“operating uncertainty using real options file“operating uncertainty using real options” to be...
TRANSCRIPT
“OPERATING UNCERTAINTY USING REAL OPTIONS”
to be held on 8 - 9 July 2009
on the island of San Servolo in the Bay of Venice, Italy
Anti-Pollution Protest and Investment in Pollution Abatement Facilities
Walailuck Chavanasporn, Kannika Thampanishvong
and Wen-Kai Wang
Venice Summer Institute 10th
6 - 11 July 2009
CESifo, the International Platform of the Ifo Institute of Economic Research and the Center for Economic Studies of Ludwig-Maximilians University
Anti-Pollution Protest and Investment in
Pollution Abatement Facilities
BY WALAILUCK CHAVANASPORN, KANNIKA THAMPANISHVONG�,
WEN-KAI WANG
School of Economics and Finance, University of St Andrews
Abstract
In this paper, we study the �rm�s response to the protests by resi-
dents who are adversely a¤ected by the pollution resulted from the �rm�s
production by making a decision on the optimal time to invest in the pol-
lution abatement facilities. This paper emphasises that if the �rm bases
its investment decision on the neoclassical rule of investment, i.e. the net
present value (NPV) rule, this could lead to a misleading decision because
such rule does not take into account the irreversible nature of this type of
investment, the uncertainty over future outcomes, and the possibility of
delaying an investment to acquire more information. By using the Real
Options approach to investment, this paper shows how the interaction
between uncertainty and irreversibility a¤ects the optimal timing for the
�rm to invest in the pollution abatement facilities.
JEL Classi�cation: Q53, H23, C61, G11
Key Words: Real Options Theory, pollution, protest, irreversible
investment, uncertainty, abatement
�Corresponding Author: Kannika Thampanishvong, School of Economics and Finance,University of St Andrews, 1 The Scores, Castlecli¤e, St Andrews, Fife, United KingdomKY16 9AL, E-mail: [email protected], Tel: +44 (0) 1334 462424, Fax: +44 (0) 1334462444
1
I. INTRODUCTION
This paper is devoted to consider the speci�c problem faced by many �rms,
particularly in the developing countries, whose operations generate various forms
of pollution � for examples, water, noise, soil and air � into the local ecosys-
tem. These pollutions, the by-products of the industrial production process, are
hazardous not only because the local environment is ruined but also the lives
of residents are in danger. How would the residents respond to this threat im-
posed on them by the �rms? The common modes through which the residents
express their dissatisfactions to the �rms vary but the common ones include
anti-pollution protests, roads and/or factory blockage, and backlash. For con-
venience, we collectively refer to these common modes of rebellious collective
actions as the �anti-pollution protest�.
To show how relevant is the anti-pollution protest to the operation of �rms in
practice, in what follows, we give some examples of the anti-pollution protests
that took place in the fast developing countries like China and India during
the past few years. The purpose of these illustrations is to show how the anti-
pollution protests could help ensure that the concerns of the residents in that
particular locality be addressed. The �rst illustration corresponds to an event
taking place in the northern part of India in 2006. On October 4, 2006, there
were over a thousand villagers from the village of Mehdiganj in the north Indian
state of Uttar Pradesh protested at the headquarters of the Coca-Cola Company
in Gurgaon. The aim of the protest was to raise awareness of the company on
the issue of pollution �the polluted groundwater and soil as a consequence of
the Coca-Cola�s bottling plant�s operations �and demand the immediate actions
by the company to clean up its act in India or to leave India (India Resource
Centre, 2006). The second illustration refers to the event that took place in
China in 2005. According to the report by AsiaNews (2005), 60,000 Chinese
in Huaxi Village, Zhejiang Province protested in April 2005 against high, local
levels of pollution emitted by 13 chemical plants located in that particular area.
The residents were angry with the chemical plants because these plants have
2
polluted the water and ground around the village, making agriculture in that
neighbourhood virtually impossible and forcing the farmers to seek other jobs.
These illustrations clearly give us some reassurance how relevant and impor-
tant is pollution generated from the manufacturing activities to the well-being
of the residents living near the production sites. Given that the protesters are
very serious about bringing about some changes to the actions of the �rm con-
cerned and that protests pose a serious threat to the operations of the �rm,
what would be the responses of the �rm to such situation?
The protests and other forms of rebellious collective actions by the local
residents evidently call for actions by the �rm concerned to minimise the overall
emissions of pollutions into the locality. It is interesting to draw an analogy
between the responses of �rm to the Air Quality Regulations, particularly the
1990 U.S. Clean Air Act1 or U.S. Acid Rain Programme, and the �rms�responses
to the anti-pollution protests. While there have been quite an extensive studies
�both theoretically and empirically �on the former, the formal studies on the
latter has been virtually non-existent.
Air Quality Regulations allow �rms to choose between purchasing the right
or allowances to emit speci�ed quantities of pollution, substituting inputs and
making investment in the equipment that help abate pollutions (Herbelot, 1992)2 .
Since allowances for the sulfur dioxide emissions are fully tradable, the coal-�red
electricity generators may purchase additional allowances in a given year, sell
unused allowances or bank them for future use. Alternatively, the electricity gen-
erators have an option to switch the types of fuels to low sulfur coal or natural
gas, which also help abate emissions of sulfur dioxide. Last but not least, the
U.S. power plants could retro�t their plants for �ue gas desulfurization (Halkos,
1993), a method also known as scrubbing or an installation of the scrubbers,
which are pieces of equipment which help remove sulfur dioxide from exhaust
before it is released into the air. Installation of scrubber, the last approach of
making sure that less sulfur dioxide is produced with each unit of output, clearly
requires the power plants to make an irreversible investment, at a time when
3
future prices cannot be known with certainty (Hunter and Mitchell, 1999; Ins-
ley, 2003; Löfgren et al., 2008; Tuthill, 2008). Since the real options approach
is commonly used to examine the �rm�s decision to invest in a large capital
project with uncertain bene�ts/costs, many papers which study the optimal de-
cisions of the electricity generators in the U.S. in retro�tting their plants with
the scrubbers to reduce pollution have followed this approach. This approach
has indeed been developed rapidly over the past few decades. Some examples
are provided in Dixit and Pindyck (1994), Trigeorgis (1996) and Schwartz and
Trigeorgis (2001).
While complying with the Air Quality Regulations give �rms di¤erent choices
of abating sulfur dioxide emissions, �rms do not have much alternative in re-
sponding to the anti-pollution protest by the residents. From the illustrations we
provide at the beginning of this section, it is evident that, facing with the prob-
lem of anti-pollution protest, �rms have no choice but to undertake abatements,
either by cleaning up or making investment in pollution abatement equipment
and/or facilities.
According to Dixit and Pindyck (1994), agents that base their investment
decision on the cost-bene�t analysis suggested by the neoclassical theory of
investment could lead to a misleading decision since the standard framework
ignores three characteristics that are crucial in determining the optimal decisions
of investors. First, there is uncertainty over the future costs and bene�ts of
investment. It is typically very di¢ cult for the agents to gauge/estimate the
future rewards or costs from the investment. The best the agents could possibly
do is to assess the probabilities of the alternative outcomes that could imply a
larger or a smaller pro�t (or loss). Second, there is irreversibility associated with
the agents�investment, which arises from the costs of investment. The initial
cost of investment incurred by the agents is at least partially sunk. Third, in
most cases, there is some leeway about the timing of the investment. In other
words, it could be feasible for the agents to delay their action and wait for new
information about the future.
4
These features clearly characterise the decision faced by the �rms regarding
the investment in pollution abatement facilities to control emissions of pollu-
tion into the local environment and to respond to the anti-pollution protests
by the residents. The type of pollution abating equipment/facilities we refer
to here shares lots of similarities to the scrubbers in the sense that they are
irreversible investment, are costly to install and maintain. This type of equip-
ment/facilities would help minimise the risk that the �rms will need to deal
with the anti-pollution protests in the future. The perceived bene�t of such
facilities then depend on the �rm�s expectations regarding future outputs and
pro�ts. In making a decision to invest in the abatement facilities, the �rms
need to be convinced that this option would be preferable to simply allowing
the residents to protest and obstruct their production. The irreversible nature
of abatement facilities investment and the uncertainty surrounding the outputs
and pro�ts make this problem deems suitable for an application of the real op-
tions approach. The model we consider in this paper is based on the framework
which is useful for identifying critical aspects of optimal decisions of the �rms
under uncertainty, when investment is irreversible and information is revealed
over time. We hope that this analysis would have some useful implications for
policy implementations to deal with the anti-pollution protest.
In this paper, we study how, in the midst of the protest by the residents,
irreversibility and uncertainty interact in a¤ecting the �rm�s timing of invest-
ment in the pollution abatement facility that generates less emissions into the
local environment. We introduce the analytical framework in Section 2. We
show how the �rm�s investment problem can be treated as an optimal stopping
time problem and discuss how the model could be solved. Section 3 is devoted
to present the numerical results of the model, while Section 4 concludes and
discusses some policy implications.
5
II. THE MODEL
The Environment
Consider a �rm which manufactures an output, x (t), where x (t) > 0. We
simplify our analysis by assuming that the production of x (t) requires only
physical capitals, K (t; x (t)). Suppose that the �rm�s production of output at
period t, x (t), follows the stochastic di¤erential equation given by:
dx (t) = (K (t; x (t))� �x (t)) dt+ �x (t) dW (t) , x (0) = x0; (1)
where � > 0 is the rate of depreciation, � is an uncertainty parameter, andW (t)
is a Wiener process which represents an uncertainty which could a¤ect the �rm�s
production. An example of uncertainty could be an unexpected decline in the
amount of foreign direct investment (FDI) from other countries3 , causing the
�rm to lack capital/resources to be used in the production of x (t).
We suppose that the price of the output, x (t), is denoted by pb, where
pb > 0. The relationship between K (t; x (t)) and x (t) can be described by
K (t; x (t)) = �x (t), where � < pb. We assume that, with the use of the exist-
ing facility, the production of x (t) generates some pollutions, which could take
the forms of noise pollution, water pollution, etc. as the by-products and such
pollutions are emitted into the local community. The pollutions resulted from
the �rm�s production are assumed to cause damages �both direct and indirect
� to the residents who are living in the neighbourhood of the �rm. In this
context, the direct damages correspond to damages on the residents�physical
well-being, which could include illness or discomfort, while the indirect damages
are associated with damages on the residents�personal belongings, property or
crops. As discussed in the introduction, in the midst of such situation, the res-
idents could express their dissatisfactions on the �rm�s actions by participating
in various forms of rebellious collective actions including anti-pollution protests,
roads and/or factory blockage, and backlash. By obstructing the factory and
interrupting the operation of the �rm, we assume that the anti-pollution protest
6
by residents imposes some adverse e¤ects on the production of x (t).
Formally, how do we model the impact of the anti-pollution protest by the
residents on the �rm? We capture the impact of protest on the �rm�s production,
x (t), through the state equation. In the midst of the anti-pollution protest, the
production of the �rm follows the following di¤erential equation:
dx (t) =��x (t)� �x (t)� �x2 (t)
�dt+ �x (t) dW (t) , x (0) = x0;
where the term ��x2 (t) captures the impact of protest on the production.
Suppose � = �� �. This di¤erential equation can be rewritten as:
dx (t) = (�� �x (t))x (t) dt+ �x (t) dW (t) , x (0) = x0; (2)
How could the �rm deal with the angry anti-pollution protesters? In prac-
tice, we can observe that �rms can deal with the anti-pollution protesters in a
number of ways. For instance, in the case in which the damages caused by the
by-products from the �rms are not so large and the number of residents who
are a¤ected is relatively small, �rms can choose to make lump-sum payments
to compensate these victims of their actions for the damages they receive. Yet
other �rms might choose to clean up the polluted or contaminated areas to make
the protesters stop protesting; however, this approach only helps the �rms tackle
the problem temporarily. In other words, only the pollution resulted from the
past productions is being cleaned up. The pollution that would be emitted into
the local environment as the by-products of the �rms�current and future rounds
of production would necessitate the �rms to take similar actions in the future.
Since cleaning up is usually quite costly, the �rms need to seriously take this into
account when they makes a decision on this matter. Alternatively, some �rms
can try to get around the problem in a very short-sighted way; for instance, if
contamination in the water is the source of the problem that triggers the resi-
dents in Village X to go to protest, these �rms might instead try to divert the
drainage of wastes and chemicals to other villages, such as Village Y or Village
7
Z that are close-by. By doing so, these particular �rms do not directly tackle
the problem as, in the near future, the residents in Village Y or Village Z who
get a¤ected by the �rms�actions might initiate similar kinds of anti-pollution
protests, demanding the �rms to take some serious actions to solve the problem.
With these di¤erent ways of addressing the problem in mind, in this paper,
we restrict our attention to a (sunk) investment by the �rm in the pollution
abatement facilities, which generate less emissions into the local environment
as, from our opinion, this is the way the �rm can address the problem at its
root. Suppose the sunk cost incurred by the �rm from investing in the pollution
abatement facilities is given by I, where I = I1x+ I0, I1 > 0 and I0 > 0. Since
investing in the pollution abatement facilities involve sunk costs, the �rm faces
an irreversible investment problem.
Before the �rm decides to adopt the pollution abatement facilities, the �rm
faces the following problem:
max�E
8<:�Z0
e�rtpbx (t) dt+ e�r� (Va (x (�))� I (x (�)))
9=; ; (3)
subject to equation (2), where r denotes the discount rate, � is the time that the
�rm invests in the pollution abatement facilities, I (x (�)) is the cost incurred by
the �rm from investing in pollution abatement facilities and Va (x (�)) is �rm�s
value function after it invested in the pollution abatement facilities.
What happen after the �rm makes an investment in the pollution abatement
facilities? After the �rm invests in the pollution abatement facilities, we assume
that the residents stop protesting as the level of pollution emitted by the �rm
into the local environment substantially declines to the level that is no longer
hazardous to the residents. As a consequence, protest no longer has any adverse
impact on the �rm�s production, i.e. � = 0. In this sense, an investment in the
pollution abatement facilities is bene�cial for the �rm. However, the �rm needs
to face the cost of C (t; x (t)) in maintaining the new facilities. Since it is not
rational for the �rm to bear the entire burden of an increase in costs by itself,
8
the �rm should have an incentive to pass on this additional cost to the �nal
consumers who purchase the �rm�s products by raising the retail price; thus,
the new price of the �rm�s products after the pollution abatement facilities are
put in place is given by pa, where pa > pb. After investing in the pollution
abatement facilities, the present value of the �rm�s pro�ts is given by:
E
8<:1Z�
e�rt (pax (t)� C (t; x (t))) dt
9=; ; (4)
and the �rm�s production of output, x (t), now follows the following stochastic
di¤erential equation:
dx (t) = �x (t) dt+ �x (t) dW (t) : (5)
We assume that C (t; x (t)) is linear in x (t), or formally, C (t; x (t)) = Cx (t),
where 0 < C < pa. Since equation (1) has a Markovian structure, it follows
that the optimal stopping time problem takes the form of:
� = inft�0fx (t) = x�g ; (6)
where x� is the output threshold. The problem of �nding the optimal time for
this particular �rm to invest in the pollution abatement facilities is essentially
equivalent to �nding the output threshold, x�. It is important to note that equa-
tion (2) follows a Geometric Mean Reverting (GMR) process and each moment
has been derived in Ewald and Yand (2007). The dynamic of equation (2) is
tied to the mean reversion level �� and �, a parameter which captures how fast
the economy reacts to the disturbance from �� . Equation (5), however, follows
a Geometric Brownian Motion (GBM). It is commonly known that the GBM
process is unbounded. The following are some properties of the GMR and GBM
processes. Both of these processes are non-negative; moreover, they have the
property that, once they reach 0, they remain there. Interpreting such property
in the context of our analysis would be the situation in which the �rm enters
9
into a bankruptcy.
Solving the Model
We begin by solving system (4) subject to (5), which characterises a situation
after the pollution abatement facilities have already been adopted by the �rm.
To solve the model, we apply the Hamilton-Jacobi-Bellman (HJB) equations.
The HJB equation for the system for this system (4) subject to (5) is given by:
rVa (x) = (pa � C)x+�2x2
2V 00a (x) + �xV
0a (x) ; (7)
where Va (x) is the value function for the �rm after it invested in the pollution
abatement facilities.
The solution for equation (7) takes the form of A1x2+A2x+A3. By substi-
tuting the solution into equation (6), we obtain A1 = A3 = 0 and A2 =pa�Cr�� .
It is crucial to note that our analysis is based on the condition that r > �. This
condition requires that the discount rate, r, be su¢ ciently large to guarantee
that the value function, Va (x), exists. If r � �, the value function, Va (x), does
not exist since the integration may be in�nity. The reason is that the �rm�s
instantaneous pro�t function, (pa � C)x, is increasing in x and equation (5)
is unbounded4 . Therefore, it is important that we suppose that the discount
rate is su¢ ciently large to ensure that the value function of the �rm is a �nite
number.
Next, we proceed with an analysis of the situation before the �rm invested
in the pollution abatement facilities. This corresponds to system (3) subject to
(2). The HJB equation for this system is given by:
rVb (x) = pbx+�2x2
2V 00b (x) + (�� �x)xV 0b (x) ; (8)
where Vb (x) is the �rm�s value function before it invested in the pollution abate-
ment facilities.
10
The free boundary conditions are
Vb (x�) = Va (x
�)� (I1x� + I0) ; (9)
and
V 0b (x�) = V 0a (x
�)� I1; (10)
where condition (9) and (10) are called the �value-matching� condition and
the �smooth-pasting� condition, respectively. The value-matching condition
requires that, at the time the �rm invests in the pollution abatement facilities, its
payo¤ needs to be equal to its payo¤ from investing in the pollution abatement
facilities net of the sunk cost it incurs from such investment, I (x�). The smooth-
pasting condition ensures the continuity of the slopes of Vb (x�) at x�.
Moreover, since both instantaneous pro�t functions are 0 at x = 0, we obtain
the third condition:
Vb (0) = 0. (11)
Note that, when pb = 0, according to Dixit and Pindyck (1994), the solution
takes the form of hx�H (x), where h is a constant and � is given by an appro-
priate positive number. H (x) is a solution to the Kummer�s M function. On
the contrary, when pb 6= 0, equation (8) is non-homogeneous and variation of
parameters leads us to the solution of the form: h (x)x�H (x), for some h (x).
It becomes evident that the problem becomes more complicated to solve. How-
ever, we take advantage of condition (11) and apply the shooting method. The
idea underlying this method is as follows. First, we choose x and suppose that it
is the optimal stopping time. We then apply the shooting method together with
the boundary equations (9) and (11) to solve equation (8), and test whether or
not the solution satis�es condition (10). If condition (10) is not satis�ed, we
then give a small " and move x to x+ " and repeat the procedure until we �nd
x�.This completes the description of how the problem of optimal stopping time
could be solved. In Section 3, we present the numerical results.
11
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.450
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
x
V(x)
r = 0.2 ; α = 0.25 ; δ = 0.1 ; β = 0.3 ; σ = 0.1 ; p b = 0.3 ; p a = 0.6
VbVa
Figure 1: Numerical Result: The �rm�s value functions and the threshold
III. THE NUMERICAL RESULTS
In this section, we present the numerical results of the model discussed
in the previous section. When choosing the values of parameters to be used
in our numerical simulations, we need to ensure that the following condition,
(�� �)� � �2, is satis�ed in order to ensure that the process is non-negative.
Figure 1 plots the value function, V (x), as a function of output, x, using
the following parameter values: r = 0:2, � = 0:25, � = 0:1, � = 0:3, � = 0:1,
C = 0:1, I1 = 2, I0 = 1, pb = 0:3, and pa = 0:6. The thick line in the �gure
represents the �rm�s value function after investing in the pollution abatement
facilities, Va, while the dash line represents the �rm�s value function before it
invested in the pollution abatement facilities. The result shown in the �gure
shows that it is optimal for the �rm to invest in the pollution abatement facilities
at the threshold, x� = 0:2848. If the �rm invests in these particular facilities
before x� is reached, it would not be worth it for the �rm to do so; on the
contrary, if the �rm chooses to delay its investment in these facilities further,
the �rm would forego some bene�ts.
There are several parameters in the model that could a¤ect the �rm�s value
function, V (x), and thus the threshold, x�. In what follows, we examine how Vb,
12
0 0.1 0.2 0.3 0.4 0.5 0.60
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
pa = 0.5 →
pa = 0.55 →
pa = 0.6 →
r = 0.2 ; α = 0.25 ; δ = 0.1 ; β = 0.3 ; σ = 0.1 ; p b = 0.3
x
V(x
)
VbVa
Figure 2: How the �rm�s value function and the threshold are a¤ected by changesin pa
Va and x� are a¤ected by changes in the values of the following parameters: (i)
the price of the �rm�s product after the �rm invested in the pollution abatement
facilities, pa; (ii) the impact of protest on the �rm�s production, �; (iii) the
volatility parameter, �; (iv) the sunk costs from investment in the pollution
abatement facilities, I0 and I1; and (v) the cost of maintaining the pollution
abatement facilities, C.
We begin by examining how changes in the price the �rm charges after it
invested in the pollution abatement facilities, pa, a¤ect its value function and
the threshold. Figure 2 depicts the results of our numerical simulations on V (x)
and x� for three di¤erent values of pa, namely 0.5, 0.55 and 0.6. It is clear from
the �gure that, a higher is pa, a smaller x� would be. The interpretation for
this result is that, if the �rm can increase the price of its product to pass on
the additional cost it incurred from the investment in the pollution abatement
facilities to the consumers, this would shorten the time the �rm has to wait to
have su¢ cient money to cover the large cost of installing the pollution abatement
facilities.
Next, we examine how the �rm�s value function and the threshold change
13
0 0.1 0.2 0.3 0.4 0.50
0.5
1
1.5
2
2.5
3
← β = 0.1
← β = 0.2
← β = 0.3
r = 0.2 ;α = 0.25 ;δ = 0.1 ;σ = 0.1 ; pb = 0.3 ; pa = 0.6
x
V(x
)
VbVa
Figure 3: How the �rm�s value function and the threshold are a¤ected by changesin �
when we allow for a variation in �, which captures the impact of protest on the
�rm�s production. The result of this numerical simulation is shown in Figure
3. In the �gure, we conduct the numerical simulations on V (x) and x� using
three di¤erent values of �, given by 0.1, 0.2 and 0.3. It is evident from the �gure
that, if protest imposes a larger adverse impact on the �rm�s production, the
�rm would try its best to ensure that the protest subsided as quickly as possible
by investing in the pollution abatement facilities. Therefore, a larger is �, a
smaller would be the threshold x�.
In Figure 4, we present the results from our simulations on V (x) and x� with
three di¤erent values of �, which is the volatility parameter. It is clear from the
�gure that x� becomes higher as � increases. Therefore, the �rm needs to wait
longer to invest in the pollution abatement facilities. This result is intuitive
as, the more uncertain is the future cost and bene�t from its investment in the
costly facilities, the �rm should have an incentive to delay its investment.
In what follows, we examine how the changes in the costs associated with the
pollution abatement facilities a¤ect the �rm�s decision to invest in the facilities
and its value function. Figure 5 shows how an increase in the sunk cost, I1,
14
0 0.05 0.1 0.15 0.2 0.25 0.3 0.350
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
← σ = 0.1
← σ = 0.15
← σ = 0.2
r = 0.2 ; α = 0.25 ; δ = 0.1 ; β = 0.3 ; pb = 0.3 ; pa = 0.6
x
V(x
)
VbVa
Figure 4: How the �rm�s value function and the threshold are a¤ected by changesin �
a¤ects V (x) and x�. It is clear from the �gure that a larger sunk cost parameter,
I1,would cause x� to be higher. This indicates that the �rm tends to delay its
investment in the pollution abatement facilities if the sunk cost involved in the
investment in such facilities is high.
In Figure 6, we show how a variation in the value of I0, another sunk cost
parameter a¤ects the threshold and the �rm�s value function. The �gure clearly
shows that an increase in I0 leads to a higher x�. This result is intuitive since
if the �rm realises that the sunk cost it needs to incur from investing in the
pollution abatement facilities is higher, it should have an incentive to delay its
investment even further.
Last but not least, we show in Figure 7 how an increase in the maintenance
cost, C, a¤ects V (x) and x�. From the �gure, it is quite obvious that if the
cost of maintaining the pollution abatement facilities is high, the �rms would
have an incentive to wait longer in order to ensure that it has enough money
to invest in the facilities that help abate the pollution. It is important to note
that the �rm�s value function at x�, V (x�), changes slightly compared to the
15
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.40
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
I1 = 2 →
I1 = 2.5 →
I1 = 3 →
r = 0.2 ; α = 0.25 ; δ = 0.1 ; β = 0.3 ; σ = 0.1 ; pb = 0.3
x
V(x)
VbVa
Figure 5: How the �rm�s value function and the threshold are a¤ected by changesin I1
0 0.1 0.2 0.3 0.4 0.50
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
I0 = 1 →
I0 = 1.5 →
I0 = 2 →
r = 0.2 ; α = 0.25 ; δ = 0.1 ; β = 0.3 ; σ = 0.1 ; pb = 0.3
x
V(x)
VbVa
Figure 6: How the �rm�s value function and the threshold are a¤ected by changesin I0
16
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.40
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
C = 0.1 →
C = 0.15 →
C = 0.2 →
r = 0.2 ; α = 0.25 ; δ = 0.1 ; β = 0.3 ; σ = 0.1 ; pb = 0.3
x
V(x)
VbVa
Figure 7: How the �rm�s value function and the threshold are a¤ected by changesin C
changes in the threshold, x�.
IV. CONCLUDING REMARKS
With a relatively less stringent environmental standard or legal control on
pollution in some developing countries, this gives lots of leeway for the �rms in
these countries to continue to use cheap but dirty production facilities, which
generate quite a large amount of pollution into the local environment. In re-
sponse to the damages resulted by the �rms�actions, the anti-pollution protests
have become increasingly common modes through which the residents who are
the victims of pollution express their dissatisfaction. By blocking the factories
or obstructing the workers from performing their usual tasks, the anti-pollution
protest could have a detrimental impact on the �rms�production. In this paper,
we study the optimal timing for the �rms to invest in the expensive pollution
abatement facilities that would help reassure the �rms that the threat of protest
is subsided, taking into account the irreversibility nature of such investment, the
uncertainty and the possibility of delaying the investment.
17
The results from our analysis show that a number of parameters play a
crucial role in determining the optimal timing for the �rms, particularly the sunk
costs the �rms would incur from making such investment and the maintenance
costs for the pollution abatement facilities. Our numerical results clearly show
that, in general, the high costs associated with the investment in the facilities
that help abate pollution could cause the �rms to delay their investment. This
provides a scope for policy intervention by the governments in such countries.
To encourage the �rms in their countries to take a quick step against pol-
lution that results from their actions, the governments could consider o¤ering
some �nancial supports to these �rms so that they could a¤ord such an expen-
sive investment and no longer be able to use this as an excuse for not making an
investment in the pollution abatement facilities. The �nancial supports could
take various forms, such as loans dedicated for the purchase of pollution abate-
ment facilities or subsidies which allow these �rms to buy the expensive facilities
at the lower price. Even though this could essentially imply that a large amount
of taxpayers�money need to be devoted for this purpose, from our opinion, this
is one of the most direct solutions to the problem. However, the governments
need to closely monitor the retail prices of the �rms�products after the new fa-
cilities have been installed. The governments need to make sure that the �rms
that undertake investment in the expensive pollution abatement facilities do not
take too much advantages of their consumers by charging too high retail prices.
Notes
1According to Insley (2003), Title IV of the 1990 Clean Air Act mandated a 10 million ton
(or equivalent to 50 percent) reduction from 1980 levels of acid rain precursor emissions from
electric utilities, particularly the sulfur dioxide (SO2).
2Hunter and Mitchell (1999), Insley (2003) and Löfgren et al. (2008) provide a very good
review of literature in this area.
3For illustration, let us consider the case of the car manufacturing �rms in Thailand, which
heavily rely on the FDI from the Japanese companies such as Toyota, Honda and Mitsubishi. If
the headquarter of Toyota Company has a policy to substantially reduce its direct investment
18
in Thailand because of the political instability, the impact this would cause on the production
of cars in Thailand would be quite sizable.
4The solution for equation (5):
x (t) = x0e
����2
2
�t+�W (t)
and it is well-known that the
1Z0
ef(x)dx does not exist if f (x) is positive.
References
[1] AsiaNews (2005). �Sixty thousand people protest
against pollution�, available at the following URL:
http://www.asianews.it/view4print.php?1=en&art=3036 (April 14)
[2] Dixit, A. and Pindyck, R. (1994). Investment Under Uncertainty. Princeton
University Press.
[3] Ewald, C.-O. and Yand, Z. (2007). �Geometric Mean Reversion: Formula
for The Equilibrium Density and Analytic Moment Matching.�Available on
SSRN at http://papers.ssrn.com/sol3/papers.cfm?abstract_id=999561.
[4] Halkos, G.E. (1993). �Sulphur abatement policy�, Energy Policy, pp.1035-
1043.
[5] Herbelot, O. (1992). Option valuation of �exible investments: the case of
environmental investments in the electric power industry. Ph.D. Disserta-
tion, Massachusetts Institute of Technology.
[6] Hunter, G.W. and Mitchell, G.T. (1999). �Options For Polluting Firms:
Bankable Permits or Abatement�, University of California, Santa Bar-
bara Working Paper WP23-98R. Available at the following URL:
http://www.econ.ucsb.edu/papers/wp23-98R.pdf.
[7] India Resource Centre (2006). �Major Protest Against Coca-
Cola in India�. Press Release available at the following URL:
http://www.scoop.co.nz/stories/WO0610/S00091.htm
19
[8] Insley, M.C. (2003). �On the Option to Invest in Pollution Control under
a Regime of Tradable Emissions Allowances�, The Canadian Journal of
Economics, vol.36(4), pp.860-883.
[9] Löfgren, Å, Millock, K. and Nauges, C. (2008). �The e¤ect of uncertainty
on pollution abatement investments: Measuring hurdle rates for Swedish
industry�, Resource and Energy Economics, vol.30, pp.475-491.
[10] Schwartz, E. and Trigeorgis, L. (2001). Real Options and Investment un-
der Uncertainty, Classical Readings and Recent Contributions, Cambridge,
MA: MIT Press.
[11] Trigeorgis, L. (1996). Real Options, Managerial Flexibility and Strategy in
Resource Allocations, Cambridge, MA: MIT Press.
[12] Tuthill, L. (2008). �Investment in Electricity Generation Under Emissions
Price Uncertainty: The Plant-Type Decision�, Oxford Institute for Energy
Studies EV 39.
20