Oil, Gas & Energy Law Intelligence

www.gasandoil.com/ogel/ Application of Portfolio Management

Issue : Vol. 4 - issue 1 to Optimize Capital Allocation in Oil
May 2006
and Gas Projects
by B. Lubiantara

About OGEL

OGEL (Oil, Gas & Energy Law Intelligence): Focussing on

recent developments in the area of oil-gas-energy law,
regulation, treaties, judicial and arbitral cases, voluntary
guidelines, tax and contracting, including the oil-gas-energy
geopolitics.

For full Terms & Conditions and subscription rates, please visit
our website at www.gasandoil.com/ogel/.

Open to all to read and to contribute

Our aim is for OGEL to become the hub of a global

professional and academic network. Therefore we invite all
those with an interest in oil-gas-energy law and regulation to
contribute. We are looking mainly for short comments on
recent developments of broad interest. We would like where
possible for such comments to be backed-up by provision of
in-depth notes and articles (which we will be published in our
'knowledge bank') and primary legal and regulatory
materials.

Please contact Editor-in-Chief Thomas Wälde at

[email protected] if you would like to participate in this
global network: we are ready to publish relevant and quality
contributions with name, photo, and brief biographical
description - but we will also accept anonymous ones where
there is a good reason. We do not expect contributors to
produce long academic articles (though we publish a select
Editor-in-Chief number of academic studies either as an advance version or
Thomas W. Wälde
an OGEL-focused republication), but rather concise
[email protected]
Professor & Jean-Monnet Chair comments from the author's professional ’workshop’.
CEPMLP/Dundee
OGEL is linked to ENATRES, the electronic energy law, policy
and economics information and discussion forum moderated
© Copyright OGEL 2006
OGEL Cover v1.2 by Thomas Wälde.
 APPLICATION OF P ORTFOLIO M ANAGEMENT TO OPTIMIZE
C APITAL ALLOCATION IN OIL AND G AS PROJECTS
BENNY L UBIANTARA

ABSTRACT

Oil and gas companies are constantly faced with decisions on how to invest limited
amounts of capital in order to maximize return. The traditional approach to select the
projects is usually to rank all available projects using common measures such as NPV,
IRR, or Profit to Investment ratio (PI) until all the available capital is exhausted. The
main weakness of this traditional approach is that it maximizes expected value but
ignores risk.

In the capital market world, one method of capital allocation that takes explicit account of
risk is the modern portfolio theory, which was initially developed by Markowitz in the
1950s and has been used extensively in stock market investment. The modern portfolio
theory allows one to choose sets of efficient portfolios with either the highest level of
expected return for a given level of risk, or the lowest level of risk for a given expected
return.

This paper shows how to apply the modern portfolio theory concept to the problem of
capital allocation in oil and gas projects.

INTRODUCTION consists in maximizing the expected

utility over the set of feasible portfolios.
Generally, the term “portfolio” refers to
The second approach, first proposed by
an investment mix in the financial
Markowitz, is very intuitive and reduces
market. In the case of real assets, a
the portfolio choice to a set of two
portfolio is defined as a set of possible
criteria – reward and risk – with possible
funding opportunities. The portfolio
tradeoff analysis. This paper will discuss
optimization problem in practice is to
the second approach.
select an investment mix to optimize the
satisfaction of the investors.
P ORTFOLIO THEORY
Traditionally, the portfolio selection In practice, a business will normally
problem is modeled by two different invest in a portfolio of investment projects
approaches. The first one is based on an rather than in a single project. The
axiomatic model of risk-averse problem with investing all available funds
preferences, where decision makers are in a single project is, of course, that an
assumed to possess an expected utility unfavorable outcome could have
function and the portfolio choice disastrous consequences for the business.

1
By investing in a spread of projects, an assets. It holds that if a company invests
adverse outcome from a single project is in many independent assets of similar
unlikely to have major repercussions. size, the risk will tend asymptotically
Investing in a range of different projects is towards zero. For example, as companies
referred to as diversification, and by drill more exploration wells, the risk of
holding a diversified portfolio of not finding oil reduces towards zero.
investment projects, the total risk Consequently, companies that endorse a
associated with the business can be strategy of taking a small equity in many
reduced. Many finance textbooks explain wells are adopting a lower risk strategy
the total risk relating to a particular than those that take a large equity in a
project into two elements: diversifiable small number of wells. However, the
risk and non-diversifiable risk (Figure 1). economic returns on independent assets
are, to a greater or lesser extent,
dependent on the general economic
• Diversifiable risk is that part of conditions and are non-diversifiable.
the total risk which is specific to Under these conditions, simple
the project, such as reserves,
diversification will not reduce the risk to
changes in key personnel, legal
zero but to the non-diversifiable level.
regulations, the degree of Markowitz diversification relies on
competition, and so on. By combining assets that are less than
spreading the available funds
perfectly correlated to each other in order
between investment projects, it is
to reduce portfolio risk. Markowitz
possible to offset adverse diversification is less intuitive than simple
outcomes occurring in one diversification and uses analytical
project against beneficial
portfolio techniques to maximize portfolio
outcomes in another.
returns for a particular level of risk. This
• Non-diversifiable risk is that part approach also incorporates the fact that
of the total risk that is common assets with low correlation to each other,
to all projects and which, when combined, have a much lower risk
therefore, cannot be diversified relative to their return.
away. This element of risk
arises from general market
conditions and will be affected Using these principles, portfolio
by such factors as rate of optimization is a methodology from
inflation, the general level of finance theory for determining the
interest rates, exchange rate investment program and asset weightings
movements, and so on. The most that give the maximum expected value for
critical non-diversifiable risk for a given level of risk or the minimum level
exploration companies is the oil of risk for a given expected value. This is
price. achieved by varying the level of
investment in the available set of assets.
The efficient frontier is a line that plots
There are two types of diversification: the portfolio, or asset mix, which gives
simple and Markowitz. Simple the maximum return for a given level of
diversification (commonly referred to as risk for the available set of assets.
market or systematic diversification in the Portfolios that do not lie within the
stock market) occurs by holding many efficient frontier are inefficient in that for

2
the level of risk they exhibit, there is a The first step: Collect information about
feasible combination of assets that results all the variables that affect the calculation
in a higher expected value and another of the cash flow of one of the projects,
which gives the same return at lower risk and estimate their probability
(Figure 2). distributions.

The second step: For each project, using

MEASURES OF RISK the Monte Carlo simulation, a number of
cumulative discounted cash flows and
When “risk” is discussed in relation to
their matching discounted investments
the overall return on an investment, the
can be generated simultaneously. From
word has come to mean the anticipated
these points, the NPV and the semi-
potential variability of the actual return
standard deviation of the NPV of each
from its expected value. Other things
project can be calculated.
being equal, the investment with the
greatest range of possible results is said
The third step: Determine the coefficient
to be the riskiest (Figure 3). The most
of correlation between projects. (The
widely used measure for risk is the
value of this coefficient can range from
variance (sometime standard deviation is
+1 in the case of perfect correlation,
used).
where the two assets move together, to –1
in the case of perfect negative correlation,
In a risky environment, variables are not
where the two assets always move in
known with certainty but follow
opposite directions. The coefficient is 0
probability distributions. With risk, the
when there is no association between the
outcome of a capital investment decision
assets and they are said to be
will be a particular value, but that value
independent.)
cannot be known in advance. When risk
is present, only the range of values and
The fourth step: Perform the optimization
the associated probabilities can be
using the algorithm explained below and
known or predicted.
create the efficient frontier curve.

P ORTFOLIO OPTIMIZATION
P ROPOSED METHODOLOGY
Markowitz introduced an intuitive model
The variance is a useful approximation of
of risk and return for portfolio selection.
risk: the aim is to maximize the expected
This model is useful to guide one’s
return under a certain level of variance,
intuition, and because of its simplicity, it
which is equivalent to minimizing the
is also commonly used in practical
variance under a certain level of expected
investment decisions.
return. In order to determine the variance,
the Monte Carlo simulation is used. In the
Given a set of n assets, A=(a1, a2,…an)
case of upstream oil and gas investment,
such that each represents a kind of asset
instead of using the variance or standard
(it could be projects, wells, fields,
deviation, the semi-standard deviation is
prospects, etc.), each asset ai has
preferable to use as a measure of risk.
associated a real valued expected return
The proposed methodology is as follows:
(per period) ri, and each pair of assets
<ai, aj> has a real-valued covariance sij.

3
The matrix s n*n is symmetric and each NPV, and Profit to Investment (P/I) of
diagonal element s ij represents the each project is shown in Table-1, the
variance of asset ai. A positive value D budget available is only 2,000 MM $,
represents the maximum acceptable risk. and the management needs to decide the
(Or a positive value R represents the allocation of this capital budget among
desired expected return.) the projects.

A portfolio is a set of real values Table-2 shows the approach using the
X=(x 1,x2,…xn) such that each x i traditional approach to select or rank the
represents the fraction invested in asset projects. Since the budget is only 2,000
ai. The value Sni=1Snj=1s ijx ixj represents MM $, therefore the projects selected are
the variance of the portfolio. The F, G, H, E, I, and C (note that project C
formulation problems are: is funded only 71%).

OP1: As mentioned before, this traditional

Given risk, maximize return. approach is to maximize the return but
The basic problem: ignore the risk; Table-3 shows the risk of
max Srix i each project (the risk is the semi-
s.t. Sni=1Snj=1s ijxixj <=D standard deviation of NPV obtained
Sxi = 1 from the Monte Carlo Simulation). For
0<= xi <= 1 (i=1,…n) illustrative purposes, the coefficient of
OP2: correlation between projects is assumed
Given return, minimize risk. to be zero (0), meaning that the projects
Min Sni=1Snj=1s ijxixj are independent of each other.
s.t. Srix i = d
Sxi = 1 Portfolio optimization is performed
0<= xi <= 1 (i=1,…n) using Excel Solver to create the efficient
frontier curve (Figure 4).
OP3:
Minimizing the risk and maximizing the ANALYSIS
return In Figure 5, it can be seen that the
Min ?Sni=1Snj=1s ijxix j – (1-?) Sni=1rix i portfolio X (maximized return), will
s.t. Sni=1xi = 1 have the NPV = 1,090.32 MM $ and
0<= xi <= 1 (i=1,…n) Risk = 268.59 MM $; another portfolio
where ? is a parameter 0<?<1. By (portfolio Y) will have the NPV =
varying the parameter ? between zero 1083.52 MM $ and Risk = 228.53 MM
and one, the efficient frontier is $. In this case, we can see that the
computed. management could choose portfolio Y
by scarifying the NPV by only 6.8 MM
The optimization algorithm above is $; meanwhile, the portfolio will
solved using Excel Solver. significantly reduce the risk by 39.8 MM
$. Furthermore, if the management has
ILLUSTRATIVE CASE concerns about the risk, and can only
Assume there are 10 oil and gas projects accept a portfolio risk no higher than
(A to J) available with total investment 200 MM $, then the management could
equal to 3,610 MM $. Investment cost,

4
consider portfolio Z, which has portfolio Chemical Engineers, New York,
risk = 186.28 MM $. 1991.

The percentages funded of projects in 4. Carter, David; Dare, William;

each portfolio (X, Y, Z) can be seen in Elliot, William, Determination of
Table-4. Mean-Variant Efficient
Portfolios Using an Electronic
CONCLUSION Spreadsheet, Oklahoma State
University, 2001.
• Portfolio optimization will provide a
trade-off between risk and return. 5. McMillan, Fiona, Risk,
This approach can also quantitatively Uncertainty and Investment
measure the risk versus return and its Decision Making in Upstream
implications on the portfolio Oil & Gas Industry, University
selection. of Aberdeen, 2000.

• The creation of the efficient frontier

curve from the portfolio optimization
does not provide the “final answer.”
It provides a perspective for
management to make better long-
term decisions in line with the
corporate strategy.

• Portfolio optimization empowers the

decision makers to measure and
manage the project risks.

REFERENCES
1. Bodie, Z.; Kane, A.; Marcus, A.J,
Essentials of Investments. 3rd ed.
McGraw-Hill, 1997.

2. Adam, S.T.; Albers, J.A.;

Howell, J.L.; Lung, J. and
McVean, J., “Portfolio
Management for Strategic
Growth,” Oilfield Review, Vol.
12, No. 4 (Winter 2000/2001).

3. Herbst, Anthony, “The Many

Facets of Risk in Capital
Investment Evaluation,” in
Investment Appraisal for

5
 TABLES
Table-1
Investment
Project Cost NPV P/I
MM $ MM $
A 300 100 0.33
B 450 200 0.44
C 325 146 0.45
D 400 166 0.42
E 250 127 0.51
F 300 200 0.67
G 385 250 0.65
H 385 200 0.52
I 450 210 0.47
J 365 150 0.41
Total 3610

Table-2 (project ranking)

Investment
Project Cost NPV P/I
MM $ MM $
F 300 200 0.67
G 385 250 0.65
H 385 200 0.52
E 250 127 0.51
I 450 210 0.47
C 325 146 0.45
B 450 200 0.44
D 400 166 0.42
J 365 150 0.41
A 300 100 0.33
Total 3610

Table-3 : Project Risk Measures

Semi-Std
Project Deviation
MM $
A 20
B 40
C 35
D 25
E 30
F 150
G 125
H 100
I 150
J 125
Total 800

6
Table-4 : Portfolio
(based on Budget 2,000 MM $)
% funded % funded % funded
Project Portfolio X Portfolio Y Portfolio Z
A 0 0 0
B 0 51 100
C 71 100 100
D 0 0 6
E 100 100 100
F 100 100 63
G 100 100 100
H 100 100 75
I 100 27 19
J 0 0 0

NPV (MM $) 1090.32 1083.52 1049.95

Risk (MM $) 268.59 228.83 186.28

7
PICTURES

Figure-1 : The effect of Diversification

Total risk

Risk

Diversifiable
risk

Non-diversifiable risk

Number of assets

Figure-2 : Efficient Frontier

Efficient Frontier for Alternative

Portfolios

8
 Figure 3 : Measures of Risk

Measures of Risk

Probability
Low risk

High risk

0 Return

Figure 4 : Efficient Frontier Curve

Efficient Frontier
1150

1100
Return (MM $)

1050

1000

950

900

850

800
50 100 150 200 250 300

Risk (MM $)

9
Figure 5 : Portfolio Alternatives

Efficient Frontier
1400
Portfolio Y :
1300 Return = 1,083.52 MM $ Portfolio X :
Return (MM $)

Risk = 228.83 MM $ Return = 1,090.32 MM $

Risk = 268.59 MM $
1200 Portfolio Z :
Return = 1,049.95 MM $
Risk = 186.28 MM $
1100

1000

900

800
50 100 150 200 250 300

Risk (MM $)

10
FIGURES

Figure-1 : The effect of Diversification

Figure 3 : Measures of Risk

Measures of Risk
Total risk

Risk

Probability
Diversifiable Low risk
risk

High ri
Non-diversifiable risk

Number of assets 0

Figure-2 : Efficient Frontier

Efficient Frontier for Alternative

Portfolios Figure 4 : Efficient Frontier Curve

Efficient Frontier
1150

1100
Return (MM $)

1050

1000

950

900

850

800
50 100 150 200 250

Risk (MM $)

11
Figure 5 : Portfolio Alternatives

Efficient Frontier
1400
Portfolio Y :
1300 Return = 1,083.52 MM $ Portfolio X :
Return (MM $)

Risk = 228.83 MM $ Return = 1,090.32 MM $

Risk = 268.59 MM $
1200 Portfolio Z :
Return = 1,049.95 MM $
Risk = 186.28 MM $
1100

1000

900

800
50 100 150 200 250 300

Risk (MM $)

12