See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/249922608

Investment decision in oil and gas projects using real option and risk
tolerance models

Article  in  International Journal of Oil Gas and Coal Technology · January 2008

DOI: 10.1504/IJOGCT.2008.016729

CITATIONS READS

2 236

2 authors, including:

Gabriel Costa Lima

University of Campinas
58 PUBLICATIONS   117 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

A Model for the Valuation of an Inspection Program in Offshore Oil Production Facility View project

All content following this page was uploaded by Gabriel Costa Lima on 12 February 2016.

The user has requested enhancement of the downloaded file.

Int. J. Oil, Gas and Coal Technology, Vol. 1, Nos. 1/2, 2008 3

Investment decision in oil and gas projects using real

option and risk tolerance models

Gabriel A. Costa Lima and Saul B. Suslick*

Center of Petroleum Studies and Institute of Geosciences,
State University of Campinas – Unicamp,
P.O. Box 6152,
Campinas, SP 13083-970, Brazil
E-mail: gabriel@ige.unicamp.br
E-mail: suslick@ige.unicamp.br
*Corresponding author

Abstract: This paper presents a model for valuation and decision making,
integrating discounted cash flow, real-options pricing and preference theory,
aiming to cover the following questions: i) what is the current value of a
project?; ii) what is the optimal investment rule?; iii) what is the optimal
working interest? The traditional model suggests that, when the project value is
above its investment cost, the corporation should invest immediately and incur
in 100% working interest. The real option pricing suggests that the corporation
should only invest if the project’s current value is at least 1.85 times investment
cost. The preference theory suggests funding only 44.38% working interest,
and partners must acquire the remaining 55.62%. These tools must be
integrated in order to allow a more realistic treatment of risk. In general, when
the uncertainty (volatility) of cash flow components increases, the two models
give more divergent results.

Keywords: uncertainty; real options; capital budgeting; preference theory;

petroleum exploration; production.

Reference to this paper should be made as follows: Costa Lima, G.A. and
Suslick, S.B. (2008) ‘Investment decision in oil and gas projects using real
option and risk tolerance models’, Int. J. Oil, Gas and Coal Technology,
Vol. 1, Nos. 1/2, pp.3–23.

Biographical notes: Gabriel A. Costa Lima is a Mining Engineer, PhD in

Business and Mineral Resources Policy. Currently, he is an Visiting-Professor
at State University of Campinas and Researcher at Center for Petroleum
Studies (CEPETRO) of the same university. His areas of professional interest
are those associated with valuation, uncertainty modelling, real options
analysis, risk analysis and mineral economics.

Saul B. Suslick is a Professor at the Department of Geology and Natural

Resources, Institute of Geosciences, State University of Campinas – Unicamp.
He is responsible for the Petroleum Economics and Economic Evaluation
Methods for E&P Projects in the Science and Petroleum Engineering and
Geosciences Graduate programs at the Unicamp. His main areas of interest are
Decision and Risk Analysis Techniques, Development of Quantitative Methods
for Mineral and Petroleum Applications (resources appraisal, mineral
forecasting, estimation of environmental impacts, etc.) and Mineral and

Copyright © 2008 Inderscience Enterprises Ltd.

4 G.A. Costa Lima and S.B. Suslick

Petroleum Economics Studies (oil taxation models, mineral market analysis,

etc.). Currently, he is Director of the Center of Petroleum Studies – CEPETRO
at Unicamp.

1 Introduction

The valuation and decision making of capital-intensive projects have been a topic of
concern for large corporations and governments in the oil and gas business, in part
because the resource is finite and investment is high – for example, irreversible
investments in the development phase may reach billions of dollars. Several studies
concerning modern investments theory, such as those in the oil and gas industry, have
three important characteristics according to Dixit and Pindyck (1994):
1 uncertainty over the future operational cash flow
2 irreversibility of the investment
3 value of the timing or some leeway to implement decisions.
The irreversibility of investment gives rise to implications in the financial status of
corporations, depending on capital exposed to risk and the magnitude of budget of the
company. For example, consider an investment of US$ 100 million in a risky oil project.
For a company with a budget as high as US$ 1 billion, this risky investment might be
tolerated. On the other hand, for a firm with a budget of, say, US$ 300 million, managers
might postpone this risky project or sell a share of this project, since the potential loss is
too high compared to its budget. Since unsuccessful results of an investment may give
rise to serious financial impacts on the corporation, the irreversibility of investment is
one of the main driving forces of investor’s risk aversion.
The irreversibility of most part of the investment is a frequent characteristic of the
capital intensive industries. For example, in case of a crisis in the oil industry,
equipments such as rigs, platforms, infrastructure, etc. will have a large reduction of
value in a secondary market. Even equipment of more general applications will
experience a value reduction due to a decrease of the oil-related assets.
In order to manage a project under a scenario of future uncertainty, coupled with
investment irreversibility, the manager needs managerial flexibilities (real options) to
adapt the project to new market conditions. This is clearly stated in Trigeorgis (1996)
that without this managerial proactive role, uncertainties will tend to generate a
symmetric distribution for Net Present Value (NPV), whereas, with active management,
the NPV will have an asymmetric distribution to the right. This positive asymmetry in
the NPV generated by managerial flexibility has value and can be properly estimated
using option pricing methods.
If the management has the option to choose the time to invest, this flexibility has
value and must be considered in the model of valuation and decision making. Since the
option to invest is alive until the time the decision is implemented, if the corporation
invests today, it kills the opportunity of investing in the future, when the market
conditions, technology, as well as other market components may be better. Then, by
investing today, the corporation incurs in an opportunity cost by not waiting to allocate
this capital in the future. In addition, because the corporation invests in a
 Investment decision in oil and gas projects 5

capital-intensive project, there is a long period of time (which may reach 20 years or
more) of cash-flow exposure to risk, so that firms may prefer to participate with less than
100% in the project. Therefore, in the process of investment valuation and decision
making, corporations must consider, in an integrated way, the process of valuation,
strategic decision making and tolerance to risk exposure. Three questions are of
paramount importance:
1 What is the current cash flow value of the investment?
2 What is the optimal decision rule to invest, that is, should the corporation invest
now or in the future?
3 What is the optimal working interest for the project?
The solutions to these questions have been achieved using the traditional model of
valuation and decision making, based on the NPV of the project’s cash flows, that is,
invest as long as NPV is positive and incur in 100% of working interest.1
But, contrarily, McDonald and Siegel (1986), Dixit and Pindyck (1994), Trigeorgis
(1996) and Copeland and Antikarov (2001) pointed out that results from the traditional
model do not properly account for the role of uncertainty, irreversibility and managerial
flexibilities. Alternatively, they suggest the use of option-pricing techniques.
In addition, real-world practice shows that most large projects are financed by a pool
of companies, what is different from suggestions from the NPV approach. A model to
find the optimal working interest (W) is suggested by Cozzolino (1980), Walls (1995),
Walls and Dyer (1996), Lerche and Mackay (1996) and Wilkerson (1998) for petroleum
exploration and production projects.
But, these two approaches have been applied in a separate way. Costa Lima (2004)
suggests the use of an integrated model considering Monte Carlo simulation,
option-pricing and optimal working-interest to give decision makers more realistic
information to make their choices. The risk of project is estimated by considering
uncertainty in future cash flow inflow and outflow through Monte Carlo simulation.
The strategic decision rule is achieved according to the option-pricing model.
The optimal working interest in joint venture projects is found according to the optimal
working interest from concepts of preference theory.
This paper is structured in three sections. Section 1 presents some details on the
integrated methodology for valuation and strategic decision making. Section 2 applies
the model to decision making of capital-intensive project for heavy-oil deep-water
production. Section 3 presents some discussions and implications.

2 The integrated model

2.1 NPV of the project’s cash flow

The first step in the process of investment decision making is the estimation of indicators
from the project’s future cash flows. Traditionally, the NPV is the main indicator for
investment valuation and decision rule, since it has been recognised in financial literature
(Brealey and Myers, 1992) as theoretically correct to measure the creation of value for
6 G.A. Costa Lima and S.B. Suslick

stakeholders. If a project generates over time a stream of operational cash flow (X(t)) and
requires a stream of investments (I(t)), its NPV is (Equation (1)):
N E [ Xt − I t ]
NPV = ∑ (1)
t =0 (1 + µ)t

where µ is capital cost of each cash flow. Equation (1) provides the expected value of the
NPV under static scenarios for price, cost, production, etc. Since there is uncertainty
about the future, the NPV may change as long as fluctuations in price, cost, production
rate, etc. take place. Then, NPV is a random variable whose future values will come from
a probabilistic distribution. Equation (1) gives the expected value of the NPV, but this
is not sufficient to make a decision since this indicator does not account for the impact
of uncertainty on future cash flows, except via risk premium of the discount rate.
In order to estimate the risk of the project, a Monte Carlo simulation of the NPV may be
carried out.
The traditional decision-making model according to the NPV only considers the risk
of a project by means of a premium in the discount rate – the riskier the project, the
lower the NPV. This model has three main drawbacks:

1 the potential gains from the positive side of uncertainty are not taken into
account

2 the value of waiting to invest in the future if current market conditions are bad
is ignored

3 the optimal working interest is always 100%, irrespective of magnitude

of investment.
These issues will be discussed over the next sections.

2.2 Optimal working interest (W)

In order to discuss the problem of optimal working interest in a risky project, consider an
example involving the allocation of US$ 600 million between two alternative
investments:

1 Invest all of the money in one large project with an expected NPV of
US$ 300 million and a standard-deviation (risk) of US$ 141.42 million.

2 Invest all of the money in five projects, that is, the corporation allocates 20% of
its budget to each project. In this case, the corporation’s rate of return will have
the following return/risk profile2: E[NPV] = US$ 300 million and
σ [NPV] = US$ 63.25 million.
These two alternatives have the same expected return, but different risk levels. Just by
increasing the number of non-correlated projects, the return of portfolio remains the same
and its global risk drops – this is the remarkable role of diversification. In Table 1, the
effects of diversification through different working interest values on risk and return of
portfolios are shown.
 Investment decision in oil and gas projects 7

Table 1 Effects of diversification on portfolio risk and return level

Number of projects Working interest (W) Portfolio risk Portfolio return

(US$ × 10 ) (US$ × 106)
6

1 100.00% 141.42 300

2 50.00% 100.00 300
– – – –
– – – –
– – – –
10 10.00% 44.72 300

The choice of working interest depends on the investor’s risk tolerance, since all
portfolios give the same return of US$ 300 million. The investor’s risk tolerance will
depend on the amount exposed to risk compared to the investor’s total stock of wealth
and his risk preference characteristics.
The theoretical foundations of decision making involving choices under uncertainty
(such as selecting working interest in a project of high CAPEX) is the preference theory
developed by Von Neumann and Morgenstern (1953) which advocates that the
usefulness of things determines their attractiveness. Basically, choices under uncertainty
or risk attitudes of decision makers are organised into three main groups:
1 risk-neutral
2 risk-averse
3 risk-prone.
According to Luenberger (1998), the risk-neutral individual is one to whom uncertain
outcomes are valued as expected values, paying no attention to the potential of gains and
losses. A risk-prone individual is one to whom uncertain outcome is valued by more than
the expected value. A risk-averse investor has more concern with the potential of loss
than that of gains and, therefore, this individual values uncertain outcome by less than
expected value.
In this paper, to find the optimal working interest in this model, it is assumed that
individuals as well as corporations are risk-averse in capital-intensive projects. This is
the case of large heavy-oil projects, where investment irreversibility contributes to
motivate management towards limiting risk exposure by taking, for example, less than
100% working interest in the project, depending on the interaction among project risk
profile, magnitude of the corporation’s financial budget and decision maker’s risk
attitudes (Campbell et al., 2001; Cozzolino, 1980; Nepomuceno et al., 1999; Newendorp
and Schuyler, 2000; Walls and Dyer, 1996). For the petroleum industry, Cozzolino
(1980) and Walls (1995) suggest the use of the exponential utility function to model the
choices made under uncertainty. The complete exponential utility equation is:

U ( x ) = a − b × e− Xi / T (2)

where a and b are constants, Xi is the random monetary quantities and T is the
corporation’s risk tolerance. The absolute value of utility (positive or negative) is just a
number without much significance and is not enough for taking decisions. On the other
hand, utility values are suitable for comparisons – for example, if U(X) = 11 and
U(Y) = 8, then, X is preferable to Y.
8 G.A. Costa Lima and S.B. Suslick

On the other hand, the certainty-equivalent or Risk-Adjusted Value (RAV) is the

main concept to estimate the optimal level of financial participation or working interest
(W). According to Luenberger (1998), RAV is that value whose utility is equal to the
expected value of utility of K possible monetary outcomes (Xi) with probability pi. From
Equation (2), if a = 0 and b = 1, the RAV is found from:
k
− e−(RAV /T)
= ∑
i= 1
(
pi − e
− (W × X i / T )
) (3)

where T is the corporation’s risk tolerance. From Equation (3), we get:

⎛ K − ((W × X i ) / T ) ⎞ ⎟⎟
RAV = − T × l n ⎜⎜ ∑ p i × e (4)
⎜⎝ i = 1 ⎠⎟⎟

The model to estimate the optimal working interest is given in Equation (4), that is, the
decision-maker should select W that maximise the RAV. In practice, W is found
numerically. The use of Equation (4) requires two inputs:
1 corporate risk tolerance (T)
2 probability distribution of Xi – for example, the NPV of the project.

2.3 The optimal investment decision-rule

The optimal investment decision rule must take into account not only the expected future
cash flows but also the strategic or operational options that are available for management
over the life of a project. As in the case of financial options, these real options are
understood as a right, but not as an obligation to be implemented and have value to be
added to the traditional NPV. As described by Trigeorgis (1986), the true project value is
always equal to or greater than its traditional NPV.
Most modern investments in exploration and production share the following
characteristics:

1 future uncertainty in variables such as cost, price, exchange rate, etc.

2 irreversibility of investment, leading to new considerations in the process of the

project’s acceptance and management of risk exposure

3 timing or strategic real option to implement decisions.

Dixit and Pindyck (1994) consider that such investments should be analysed using the
flexible approach of real-option pricing. Costa Lima and Suslick (2002) pointed out that
the interaction of uncertainty and timing may aggregate value to the project, in part
because the investor’s loss is limited to the irreversible investment, whereas the upside
potential is theoretically infinite.
The first step in valuation and decision-rule from the option-pricing model is the
modelling of the dynamics of the asset’s future price. Following Black and Scholes
(1973), we assume that the changes in the present value of future cash flows evolve as a
risk-neutral geometric Brownian motion:
dV = (r − δ )Vdt + σVdZ (5)
 Investment decision in oil and gas projects 9

where V is the present value of project’s cash flow, r is the risk-free interest rate, δ is the
dividend rate, σ is the volatility and dZ is Wiener’s increment or white noise.
In order to find the optimal investment decision-rule, let F(V) be the value of the
option to invest in an oil and gas project. If we assume that option maturity is at least
four years, the investment option can be regarded as independent of time. Then, by
following standard procedures in financial economics, McDonald and Siegel (1986) and
Dixit and Pindyck (1994) showed that, under the assumption of market efficiency and
non-arbitrage opportunities, F(V) must satisfy the following ordinary differential
equation (PDE)

1/ 2σ 2V 2 FVV′′ (V ) + (r − δ )VFV′ (V ) − rF (V ) = 0 (6)

In order to use Equation (6), the following boundary conditions are considered:

F (0) = 0 (7)

NPV (V ∗ ) = F (V ∗ ) = V ∗ − I (8)

FV′ (V ) = 1 (9)

Equation (7) means that if V = 0, F(0) = 0, that is, there is no chance of increase in
V in the future, which is considered an absorbing property of the GBM. Equation (8)
means that the corporation should invest as long as V reaches a trigger value (V ∗)
and not merely V ≥ I. Equation (9) is the smooth-pasting condition or the last
boundary condition for optimisation (see Dixit and Pindyck, 1994, Chapter 4).
The solution of F(V) is:

1 (r − δ ) ⎛⎜⎛⎜ (r − δ ) 1 ⎞⎟ 2r ⎞⎟⎟
2 1/ 2

β= − + ⎜⎜
⎜⎜ σ 2 − ⎟ + ⎟ (10)
2 σ2 ⎜⎝
⎜⎝ 2 ⎠⎟ σ 2 ⎠⎟⎟

β
V∗ = I (11)
(β −1)

⎛ ∗ ⎞
⎜ V − I ⎟⎟ β
F (V ) = ⎜⎜⎜ ⎟⎟V if V < V ∗ (12)
⎜⎜ (V ∗ )β ⎟⎟⎟
⎝ ⎠

F (V ) = NPV(V ) = V − I if V ≥ V ∗ (13)

where β is positive root of a quadratic equation derived from the ordinary differential
Equation (6), I is the present value of investment cost and V ∗ is the trigger value, that is,
the minimal project value to invest immediately. Equation (12) is the value of the
investment option if V < V ∗ and Equation (13) gives the value of investment option if
V ≥ V ∗, that is, Equation (13) has the same value as the traditional NPV. In order to use
the set of Equations (10)–(13), as analogous to the case of financial options, the
real-options analyst needs those five input parameters of Table 2.
10 G.A. Costa Lima and S.B. Suslick

Table 2 Analogy of the determinants of financial and real-option pricing models

Financial options Real options Symbol

Underlying asset Present value of project’s cash flow ($) V
Exercise price Present value of investment cost ($) I
Financial asset’s future volatility Future volatility of project’s cash flow (%) σ
Financial asset’s dividend rate Future dividend from project’s cash flow (%) δ
Risk-free interest rate Risk-free interest rate (%) r

The estimation of these parameters is a condition to valuation and decision making in

projects. In the case of financial options, except for future volatility, they can be
estimated from past market data. On the other hand, in the case of projects, the procedure
is much more complex, since these parameters, especially dividend and volatility, must
be estimated by considering the dynamics in future cash flows. For example, Copeland
and Antikarov (2001) suggest the use of Monte Carlo simulation to estimate the future
volatility of projects, whereas Costa Lima and Suslick (2006) derive an analytical
expression for volatility of projects considering two sources of uncertainty.

3 Analysis of a capital-intensive project

Recently, significant offshore heavy oil discoveries were made in ultra-deep-water of

Brazil and West Africa. Among the new technologies required for commercial
production of these heavy oil reservoirs in deep water, new artificial lift devices and long
horizontal wells length can be detached, completed with efficient sand control
mechanisms (Pinto et al., 2003). Besides providing commercial value for the heavy oil
wells, it is expected that these new technologies will create a new value for such
resources. Indeed, heavy oil reserves can become more available over time in these
environments if the cost-reducing effects of new technologies (reduction of CAPEX and
OPEX) more than offset the cost-increasing effects of depletion. Offshore heavy oil can
be considered good assets for future revenues if new technologies are available to
transform the potential reserves of heavy oil into viable projects.
These heavy-oil projects are a good sample to evaluate the performance of the
proposed methodology. Most heavy-oil development projects in this type of environment
are typically capital-intensive, where irreversibility, uncertainty, timing and risk-aversion
are present. In order to discuss some numerical results, consider a heavy-oil project with
the geological, economic and financial characteristics shown in Table 3.
The cash flow of this project is in Appendix. For simplicity, we use a simple linear
taxation of 50% representing the ‘government take’ of this project based on its specific
characteristics, such as water depth, operational conditions, etc. Under a static scenario of
price, cost, production and fiscal regime, using Equation (1), we have E[NPV] =
US$ 391.38 million. According to the traditional model of valuation and decision
making, this project should be “accepted immediately because it creates value for
stockholders and the corporation should incur in 100% of its funding and get 100% of its
profits”. On the other hand, over the entire project life, the true values of price, cost,
production and tax may differ from those expected ones from the static scenario and the
NPV may become more positive or even negative.
 Investment decision in oil and gas projects 11

Table 3 Geological, technical and economic characteristics of the project

Technical and economic characteristics Properties value

Oil Reserve (MMbbl) 389.30
Water depth (metres) 1500.00
Oil quality 25° API
Oil production peak(MMbbl/year) 68.50
CAPEX (US$ MM) 887.02
OPEX (US$/bbl) 12.00
Oil spot price (US$/bbl) 30.00
Discount rate (%) 13.00%
Government take (%) 50.00%

However, in practice, the real NPV may be quite different and, actually, we will know its
true value only when the oil production reaches economic exhaustion, that is, the real
NPV is a random variable whose expected value is US$ 391.38 million. Since the NPV is
a random variable, the optimal decision making should not be based solely on expected
values, but should also consider its uncertainty because even a positive expected NPV
may become negative if market conditions change over time. As a result, a natural
complement to a static NPV is its risk analysis.

3.1 Estimation of the risk of the project

In finance, broadly speaking there are two types of risk: systematic and unsystematic.
There is a consensus that there is no reward for unsystematic risk because it can be
eliminated by diversification in well developed markets. It is important to note that this is
true for investors in markets with a large menu of assets, what is not always the case in
specific domestic markets of the major oil companies. For example, in non-mature
financial markets such as in Brazil,3 the total risk (unsystematic) is what is of concern.
In this paper, we consider risk as the possibility of an unfavourable outcome, such as
the chance of a negative NPV. In this sense, a risk analysis consists of finding the
probabilistic distribution of the NPV from uncertainty in its primary variables such as
price, cost and production, among others.
Although the NPV of this project is highly positive, it has some risk because of a
possible change in market conditions, which is a fair and rational motivation for risk
analysis. For this project, we use the following assumptions:

• Future values of oil production cost (OPEX): this variable is modelled as a

triangular distribution, whose most likely value is US$ 12/bbl, whereas the
optimistic cost is US$ 6/bbl and the pessimistic cost is US$ 18/bbl.4

• Future values of oil price: this variable is modelled according to a lognormal

distribution where mean price is US$ 30/bbl and standard deviation is
US$ 15/bbl.

• Future values of production: oil production is expected to reach a yearly peak

production of 68.5 million bbl and drop year after year to the ending value of
2.9 million bbl.5 The uncertainty in oil production is modelled considering that
12 G.A. Costa Lima and S.B. Suslick

future production will be distributed as triangular distribution, with pessimistic

yearly production equal to 50% of the most likely value, whereas the yearly
optimistic production is 50% above the most likely production.
This modelling assumes that the components of the project’s cash flow are statistically
independent, although in reality, for oil production projects, this assumption is incorrect,
since most variables are directly linked to oil price to some degree, apart from their
dependency on the non-linearity in tax structure, logistics, etc.
The next step consists of simulating thousands of possible paths for these uncertain
variables using the Monte Carlo technique. Then, after carrying out a simulation with
10,000 iterations, we get the histogram of the NPV shown in Figure 1.

Figure 1 Cumulative frequency of the NPV and project’s risk level

Simulation of the NPV shows that its values range from US$ −1.3 billion to US$ 2.0
billion, but there is a concentration of values around the mean of US$ 276.07 million.6 In
practice, most of these extreme values of the NPV have little significance, since they
occur in very unlikely situations, that is, high cost and low price and vice-versa.
From Figure 1, the probability of a negative NPV is 30.95%. This means that, in the
case of unfavourable events, the corporation may not be able to recover its full
investment allocated to the project. Since investment is irreversible, the corporation must
adopt two complementary policies in order to decide if it should invest in this project
or not:
• firstly, the irreversibility of investment, together with the ability to invest in the
future in a scenario of less risk, gives rise to a new decision-making rule, which
is according to the real-option pricing approach
• secondly, the irreversibility of investment, together with corporation risk
aversion, may imply in a working interest of less than 100% in the project.
Another possibility is to make price hedging, but this will also reduce profits because of
the cost of hedge. Over the next sections, these two approaches will be applied to the
case of an offshore oil project.
 Investment decision in oil and gas projects 13

3.2 The optimal rule of investment (F, K, V∗)

Traditionally, the NPV has been the indicator for static valuation and decision-making.
If it is even merely positive, the benefits are in excess over the costs and the project
should be accepted. Under the static approach, this logic is correct but, if we take the
effect of investment irreversibility and future uncertainty, this approach is no longer
right. In Section 2.3, the application of the theory of exercising a financial option to the
case of investment decision in projects is proposed. According to this theory, the option
is not exercised simply if the value of underlying asset is just above its exercise cost, but
only if it is sufficiently high. Analogously, the corporation should exercise its option to
invest only when project value is sufficiently above its investment cost. Therefore, there
is a critical project value (V ∗) to trigger investment. The required input parameters of
Equations (11)–(13) are shown in Table 4.

Table 4 Inputs of decision-making based on real-options models

Input parameter Symbol Value

Project’s cash flow value V(MM US$) 1278.41
Project’s investment value I(MM US$) 887.02
Project’s dividend rate δ 11.95%
Project’ future volatility σ 51.00%
Risk-free interest rate r 5.00%

The optimal decision rule based on the theory of an option-pricing model requires,
among other factors, characteristics of project cash flow and of the market interest rate.
The risk-free interest rate is assumed to be that paid by government to its bondholders
whose maturity is similar to that of the managerial flexibility. It is not easy to find the
project’s volatility and dividend rate. The estimation of future project volatility is always
a complex task (even for financial assets when there is a long time series available to be
used as a proxy) because:
1 there are no marked transactions for project cash flows
2 petroleum projects are unique, so that each reservoir has its own geological,
technical, operational and environmental particularities.
As an alternative solution, in this paper, we use Monte Carlo simulation in order to
access the project volatility as the standard deviation of rate of return of the project, as
suggested by Copeland and Antikarov (2001). Firstly, we calculate the expected project
NPV under static scenarios, which we call NPV0. But, over the project’s entire
operational life, there is a chance of an infinite number of different cash flow trajectories.
These different cash flow trajectories are the cause of the project’s volatility, due to
possible fluctuations in the expected value of the NPV from future cash flows.
The expected rate of return of the project (µV) at time 1 is:
CF0 (1 + µ) + NPV1
E [ µV ] = −1 (14)
NPV0

The expected project rate of return is very close to the cost of capital (or discount
rate), that is, 13%. But, since cash flows in the future are uncertain, there is a chance that
thousands of µV may occur. After a simulation, we get a distribution of the rate-of-return
14 G.A. Costa Lima and S.B. Suslick

with the following two moments: E[µV] = 13.00% and σ [µV] = 51.00%. Note that
expected rate of return is close to the discount rate and the standard deviation of rate of
return is the volatility of this project. According to Costa Lima and Suslick (2006),
for typical oil projects, if the volatility of oil price is around 20%, the volatility is
much higher – around 51.00% as in this paper.7 These numbers show that project
volatility is usually much higher than price volatility. As a result, the common
assumption that project’s volatility is equal to price volatility may give rise to significant
errors, undervalue projects, and create wrong critical values to exercise the option
to invest.
The estimation of a project’s dividend yield is also complex. In an investment in
stocks, investors receive dividends as a result of profit distributed by a corporation.
The dividend of project can be understood as the cash flows per unit of time (month,
year, etc.), but this simple approach has many shortcomings, especially in the case of
long-lived projects with irregular cash flows. In this paper, we estimate project dividend
rate (δV) as a weighted average between cash flows and production:

∑ Q ( t ) × f (t )
N
t =0
δV = (15)
∑ Q(t )
N
t =0

where Q(t) is the annual production and f(t) is ratio of the yearly cash flow to the sum of
total cash flows.8 Dividends for exhaustible resources, such as oil projects, can be
associated with the reserve’s production flow over time. Using Equation (15), dividend
rate is δV = 11.91% and it means a fraction of the project’s total value that is generated
each year, that is, nearly 12.00% of the asset’s value is produced.
Now, we have all inputs necessary to estimate the investment option value, optimal
decision rule and value of waiting to invest in the future. From Equation (11), V * is
US$ 1821.09 million. This means that the corporation should exercise its option to invest
only if the current value of the project is at least US$ 1821.09 million. Since the current
value of this project is US$ 1278.04 million, the optimal decision is to wait to invest in
the future. In other words, it means that the “decision to invest immediately requires
V equal to at least 2.05 times the investment cost and not merely a positive NPV”.
The value of the option to invest (F), with the flexibility of choosing to exercise the
right to invest at some moment in the future during option’s maturity, is
US$ 491.29 million, which is higher than the traditional NPV ($ 391.38 million). The
value of waiting to invest in the future (K) is US$ 100.76 million, which comes from the
uncertainty in future cash flows, that is, volatility of projects. According to this new
decision rule, even if the NPV is positive today, the act of waiting is able to create more
value for shareholders. Thus, the optimal policy is to invest when the flexibility to wait
has no more value (K = 0), that is, when F = NPV. This condition will hold and be true
when the project’s current value is sufficiently above its cost, that is, V > V* > I.
The volatility of projects plays a very important role in both option valuation
(Equations (12) and (13)) and decision making (Equation (11)). According to the classic
investment theory, project value is represented by its NPV and the optimal decision is:
invest as long as V > I, that is, a positive NPV. Contrarily, in the real-options model,
the investment right is exercised as long as V ≥ V* and both F and NPV have the same
value, that is, the strategic value of the option to invest (F) and the intrinsic option
value (NPV) are equal. Then, F may be considered as the intrinsic option value plus its
time value. Another understanding is: when V = V *, we have that F(V*) = V * − I and,
 Investment decision in oil and gas projects 15

consequently, V * = F(V *) + I, that is, we can see F as part of the cost to invest now
instead of waiting for a better opportunity in the future. Therefore, because of volatility,
we always have V* > I and not V * ≥ I.
From Equation (1), the NPV is US$ 391.38 million and, according to traditional
decision making, the corporation should invest right now. On the other hand, from
Equation (12), the option value, which captures the value of the flexibility to invest not
only now, but also in the future, is US$ 491.29 million and clearly the option value is
above the NPV stand alone.
In the real option pricing model, F (investment option) and K (value of waiting
option) are non-negative functions, whereas the NPV can be negative if I > V. The entire
results of both approaches in valuation and decision making can be seen in Figure 2
through a sensitivity analysis of the NPV, F and K (option value) to project current value.

Figure 2 Sensitivity of NPV, F and K to project current value

As the current value of project increases, Figure 2 shows an increase of NPV, F and a
decrease in K. There are three important regions for decision making:

• I: 0 ≤ V ≤ US$ 887.02 million. The intrinsic value of the option (NPV) is

negative and the corporation will not invest. Meanwhile, F is positive. Why?
Since the future is uncertain – volatility of project is 51%, there is a potential
that project value increases to US$ 1821.09 million or more. Then, the
investment option is much higher than the intrinsic value.

• II: US$ 887.02 million ≤ V < US$ 1821.09 million. The NPV is positive, but not
sufficiently high. Once again, the effect of irreversibility and uncertainty implies
that the optimal policy to maximise the option value is to wait until the project
value reaches the value of US$ 1821.09 million. The option value is still high,
but the option should therefore be exercised in the future.

• III: V ≥ US$ 1821.09 million. The investment option must be exercised

immediately. Observe Figure 2 to confirm that the option value of waiting (K)
has no more value if V is above US$ 1821.09 million.

In general terms, the results of valuation and decision making for this project,
synthesised in Figure 2, are in agreement with the literature. For example, Costa Lima
(2004) has shown that for some oil projects, V* is on the average 2.14 times the
16 G.A. Costa Lima and S.B. Suslick

investment cost. These findings are in agreement with Dixit and Pindyck (1994, p.136).
In addition, it is worth mentioning that for reasonable input parameters of the real-
options model, particularly volatility and dividends, V* may be two times the investment
cost or even higher.

3.3 The optimal working interest in the project (W)

In many cases, even though the project value is sufficiently above its investment cost,
many corporations prefer to develop it in partnership (such as a joint venture), which
seems to contradict the optimal decision rule of the option-pricing theory. But this
practice in capital-intensive projects has some reasons due to the following
characteristics:
1 the magnitude of investment cost
2 technology availability to develop the project alone
3 search for synergy among different business units.
In such situations, a common question is: What is the optimal level of financial
participation (working interest) in this project?
The answer cannot be found through pure NPV analysis since, from this approach,
the corporation should incur in 100% of investment and revenue as long as the NPV is
positive. An alternative solution can be found using the theory of finding the optimal
working interest to maximise the RAV in Equation (4). This equation requires two main
inputs:
1 the probabilistic distribution of NPV
2 the corporation risk tolerance (T).
The corporation risk tolerance is difficult to estimate. Wilkerson (1988) argues that T can
be estimated as a fraction of, for example, corporation market value, budget and other
strategic variables. Walls (1995) shows empirically that T is around 25% of the capital
allocated to petroleum explorations. Meanwhile, for other phases in the E&P chain,
T may assume different values.
In this paper, it is assumed that the corporation budget is US$ 300 million and its risk
tolerance is US$ 120 million, that is, T = 40% of the budget. From Equation (4), the
analyst can estimate the RAV for this project, considering different assumptions and
characteristics of the corporation and of the project. As shown in Equation (4), RAV
depends on utility function, corporation risk tolerance and project risk profile – in some
cases, the RAV is an increasing function of W, whereas, in others, it is a decreasing or
even constant function. The main objective is to estimate W, numerically or analytically,
in order to maximise the RAV, given the corporation’s profile. For this case study of oil
project, the sensitivity of RAV to W is shown in Figure 3.
As shown in Figure 3, an increase in W will increase the satisfaction of decision
maker up to around 44.38%. Higher values of W reduce the satisfaction of decision
maker through the decrease in RAV. In addition, if W is higher than 93%, the RAV
becomes negative. Over the interval between 0 and 100%, the RAV is a non-linear and
concave function of W. The theory of maximising the RAV shows that, contrarily to the
traditional view, the “corporation will not take 100% working interest in the project’s
cost and revenue, but only 44.38% because this value gives the highest RAV”.
 Investment decision in oil and gas projects 17

Note that the curve of the RAV to W takes into account the uncertainty in price,
production and operational cost together with the corporation’s utility function.
Nevertheless, it is important to observe that, in order to get realistic results, it is
important to search for the best estimation of T because of its impact on W, that is, an
increase in T tends to increase W in the project and vice-versa. In Figure 4, an extended
sensitivity of the RAV to W for different values of T is presented.

Figure 3 Sensitivity of RAV to optimal working interest (W)

Figure 4 Sensitivity of RAV to optimal working interest: different values of corporation’s

risk tolerance

Note that, in the base case, T = US$ 120 million and W = 44.38%. If T increases to US$
200 million, W is also increased to 75%. Analogously, if T decreases to US$ 80 million,
W is also reduced to only 30%. In terms of valuation and strategic decision making, by
taking W = 44.38%, the company will spend around US$ 393.66 million in the
investment and receive a NPV of only US$ 173.69 million. If a negative outcome occurs,
the amount under risk is limited to that fraction of the investment, that is, US$ 393.66
18 G.A. Costa Lima and S.B. Suslick

million. This policy of reduction in the project’s NPV may be understood as the price to
limit risk exposure or as a hedge scheme against possible catastrophic losses.
The result from preference theory and risk-adjusted analysis intends to complement
the results of the NPV and option pricing because of two main reasons:
1 the investment cost may be quite high if compared to the corporation’s budget
2 even in the case of a project with a current value equal to two or more times its
cost, it may become profitless since volatility over the future may, for example,
generate a scenario of high cost and low prices.
In this case, a rational policy intended to reduce the risk of losses and, consequently, of
financial troubles, may be the choice of a small fraction of the project investment and
revenues.
This practice is very common in the oil exploration and production business and this
framework may be a valuable tool for the decision process in capital-intensive projects.
Finally, by taking a working interest of less than 100% in the project, the corporation
may invest in more than one project and, consequently, reduce the corporation’s
portfolio risk.

4 Discussions and implications

The proposed model provides a complement to results of the classic NPV in valuation
and decision making. It integrates valuation, risk quantification, strategic flexibilities and
decision making, using theories of preference and real option pricing. Although more
sophisticated, this integrated model requires two input parameters that are difficult to
estimate: corporation risk tolerance (T) and project volatility (σ).
In the present case study, we have pointed out the need to complement results of the
traditional cash flow when analysing investment in capital-intensive projects, such
as those of oil and gas industry. Table 5 presents a comparison between the traditional
and the proposed valuation and decision model.

Table 5 Results from the two valuation and decision models (some approximations have
been used to the nearest round number)

Subjects Main indicators Traditional approach Proposed model

Valuation V (million US$) 1278.41 1278.41
and risk I (million US$) 887.02 887.02
NPV (million US$) 391.38 276.07
F (million US$) – 452.34
Decision Risk (%) 30.95% 30.95%
making NPV (million US$) 391.38 491.29
V* (million US$) 887.02 1821.09
W (%) 100.00% 44.38%
*
Optimal VF Since NPV is positive, invest Even though NPV is positive,
decision now and take 100% working invest in the future (when V is
interest US$ 1821.09 million) and
incur in a working interest of
44.38%
 Investment decision in oil and gas projects 19

According to the traditional cash flow model, the NPV is the indicator for valuation and
decision making. The project NPV is US$ 276.07 million and the corporation should
invest immediately and incur in 100% of the project. The proposed model has a
complementary structure, where F is the valuation indicator and the decisions are made
according to the results of V* and W.
The numerical analysis summarised in Table 5 shows that, although the NPV is
positive, the option to invest is not deep-in-the-money. Therefore, the optimal decision is
to wait for an increase in V to, at least, US$ 2.10 billion. This may happen, for example,
due to oscillation in price, costs and production, as well as in other strategic variables.
By choosing only projects whose V is at least equal to V*, the corporation follows a more
conservative and risk-limiting policy, which is an advance and is radically different from
the traditional NPV. Nevertheless, this procedure does not assure that the option exercise
is profitable, because the result of the exercise of real options, contrary to financial
options, is not immediate, but depends on a cash flow from many years of operations.
In this context, since the project’s investment is high compared to the corporation’s
risk tolerance, the corporation’s optimal working interest is 44.38%, whereas other
partners must fund the remaining 55.62% of the investment. This framework derives
from the chosen parameters, which may not be constant over time. If some of them
undergo oscillations, for example, volatility or operational cost, global output may be
very different, requiring that the analysis must be updated continuously in order to
improve the corporation’s valuation and decision making in the search for an efficient
resource allocation and value creation for stockholders.

5 Conclusions

This paper presented an integrated framework, combining theories of real-options and

preference, useful for valuation and decision making under uncertainty for
capital-intensive projects. The outputs of the real-options model may be used to
complement those of the traditional NPV, especially by incorporating the value of
uncertainty and irreversibility.
The results from the preference theory allow the corporation to estimate its optimal
level of financial participation in a risky project that is compatible with the utility
function of the decision maker. This integrated model represents a significant gain when
compared with the traditional one (based solely on expected values) by considering the
relationship among irreversible investment, risk tolerance, timing and risk-aversion.
In addition, findings from this new model is that its results tend to diverge from the
traditional valuation and decision-making model as the level of uncertainties increases as
in the case of heavy-oil in deep water projects where the timing and frequency of the
unknown technologies for cost reductions are speculative and not fully dominated by oil
industry.

Acknowledgements

The authors would like to thank Margaret Armstrong and Alan Galli (Ecole des Mines
de Paris) and an anonymous referee for helpful comments on the previous version of this
paper and CAPES, CEPETRO/UNICAMP, CNPq and PETROBRAS for financial
support to carry out this research.
20 G.A. Costa Lima and S.B. Suslick

References
Black, F. and Scholes, M. (1973) ‘The pricing of options and corporate liabilities’, Journal of
Political Economy, Vol. 81, pp.637–659.
Brealey, R.A. and Myers, S.C. (1992) Corporate Finance, 4th edition, New York: McGraw-Hill.
Campbell Jr., J.M., Campbell Sr., J.M. and Campbell, R.A. (2001) Analyzing and Managing Risky
Investments, John M. Campbell.
Copeland, T. and Antikarov, V. (2001) Real Options: A Practitioner’s Guide, New York:
Texere LLC.
Costa Lima, G.A. (2004) ‘Using real options and preference theory in projects of production and
exploration of petroleum’, PhD Dissertation, State University of Campinas, Campinas, Brazil
(in Portuguese).
Costa Lima, G.A. and Suslick, S.B. (2002) ‘Impacts of environmental constraints on the optimal
decision-making process applied to Brazilian petroleum industry’, Journal of Canadian
Petroleum Technology, Vol. 41, pp.37–43.
Costa Lima, G.A. and Suslick, S.B.(2006) ‘Estimating the volatility of mineral projects considering
price and operating cost uncertainties’, Resources Policy, Vol.31, No. 2,
pp.86–94.
Cozzolino, J.M. (1980) ‘Controlling risk in capital budgeting: a practical use of utility theory for
measurement and control of petroleum exploration risk’, The Engineering Economist, Vol. 25,
No. 3, pp.161–186.
Dixit, A.K. and Pindyck, R.S. (1994) Investment under Uncertainty, New Jersey: New Jersey
University Press.
Lerche, I. and Mackay, J. (1996) Economic Risk in Hydrocarbon Exploration, San Diego:
Academic Press.
Luenberger, D. (1998) Investment Science, Oxford University Press, p.594.
McDonald, R. and Siegel, D. (1986) ‘The value of waiting to invest’, Quarterly Journal of
Economics, Vol. 4, No. 101, pp.707–727.
Nepomuceno, F., Suslick, S.B. and Walls, M. (1999) ‘Managing technological and financial
uncertainty: a decision science approach for strategic drilling decision’, Natural Resources
Research, Vol. 8, No. 3, pp.193–203.
Newendorp, P.D. and Schuyler, J. (2000) Decision Analysis for Petroleum Exploration,
2nd edition, Aurora, Colorado: Planning Press, p.606.
Pinto, A.C.C., et al. (2003) ‘Offshore heavy oil in Campos Basin: the Petrobras experience’, 2003
Offshore Technology Conference, Houston, TX.
Trigeorgis, L. (1996) Real Options – Managerial Flexibility and Strategy in Resources Allocation,
3rd edition, Cambridge: MIT Press.
Von Neumann, J. and Morgenstern, O. (1953) Theory of Games and Economic Behavior,
3rd edition, Princeton, NJ: Princeton University Press.
Walls, M. (1995) ‘Corporate risk tolerance and capital allocation: a practical approach to
implementing an exploration risk policy’, Journal of Petroleum Technology, Vol. 47, No. 4,
pp.307–311.
Walls, M. and Dyer, J. (1996) ‘Corporate risk propensity and firms performance: a study of
petroleum exploration industry’, Management Science, Vol. 42, No. 7, pp.1004–1021.
Wilkerson, J.P. (1988) ‘A case history application of risk preference techniques to
evaluate petroleum exploration prospects’, PhD Dissertation, University of Texas at
Austin/USA.
 Investment decision in oil and gas projects 21

Notes
1
The model assumes that a corporation has enough credit to finance the project, irrespective of
capital limitations.
The expected NPV is: E[NPV] = 0.2 × 300 + 0.2 × 300 + 0.2 × 300 + 0.2 × 300 + 0.2 × 300 =
2

US$ 300 million. The risk of NPV is: σ[NPV] = [(0.2) × 2 × 10 + (0.2) × 2 × 10 + (0.2) × 2
2 16 2 16 2

× 10 + (0.2) × 2 × 10 + (0.2) × 2 × 10 ] = US$ 63.25 million.

16 2 16 2 16 1/2

3
Petrobras is the most traded asset in the Sao Paulo Stock Exchange and is present in the majority
of institutional and private portfolios of assets.
4
These assumptions considering an oscillation of 50% are intended to show how to make decisions
in a high environment of uncertainty when dealing with heavy-oil projects in deep waters.
In practice, according to American Association of Cost Engineers (AACE) recommendation,
making decision with and oscillation range of no more than 10%.
5
In practice, a petroleum engineer knows that oil production declines exponentially, at least when
approaching exhaustion or higher production costs due to heavy-oil production profile.
6
The simulated expected value will differ from the expected value because as predicted by the
Jensen’s inequality (see Dixit and Pindyck, 1994).
7
This high project volatility may also reflect the cost structures of the project.
8
The total discounted cash flow is $ 1,278,416,393.97 and the first operational cash flow is
268,177,684.38. Then the first dividend is: f(1) = 268,177,684.38/1,278,416,393.97 = 20.98%.
The other values of dividends are estimated accordingly. An estimation of the true dividend of
the project is found using Equation (15) which is a weighted average.
22 G.A. Costa Lima and S.B. Suslick

Appendix
 Investment decision in oil and gas projects 23

Nomenclature

EBTI Earnings Before Tax and Interest

EATI Earnings After Tax and Interest
NCF Net Cash Flow
Div Dividends

View publication stats