.25 0.5 0.25

(a) Which fund would maximize the expected return?

(b) What is the maximum amount that should be paid for a perfect forecast of the economy
(i.e., what is the EVPI)?
(c) Suppose there is uncertainty about the return of Fund A in a strong economy. The return
could be higher or lower than $8,000. What value for this return would cause an investor to
be indifferent between Fund A and Fund B (i.e., the EMVs would be the same)?
(d) Perform a sensitivity analysis on the probability of a strong economy (p) while keeping
the probability of an average economy fixed at 0.5. For which values of p will the optimal
decision be Fund A? For which values of p will the optimal decision be Fund B?
(e) Use the Excel “what-if” analysis to conduct the sensitivity analysis in (d). Produce a table
for p values of 0, 0.1, 0.2, 0.3, 0.4, and 0.5. (You do not need to submit the Excel file. Write
the table with different EMVs).
Solution
(a) EMV(A) = 8000*0.25+2000*0.5-6000*0.25 = $1500;
EMV(B) = 3000*0.25+1000*0.5+0*0.25 = $1250.
So, Fund A would maximize the expected return.

(b) EVPI = (8000*0.25 + 2000*0.5 + 0*0.25) – 1500 = $1500.

OR EVPI = min EOL = 0*0.25+0*0.5 + 6000*0.25 = $1500.
Hence, the maximum amount that should be paid is $1500.

(c) Let X represent the value of the return of fund A in a strong economy such that EMV(A) =
EMV(B). Hence, we have 0.25X + $1000 – $1250 = $1250; thus, X = $7,000.

(d) Let p be the probability of a strong economy.

8000p + 2000*0.5 – 6000*(1-0.5-p) = 3000p + 1000*0.5 – 0*(1-0.5-p). Thus, p=0.2273.
When the probability of a strong economy is greater than p=0.2273, the decision will be
Fund A. Otherwise (p < 0.2273), the decision will be Fund B.
The sensitivity analysis is presented in the graph below.
(e)
Fund A Fund B
p 1500 1250
0 -2000 500
0.1 -600 800
0.2 800 1100
0.3 2200 1400
0.4 3600 1700
0.5 5000 2000

Sensitivity Analysis
6000
5000
4000
3000
2000 Fund A
1000 Fund B

0
0 0.1 0.2 0.3 0.4 0.5
-1000
-2000
-3000
Problem 3 (Susan Solomon)
Susan Solomon is considering opening her own independent coffee shop. Susan’s problem is
to decide on the best size for her shop. The annual returns will depend on both the size of
her shop and a number of market factors related to the beverage industry and coffee
demand. After conducting a careful analysis, Susan developed the following decision table:
SIZE OF COFFEE GOOD FAIR POOR
SHOP MARKET($) MARKET($) MARKET($)
Small 30,000 10,000 –10,000
Medium 50,000 25,000 –20,000
Large 120,000 25,000 –50,000
Very large 200,000 15,000 –100,000

(a) What is the maximax decision?

(b) What is the maximin decision?
(c) What is the equally likely decision?
(d) What is the criterion of realism decision? Use an α value of 0.4.
(e) Develop an opportunity loss table.
(f) What is the minimax regret decision?
Solution
(a) What is the maximax decision?
SIZE OF COFFEE SHOP Maximum in a row
Small 30,000
Medium 50,000
Large 120,000
Very large 200,000
The maximax decision is a very large station.

(b) What is the maximin decision?

SIZE OF COFFEE SHOP Minimum in a row
Small -10,000
Medium -20,000
Large -50,000
Very large -100,000
The maximin decision is a small station.

(c) What is the equally likely decision?

SIZE OF COFFEE SHOP Average Payoff
Small 10,000
Medium 18,333
Large 31,667
 Very large 38,333
The equally likely decision is a very large station.
(d) What is the criterion of realism decision? Use an α value of 0.4.
SIZE OF COFFEE SHOP Weighted Average Payoff
Small 6,000
Medium 8,000
Large 18,000
Very large 20,000
The criterion of realism decision is a very large station.

(e) Develop an opportunity loss table.

SIZE OF COFFEE SHOP GOOD MARKET($) FAIR MARKET($) POOR MARKET($)
Small 170,000 15,000 0
Medium 150,000 0 10,000
Large 80,000 0 40,000
Very large 0 10,000 90,000

(f) What is the minimax regret decision?

SIZE OF FIRST STATION Maximum Regret
Small 170,000
Medium 150,000
Large 80,000
Very large 90,000
The minimax regret decision is a large station.
Problem 4 (Thompson Lumber)
Consider the following payoff table for the Thompson Lumber Company example.

State of Nature
Alternative Favorable Market Unfavorable Market
Construct a large plant 200,000 -180,000
Construct a small plant 100,000 -20,000
Do nothing 0 0

Before deciding on building a new plant, John has the option of hiring XYZ, Inc. to conduct a
market survey at a cost of $20,000. John has a prior belief that P(Fav) = 0.5, and historical
data from XYZ, Inc. is as follows:

Positive Negative Total

Survey Result Survey Result
Favorable Market 30 20 50
Unfavorable Market 10 40 50
Total 40 60
(a) Draw a portion of a decision tree to represent the situation after conducting the market
survey. Please keep two decimal places while calculating probabilities.
(b) Should John hire XYZ to conduct the survey?
(c) What’s the efficiency of the sample information?
Solution
(a)

(b) We should not hire ABC, Inc.

(c) The expected value of sample information (EVSI) is $(33,760 + 20,000 – 40,000) =
$13,760

The value of perfect information (EVPI) is

$200,000 × 0.5 + $0 × 0.5 − $40,000 = $60,000.
Hence, the efficiency of the sample information is
13,760
× 100% = 23%.
60,000
Problem 5 (Casino)
Mike visits a casino and decides to spend $100 on a slot machine for entertainment. The
game requires a minimum deposit of $100, which will be returned to him at the end, to
participate. After depositing the money, Mike can choose to pull the machine's lever once or
twice. Each time he pulls the lever, he has a chance to win $500 or lose $50. He can cash out
his balance either after the second pull or stop after the first pull if he prefers.
Mike believes that the probability of winning during each pull is approximately 0.1.
Additionally, Mike is a risk avoider, and his utility associated with having a monetary value
of $x for this decision is represented by the function 𝑈($𝑥) = 6𝑥 ⁄1000.
Draw a decision tree for Mike and determine the expected utility of participating in this
game. What is the best strategy for Mike if he plays this game? If Mike has the option to
avoid playing on this slot machine, should he decide to play? Please keep two decimal
places while calculating probabilities.

The expected utility of playing this game is 0.1 × 6600/1000 + 0.9 × 650/1000 = 0.28
If the first pull is a success, Mike should stop. If the first pull is a failure, Mike should also
stop.
The expected utility of not playing this game is 6100/1000 =0.32. Hence, Mike should not
play this game in the first place.