Diamond

Energy Case Report

Background

Diamond energy resources was a wholly owned subsidiary of the Australian based mining company
Diamond energy Inc. It owned six mines and a combined area of 4,700 hectares of coal mining
concession area and an esHmated 40 million tons of coal reserves. Its coal producHon was mainly sold to
domesHc and internaHonal companies like Glencore, INCO, and Trafigura. The company had been
performing well financially since it began operaHons with operaHng margins and return on invested
capital averaging 21% and 16% in the last five years respecHvely.

In 2015, the CEO of Diamond Energy Resources had received an acquisiHon plan to acquire a new mine.
This acquisiHon would be strategically beneficial to the company as it would mean expanding their coal
export market. For this acquisiHon to materialize, the CEO had to take the approval of the parent
company for the influx of capital and also an approval from the bank to secure a loan necessary for the
acquisiHon. This basically is the objecHve of the case - to value the acquisiHon based on reasonable
assumpHons with respect to forecast percentages, discount rates and any other factors that might
influence the NPV.

Part A

The enHre analysis of this case is done using four different frameworks as detailed below. These
frameworks are designed in such a way as to calculate NPV of the new mine using various different
assumpHons with respect to growth percentages, discount rates and WACC, and a comprehensive
simulaHons framework using equal and weighted probabiliHes.

Framework 1

To begin with, I considered 3 different scenarios for the forecast percentages. These were to calculate
the free cash flows with 5% increase in factors (Price of Coal, Investment, OperaHng Expenses and
ProducHon), 5% decrease and a 7.5% increase. The opHmisHc growth percentages could be
substanHated with the increase in demand for coal in India, China and parts of south-east Asia.

For calculaHng WACC, I considered 4 different scenarios varying the betas, risk free rates and market risk
premiums and considering the data given in the case for cost of debt and cost of equity. For 3 scenarios I
chose the Debt to value and equity to value raHo of the new project and for the other scenario I chose
the exisHng debt-to-value raHo of the company. The WACC calculaHons are detailed in Exhibit A.

I calculated the free cash flows by varying 1 factor and keeping the other 3 constant at a Hme. I
considered the capital investment and change in working capital given in the case. Taking into account
the increase in operaHng expenses and the royalHes, I was able to calculate the free cash flows for all the
different scenarios. Using the four discount rates, I discounted these free cash flows to arrive at a set of
52 NPVs. The NPV distribuHon of the 52 NPVs is given below.
Framework 2

This framework is where simulaHons is introduced into the analysis. I created a simulaHons page and
created 2 secHons in it. ‘Underlying Random Numbers’ and ‘Resultant Scenario Choices’. These resultant
scenario choices would point to 1 of the 4 scenarios I had considered in framework 1 with increase and
decrease in growth percentages. For the underlying random numbers for each of the 4 factors (Price per
ton, Investment, ProducHon and OperaHng Expenses), I generated random numbers using the excel
funcHon rand()which generates a random number between 0 and 1. Using this, I calculated the
resultant scenario choices by mulHplying the underlying random number with the number of scenarios
for that parHcular factor. I used the excel funcHon roundup() to come up with a whole number. I
generated 500 simulaHons. These resultant scenario choices will be the basis for the analysis. An
example of the simulaHons generator is given below.

Now, using these scenario choices, I generated 500 simulaHons for each of the four factors from 2015 to
2026. If the resultant scenario choice for a parHcular simulaHon and for a parHcular factor read ‘3’, that
would mean the resultant value of that factor would be the 3rd scenario in my original forecasHng
percentages sensiHvity analysis table (Exhibit B). I calculated the values for each of the factors for all the
years using a nested IF condiHon where the logical test was to check if the number on the resultant
scenario secHon (menHoned in the image above) is the same as the scenario in my original sensiHvity
table (Exhibit B) with 5% increase, 5% decrease and 7.5% increase scenarios. The true value would be
the corresponding value for that factor in that scenario; and the false value will be another if condiHon
progressing through the remaining scenarios. The key aspect in this framework is that the scenarios are
basically given equal weightage as the random funcHons picks the scenario without any probabiliHes.

Using this method for all the factors, I was able to come up with 500 simulaHons for each of the factors
across the Hmeline. Using these new values generated using the simulaHons, I was able to calculate the
free cash flows for all the simulaHons. Now, as for the discount rate, while there is a way to consider the
discount rate as the fidh factor and assign the random numbers for those as well, I decided to keep the
original four discount rates (Exhibit A) and to calculate the NPV for those set of discount rates. I ended
up with a set of 2,000 NPV, that is, 500 free cashflows across 4 different discount rates. The NPV
distribuHon for this framework is given below
Framework 3

This framework changes one key aspect from the previous one - the probabiliHes assigned to the
scenario choices. Since the resultant scenario choices in the previous framework are chosen by the
underlying random numbers, it is evident that all the factors involved have equal probabiliHes. This
framework is designed in such a way that we can assign our desired probability weights to each of the
factors. These probabiliHes are picked at random for the sake of simplicity in this case. For example, for
price of coal per ton, I have chosen 15%, 25%, 35% and 25% probabiliHes for scenarios 1,2,3 and 4
respecHvely. Likewise, probabiliHes are assigned to all of the factors.

Although we can assign random probabiliHes, an alternaHve way to assign truly random probabiliHes is
through the ‘Solver’ funcHon in excel where we can set the sum of probabiliHes to 1 by changing the
probabiliHes for each of the scenario. While this method might yield truly random probability values, the
drawback is we might end up with very low or very high probabiliHes for some scenarios.

Now that the probabiliHes are defined, we can arrive at our resultant scenario choices by using a nested
IF condiHon where the logical test is to compare the underlying random number corresponding to that
parHcular factor with the probability assigned. If the random number is less than the probability
assigned, then scenario 1 is chosen. Else, another IF condiHon is there for the false value which checks to
subsequent probability values and so on. By doing this, we arrive at a list of resultant scenario choices,
now no longer dependent only on the random numbers, but are tethered to the probability weights
assigned to them. Once the resultant scenario choices are calculated for all 500 simulaHons, the rest of
the analysis follows the same was as Framework 2. Ader calculaHng the free cash flows for all the 500
simulaHons, I calculated the NVPs with the same set of 4 discount rates. The frequency distribuHon for
the NPVs are given below.

Framework 4

Both of the previous simulaHons frameworks considered all the factors independently. In reality, there
might be a relaHonship between some or many of the factors considered in the case. For example, when
the price per coal increases, there might be a decrease in demand which might lead to a decrease in
producHon. Clearly, in this case, price and producHon are related. Hence while simulaHng the values, it is
important to consider the relaHonship between the varying factors. This model is designed in a way such
that we can define the relaHonship between factors in a way that will result in a scenario choice for one
factor based on the other factor it is depending on.
I have considered two relaHonships in this framework. Firstly, I have considered an inverse relaHonship
between price of coal and the producHon. This is because of the fact that when price increase, there
might be less demand which could lead to a reducHon in producHon of coal. This relaHonship is
implemented using a nested IF condiHon relaHng to the original four scenarios defined with 5% increase,
5% decrease and 7.5% increase (Exhibit B). For the base case, there will be no change in the price and
producHon equaHon. Using these relaHonships and the nested IF condiHon, the resultant scenario
choices for producHon is generated for all 500 simulaHons. The resultant scenario choices for prices is
sHll calculated using a random number funcHon and the roundup funcHon.


The second relaHonship I have considered is a posiHvely correlated relaHonship between price of coal
and the operaHng expenses of the firm. This is because of the assumpHon that in the case where the
operaHng expenses are increase, it is natural to increase the price of the coal to maintaining the profit
margin. In a way, operaHng expenses are driving the price of the coal. This relaHon is also implemented
using a nested IF condiHon as menHoned above. Using the nested IF condiHon, I have generated the
resultant scenario choices. For the investment factor, I have again used the rand() funcHon to generate
random numbers and the roundup() funcHon to come up with scenario choices.

Ader calculaHng the resultant scenario choices, the rest of the process in the analysis is similar to
previous frameworks. I calculated the free cash flows for each of the 500 simulaHons based on the
scenario choices using a nested IF condiHon. This again referenced the original sensiHvity analysis table
for each of the factors involved (Exhibit B). I calculated the NPVs using the four discount rates. The
frequency distribuHon of the NPVs for this framework is given below. It is noHceable that many scenario
choices are repeated in this framework because unlike the previous frameworks, the resultant scenario
choices are no longer only based on random numbers rather are based on the relaHonship between the
two factors. Hence, if there are scenarios where the two factors are strongly correlated, then there might
be repeHHon of factors which leads to a more Hght and coherent distribuHon of NPVs. It is also
noHceable that the distribuHon varies from the previous framework because of the linkages provided
while generaHng the resultant scenario choices.

Part B

Part B of this report discusses the implicaHons of use of simulaHons, analysis of simulaHons in a
valuaHon segng and the advantages and disadvantages of using simulaHons in capital budgeHng
decisions.
Advantages of SimulaCons in ValuaCon

SimulaHons can be immensely helpful while considering uncertainty in any applicaHon. For instance, this
case presents us with a lot of room for uncertainty with respect to growth and discount rates. In a
situaHon like this, only through simulaHons can we come up with a range of all the possible NPVs. The
higher the number of simulaHons the beher because of the law of large numbers. The mean values of
the simulaHons (in our case, NPV) will approach the actual average values with increasing simulaHons.

SimulaHons can also be very helpful in situaHons where there is limited data since we can generate data
based on our base case assumpHons with simulaHons. It can also be very flexible in that we can use
simulaHons to come up with specific models tailored to our needs like the three frameworks in this case.
We can also assign probability values as we did in framework 2 for the simulaHons model. Now these
probability values can be based off on historical data so that the simulated values make much more
sense.

Disadvantages of SimulaCons in ValuaCon

One clear disadvantage of using simulaHons in capital budgeHng decision is the fact that the simulaHon
model is only as good as the underlying assumpHons. In this case, the original sensiHvity analysis table. A
skewed set of input variables might render the simulaHons model less effecHve than necessary.

The simulaHons model might also be affected by the probability weights that are assigned to the factors.
Or if there any no probability weights to begin with; in which case the simulaHon model will solely be
based on random numbers.

In case like this, the point esHmate of NPVs with just one base case does not make much sense given the
uncertainty involved in the case. Only by using a simulaHons model we can come up with the distribuHon
of NPVs considering all the different factors affecHng it.

The range of values (NPVs) generated by the simulaHons model will be beneficial for all the parHes
involved in the acquisiHon deal. The company can understand the impact of various forecasHng
percentages, discount rates and relaHonship between factors (linked scenarios) and see how the NPV of
the project varies. Depending on how risk averse they are and the parent company is, they can come to a
decision by considering the volaHlity levels projected in the simulaHons. SimulaHons analysis are also
most useful when there are mulHple scenarios to be considered with the given factors like we have in
this case. Only then we can arrive at a comprehensive sensiHvity analysis which will entail most if not all
of the possible scenarios.

SimulaCons Explained

The underlying concept of a simulaHons model is to come up with hypotheHcal values for variables
which will provide a sense of range of what the actual values might be. These hypotheHcal values are
generated using the random number generator. Here, we have also considered frameworks (3&4) where
in the resulHng values are not only relied upon the random numbers but also with weighted probabiliHes
and linked scenarios.

In this case, we have four factors which are varying apart from the discount rates. These frameworks
above consider a range of values (500 simulaHons) which represent the range of values those factors
might take in real life. By doing so, we can account for all the different possibiliHes. The important factor
in all these analysis is how the final value (NPV) varies with varying factors and different discount rates.
The objecHve of these kind of simulaHon exercises is that when we are to make a decision on choosing
the factors/discount rates or to see how the variaHon in one factor might lead to a variaHon in our end
result (NPV), we can make an informed decision based on the simulated values.

The key takeaway from this exercise is that although simulaHons presents us with a range of hypotheHcal
values, they are sHll grounded based on the underlying assumpHons; and that using the probabilisHc
simulaHons model gives more sensible results than a simulaHon model completely based on random
numbers. Framework 4 takes the exercise a step further by considering the factors’ inherent relaHonship
into account. This addresses the fact that financial factors such as price, expenses and investments etc
are not independent rather they are linked to one another.

Exhibit A - WACC CalculaHons

Exhibit B - Forecast Parameters SensiHvity Analysis Table

Appendix
An alternaHve way to perform simulaHons other than using excel is to use a programming language
(Python or R) where we can write a simulaHon funcHon defining the parameters and call the funcHon for
all the variables we need simulated values for. Python libraries which are used are numpy and pandas.