Decision tree analysis case study

Table 1: Summary of literature reviewed

Authors

(year)

Main

objective

Methodology/

Research

approach

Decision

analysis

technique(s)

Key findings

Dey, K.

(2012)

To highlight

the usefulness

of decision

analysis within

the risk

management

process.

A case study

approach with

the use of

secondary

research.

Analytical

hierarchy

process and

the use of

decision tree

analysis.

The paper argues the case for a

decision support system that

will strengthen the process of

project risk management. A key

recommendation is a decision

support system which includes

an analytical hierarchy process

to aid risk assessment and

decision tree analysis to identify

alternative courses of action.

etc.

Part 2: Case Study: Vitam Machines Assembly Company (VMAC)

(approximately 1000 words)

Notes and introduction to the case study: This case study is based on a real case; of

course anonymized for obvious reasons. The case study considers the problem of

Vitam Machines Assembly Company (VMAC), a company considering relocating its

manufacturing facilities from the UK to an overseas country. In making its decision,

the company needs to take into account a number of political risks that it will face if

it decides to go ahead with the relocation. The decision problem is made complex by

the large number of combinations of possible events that can occur and the

challenges that arise from the need to structure the problem in a way which makes

analysis of the problem tractable. Note that all of the monetary values presented in

the case have already been expressed as present values to avoid the additional

complication of applying discounted cash flow analysis to the data. Carefully read

the case study and answer the questions that follow.

The Vitam Machines Assembly Company (VMAC), which has its headquarters in the

UK, is considering opening a manufacturing plant in an overseas country and

transferring much of its current UK-based production to the new plant. After extensive

data collection and visits by managers to a number of possible countries, Almeria has

been identified as the most promising country for a new plant. A site near the capital,

Lasia, appears to be highly suitable and a new state-of-the art manufacturing facility

could be constructed there very quickly.

The decision on whether to go ahead with the move to Almeria will be based on the

level of monetary savings in production costs that it is hoped would be generated over

the next 10 years by opening a plant there. However, there are a number of risks

associated with these savings and, for simplicity, the level of savings is categorised as

either high, medium or low. If a move to Almeria does go ahead, VMAC will review

the success of its investment after the first five years and will have the option of

5

withdrawing from that country and returning operations to the UK if this appears to be

appropriate.

Almeria has a relatively new democracy which was created following the overthrow of

a military dictatorship that had ruled the country for nearly thirty years. However, there

is considerable poverty and unemployment rates have recently been as high as 38%.

The current government is therefore keen to attract foreign investors, but it only has a

narrow majority in the country’s parliament. Despite the efforts of the government,

widespread corruption has persisted and Almeria is ranked 5th in the World league table

of corruption. Corruption is partly responsible for the neglect of the country’s road and

rail systems which are now amongst the worst in the region.

If a decision is made to relocate to Almeria there is a risk that a new government will

come into power and nationalize all foreign investments. There is thought to be only a

0.05 probability of this happening during the first five years, but if it did occur, the loss

of assets would cause VMAC to be worse off by $75 million (in present value terms)

compared to the returns that would have been generated by continuing manufacturing

in the UK. Nationalization would also cause VMAC’s association with the country to

end immediately.

Insurance can be purchased to cover the political risk of nationalization for the first five

years of operations by paying a total premium which has a present value of $16 million.

(Note that the insurance can only be purchased at the start of the five years). If the

company does purchase political risk insurance and nationalization occurs in the first

five years then the insurance will only cover the loss of assets. It is expected that any

savings generated before nationalization would be canceled out by the costs of

relocation and so would have present value of $0. If nationalization does not take place

it is thought that there is a 0.6 probability that in the first five years the investment

would generate high savings having an estimated present value of $85 million. There is

also an estimated 0.25 probability that medium savings, with a present value $48

million, would be earned in the first five years and a 0.15 probability these savings will

be low and only amount to $5 million.

At the end of the first five years the company would have to decide whether to continue

to operate the plant in Almeria for another five years or whether to transfer operations

back to the UK. However, this decision will only be considered if the savings in the

first five years have been low. If a decision to withdraw is made then the plant will be

sold for a return with an estimated present value of $10 million. If VMAC decide to

continue operations in Almeria for a further five years the risk of nationalization during

this period is estimated to be 0.15.

The total insurance premium to cover these risks for the second five years would have

a present value of $12.8 million. If insurance is purchased and nationalization occurs in

the second five years then it is assumed that gross savings made before nationalization

will again be cancelled out by the costs arising from the disruption. For simplicity, the

present values of other costs and savings occurring under each set of conditions in the

second five years are assumed to be the same as those in the first 5 years, with a 20%

reduction to take into account the time value of money. However, it is thought that the

probabilities of high, medium and low returns in the second five year period will be

dependent on the level of returns achieved in the first five years as shown in the table

below.

6

Second five years

High Medium Low

High 0.60 0.30 0.10

First five years Medium 0.10 0.80 0.10

Low 0.03 0.07 0.90

For example, the table shows that if savings in the first five years have been high then

there is a 0.60 probability that high savings will be maintained in the next five years, a

0.3 probability that only medium savings will be generated and a 0.10 probability that

savings will be low. The other two rows can be interpreted in a similar manner. It can

be assumed that if the company stays in Almeria for ten years, it will sell the plant at

the end of this period and hence generate extra returns with a present value of $6

million.

Question 1

Using decision tree analysis, structure the decision problem faced by VMAC

Guidance notes:

 Present your decision tree diagrams, clearly showing all probabilities and net values at the end of the branches

 Follow the conventions of constructing decision trees, such as the basic shapes distinguishing decision nodes from chance nodes

More guidance notes on applying decision tree analysis on the VMAC decision

problem

A decision tree model can be used to analyze this decision. Because of the size

of the problem, it is suggested that you break down the decision trees into four

sub-trees as follows:

Decision Tree 1: Depicts the decision that face the company in planning for the

first 5 years of potential operation in Almeria

Decision Tree 2: Depicts the decision for second 5 years if savings in the first

5 years are high

Decision Tree 3: Depicts the decision for second 5 years if savings in the first

5 years are medium

Decision Tree 4: Depicts the decision for second 5 years if savings in the first

5 years are low

Decision trees 2, 3 and 4 are best constructed first so that the optimal expected

savings that they indicate can be ‘rolled back’ and added to the savings for the first

five years in decision tree 1.

Note: You are strongly encouraged to use PrecisionTree® software to facilitate your

analysis and solution of this case study problem. However, you will not be penalized

for not using the software.

7

Question 2

Recommend the policy or strategy that the company should pursue to maximize

expected savings.

Guidance notes:

 Clearly show all your workings

 Explain/justify your recommendations

Question 3

Conduct a sensitivity analysis of your model and discuss the implications to the

VMAC’s decision situation.

Guidance notes:

 Justify your selection of input variable(s) used as the basis of your sensitivity analyses

 Clearly show and explain the impact of your selected variables, with the aid of graphical representations

Question 4

Outline the usefulness/strengths (and any limitations) of using decision trees (including

sensitivity analysis) to address VMAC’s decision problem. In other words, outline the

strengths and limitations of your decision tree analysis in terms of the usefulness of the

guidance that it would provide to VMAC’s managers.