3rd
Annual Real Options
Conference
1999 Leiden
Netherlands Institute for Advanced Studies
Monday: Introductory Workshop/Tutorials
| Theoretical Issues in Real Option Valuation | New Product Development, Infrastructure and Firm Valuation |
Empirical Evidence | Valuing Product(ion)
Flexibility | Real Options and Learning | Valuing Natural Resource Investments
Tuesday: Case Studies in Real Options | Sharing
Experiences in Oil Development/Natural Resources and Panel | Competition and Strategy | Sharing
Experiences in Pharmaceuticals and Panel Discussion | Real
Options, The Environment and Agency Issues | Keynote
Address: Stephen A. Ross
1A. Introductory Workshops/Tutorials
Using Spreadsheets and Real Options Software Tools in the Petroleum
Industry
Paper, Spreadsheet
Marc Paulhus and Gordon Sick,
University of Calgary
There is a multitude of sources of uncertainty affecting the operations of
a petroleum company. This paper sets up a real option analysis approach to
valuing a general oil-field development project that integrates several types
of risk into a multinomial tree/forest pricing scheme. The approach blends
the decision tree model commonly used for binary one-time risks with binomial
tree models for the generalized diffusion process(es) followed by the
underlying commodity prices. The project specification part of the model has
been developed in conjunction with Tuffs, an exploration geologist at Union
Pacific Resources. We model the complex inter-dependencies of the cash flows
as realistically as possible. Parameters include a finite quantity of
reserves, variable rates of production, and cost of production as a function
of remaining reserves. A detailed sensitivity analysis is performed with the
aim of identifying dependencies of the outcome on the input variables.
The analysis will demonstrate the Real Options Software Engine (ROSE)
package.
Using Spreadsheet Add-Ins to Value Managerial Flexibility
By Wayne Winston, Indiana
University
We will discuss how the EXCEL simulation add-in @RISK and the EXCEL
simulation/optimization add-in RISKOPTIMIZER can be used to easily and
efficiently value many real options. We begin by modeling the value of an
underlying asset as a Lognormal random variable and value a European
expansion option and contraction option by simulation. Next we will show how
to value a European real option when the underlying asset is modeled as a
jump diffusion process.
Next we will show how to use RISKOPTIMIZER to value American real options.
Our example will be the classic model involving start-up and shutdown of a
gold mine. RISKOPTIMIZER uses genetic algorithms to search for a set of
changing cells that optimize the mean value of a simulation. These changing
cells can correspond to asset prices that trigger a reopening or a shut down
of the gold mine.
Finally, we will briefly discuss how RISKOPTIMIZER can be used to value
managerial flexibility in classical capital budgeting problems when
uncertainty involves factors other than price. Our example will concern the
timing of entering a product into a new market. If we enter now we obtain a
larger market share, but we do not yet know whether the market is large
enough to be profitable. If we wait a while, we will have more information
about the size of the market, but if we enter we lose the first-mover
advantage and gain a smaller share of the market.
1T. Theoretical Issues in Real Options Valuation
Dan Calistrate, Marc Paulhus, University
of Calgary and Gordon
Sick, University
of Calgary and NIAS
One of the main challenges in pricing complex real option projects is in
the choice of the binomial/multinomial numerical scheme for representing the
various possible underlying stochastic processes. A computationally simple
scheme is desirable. On the other hand, correct modelling of the typical
commodity price or interest rate process must incorporate general assumptions
on the diffusion parameters such as mean reversion, variable volatility with
time, etc. This paper introduces a new binomial tree model which converges in
distribution to a general diffusion:
dS = mu(S,t)dt + sigma(S,t)dW.
The model is close in spirit to the work of Nelson & Ramaswamy. However
there are advantages over the N&S approach. The discrete volatility fit
is precise, allowing for fast convergence to the continuous distribution
while retaining computational simplicity through recombination in the tree.
Also, no transformation of the original process is necessary - for certain
diffusions, the existence of a readily integrable transform (as required in
the N&S model) may not be guaranteed. The one-dimensional binomial scheme
can be successfully applied to pricing projects contingent on multiple
diffusion processes after an appropriate orthogonalization of the factors.
The benefits of such an approach are particularly apparent when the
correlation parameters vary with time. In most existing models, this prevents
the trees from recombining.
Valuing Real Option when Time to Maturity is Uncertain
Tony Berrada, University
of Geneva
Recent literature in real options has shown how similar investment
opportunities and financial options are. The analogy is clear on the
underlying risk as a key variable and the irreversibility of investment,
however the relation between the contractual time to maturity of a financial
option and the time period over which the opportunity is available is less
evident. The assumption often made for investment opportunity is an infinite
time to maturity, which simplifies the computation of the option value.
However, when competition is considered in a project for a new market for
example, the time to maturity could be the time before a competitor enters
the market with a similar or substitute project. Therefore assuming that the
time when the competitor will decide (or will be ready) to enter the market
is known with certainty is a rather strong assumption. Furthermore the
trigger value, i.e. the underlying value at which the project should be
undertaken, is an increasing function of the time to maturity (in the simple
case of constant risk free rate and investment cost). Therefore, assuming an
infinite lifetime for the option overestimates this trigger value. If the time
to maturity was not known with certainty the trigger value would be altered,
and investment would be undertaken at a different time. In this paper we
model the uncertainty over the time to maturity and study its impact under
different assumptions for the level of uncertainty. We provide a lower and
upper bound for the value of an american contingent claim when the time to
maturity is uncertain but the density is known. We also propose an
approximation of the value of the american contingent claim using a
particular assumption on the form of the exercise boundary. The results
obtained confirm the fact that the trigger value (for an investment
opportunity, i.e. a call option) is lowered by the uncertainty on the
maturity of the option.
2A. New Product Development, Infrastructure, and Firm
Valuation
Onno Lint, Erasmus University Rotterdam, Eindhoven
University of Technology and Enrico Pennings, Catholic University of Leuven,
Erasmus University Rotterdam
Current phase-review processes for new product development cannot properly
capture the economic value of managerial flexibility to continue or abandon a
project at different stages of development. Due to this, the discussion
whether it is fruitful to skip phases in the development process in order to
create time-to-market advantage is left open. By applying insights from the
valuation of real options, this article proposes a framework for the
assessment of new product development at different stages, derives criteria
to speed up the development process or not and introduces the options
portfolios, which serve as a basis for assessment and as a tool for choosing
an optimal set of business initiatives from a variety of feasible
alternatives.
The implementation of the option approach may also entail initial
obstacles. In particular the estimation of uncertainty, the key variable in
the option approach, will challenge management. Sales predictions from senior
managers can be substantially different and, hence, are subject to
uncertainty. By a first step in implementing options-based aproaches to NPD
at Philips Electronics we found that sales predictions and business events
serve well to estimate the uncertainties described. Second, managerial
experience can help determining the uncertainty of a project by judging the
uncertainty of analogous projects in the past. Finally, at Merck, average
stock volatility of companies in the same business as intended for the new
product serves as a proxy for project uncertainty.
Helen Weeds, Fitzwilliam
College, University
of Cambridge
This paper considers optimal investment behavior when a firm faces both
technological and economic uncertainty, in the context of an innovative
research project. Specifically, the value of the prize for successful
innovation follows a stochastic process, while discovery itself takes place randomly
according to a Poisson arrival. The firm's optimal investment strategy, in
the form of a pair of trigger points for investment and abandonment, is
derived. As in the Dixit (1989) model of product market entry and exit, the
investment trigger exceeds the Marshallian entry point given by the sum of
variable costs and interest charges on capital costs. However, the
abandonment trigger may in some cases exceed the Marshallian exit point, in
contrast with the Dixit result, giving rise to reverse hysteresis. Thus, a
firm will abandon research projects rapidly as their profitability falls, in
some cases despite the existence of positive expected profits.
In addition, the model provides a unified framework encompassing two
classic real options models as limiting cases. As the expected rate of
innovation becomes large the outcome converges to the investment trigger for
the McDonald & Siegel (1986) model of a single irreversible investment
opportunity. As discovery becomes slow, on the other hand, the Dixit (1989)
entry / exit model is approached.
Real Option Valuation of Strategic Platform Investments
Enrico Perotti and Silvia Rossetto, U. Amsterdam,
Netherlands
In this paper we investigate the theme of real option valuation of
platform investment in a context of strategic differentiation.
Platform investment has achieved mainstream interest in management
science, but to date there are hardly any theoretical analysis. We propose a
definition of platform as a market-opening technological innovation that
creates a new product market. A classic example is the establishment of
operating systems such as DOS or Windows. The innovating firm has the ability
to choose the size of the platform, which in itself defines the potential
range of product differentiation. The innovating firm must incur the cost of
the creation of the platform at a time when there is still large demand
uncertainty, while later entrant will be able to operate on the established
platform by creating compatible products. On the other hand, the firm controlling
the platform has some critical advantage on product and productive processes.
In our model we consider a world characterized by dynamic uncertainty on
future consumer demand. In the product market the innovative firm which
creates the platform has superior strategic powers. We model first the
initial decision on the platform size, which indirectly affects the extend of
comparative advantage gained by the innovating firm; we then study the entry
decision on specific products within the platform decision of the innovating
firm and of a potential entrant. We compute the value of waiting to invest,
the point in time when each of the two firms chooses to invest and on which
type of product (differentiation). In future research we will consider the choice
of investing in a competing platform versus in a platform-compatible product.
The model allows to determine the impact of controlling the platform on ex
post competition (as in the case of Microsoft and Windows) and the impact of
uncertainty over both the initial platform investment and the extend of later
entry. The possibility of waiting to invest is valuable, and increases in
value with demand uncertainty. Less obvious is the impact of uncertainty on
platform size and ex post entry.
The platform size (which is related to the potential range of
differentiated products) is influenced by the hypothesis that the greater is
the differentiation, the weaker are the cross-price effects. We show that the
firm in charge of the platform may become dominant in some products as later
entrants face either higher productive costs or a delay disadvantage relative
to the platform-owner, and choose therefore a more limited range of entry
among products.
The outcome of this decision process is not trivial, as it depends on
product features and expected market demand evolution.
Valuation of a Biotechnology Firm: AnApplication of Real-Options
Methodologies
David Kellogg (Sprint USA), John M.
Charnes and Riza Demirer (University of Kansas)
Much of the value contained in early stages of pharmaceutical projects is
in the promise of developing a blockbuster drug. This is especially apparent
for firms in the biotechnology industry. Many biotech firms have significant
valuations, yet do not have product revenue because their products are in
early stages of development. In the past ten to fifteen years investors have
bid up the stock prices of these firms, and their prices have remained high
relative to their discounted cash flow valuations.
In this paper we compare the value of a biotechnology firm, Agouron
Pharmaceuticals, Inc., to the sum of the values of its drug development
projects. Each project's value is found using the decision tree (DT) and
binomial-lattice methods. An influence diagram (ID) method is also used and
discussed. We compare our computed values of Agouron to actual market values
at selected points in time during the development of Viracept¬, a drug used
to treat HIV-positive patients.
The approach and results are of interest to stock analysts because it
provides a means to value biotechnology companies that have no current
revenue. Financial analysts in pharmaceutical companies can use these methods
to value projects and compare their relative worth for capital budgeting
purposes. Executive management of pharmaceutical firms can use these methods
to better understand the values of their projects and convey them to
investors. Finally, academic readers can find an interesting case study that
demonstrates the use of real-option valuation methodologies.
The decision tree method considers the net present value (NPV) of eleven
possible outcomes of each project and calculates an expected NPV by
multiplying the NPV by its probability of occurrence. The eleven possible
outcomes are failure at each stage of development, and five levels of
commercial success.
Values for Agouron were also found using a binomial lattice with the
addition of a growth option. The growth option was added because the
development of an initial new molecular entity (NME) is similar to purchasing
a call option on the value of a subsequent NME. Without engaging in
development of the initial NME, it is not possible to develop the subsequent
NME. Intermediate results from the decision tree method were used to
calculate the beginning asset values and volatilities for valuing both the
initial NME and the growth option with the binomial lattice method.
We show how a real options approach can be used to value a biotechnology
firm. Usage of average assumptions and binomial lattices seems to work better
when projects are in earlier stages of development and less is known about
the drug. As projects move into later stages, more specific information
regarding time to launch, market size and probability of success should be
utilized in DT/ID models to reflect the value of the firm more accurately.
2T. Empirical Evidence
M. Diane Burton, Alberto Moel, and Peter Tufano, Harvard
University
Understanding how managers make decisions and react to external stimuli is
an important topic that academics in many disciplines study. Economists tend
to build normative models of rational behavior, such as real options models,
that represent how value-maximizing agents should optimally choose to
exercise the options that business presents. Sociologists tend to build
positive models of behavior, characterizing how individuals and organizations
are likely to respond to the world around them. Far too often, these
single-disciplinary research efforts are unconnected.
In our project, we seek to use insights from both economics and sociology
to study one particular managerial reaction to the environment: the decision
by gold mining firms whether to open or shut mines in response to changes in
gold prices. In previous research (Moel and Tufano 1999), we analyzed this
decision using the real option framework of Brennan and Schwartz (1985).
Using a hand-collected database, we tracked the annual opening and closing
decisions of developed North American gold mines in the period 1988-1997. Our
analysis provided strong support for the real options approach as a useful
model to describe and predict a mine's opening and shutting decisions. In
particular ,we found that the decision to open and close demonstrated
hysteresis, and was sensitive to both the level and volatility of gold prices
as well as to mines' operating costs.
The prior paper took the mine as the unit of observation, and tested if
mine and market characteristics affected the decision to close. We used a
narrow economic perspective, and tested whether managers behaved
"rationally" in response to market and firm conditions. There are,
however, entirely different ways of looking at these decisions, where the
unit of analysis is not the rational decision manager, but the context within
which the manager operates. In particular, the fields of sociology and
organizational behavior have developed models of managerial decision-making
where the organizational structure and context dictate the boundaries of the
manager's decision-making capabilities. Sociologists and organizational
theorists have, over the past twenty years, argued that economic transactions
are embedded in social relations (Granovetter 1985). These social relations
often account for what economists observe to be "non-rational"
behavior. In this new paper we propose to explore this line of research and
examine the social and organizational factors that influence a firm's
decision to close or reopen a mine.
In this current project, we recognize that important decisions such as
whether to close a mine are made by managers, who often work in larger
organizations (firms), and whose decision-making may reflect more than narrow
economic criteria. Drawing upon the literature in sociology which deals with
factors that might affect how firms respond to external stimuli, as well as
the related behavior finance field, we will seek to understand how the
following factors could influence the decision whether or not to close a gold
mine. This is work-in progress, and we are unsure how much of this data we
can collect:
- Stakeholder concerns:
Whether the location of the decision-maker affects the decision to close
a mine and temporarily layoff workers. Related to the number of workers
to be laid off, local labor market conditions, and the union status of
the workforce.
- Regulatory costs:
Whether the decision to close varies between the US
and Canada,
where laws and environmental regulations differ.
- Financial impediments:
Whether the existence of prior debt and hedging contracts are related to
the decision whether or not to close a mine.
- Decision maker
characteristics: Whether the background of senior executives (mining v.
financial background, age, compensation) affects the decision to close.
- Organization
structure: Whether the organizational structure (single line or
multidivisional), size and age of the firm affect the decision to close
- Prior experience:
Whether prior experience in closing a mine affects the current decision
to close.
- Overall organization
profitability: Whether firm profitability affects the decision to close.
Growth Option Company Acquisitions: In Search Of An Optimal Option
Portfolio
Tomi Laamanen, Helsinki
University of
Technology
Previous work shows that strategic synergy advantages in company
acquisitions are often not immediately realized, but rather affect the
combined growth options of acquiring and target firms. This paper sets out to
explore the stock market responses to changes in the composition of a firmÍs
growth option portfolio. It is hypothesized that with a small number of
growth options, the growth opportunities are constrained setting a limit to
the expected value of the firm. With a large number of growth options, the
value-increasing effect of variability gradually levels off and option
interactions start to dominate decreasing the expected value of the growth
option portfolio. More specifically: (1) The higher the growth expectations,
i.e. the higher the market-to-book ratio of a firm, the less positive the
stock market reaction to acquisitions of additional growth options. (2) The
lower the growth expectations, i.e. the lower the market-to-book ratio of a
firm, the more positive the stock market reaction to acquisition of
additional growth options. Furthermore, (3) the higher the market-to-book ratio,
the more positive the stock market reaction to focus decisions such as
divestments and (4) the lower the market-to-book ratio, the less positive the
market reaction to divestment decisions.
The hypotheses are tested empirically by studying the abnormal returns
from acquisitions and divestments of three major Finnish companies possessing
different growth option characteristics. Altogether 188 acquisition and
divestment events during a period from 1987 to 1997 are studied. The data
would seem to provide support for the hypotheses concerning the relationship
between the cumulative abnormal returns in acquisitions and the
market-to-book ratio. In connection with divestments, the results are opposite
to the hypotheses. Divestments would seem to reduce the book value more than
the market value resulting into increasing market-to-book ratios when the
initial market-to-book ratios are low. The results also corroborate partly
the earlier results concerning the cumulative abnormal returns to
shareholders of the acquiring and target firms. As expected, the stock market
reactions to divestments and acquisitions differ. The differences are not,
however, identical to what has already been documented in the literature. In
addition to being dependent on the nature of acquisition or divestment, the
stock market reactions to acquisitions and divestments seem to be company
specific and time period specific. A signaling explanation is put forward to
explain the stock market reactions to divestments.
Bart Lambrecht, University
of Cambridge and
William Perraudin, Birkbeck
College, London
and CEPR
This paper introduces incomplete information and strategic behaviour into
an equilibrium model of firms holding real options.
The main contributions of our study are, first, to provide a practically
useful and yet consistent way of combining real option values and threats of
preemption with incomplete information in a capital budgeting framework.
Incorporating incomplete information is important as it is otherwise hard to
see what prevents firms from colluding and splitting the surplus efficiently.
Given some basic data about a firm's investment costs, revenue potential
after investment, and conjectures about the distribution of costs faced by
competitors, our techniques permit one to derive an optimal investment
strategy in a simple fashion.
Second, we explore the implications of our analysis for equity returns.
Our model allows one to decompose equity volatility into price changes
reflecting changes in publicly observable profit variables and price changes
caused by the resolution of uncertainty about competitors. Parametrizing our
model so that in simulations it yields second and higher moments that
resemble those of US biotech stocks, we show that 'competitor risk' may
comprise a substantial fraction of these firms' total volatility especially
close to the firms' initial listing dates. As we show, the degree to which
the threat of preemption may accelerate investment according to our estimates
is comparable to the firm facing an expected time to ruin on its individual
real options of approximately half a year.
Third, we show that incomplete information and competitive pressure
interact in a way that crucially affects real option values. For firms of a
similar type, if information is complete, preemption will lead to the total
destruction of option value. Under incomplete information, these firms will
be better off. On average, if one integrates across firms of all different
types, incomplete information leads to a reduction in firm value, however.
3A. Valuing Product(ion) Flexibility
J. Bengtsson, LIT, Sweden
Managers in the manufacturing industry of today ask the question how to
justify the higher costs often associated with a flexible manufacturing
system. Even though they are aware of all the advantages of a flexible
manufacturing system they face a problem since the traditional capital
budgeting approaches undervalue the potential of a flexible system. Using the
traditional approach might in the worst case result in rejection of a
profitable project, wherefore other methods should be used to evaluate
projects with embedded flexibility.
This paper considers applications of real option thinking and valuation of
manufacturing flexibility using option-pricing theory and will present
results and experiences from research carried out with industry. Three
Swedish midsize companies constitute the empirical base. A case description
of each of the three companies will also be presented in the paper. The
descriptions will show how the companies view flexibility today and how this
can be looked at from an option-theoretic point of view. Also, the different
kinds of external and internal uncertainties that the companies are, or could
be, affected by will be described as well as the means to achieve the desired
flexibility.
The companies are producing totally different products; electrical motors,
casting reels and industrial robots, to industry and consumer customers.
While all industrial robots produced are customer specific and require a
flexible production process, the electrical motors and casting reels have a
big proportion of itÍs annual volume in standardised products. For the two
latter products it might thereby be an alternative to separate standardised
product and customer-specific and produce these in different line where the
line producing the customer-specific products has a flexible process.
Different production set-ups can be evaluated
using option pricing to find an optimal set-up, which maximise NPV.
The paper will focus on flexibility in the assembly part of production.
Assembly is a central operation to the companies that are studied and
flexibility of the whole production line is often constrained by the
flexibility in the assembly stations. All three companies strive to achieve
the same types of flexibility, volume and product-mix, to cope with
uncertainties in demand. However, the ways to achieve these types of
flexibility differ between the companies due to differences in products and
ways to produce these. For example, one company is solely carrying out manual
assembly while another uses automatic machines. Thus different kinds of
options have to be identified and different kinds of exercise patterns may be
shown. We will highlight some practical problems associated with the
different steps from identifying options, finding sources of uncertainty,
gathering appropriate data from the financial markets, etc.
Jussi Keppo, Columbia
University and
Sampsa Samila, Columbia
University
This paper aims to lay a theoretical foundation on how
individuals value durable products. In this paper we emphasize ownership and
analyze why customers want to own products and the value they give to
ownership. Specifically, we argue that the ownership of a product represents
a bundle of options. At any given point in time, the owner of a product has
the option to choose whether she wants to use the product or not. In
addition, this paper extends the current literature on product value by
taking explicitly into account several important characteristics of modern
products, namely modularity, and systemic as well as network effects. This
approach has the significant benefit that it allows consideration of
uncertainty about the future use of the product. The model shows that the
value of a product is sensitive to the changes in uncertainty, especially
when the variable costs are high compared to the utility. This uncertainty
about future utility depends on the uncertainty about future needs and wants,
about the quality of the product, and about the availability and quality of
future upgrades. However, the value is also dependent on the uncertainty
about the future variable costs.
3T. Real Options and Learning
Paul D. Childs, Steven Ott, University
of Kentucky and
Timothy J. Riddiough, Massachusetts Institute of Technology
Unlike the assets that underlie exchange-traded claims, many real assets
are infrequently traded so that asset values cannot be continuously and
precisely observed. If the exact value of the asset that underlies the
contingent claim is not known with certainty, both the valuation and any
exercise decision must be made with an imperfect estimate of real asset
value. In this setting, the claimholder has an incentive to more precisely
determine asset value by acquiring additional information about the
underlying asset value.
The framework for this paper is that the value of underlying assets for
some contingent claims is partially obscured by noise. We determine
distributional parameters for the conditional expected asset value where the
noise and the underlying asset value dynamics follow normal, lognormal and
mean-reverting processes. We also examine the effects of two types of noise:
an initial level of noise present when the underlying asset value is
originally observed or estimated, and a dynamic process that accumulates
noise after the initial observation. When a costly information acquisition
technology does not exist, the initial level of noise volatility does not
affect option value or exercise policy, while noise that accumulates always
results in lower option values.
To study information acquisition policy, we examine the case of a borrower
who holds the default (put) option inherent in a risky discount debt
contract. Costly information acquisition technology will be used to reduce
potential errors in exercise policy at the debt payoff date if the technology
is sufficiently inexpensive and the conditional expected asset value is
sufficiently close to the face value of the debt. Comparative static results
reveal that when asset volatility is greater than the accumulating noise
volatility, the optimal level of information acquisition is most sensitive to
noise volatility. Initial noise volatility has the greatest impact for
short-lived claims while accumulating noise volatility has the greatest
impact for longer-lived claims.
We provide solutions for the value of the debt given that market
participants anticipate acquisition of information. Under the more general
setting where there are multiple opportunities to gather information, it is
optimal to acquire information in smaller increments to reduce the potential
of ex-post overinvestment and underinvestment in information acquisition.
Nevertheless, the cumulative level of information acquisition is often higher
relative to the case when information can only be gathered once.
Real Options with Random Controls and the Value of Learning
Spiros Martzoukos, U. Cyprus
In this paper we propose a conceptual framework for valuation of real options
in the presence of controls with random outcome (learning) in continuous
time. The controls affect the value of the underlying asset, and are incurred
at some cost. They represent optional efforts by management to add value to
the underlying real investment over which it has monopoly power, albeit with
uncertain results. A special case of this framework captures costly learning,
like in R&D projects. A solution methodology is demonstrated and the
impact of such uncertain actions is seen to be relatively more significant in
the case of less profitable investment options.
4. Valuing Natural Resource Investments
Morten W. Lund, Natural Gas Marketing & Supply, Statoil,
Norway
The Norwegian offshore activity has evolved rapidly and Norway
is today the 7th largest oil producer in the world. As a consequence the oil
and gas industry has become a very important element of the Norwegian
economy, and the focus on improved development strategies has strengthened
over the last 10-15 years, putting emphasis on the need for so-called
flexible development strategies. A number of contributions have addressed the
subject of investment under uncertainty in the oil industry. Especially the
development of contingent claims analysis and its applications to real
investments have provided increased insight into this topic. However, most of
the published examples greatly simplify the project description, both
regarding the number of stochastic variables and the operator's decision
making freedom. Due to the simplifications it is hard to discern the benefit
of real options in a realistic oil field development project from
contributions reported in the literature.
The model described in this paper seeks to provide a more complete and
realistic description of an oil field development project by capturing the
main types of options present in the projects. Compared to related models the
presented framework represents a significantly extended approach. Both the
degree of decision making freedom and the number of stochastic variables are
increased, implying a much more computationally demanding model.
The model is a finite horizon Markov decision process and includes three
stochastic variables; the oil price, the well rate, and the reservoir volume.
These variables are of major importance to the production profile and the
cash flow of the field. The size of the model, measured by its state space,
the number of alternative decisions and the number of stages, depends on the
field development project being addressed as well as e.g., the choice of
stochastic process for the oil price. Nevertheless, reported solution times
for runs made on a PC and on a Unix machine are considered acceptable for
practical decision making situations.
To illustrate the qualities of the developed framework the model has been
implemented for a small oil field about to be developed on the Norwegian
continental shelf. Project data are, however, somewhat modified in order not
to reveal any restricted information.
The case study reveal several interesting consequences of going from a
deterministic to a stochastic evaluation of
the oil field development project, and clearly illustrate the importance of
giving due attention to flexibility in future oil field development projects.
The value of flexibility for the addressed project is substantial, both in
relative and absolute terms.
The model outlined in this paper represents a first approach to provide a
comprehensive decision support model for oil field development projects. It
should be conceived of as a prototype, and the possibilities for refinement
and expansion are thus abundant. Such improvements are not considered in this
context.
Marco Antonio G. Dias (Petrobras, Brasil) and Katia Maria C. Rocha
(IPEA, Brasil)
Petroleum firms acting in exploration and production (E&P) routinely
need to evaluate concessions and to decide
the investment timing for its project portfolio. In some countries, the
exploration concession has features such as the possibility of extension of
the exploratory period. The holder of a petroleum exploration concession has
an investment option until the expiration date fixed by the governmental
agency, and these rights can be extended by additional cost. The additional
costs can be some fee (extra-tax) to governmental agency and/or additional
exploratory/appraisal investment.
The value of these rights and the optimal investment timing thresholds are
calculated by solving a stochastic optimal control problem of an American
call option with extendible maturities, using the dynamic programming
framework. The thresholds for both the optimal extension of the option and
the immediate development investment are presented and discussed.
The uncertainty of the oil prices is modeled as a mix jump-diffusion
process. Normal information generates continuous mean-reverting process for
oil prices, whereas random abnormal information generates discrete jumps of
random size. We adapt the Merton (1976) jump-diffusion idea to the oil prices
case. Normal information means both marginal interaction between production
and demand (inventory levels is an indicator) and depletion versus new
reserves discoveries (the ratio of reserves/production is anindicator).
Abnormal information means very important news, causing in a short time
interval, a large variation (jumps) in the prices. These jumps, which has
been observed sometimes in oil prices history, are modelled with a Poisson
process.
The paper's new contributions are two: the framework of options with
extendible maturities for real assets and the utilization of a mixed
stochastic process (mean-reversion with jumps) to model the petroleum prices
which, despite of its economic logic, has not been used before in petroleum
economic literature. Comparisons are performed with the popular geometric Brownian
process for both the concession option value and the optimal investment
timing policy. The role of the convenience yield is also discussed for both
stochastic process, which has an important impact in the optimal development
investment threshold. Analysis of alternative timing policies for the
petroleum sector is presented, and also the comparative statics for the main
parameters of the model.
Optimal Exploration Investments under Price and Geological Uncertainty: A
Real Options Model
Gonzalo Cortazar, P.
Universidad Católica de Chile, Eduardo S. Schwartz, University of California
Los Angeles, and Jaime Casassus, P. Universidad Católica de Chile
We present a real options model for valuing natural resource exploration
investments when there is joint price and geological uncertainty. Price
uncertainty is modeled by a brownian motion, while geological risk is on
reserves, development investments and cost structure, with uncertainty
declining as exploration investments are undertaken. The model considers that
in case of finding an economically feasible mine, there may be a development
investment, to be followed by an extraction phase. All phases are optimized
contingent on price and geological uncertainty.
Several real options are considered in the model. The exploration
investment schedule is considered flexible and may be stopped and/or resumed
at any moment depending on cash flow expectations, which in turn depend on
current commodity price and geological expectations. The model allows for
several exploration phases, each one with its own investment schedule and
probabilities of success. In the event of an exploration success the model
considers a timing option for the development investment, and closure,
opening and abandonment options for the extraction phase.
The model has the virtue of maintaining a relatively simple structure by
collapsing price and geological uncertainty into a one-factor model for
expected value. The model was applied to value some real exploration
prospects for a major copper company. Results for a case are presented.
1. Case Studies in Real Options
Alberto Micalizzi, Bocconi
University
This case is taken from Schering Plough in the pharmaceutical sector, a
sector in which the concept of value creation is currently under significant
reconsideration. Enabling technologies and optimization of stop-go decisions
must be seen as a means through which managers can improve the R&D
efficiency. It follows that more and more attention is being devoted to:
- Increasing market
opportunities for each product
- Selecting the best
candidates to develop
- Boosting the customer
satisfaction
The fact that the fixed cost investments are focused increasingly on the
late stage of clinical trials gives more importance to the product launch as
an irreversible investment decision. In particular, the costs of the last
phase of clinical trials necessary to bring new products to market are
typically undertaken three to five years before product launch. These
resource-intensive trials can significantly impact the total value of the
project.
The firm is faced with an irreversible decision of whether or not (and
when) to invest in the third stage of clinical trials of a new anti-asthma
product, Newprox. The potential worldwide market for Newprox is approximately
one billion dollars.
The case is enriched by the fact that the firm plans the launch of a product,
Minprox, designed to treat nasal congestion caused by allergies. This product
uses the same molecule in Newprox but the market for Minprox could be
significantly small, which would result in a negative NPV.
There are two important aspects worth considering about the Minprox
project and its link to Newprox. First, the launch of Minprox would underline
the continuity of investments in company image and would represent a bridge
between the anti-allergy segment (where Schering Plough has been present for over
ten years) and the anti-asthma segment (where Schering Plough is a newcomer
). Second, Minprox represents an important source of information that allows
SP to postpone and condition the launch of Newprox. As a matter of fact, SP
bases the decision to invest in the last stage of clinical trials of Newprox
on the potential success of Minprox.
Airbus: Valuing Options in the Airline Industry
John Stonier, Airbus Industrie and Alexander Triantis, University
of Maryland
Real options may arise naturally in the course of
business, such as the option to wait to invest. Alternatively, they may be
granted or purchased in the form of contractual options. This paper presents
the authors' experience with the valuation and use of contractual aircraft
delivery options in the aircraft manufacturing and airline industries. The
relative value of a contractual option compared to an airline's natural
option to wait to invest is captured using a contingent claims model. Some of
the interesting issues that arise in the analysis include mean reversion, the
estimation of underlying present values and volatilities, and the appropriate
specification of stochastic exercise prices and queue lengths for delivery.
In addition to a standard delivery option, we also examine conversion rights
- options to take delivery when there is a choice of aircraft type (i.e.,
options on the maximum). The use of contractual delivery options raises many
interesting strategic issues from the perspective of both the aircraft
manufacturers as well as the airlines. We explore implications for risk
management, value creation through flexibility, competitive strategy, and
optimal contract design.
2A. Sharing Experiences in Oil Development/ Natural
Resources & Panel
Soussan Faiz, Texaco Inc., USA
The Real Options methodology is emerging as the state-of-the-art technique
for asset valuation among practitioners. The increased collection of
comprehensible literature on real-life applications is paving the way for
using "option-pricing" methods, customary on Wall Street, in
situations within firms of all sizes on Main Street.
Separately, the concept of "efficient frontier" is also making
headway for applications in strategic diversification. Various products
within the marketplace now enable a portfolio manager to optimize an asset
(or business) mix under different risk and return trade-offs, short- and
long-term objectives, and key operational and fiscal constraints.
There is, however, a large gap in the literature and in practice on
combining Real Options with Portfolio Optimization. Often a portfolio
contains opportunities that are dependent on common macroeconomics (e.g., oil
price), have possible other interplays, and include various embedded options
through their lifecycles. The ability to dynamically optimize asset
combinations and determine the percentage stake and staging of each
opportunity is of critical interest to senior executives.
This presentation conveys some insights from a Real Options application
and, separately, from a Portfolio Optimization case study. In the absence of
established guidelines to link the two techniques, however, it invites the
Real Options community to focus its attention on closing this important gap
and providing a seamless solution to a portfolio manager.
Peter H.L. Monkhouse, Rio Tinto plc, UK
This presentation outlines the experiences of Rio Tinto plc in applying
the real options framework to the valuation of mining properties. It sets out
our successes, the difficulties we have encountered, and one major problem
that is currently halting the quantitative implementation of the real options
approach within Rio Tinto.
As a company we have been using the real options approach for about 10
years. We began by treating the project value as the stochastic variable.
This approach has two major deficiencies. It is difficult to parameterise the
stochastic process and it is poorly suited to valuing multiple real options,
which is a characteristic of almost all mining projects.
In an effort to overcome these difficulties we moved to treating the
commodity price as the stochastic variable. This allows us to use futures
data to parameterise the stochastic process, and coupled with the concept of
switching between modes, allows the valuation of multiple American real
options. Using this approach we have constructed a lattice-type model within
Excel. The model uses a two-factor price process and can value three American
real options simultaneously. In a mining sense, the model also allows for
finite reserves, and for ore grades and production rates to vary with cumulative
production.
In using this model to value mining projects we came across a number of
disappointing and problematic issues. The first issue was that many of the
commonly quoted options often have limited value. The second issue was just
how difficult it was to reconcile our real option values with our
conventional values that use a risk-adjusted discount rate. Disconcertingly,
this analysis highlighted the possible importance of interest-rate risk.
Allowing for both stochastic interest rates and stochastic commodity prices
has received only limited attention in the literature. As such, it is
difficult to know if the usual assumption of constant interest rates is
reasonable, thereby detracting from the confidence we have in our calculated
real option values. The third issue was the treatment of corporate debt and
investor-level taxes within the real option framework. Again this is an area
that seems to have received only limited attention in the literature, but is
of importance if we are to have confidence in our calculated real option
values. Finally an issue we have identified but not addressed is the
reconciliation of real option values to quoted share prices.
In attempting to value real options, the most significant problem is
estimating a long-term forward curve. While we are endeavouring to
parameterise the two-factor price process by Kalman filtering futures price
data, we are currently unable to get the filter to converge to a stable
solution, even after scaling the variables. We believe that unless we can
estimate a stable and plausible forward curve, the real options approach will
have limited quantitative application within Rio Tinto.
2T. Competition and Strategy
Helen Weeds, Fitzwilliam CollegeUniversity of Cambridge
When an irreversible decision is taken under uncertainty, real options
theory tells us that there is an option value of delay. From the game
theoretic analysis of situations in which a small number of agents compete
for the same prize, where there is an advantage to the first-mover, it is
well-known that the fear of pre-emption causes agents to move sooner than
would otherwise be the case. Thus, it is generally presumed that in
imperfectly competitive settings the fear of pre-emption undermines option
values and the predictions of the real options approach will no longer have
any relevance.
In this paper, we present a real options model of R&D competition in
which two firms have the opportunity to invest in competing research projects
with uncertain returns. Depending on parameter values two distinct types of
equilibria may arise. One is a leader-follower equilibrium in which one firm
invests strictly earlier than its rival and option values are reduced by
competition. In the other case, however, the outcome is a symmetric
equilibrium in which both firms delay investment to a greater extent than in
the corresponding model where a single firm has the opportunity to invest, in
contrast with the expected outcome.
Sigbjorn Sodal, Agder
College, Norway
Products like pharmaceutical drugs, cars, aircraft, and computers normally
require two quite different investments: R&D and production. In one sense
or another, the first investment will represent a patent, while the second
one implies to activate the patent. Technically, the first investment can be
thought of as an option investment, and the second one as exercising the
option.
Since demand or cost variables may change, it is not always optimal to
undertake wide-scale investment in production right after the patent has been
acquired. Such changes imply, more generally, the possibility of an
increasing wedge between the net present value (in terms of sales) and the
cost of production, thereby also gains from holding on to the patent instead
of exploiting it massively right away. There will, however, also be a cost of
this kind of waiting, since the obtained revenue must be discounted more
heavily the longer production is postponed. We discuss whether the gains from
waiting exceed the costs, using a stylistic one-sector model, and deriving a
criterion for whether waiting is optimal. Typically, waiting is optimal if
demand is highly expected to increase or if it is highly volatile, if the
discount rate is small, and if the patent cost is small relative to the
production cost.
J. Tvedt (Den norske Bank,
Norway)
This paper studies the dynamics in the crude oil industry. The basic model
derives the price dynamics of an industry where fixed capital is produces by
another industry with costs of structural changes. Due to the cost of
increasing and decreasing capacity in the capital producing industry, output
prices in the industry that apply the capital will follow a long run mean-reverting
pattern. A stochastic partial differential equation for this price process is
derived.
At any point of time the equilibrium in the oil market is model as a
Stackelberg leader/follower model. The leader, OPEC, has the lowest marginal
costs whereas the marginal costs of the non-OPEC producers, the followers,
depend on future and present investments. Present investments are restricted
by the size of the investment goods producing industry. To increase or reduce
the capacity of this industry entails costs. The dynamic equilibrium in the
investment goods industry is, therefore, modelled as a stochastic optimal
control problem. Hence, uncertainty in the demand for oil, via the demand for
oil industry investments, implies hysterisis effects in the adjustment of the
capacity of the investment goods industry.
The main theoretical contribution of this paper is the effect the cartel
exercises on the investments of the non-cartel producers. In a perfectly
competitive market changes in investments are solely triggered by the current
oil price, where the oil price is determined by current demand and production
capital. However, the cartel members know that they influence the oil price
by their production strategy. Hence, when choosing the optimal production
path the cartel takes into account the effect this path may have on the
investment decision of the non-cartel producers. Hence, the cartel production
strategy is a trade-off between high oil prices and the risk of triggering
non-cartel oil production investments.
The industry level model indicates that the oil price process in the crude
oil market will be mean reverting. Hence, when valuing the option to invest
in an oil field development, the market power of the efficient OPEC oil
producers should be taken into account. Uncertainty is not only due to shifts
in demand and technological shocks, but also due to strategic changes in OPEC
production policy.
3A. Sharing Experiences in Pharmaceuticals & Panel
Discussion
3T. Real Options, the Environment and Agency Issues
Jean-Daniel M. Saphores, Université Laval and Peter Carr, NationsBanc
Montgomery Securities
This paper analyses the decision to make an investment to reduce the
emissions of a stock pollutant under environmental uncertainty, which is
introduced through stochastic variations in the stock of a pollutant. Two
types of irreversibility are present here: the first one is environmental, because
society has to live for a long time with a slowly decaying stock of
pollutant; the second is economic because the investment needed to reduce the
emissions of the pollutant are sunk. We use a continuous time formulation and
concepts from the theory of real options to formulate and solve a social
planning problem. We consider two classes of stochastic processes and obtain
analytic expressions for the option value terms and the optimal stopping
rules. Most of the previous work on this topic (e.g., Kolstad, 1996, and Ulph
and Ulph, 1997) has been done using discrete time, two-period models that
cannot fully handle the dynamics of the problem.
This paper makes two contributions. First, it shows that there is no
simple irreversibility effect for the management of a stock pollutant. The
irreversibility effect was introduced by Arrow and Fisher (1974) when they
showed that a standard (i.e. static and deterministic) cost-benefit analysis
for the development of a natural area was biased against the environment. We
show that the decision to invest to reduce the emissions of a stock pollutant
depends both on the type and level of uncertainty. Indeed, when uncertainty
is small, it may be optimal to delay or advance this decision compared to the
deterministic case, depending on the value of the parameters (mainly
pollutant decay rate, social rate of discount, and cost of reducing
emissions). This results from the tension between environmental
irreversibility (the stock of pollutant causes costly long-term social damages),
and investment irreversibility (pollution abatement investments are sunk).
Moreover, when uncertainty is large enough, we find that pollutant emissions
should be curbed immediately. This is to be contrasted with Pindyck's results
(1994): he found that an increase in the uncertainty over the valuation of
pollutant damage leads to postpone a sunk investment to reduce the emissions
of a pollutant. These results have implications for global warming.
The second contribution of this paper is a clarification of the notion of
option value in the environmental literature. Work on quasi-option value
(Arrow and Fisher, 1974) had tried to link it with the value of information,
but Hanemann (1989) showed that this interpretation does not hold when some
of the basic assumptions of the Arrow-Fisher model are relaxed. By using the
same formulation for both the static and the deterministic case, we show that
option value (in a real options framework) can be seen as the value of the
flexibility to change a decision with irreversible consequences, with or
without uncertainty.
Water Management In France:
Delegation And Irreversibility
Ephraim Clark, Middlesex
University Business
School and Gérard
Mondello, LATAPSES
The problem that we address in this paper stems from the trend to
delegation in the water management field. It refers to the municipality's
negotiating disadvantage in the face of cartelized water management firms
that makes delegation, once undertaken, virtually irreversible. We show why
the characterisitics of the delegation auction render it useless as a tool
for collective welfare maximization. We also show that the remaining tool for
achieving collective welfare maximization, i.e. the municipality's right to
revoke delegation and return to direct management, is also ineffective due to
a lack of credibility that is essentially financial in nature. Thus, if the
credibility of revocation could be restored, the municipality's bargaining
power could also be restored. Using standard methods of stochastic calculus,
we model the municipality's right of revocation as a call option held by the
municipality. We show that the key variable for the value of this option, and
thus for the municipality's position, is the exercise price, which is partly
determined by objective economic criteria and partly by legal and
institutional conventions. We show that community welfare maximisation occurs
at the point where the exercise price is determined exclusively by objective
economic criteria. Since the delegated firm as a simple agent has the right
to abrogate the contract if delegation becomes unprofitable, we then model
this right as a put option held by the firm. Its value also depends to a
large extent on the exercise price, which is partly determined by objective
economic criteria and partly by legal and institutional conventions.
Combining the exercise points of the two options enables us to determine the
price-profit interval over which delegation will be acceptable to both
parties. We conclude that the optimal interval will be the one where the
exercise prices are determined entirely by objective economic criteria.
Joril Maeland (NHH-Bergen,
Norway)
This article examines dynamic investment decisions when there is an agency
problem. A principal delegates the decision of an investment strategy of a
project to an agent. The agent has private information about the investment
cost, whereas the principal only knows the probability distribution of the
cost. The principal's problem is how to compensate the agent in order to
optimize the value of the principal's investment opportunity.
One reason for an owner of an investment possibility to delegate the
management of a project to an agent, may be that the management requires
expertise that the principal does not possess, or that is too costly for him
to obtain. In other cases it may be impossible for the principal to make the
decisions himself, but it may be possible for him to commit to a delegation
contract.
The information asymmetry creates a situation where adverse selection may
occur. The agent is compensated according to a contract. The principal
observes the outcome from the investment project, and the contracted
compensation is a function of this variable. Owing to the asymmetric
information about the investment cost, it may be optimal for the principal to
leave the agent some "information rent".
The model applies to situations where the production from the project is
sold in perfect markets, whereas there are imperfections due to the costs of
projects.
An application of the model is the case where a government owns some
natural resources. Production of natural resources involves large and
(partly) irreversible investments, and uncertainty due to future output
prices. A feature of production of natural resources is that uncertainty in
output prices usually is common knowledge, whereas investment and production
costs may be private information for those investing in and operating such
projects. To exploit the resources, the government delegates the production
of the resources to companies. The companies may have incentives to signal
higher cost than the true cost in order to obtain a larger profit within the
companies. The model presented in this paper gives the government a method of
how to find the most efficient contract between the government and the
companies to which it gives the right to invest in production of natural
resources. The contract can be in the form where the companies are paid a
compensation for the management of the resources, or it can be in the form of
a taxation system.
Keynote
Address
Dr. Stephen A. Ross, Massachusetts Institute of
Technology
Professor Ross is the Franco Modigliani Professor of Finance and Economics
at MIT and Co-chairman of Roll and Ross Asset Management Co. He previously
taught at Yale University
and at the Wharton School
of the University of Pennsylvania.
He received his Ph.D. in economics from Harvard in 1970. A Fellow of the
Econometric Society and a member of the American
Academy of Arts and Sciences, he
currently serves as an Associate Editor of several economics and finance
journals and was President of the American Finance Association.
Professor Ross is the author of more than 75 articles on a variety of
topics in economics and finance (and is the coauthor of a leading
introductory textbook in finance). He is best known for having invented the
Arbitrage Pricing Theory and the Theory of Agency, and co-developing the
risk-neutral valuation and the binomial model that enabled the practical
valuation of both financial and real options. Such models are now standards
for valuation in major securities trading firms and other corporations.
Professor Ross' other contributions to real options specifically include the
recognition that interest rate uncertainty also creates a valuable option to
wait, which he will discuss in his keynote address.
Professor Ross has been a consultant to a number of investment banks and
major corporations, has served as an expert advisor and advisor to various
government departments, and serves on the board of major international
organizations.
|
