Analysis

The incentives of private companies to invest in protected area

certicates: How coalitions can improve ecosystem sustainability
Nathalie Meiner
Institute for Environmental Economics and World Trade, Leibniz University of Hannover, Knigsworther Platz 1, D-30167 Hannover, Germany
a b s t r a c t a r t i c l e i n f o
Article history:
Received 23 December 2012
Received in revised form 10 July 2013
Accepted 23 August 2013
Available online 23 September 2013
Keywords:
Ecosystem sustainability
Non-cooperative games
Ecological footprint
Protected areas
Certication
Since the early 80's, the global demand on nature has exceeded the earth's capacity. To reduce the overuse of the
very resources on which human life depends, protected areas have been developed worldwide. Typically, nation-
al states, NGOs and charities have funded protected areas, with limited investment fromprivate companies. This
paper analyzes one option to increase private investment: an international market for protected area certicates.
Following a costbenet analysis, a three-stage coalition game is developed. The corporate dependency on eco-
systems is modeledthrough the ecological footprint. By implementing instruments suchas side payments, mem-
bership restriction and non-compliance penalties, the model shows that corporate environmental agreements
reduce the individual cost of ecological protection and enhance social welfare. The ndings are supported by a
sensitivity analysis conducted for the German tourism sector in Zanzibar.
2013 Elsevier B.V. All rights reserved.
1. Introduction
The development of a representative network of protected areas is
nowadays a well-established instrument for the conservation of ecosys-
tems that provide a multitude of provisioning, regulating, cultural and
supporting services (Duraiappah & Naeem, 2005). However, there is
limited involvement of private companies in preserving areas with
global importance (Emerton et al., 2006). To enhance private funding
for protected areas, international certication markets are developed.
Still in their infancy, these markets are being led through initiatives
suchas REDD+ andGDI (Carius, 2010).
1
There are two drivers of corpo-
rate environmental responsibility that are emphasized in literature: the
image effect of environmental commitment; and the mitigation of cor-
porate ecological risks (e.g. natural disasters, depletion of resources)
(CDP, 2012a,b; Dummett, 2006; Koellner et al., 2010). Regarding the
second driver, it should be considered that many ecosystem services
are public goods characterized by non-rivalry and non-excludability
(Pascual & Muradian, 2010). This leads to the problem that companies
that invest in the preservation of ecosystems not only have to share
the resulting benets (positive externalities) but also suffer from
the exhaustion of ecological resources caused by other market par-
ticipants (negative externalities) (Tietenberg & Lewis, 2012).
The application of game theory shows that externalities are crucial
for determining the behavior of strategic players as they constitute the
basis for the free rider problem; a situation in which the individual
self-interest of players leads to a social outcome that is not Pareto opti-
mal (Barrett, 2007). Environmental agreements (EAs) aim to overcome
such social dilemmas and are typically described by cooperative ap-
proaches that are based on the concept of the core (Chander &
Tulkens, 1997) or by non-cooperative approaches that follow internal
and external stability conditions (Barrett, 1994; Carraro & Siniscalco,
1993). The objective of cooperative game theory is to distribute coali-
tion payoffs in a manner that enables forming the socially optimal
grand coalition. The core denes the set of payoff vectors that cannot
be improved by any subgroup of players (Esteban & Dinar, 2013). In
contrast, non-cooperative games develop individual payoff functions
for each player under a given transfer scheme to predict sustainable co-
alition structures. A detailed overview of the different methods is given
in Chander and Tulkens (2008) and Finus (2003).
So far, scientic EAstudies that applied coalitiongame approaches ei-
ther focused on the design of international EAs between countries that
face global public good allocation problems (Barrett & Stavins, 2003;
Finus et al., 2009) or on the cooperation between locally affected agents
(e.g. companies, residents, neighboring countries) that sustainably man-
age common-pool resources (Abbink et al., 2005; Ambec &Ehlers, 2008).
Inorder to extend the applicability of coalitiongames and analyze the in-
centives of private companies to invest in the sustainability of ecosys-
tems, this article introduces a novel non-cooperative coalition game
model. Basically, we assume that an international market for so called
Ecological Economics 95 (2013) 148158
Tel.: +49 511 7624186; fax: +49 511 7622667.
E-mail address: meissner@iuw.uni-hannover.de.
1
The Reducing Emissions from Deforestation and Forest Degradation (REDD) scheme
was adopted in the Bali Action Plan of the United Nations Framework Convention on
Climate Change during the 13th session of its Conference of the Parties COP-13 held in
Bali in 2007 with the aimto create incentives for emission reductions in developing coun-
tries. The COP-16 agreement on REDD+, which was established in Cancn in 2010, addi-
tionally considers the role of conservation measures, sustainable forest management and
enhancement of forest carbon stocks. The Green Development Initiative (GDI) responds
to the Convention on Biological Diversity by establishing an international standard and
certication system that supports land management plans. GDI started with its two-year
pilot phase in 2012.
0921-8009/$ see front matter 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.ecolecon.2013.08.015
Contents lists available at ScienceDirect
Ecological Economics
j our nal homepage: www. el sevi er . com/ l ocat e/ ecol econ
protectedarea certicates (PACs) exists. The landbasedcerticates cover
a dened geographical area and certify that the ecological values of this
area are maintainedinaccordance withspecic standards. By purchasing
PACs, companies not only have the possibility to provide nancial sup-
port for the conservation of nature but also to offset their impact on eco-
systems and label their products accordingly. We further suppose that
companies canchoose inwhichprojects they invest, andthus whicheco-
systembenets are generated. The certicates do not grant any property
rights so that companies do not obtain an improved access to preserved
resources. In fact, the protected areas are either managed by government
agencies, environmental charities or the local community (Dudley,
2008). Given this scenario, a coalition is dened as a corporate EA that
enables signatories to collectively cope with and manage ecological chal-
lenges through a mutual obligation to buy PACs. Corporate EAs deter-
mine the number and type (e.g. origin, provided ecosystem services) of
PACs that signatories are required to buy in a certain period, and can
be part of a public or company driven environmental initiative. To appeal
to both large companies as well as small and medium sized enterprises
(SMEs), a transfer scheme is installed.
The model comprises two differences to previous EA models. First,
non-cooperative coalition formation models are typically described as
two-stage games in which players rst decide on their participation,
and second on their economic strategy (complete plan of action) and
side payment schemes (Finus et al., 2009; Pintassilgo & Lindroos,
2008). In contrast, the presented model is based on a three-stage
game approach adding another decision level at the beginning of the co-
alition formation process. The extension of the game is due to the image
effect of environmental commitment. Depending on the behavior of
their strategic opponents, companies might even be interested in buy-
ing PACs without the existence of a corporate EA. The image effect is
also the reason for the second difference in model design. Here, price
premiums reduce free rider incentives of private companies. If the
extra revenue is sufciently high, no free riding incentives might exist
at all.
The objective of the study is to evaluate the feasibility of an interna-
tional market for PACs. Chapter 2 starts with a costbenet analysis in
order to describe corporate incentives to invest in the sustainability of
ecosystems. In chapter 3, the design of the three-stage coalition game
model is specied. PAC threshold price developments are displayed
for individual as well as for common behavior of market participants
and a regulatory framework is suggested to create coalition stability.
After the theoretical construction, chapter 4 illustrates the derived
propositions in a numerical example applied for the German tourism
sector in Zanzibar by using a sensitivity analysis approach. Finally, chap-
ter 5 draws a critical conclusionandpresents anoutlook for future work.
2. Incentives of Private Companies to Invest in PACs
The process of modeling the PAC investment decision of private
companies starts with a costbenet analysis. As already mentioned,
there are two main drivers for voluntarily protecting ecosystems. First,
the image effect of environmental commitment. Through the certica-
tion of environmental performance, companies have the opportunity
to communicate their ecological objectives and increase their reputa-
tion in society (McWilliams & Siegel, 2011). If nal products are labeled
with ecological achievements (Grote et al., 2007; Rotherham, 2004), the
image effect might even allow companies to differentiate their product
portfolio, strengthen the brand and in the end realize a price premium
(Drr, 2008; Juutinen et al., 2011; Porter & Kramer, 2006). The second
driver is the protection of ecosystems to mitigate ecological risks that
have a direct inuence on the business of a company and its supply-
chain (Trucost, 2013). The higher the risks, the more a company stands
to gain from the long-term conservation of nature (CDP, 2012a,b). The
model uses the ecological footprint concept to consider a company's ex-
posure to ecological risks. The corporate ecological footprint describes
the company's dependency on ecosystems by measuring the amount
of biologically productive land and water area a company requires to
sustain its business and absorb its created waste (Wiedmann et al.,
2006). It is naturally understood that the overall capacity of ecosystems
in terms of resource production and waste absorption is limited
(Rockstrm et al., 2009). The ability of ecosystems to regenerate with-
out human intervention is expressed by the earth's biological capacity.
Both the ecological footprint and the biological capacity are measured
in global hectares (gha); a hectare with world average productivity. To
control for variation in geographic conditions and specic land area
types, local productivity is adjusted via yield factors (ratio of national
to world average yield per hectare) and equivalence factors (conver-
sion of relative productivity of land area types into world average bi-
ologically productive area) (Ewing et al., 2010; Wackernagel, 1994;
Wackernagel & Rees, 1996). Looking at the expected cost of acquiring
PACs, one has to distinguish between direct and indirect components.
Direct costs containthe expenditures for the actual establishment, man-
agement and certication of protected areas, e.g. labor, administration
and monitoring costs, as well as the costs that are connected with the
development of the initial standard and certication scheme. In con-
trast, indirect costs consider the externalities that stem from the provi-
sion of a public good(Grote, 2009; Karousakis &Brooke, 2010). Merging
benets and costs, the following payoff function can be set up.
u
ij

ij;0
p

i
; x

j
_ _
p
Z
Z
ij
p
P
i
Y
e
ij
p
Z

i

ij
b Z 1
Let C be the set of all green companies that buy PACs, F the set of all
free riders and S
i
the set of all companies in sector i. In the business as
usual scenario, company j S
i
gains a net prot
ij,0
by choosing the op-
timal factor input x
j

= (x
j1
, . . . , x
jm
) at the market equilibrium price p

i
with mN

. A part of the price p

Z
for every purchased certicate Z
ij
is
needed to assure the accuracy of the PAC label and develop an appropri-
ate standard and certication scheme. The remainder depends on the
location of the protected area and species the expenses required for
the preservation of one global hectare of land. A contributing company
j C that decides to invest in PACs can forward a part of the occurring
certication costs to its customers whose share
i
is willing to pay a
price premium p
i
P
for PAC labeled products. Assuming that a company
acquiring PACs wants to satisfy the whole demand of its environmental-
ly conscious customers, not less thanY
ij
e
=
i
f
i
(x
j

) certicates have tobe

purchased with f
i
(x
j
) accounting for the production function that is in-
creasing and concave in the number of factor inputs. The calculation of
PACs that are used for product labeling is based on two assumptions.
First, transaction costs for supplier changes are deemed to be sufcient-
ly high so that eco-friendly customers fromcompetitors cannot be lured
away. Second, companies that decide to offer ecologically labeled prod-
ucts Y
e
at a price p

+ p
P
continue to put normal products Y
o
at a price
p

up for sale. Thus, PAC investors not only serve their environmental
conscious customers
i
but also free riding customers (1
i
) that refuse
to pay a higher price. An upper limit for PAC investment does not exist.
Hence, the quantity of certicates purchasedby a green company can be
extended by the amount Z
ij
+
with Z
ij
= Y
ij
e
+ Z
ij
+
. A free riding company
j F neither suffers fromcertication costs nor does it have the possibil-
ity to get a price premium for its products (Z
ij
= Y
ij
e
= 0).
The benets of ecological risk mitigation derived from the develop-
ment of protected areas are described by the function b(Z) that is in-
creasing and concave in the number of certicates purchased by all
market participants. To get a monetary value, the benet function is val-
ued at the average price p
Z
for one global hectare of protected land. Fur-
thermore, the benet function is weighed by the corporate dependency
on ecosystems to account for the prot that can be directly assigned to a
company and for the indirect costs of providing a public good. A
company's dependency on ecosystems is expressed by the corporate
share of the total ecological footprint EF that results from multiplying
sector i's dependency
i
= ef
i
/EF by the market share of the company
149 N. Meiner / Ecological Economics 95 (2013) 148158

ij
= f
i
(x
j

)/Y
i
. Both a contributing and a free riding company benet
from the overall mitigation of ecological risks through PAC investment.
3. The Game Model
The model in this paper pursues the approach of non-cooperative
coalition theory (Carraro & Siniscalco, 1993; Finus et al., 2009;
Pintassilgo & Lindroos, 2008) and consists of a three-stage game with
one sector and n companies as strategic players that are exposed to sim-
ilar kind of ecological risks. During all stages companies opt simulta-
neously for their strategic action. Rationality and common knowledge
of rationality are assumed. The subgame perfect Nash equilibria are de-
termined via backward induction excluding non-credible solutions
(Osborne and Rubinstein, 1994). They describe all feasible quantities
of PACs demanded and the corresponding welfare for society. In stage
1 of the game, all companies individually decide if they invest in PACs
and contribute to the environment (C) or act as free riders (F). For sim-
plicity, we assume that the price for PACs is constant. Buying certicates
is a strategy with strict dominance for player j if
u
j
C; s
j
_ _
Nu
j
F; s
j
_ _
: 2
In other words, player j is strictly better-off contributing to the
environment if the expected payoff of PAC investment is higher than
the one in the free riding scenario for all opponent strategy proles
s
j
= (s
j1
, . . . , s
j(n 1)
) with s
j(.)
{C,F}.
Instage 2 of the game, contributors as well as free riders have the op-
portunity to form a coalition that requires its members to buy a certain
amount of PACs. The development of such a corporate partnership is as-
sociated with transaction costs that have to be borne by the cooperating
parties (Jobin, 2008). In order to convince former free riders to sign a
corporate EA, the regulatory framework of the model enables side
payments (Barrett, 2003). As a matter of fact, only those coalitions are
formed that fulll the conditions of group and individual rationality
(Ambec & Ehlers, 2008; Finus & Rundshagen, 1998).

hG
u
h
PC;

s
h
_ _

hG
u
h
IM;

s
h
_ _
Group rationality 3
u
h
PC;

s
h
_ _
u
h
IM;

s
h
_ _
hG Individual rationality 4
Let G be the set of all EA signatories. The condition of group rationality
requires that the payoff of the partial coalition (PC) exceeds or at
least equals the sum of those payoffs that result from individual
prot maximization (IM). However, a coalition is not formed until
its outcome is economically rational for each coalition member.
This means joining the agreement must not lead to a deterioration of
prots for any signatory h Ggiven the strategy prole of its opponents

s
h
s
h1
; ::: ; s
h n1
_ _
with s
h :
PC; IM f g. Non-signatories are
expected to act as singletons that adopt non-cooperative behavior
(Finus, 2001). In case all companies decide to buy PACs, a grand coali-
tion is realized.
The game concludes with stage 3 in which the long-term success
of the coalition is evaluated. The existence of a coalition equilibrium
depends on the internal and external stability of the corporate EA
(Pintassilgo et al., 2010; Weikard et al., 2006).
u
h
Y;

s
h
_ _
u
h
N;

s
h
_ _
hG Internal stability 5
u
k
IM;

s
k
_ _
u
k
PC;

s
k
_ _
kG External stability 6
Internal stability applies when all coalition members h G keep to the
declared regulations (Y) instead of behaving non-compliant, and thus
leaving the coalition (N) with

s
h
s
h1
; ::: ; s
h g1
_ _
, s
h :
Y f g
andgN

representing the total number of EAsignatories. Concurrently,

the condition of external stability ensures that none of the singletons
k G have an incentive to join the coalition for all

s
k

s
k1
; ::: ; s
k n1
_ _
withs
k :
PC; IM f g. In order to evaluate the inu-
ence of several instruments on the stability of the corporate EA, the
model allows for non-compliance penalties as well as membership
restriction policies (Finus et al., 2009; Wood, 2010).
3.1. Voluntary Investment in PACs
Customers spend the share of their income M for the consump-
tion of good Y. The demand is described by the CobbDouglas func-
tion p =M/Y Y [Y
P
;Y
S
]. The demand function can be restricted
to the interval from the demanded quantity at the prohibitive price
Y
P
to the saturation point Y
S
as any welfare changes are captured in
the analysis. Before the actual game starts, companies maximize
their prot

j;0
p f x
j
_ _
K
f
K
v
x
j
_ _
7
by choosing the optimal factor input x
j

at a given market price p*. As

market entry and exit are not restricted, the equilibrium prot
amounts to zero. Combining the zero-prot and rst order condition
(FOC)
p

f x

j
_ _
K
f
K
v
x

j
_ _
0; 8
p

f x

j
_ _
x
j; :

K
v
x

j
_ _
x
j; :
; 9
the market price p

and the optimal quantity Y

= f(x
j

) can be calcu-
lated. Assuming constant economies of scale with variable costs
K
v
= c

f(x
j
) and xed costs K
f
= 0, it follows that p

= c and
Y

M
c
: 10
The social welfare is dened by the sum of consumer and producer
surplus
W
0

_
Y

Y
P
M
Y
c
_ _
dYp
Z
e EFBC 11
with e(.) accounting for the risk resulting from the environmental bal-
ance between ecological footprint EF and biological capacity BC. Due
to the dependency on ecosystems, companies have to bear the propor-
tion of the total environmental burden. The consumers are affected by
the remaining part (1). If the social welfare W
0
is maximized with re-
spect to Y, the FOC results in
Y
W

M
c p
Z

e
Y

: 12
The market outcome only describes a social optimum for
e
Y
0. But as
production processes of most commodities rely on the use of natural
capital (Trucost, 2013), this will only be the case if companies and cus-
tomers start to consider the negative ecological impacts of their produc-
tion and consumption decisions.
3.1.1. Threshold Price for PAC Investment
In stage 1 of the game, the companies have to decide if they want to
invest in PACs. The threshold price p
Z
j
for buying certicates can be cal-
culated according to the FOC of equation (1). As companies only get a
price premium for those certicates they use for product labeling,
cases have to be split.
150 N. Meiner / Ecological Economics 95 (2013) 148158
i)
p
Z
j
p
P
p
Z

j

b
Z
j
for Z
j
Y
e
j
13
ii)
p
Z
j
p
Z

j

b
Z
j
for Z
j
NY
e
j
14
Companies cannot restrict the supply of eco-friendly products to an
amount smaller Y
j
e
. Hence, equation (13) values the average benet
per certicate.
Proposition 1. Given
b
Z j
N0,

2
b
Z j
2
b 0 and

2
b
Z j Zv
b 0 with v j, investing in
PACs is a strategy with strict dominance for company j when the market
price p
Z
per certicate is smaller than the threshold price p
Z
j
that
increases with the willingness to pay of customers:
p
Z
j
p
P
N0,
increases with the ecological footprint of the sector:
p
Z
j
ef
N0,
increases with the market share:
p
Z
j
j
N0,
decreases with the total ecological footprint:
p
Z
j
EF
b 0,
decreases with the aggregated quantity of PACs:
p
Z
j
Z j
b 0;
p
Z
j
Zv
b 0.
Corollary. The more PAC investments are made, the more attractive it is to
become a free rider.
Corollary. If p
Z
b p
P
, all companies benet from investing in PACs.
3.1.2. Welfare Impacts
The quantity of PACs demanded in the rst stage of the game is
dened by
Z
1

1
Y
e
Z

Y
e
1
Z

with
1

jC

j
: 15
The consequent change of the producer surplus PS
01
is composed of
three parts. First, the cost increase for green companies. Second, the re-
ceipt of a price premium for labeled products. And third, the mitigation
of ecological risks for all companies.
PS
01

_
Y
e
1
Y
P
p
Z
p
P
_ _
dZ
_
Z
1
Y
e
1
p
Z
dZ p
Z
b Z
1
16
Likewise, three effects have to be considered with regard to the
alteration of the consumer surplus CS
01
. To begin with, we have
to distinguish between two customer groups or rather demand
functions
p
e

e
M
e
Y
Y Y
P
; Y
e
1
_ _
; 17
p
o

M
o
Y
Y Y
P
; Y
o
1
_ _
; 18
with M = M
e
+ M
o
and Y

= Y
1
e
+ Y
1
o
. While demand function (17)
applies for those customers that pay a price premium for PAC labeled
products Y
e
, demand function (18) characterizes those customers that
buy the conventional product Y
o
because they are either not willing to
spend a higher amount of their income for eco-friendly products or
are bound to suppliers that do not offer the labeled alternative. Fig. 1
shows that clustering the customers into two groups positively impacts
the consumer surplus (1-2-3-4) with
e
N . However, buyers of certi-
ed products are not only willing to pay more but also they have to
(4-5-6-7). The second effect counteracts the rst one by negatively
impacting the consumer surplus. Third, the welfare of consumers
increases with the provision of a public good. Altogether, the change
of the consumer surplus amounts to
CS
01

_
Y
e
1
Y
P

e
M
e
Y
p

p
P
_ _
dY
_
Y
e
1
Y
P
M
e
Y
p

_ _
dY
p
Z
1 b Z
1
:
19
If the aggregated impact is negative, customers will refuse to pur-
chase certied products. Thus, a willingness to pay that is bigger than
the actual price premium p
P
that has to be paid on the market does
not constitute a sufcient condition for the demand of PAC labeled
products. It is supposed that companies are aware of this effect, and
consequently choose a price premium so that CS
01
0. For math-
ematical convenience, the price premium is dened in a way that
leads to an equalization of the indicated areas in Fig. 1 so that the
net consumer surplus matches the welfare gain derived from envi-
ronmental protection (p
P
= (
e
-) M
e
/Y). The welfare of the
whole society then changes to
W
1

_
Y
o
1
Y
P
M
o
Y
c
_ _
dY
_
Y
e
1
Y
P

e
M
e
Y
cp
Z
_ _
dZ

_
Z
1
Y
e
1
p
Z
dZp
Z
e EFBC b Z
1
:
20
As long as only those companies invest in PACs that fulll the threshold
price conditions (13) and (14), it applies that (for mathematical proof
see Appendix A)
W
01
W
1
W
0
0 Z
1
; 21
W
1
W
0

Z
N0 Z
1
: 22
Proposition 2. The socially optimal number of certicates is bigger than
the number of PACs purchased (Z

N Z
1
), as private companies do not
make allowances for the ecological benets provided for their competitors
and customers.
3.2. Coalition Formation Through Corporate EAs
In stage 2, bothcontributing companies and free riders have the pos-
sibility to form a coalition in which all members h agree to purchase at
least Y
h
e
certicates. Developing the partnership causes transaction
costs T
G
. Additional PAC investments that were realized in stage 1
M
M
M
M
Fig. 1. Demand functions for conventional and PAC labeled products.
151 N. Meiner / Ecological Economics 95 (2013) 148158
are assumed to remain constant at Z
h
+
. The common payoff function
equals
u
G

hG

h;0
p
Z
p
P
_ _
Y
e
h
p
Z
Z

h
_ _
p
Z

G
b Z
2
T
G
with Z
2

v CG f g
Y
e
v
Z

v
_ _
Y
e
2
Z

hG

h

2
:
23
According to the FOC, the maximum coalition PAC price can be deter-
mined by
p
Z;G
p
P
p
Z

G

b
Z
2
: 24
The development of PAC threshold prices for individual andcommon
behavior is shown in Fig. 2. As
G

h
, it follows that p
Z;G
p
Z
h
Z.
Companies no longer just value the effect their investments have on
the mitigation of ecological risks for their own business but also for
the business of their coalition partners. EA signatories are willing to in-
vest more because they know that their strategic partners are paying
for a greener environment as well. Positive externalities are reduced
andthe threshold price curve moves up fromp
Z
h
top
Z;G
(1 2). Sudden-
ly, more companies can afford to supply certied products. This results
in an overall increase of PACs and a corresponding drop of the average
benet per certicate. Thus, the common threshold price decreases
(2 3).
Proposition 3. The threshold price for buying PACs increases with the
number of coalition members until the maximum volume of certicates
Z
max
G
is reached.
3.2.1. Group Rationality
Building a coalition is only economically rational if condition (3) ap-
plies. Hence, the average prot of additionally purchased certicates
must at least equal the average cost.
p
P
p
Z

G

b
12
Y
e
p
Z

T
G
Y
e
25
The maximumtransaction costs, a coalition is willing to invest in the de-
velopment of a corporate EA, are dened by the aggregated payoff in-
crease of cooperating companies.
T
G
max
p
Z
p
P
_ _
Y
e
p
Z

G
b
12
26
Proposition 4. Building a coalition is only economically rational if
the transaction costs are smaller than the threshold T
max
G
. If building
a partnership costs more, the condition of group rationality is no longer
met.
3.2.2. Welfare Impacts
As a result of the coalition formation process, the number of PACs
demanded increases by the amount Y
e
= Y
2
e
Y
1
e
. The changes of
the producer and consumer surplus are dened by
PS
12

_
Y
e
2
Y
e
1
p
Z
p
P
_ _
dZ p
Z
b
12
Y
e
_ _
T
G
; 27
CS
12
p
Z
1 b
12
Y
e
_ _
: 28
Thus, the social welfare can be determined by
W
2

_
Y
o
2
Y
P
M
o
Y
c
_ _
dY
_
Y
e
2
Y
P

e
M
e
Y
cp
Z
_ _
dZ

_
Z
2
Y
e
2
p
Z
dZp
Z
e EFBC b Z
2
T
G
:
29
Comparing the welfare to the one realized without the existence of a
corporate EA, it applies that (for mathematical proof see Appendix A)
W
12
W
2
W
1
0 Z
2
; 30
W
2
W
1

Z
N0 Z
2
: 31
Proposition5. If p
Z
p
Z;G
, a corporate EAleads to a Pareto superior Nash
equilibrium.
3.2.3. Individual Rationality and Side Payments
As shown in equations (13) and (14), the threshold price for buying
PACs increases with the ecological footprint, and thus the market share
of a company. A common tool to establish individual rationality and
convince SMEs to take responsibility for the environment is the imple-
mentation of a transfer scheme (Finus et al., 2009; Munro, 2008). In
the model, side payments are enabled: large companies that already
invested in PACs in stage 1 of the game pay a subsidy price p
C,S
for
every certicate; and SMEs get a PAC price reduction of p
F,S
.
u
h

h;0
p
Z
p
P
p
C;S
_ _
Y
e
h
p
Z
p
C;S
_ _
Z

h
p
Z

h
b Z
2
32
u
k

k;0
p
Z
p
P
p
F;S
_ _
Y
e
k
p
Z

k
b Z
2
33
The subsidy has to cover both the PAC price reduction for all free riders
as well as the transaction costs of developing a corporate EA.
p
C;S
Z
1
p
F;S
Z
2
Z
1
T
G
34
As long as the supplemental costs are smaller than the resulting
benet increase from additional ecological protection b
FG
, paying
a subsidy for every certicate is economically rational for big compa-
nies. On the contrary, it is economically rational for SMEs to sign a
common agreement if the combination of subsidies and improved
ecosystem conservation outperforms their former net loss of PAC
Z
P
p
Z
p
Z
h
p
G Z,
p
Z
p
h,max
Z
G
max
Z
1
2
3
Fig. 2. PAC threshold price development.
152 N. Meiner / Ecological Economics 95 (2013) 148158
investment. As a consequence, the subsidy price for each certicate
has to fulll the following conditions
p
C;S
p
Z

h

b
FG
Y
e
h
Z

h
; 35
p
F;S
p
Z
p
P
p
Z

k

b
FG
Y
e
k
: 36
Proposition6. As long as Z Z
max
G
, companies that are willing to invest in
PACs have an incentive to subsidize free riders so that the latter are willing
to buy PACs as well. The success of a subsidy scheme in terms of signing a
corporate EA
increases with the ecological footprint of the sector,
increases with the willingness to pay of customers,
decreases with the total amount of PACs demanded,
decreases with the transaction costs.
Aredistribution of prots enables an economically rational society to
realize a Pareto superior Nash equilibrium. However, not all agents be-
have economically rational when it comes to fairness issues (Camerer,
2003). For instance, the willingness to subsidize competitors is likely
to decrease if corporate ecological footprints, and thus the benets
from environmental protection are assimilating.
Proposition 7. The more similar the ecological footprints of companies
are, the more likely a subsidy scheme will fail as fairness concerns start to
outweigh rational behavior.
3.3. The Long-term Success of Corporate EAs
In stage 3, signatories can either act in compliance with the corpo-
rate EA or leave the coalition. When companies stop investing in PACs,
they record a payoff change of
u
h;GF
p
Z
p
P
p
C;S
_ _
Y
e
h
p
Z
p
C;S
_ _
Z

h
p
Z

h
b
C
GF
; 37
u
k;GF
p
Z
p
P
p
F;S
_ _
Y
e
k
p
Z

k
b
F
GF
: 38
Obviously, companies have no incentive to leave the coalition if
u
h;GF
0 p
C;S
p
Z
p
P

Y
e
h
Y
e
h
Z

h
p
Z

h

b
C
GF
Y
e
h
Z

h
;
39
u
k;GF
0 p
F;S
p
Z
p
P
p
Z

k

b
F
GF
Y
e
k
: 40
The incentives of large enterprises to leave the coalition are elim-
inated if the commitment of SMEs is not able to compensate for the
large enterprises' environmental investment. Only if the subsidy
price exceeds the threshold in equation (39), the company will
stop buying PACs. To form a coalition with internal stability, it is rec-
ommended to install a monetary penalty P
C
= u
h,GF
that bridges
the gap between the payoff scenarios. Given that company k is the
only free rider that joined the coalition, equation (40) equals condi-
tion (36) which must be met in order to sign a corporate EA. Thus,
company k has no incentive to stop buying PACs. If two or more
free riders joined the coalition, the average environmental loss per
PAC resulting from non-compliance is smaller than the initial gain
of buying certicates. As a consequence, the required subsidy price
increases. In a similar manner to initially green companies, punish-
ments P
F
= u
k,GF
are necessary to avoid cheating when the subsi-
dy price does not meet the threshold in equation (40).
Proposition 8. The more similar the ecological footprints of former inves-
tors and free riders are, the bigger the incentives for initially green compa-
nies to leave the coalition.
Proposition 9. The more free riders join the coalition, the more attractive
it is for each signatory to stop buying PACs.
Due to substantial monitoring efforts, the installation of non-
compliance penalties might not always be affordable. In this case, exclu-
sive membership restriction policies can be used as an instrument to
strengthen the average environmental benet per PAC and to prevent
the coalition from falling apart (Finus et al., 2009).
Proposition 10. If monetary penalties cannot be installed, exclusive
membership enables the formation of self-sustaining agreements.
4. Numerical Analysis Using the Example of German Tourism
in Zanzibar
A numerical example is used to illustrate the propositions of the
game model and carry out a sensitivity analysis on PAC threshold prices.
The set-up of the analysis is as follows. In stage 1 of the game, German
tourist operators that sell journeys to Zanzibar individually decide on
PAC investment. In stage 2, they have the opportunity to sign an EA
and form a coalition that jointly protects ecosystems in Zanzibar. In
the last stage, the signatories caneither act under the terms of or deviate
from the corporate EA.
The tourism market in Zanzibar is modeled via data from Lange and
Jiddawi (2009) and the statistical ofce in Zanzibar (OCGS, 2007). In
2007, a total of 219,047 visitors arrived in Zanzibar of which 4.95%
came from Germany and 7.34% were classied as eco-friendly tourists.
2
In accordance with this data, it is estimated that 796 German tourists
are willing to pay a price premium for their journey to Zanzibar if their
traveling based ecological footprint is offset via PACs. Following the re-
sults from Segerstedt and Grote (2013), who evaluate the willingness to
pay of long-distance tourists from Germany to protect ecosystems at
their travel destination, a price premium of p
P
= US$60 is dened. The
price for one PAC that allows protecting one global hectare of land is set
at p
Z
= US$200. This certicate price is based on the costs for emission
offsets. According to Atmosfair's emission calculator, the total climate im-
pact of a return ight from Frankfurt international airport in Germany to
Kisauni airport inZanzibar is estimatedat 4170 kg CO
2
. If customers want
to offset these emissions, they have to pay a fee of US$31/tCO
2
, and thus
US$130 per ight (Atmosfair, 2013). However, PACs are not restricted
to sequestering carbon. It is assumed that an extra US$70 is required to
provide ecosystem services that go beyond carbon sequestration and to
develop a new standard and certication scheme.
The benets of ecological risk mitigation derived fromthe protection of
ecosystems are described by the ln-type function b(Z) = a ln(Z + 1)
that is increasing and concave in the number of certicates. The benet
function not only considers the improved quality of tourist experience
due to the preservation of biodiversity and scenic beauty, but also the
security of the tourism industry's long-term business success that re-
sults from the conservation of land.
4.1. The Ecological Footprint of German Tourism in Zanzibar
Several publications (Becken et al., 2002; Gssling, 2000; Hyer,
2000) indicate that the majority of the ecological footprint of long-
distance journeys is caused by carbon dioxide (CO
2
) emitted into the at-
mosphere during the ight. Based on this assumption, the footprint of
2
Lange and Jiddawi (2009) specify visitor types according to accommodation data. Eco-
friendly tourists are supposed to stay in small, mostly foreign owned hotels that operate
with the greatest environmental sensitivity and focus on a long-termpartnership with lo-
cal communities.
153 N. Meiner / Ecological Economics 95 (2013) 148158
air transport on fossil energy land is used as an estimate for the ecolog-
ical footprint of the tourismindustry. Fossil energy land is dened as the
area of newly planted forest that is required to absorb and store emitted
carbon. One global hectare of fossil energy land is able to sequester CO
2
released from 73.08 GJ of energy from burnt liquid fossil fuel per year
(Loh, 2000). To account for the additional warming potential of CO
2
emitted in the upper troposphere and lower stratosphere, the Interna-
tional Panel on Climate Change suggests to weigh aircraft emissions
by a factor of 2.53.0 (Penner et al., 1999). Adopting a factor of 2.7,
the energy footprint for liquid fossil fuel from air travel is adjusted to
27.07 GJ/gha per year. The ight distance between Frankfurt and
Kisauni airport is 6975 km (Atmosfair, 2013). Considering an average
energy use of air transport of 2.0 MJ per passenger kilometer as recom-
mended by Gssling et al. (2002), a total of 27.9 GJ per passenger is
needed for the transport between the two countries. In other words,
the protection of about one global hectare of land is needed to offset
the ecological footprint of one tourist. The aggregated ecological foot-
print of German tourism adds up to 11,176 gha which equals 0.82% of
the total ecological footprint of Zanzibar.
3
Using an equivalence factor
of 1.26 for fossil energy land and a yield factor of 0.2 to account for coun-
try specic productivity (Ewing et al., 2010), a protected area zone of
70,410 ha has to be established in Zanzibar to offset the ecological foot-
print of German tourism.
4
In any case, even if the footprint of air transport is used as an esti-
mate for the total ecological footprint of the tourism industry, the ben-
ets of protected ecosystems are not restricted to carbon sequestration.
Among other things, functioning ecosystems provide biodiversity and
scenic beauty of the landscape, and thus enhance the experience of
tourismin Zanzibar. It is assumed that a protected area encloses a terri-
tory which is accessible for no other sector than the tourism industry
(Dudley, 2008). As a result, German tourism operators receive a share
of = 4.95% in the overall benet of ecological risk mitigation that cor-
responds to the proportion of their client base (OCGS, 2007).
4.2. Corporate Ecological Footprints
In this paragraph, the impact of ecosystem dependency on PAC in-
vestment is analyzed. For this purpose, two scenarios that vary in the al-
location of market shares are compared. Investment decisions are
illustrated through game matrices that display PAC threshold prices. If
the threshold price is higher (lower) than the PAC market price, the
payoff increases (decreases) by protecting ecosystems. The bigger the
difference between market and threshold prices, the bigger the net
gain (loss) of investing. A detailed description of the calculation of
threshold prices, welfare impacts and subsidies is given in Appendix B.
In scenario one, the market for German tourism in Zanzibar is un-
equally divided between two tourist operators: company A serves

A
= 80% of all tourists; and company B has a market share of

B
= 20%. Consequently, the ecological footprint of company A is
four times bigger than the ecological footprint of company B. The
corporate threshold prices for buying PACs are displayed in
Tables 1 and 2. Buying PACs is a strategy with strict dominance for
tourist operator A. However, company A has no incentive to volun-
tarily purchase an extra amount of PACs that cannot be transferred
to its customers via price premiums. Assuming rational behavior of
both companies as well as common knowledge of rationality, tourist
operator B anticipates the behavior of its competitor and refuses to
invest into ecosystem sustainability in Zanzibar. The resulting
unique pure-strategy Nash equilibrium (s
A

,s
B

) = (C,F) leads to a
payoff increase of (690.3, 194.9) measured in US$1000. The social
welfare rises by US$19.6 million. In case both companies commit to
buy certicates, the common threshold price reaches US$1326.26.
The aggregated payoff increases by US$11,311; company A gains an-
other US$26,857 and company B loses US$15,546. In total, the new
equilibrium leads to a Pareto superior outcome with an improve-
ment of social welfare of US$655,946. Tourist operator B though
does not agree to invest in PACs unless a subsidy of at least US
$97.77 per certicate is paid. Considering the different market
shares, this amount is equivalent to a net subsidy of US$24.40 for
each PAC bought by tourist operator A. In fact, company A is willing
to pay a maximum subsidy of US$42.16 per certicate as this equals
the additional benet obtained from greater ecological protection. If
the costs for the development of the coalition are smaller than the
maximum transaction costs of US$11,311, both tourist operators
agree to sign a corporate EA that guarantees the redistribution of
prots. Regarding the stability of the EA, it becomes clear that none
of the two companies have an incentive to deviate from the common
agreement as payoffs cannot be improved.
In scenario two, the market is divided equally between both compa-
nies. The adjusted corporate threshold prices are displayed in Tables 3
and 4. In a similar manner to scenario one, none of the companies
have an incentive to buy certicates that cannot be used for product la-
beling. But when a price premium is gained, investing in PACs becomes
a best response to a free riding competitor. Inthe event that the compet-
itor is contributing to the environment, the best response changes to
free riding. Hence, there are two possible pure-strategy Nash equilibria
(s
A

,s
B

) = (C,F) and (s
A

,s
B

) = (F,C). It is assumed that company A buys

PACs and company B acts as a free rider with a payoff increase of
(396.1, 451.8) in US$1000. The growth of social welfare amounts to
US$18.2 million. If both companies form a coalition and buy PACs,
they earn US$448,253 each and increase the aggregated payoff by US
$48,666 and the social welfare by US$2.1 million. Economically rational
players agree on signing a corporate EA at a subsidy rate of US
$8.86131.14 for every PAC bought by company A. However, a subsidy
scheme is difcult to enforce as both companies face the same depen-
dency on ecosystems and company B already records a higher net ben-
et. It is likely that company A considers any subsidy scheme as unfair
and refuses to improve its own payoff, contradicting the assumption of
economic rationality. Only if deviating fromthe EA is not punished, will
company A join a coalition in order to leave it as soon as company B
starts investing in PACs. This complete plan of action allows company
A to reach the preferred Nash equilibrium (F,C). Company B though
will anticipate this behavior and not sign a corporate EA in the rst
place.
3
According to Ewing et al. (2010), Tanzania's ecological footprint equals 1.18 gha/cap.
Considering a population of 1,155,065 (OCGS, 2007), Zanzibar's footprint amounts to
1,362,977 gha.
4
In the recent Ecological Footprint Atlas data is neither available for the yield factor of
forest in Zanzibar nor in Tanzania. Therefore, published data for Zambia is used (Ewing
et al., 2010).
Table 1
PAC threshold prices for Z
A,B
= Y
A,B
e
in US$ with
A
= 0.8,
B
= 0.2.
A\B C F
C 364.24*/102.23 1283.70*/0*
F 0/1023.14* 0/0
Table 2
PAC threshold prices for Z
A,B
N Y
A,B
e
in US$ with
A
= 0.8,
B
= 0.2.
A\B C F
C 151.44/37.86 189.18/0*
F 0*/188.59 0*/0*
Table 3
PAC threshold prices Z
A,B
= Y
A,B
e
in US$ with
A
=
B
= 0.5.
A\B C F
C 191.14/191.14 1195.12*/0*
F 0*/1195.12* 0/0
154 N. Meiner / Ecological Economics 95 (2013) 148158
Comparing bothscenarios, it canbe summarizedthat the more equal
corporate ecological footprints are, the more likely fairness issues inhib-
it the implementation of a subsidy scheme. If market shares do not suf-
ciently differ, punishment conditions become more important. In the
givenexample, the corporate EAbecomes unstable if the subsidy paying
company has a market share smaller than 58.78%.
4.3. Membership Restriction
It is supposed that a third competitor enters the German tourism
market in Zanzibar. The former market shares transform to
A
= 50%,

B
= 20% and
C
= 30%. The adjusted threshold prices for individual
PAC investment are shown in Tables 5 and 6. Besides labeling reasons,
no investment incentives exist. Analysis of the game matrices identies
the unique pure-strategy Nash equilibrium as (s
A

,s
B

,s
C

) = (C,F,F). In
stage 2, building a grand coalition leads to the highest threshold price
of US$1326.26. Consequently, the maximum producer and consumer
surpluses are obtained if all three tourism operators sign a common
EA. The outcome of the grand coalition is Pareto optimal. However,
the corporate EA does not forma Nash equilibriumas all signatories im-
prove their payoffs by leaving the coalition. In the event that no mone-
tary penalties are installed, a corporate EAinvolving all companies is not
stable. In fact, coalition membership has to be restricted to create inter-
nal stability. The outcomes for the grand coalition and the suggested
partial collaboration between company A and C are displayed in
Appendix B. The gures show that building a partial coalition leads to
a welfare reduction of US$655,946. However, none of the companies
seek to leave the coalition. Thus, membership restriction creates a
Pareto inferior but self-sustaining EA.
5. Conclusion
The paper gives initial policy advice for upcoming international ini-
tiatives (e.g. GDI, REDD+) that focus on the protection of ecosystems.
When companies individually decide on their environmental strategy,
they only value the benet of ecological risk mitigation for their own
business and the number of PACs falls below the socially optimal
amount. This constitutes a typical problem in the allocation of public
goods in which incentives to free ride increase with the environmental
activities of strategic opponents. But when a public good is connected
with a private good via labeling, free riding incentives are partly coun-
tered by the attempt to gain a price premium.
The demand for PACs can be increased through a corporate EA that
requires its signatories to offset their impact on ecosystems. But not all
coalition structures lead to an improvement of aggregated prots. In
fact, transaction costs for developing a corporate partnership can hinder
cooperation; a problem that increases if coalitions go beyond regional
and national partnerships. Furthermore, it might be mandatory to limit
the number of participating companies as the average benet per PAC
decreases with the overall size of protected areas. If coalitions seek to in-
clude large companies as well as SMEs, side payments are necessary to
eliminate cost differences. In the paper, side payments are realized by a
subsidy scheme that calls upon large enterprises to pay a higher price
for PACs thantheir SME counterpart. The success of the proposedsubsidy
scheme depends on the distribution of ecological footprints. The more
similar the corporate ecological footprints are, the more likely fairness
issues inhibit a redistribution of prots. In such cases, it is no longer
obvious who should be the net payers for ecological protection. If an
agreement is perceived to be unfair, players can behave economically
irrational and Pareto optimal coalitions are not formed.
If a coalition is formed, it is not necessarily self-sustaining in the
long-run. Some immediately fall apart. To transform these coalitions
into robust regimes, two options are discussed. On the one hand, mon-
etary penalties can be used to hinder companies from non-compliance.
On the other hand, membership restriction policies lead to an increase
of the average environmental loss per certicate caused by free riding.
While the implementation of punishment conditions might require
complex monitoring activities, exclusive membership causes a Pareto
inferior outcome.
Althoughexperts agree that environmental issues will play anincreas-
ingly important role in the development of corporate strategy (Esty &
Winston, 2009; Laszlo, 2008), no one size ts all solution exists. In the
end, strategic advantages always depend on the complete set of circum-
stances a company faces. For instance, the basic structure of the industry
(e.g. barriers to entry and possibility of product differentiation) has a cru-
cial impact on PAC investment (Reinhardt, 1999). However, it can be
noted that coalitions reduce the individual cost of ecological protection,
no matter how few companies are committed to PAC investment. Thus,
a voluntary but binding EAbetween companies allows the society to con-
verge towards the optimum in terms of ecosystem sustainability. Given
the effort in developing land based certication schemes, it seems advis-
able to extendthe model, for example by including different sectors inthe
analysis. Furthermore, it is recommended to examine how companies
perceive the development of an international PAC market in order to sup-
port the theoretical propositions of the paper, but more importantly to
learn from the companies' experiences with existing initiatives.
Appendix A. Development of Social Welfare
In stage 1, only those companies invest in PACs that fulll the
thresholdprice conditions (13) and(14). Thus, PS
01
0. Furthermore,
CS is strictly increasing in the amount of protected areas so that
W
01
=W
1
W
0
= PS
01
+ CS
01
0 Z
1
. Fig. A.1 shows
the change of social welfare due to PAC investment. As W
01
0, the
amount of purchased certicates Z
1
lies within the interval [0,Z

]. The
socially optimal number of PACs equals Z

. Following the FOC of the

welfare change function, two conditions are derived
i)
W
1
W
0

Z
p
Z
p
P
p
Z

b
Z
0

b
Z

p
Z
p
P
p
Z
for Z Y
e
1
;
A:1
ii)
W
1
W
0

Z
p
Z
p
Z

b
Z
0

b
Z
1 for ZNY
e
1
:
A:2
Table 5
PAC threshold prices Z
A,B,C
= Y
A,B,C
e
in US$ for s
C
= {F}.
A\B C F
C 296.77*/123.65/0* 1195.12*/0*/0*
F 0/1023.14*/0 0/0/0
Table 4
PAC threshold prices Z
A,B
N Y
A,B
e
in US$ with
A
=
B
= 0.5.
A\B C F
C 94.65/94.65 189.06/0*
F 0*/189.06 0*/0*
Table 6
PAC threshold prices Z
A,B,C
= Y
A,B,C
e
in US$ for s
C
= {C}.
A\B C F
C 191.14/102.23/127.51 245.31*/0*/148.89
F 0*/156.47/233.05* 0/0*/1097.91*
155 N. Meiner / Ecological Economics 95 (2013) 148158
However, contributing companies improve their aggregated payoff
according to
i)
PS
C
1
PS
C
0
_ _
Z
p
Z
p
P
p
Z

1

b
Z
0

b
Z

p
Z
p
P
p
Z

1
for Z Y
e
1
;
A:3
ii)
PS
C
1
PS
C
0
_ _
Z
p
Z
p
Z

1

b
Z
0

b
Z

1

1
for ZNY
e
1
:
A:4
Considering that the benet function is increasing and concave in the
number of PACs, the results show that the optimal number of certi-
cates for contributing companies (
1
1) is smaller than the optimal
amount for the overall society. Thus, Z
1
b Z

and
W
1
W
0

Z
N0 for Z
1
Y
e
1
Z

: A:5
In stage 2, only those coalitions are build that fulll the threshold
price condition (24) so that PS
12
0 and W
12
= W
2
W
1
=
PS
12
+ CS
12
0 Z
2
. The maximum welfare increase of soci-
ety is dened by
W
2
W
1

Z
p
Z
p
P
p
Z

b
12

Z
0

b
12

Z

p
Z
p
P
p
Z
for ZNZ
1
:
A:6
In contrast, coalition members maximize their aggregated payoff
according to
PS
G
2
PS
G
1
_ _
Z
p
Z
p
P
p
Z

2

b
12

Z
0

b
12

Z

p
Z
p
P
p
Z

2
for ZNZ
1
:
A:7
The demand for PACs (
2
1) remains below the social optimum so
that Z
2
b Z

and
W
2
W
1

Z
N0 for Z
2
Y
e
2
Z

: A:8
Appendix B. Data of the Numerical Example
The following calculations apply to scenario one with
A
= 0.8,

B
= 0.2,
G
= = 0.0495, p
P
= US$60, p
Z
= US$200, Y
e
= 796,
Y
A
e
= 637 and Y
B
e
= 159. The maximum benet of the protected
area scheme in Zanzibar equals the difference of the ecological
footprint (EF = 1,362,976.7 gha) minus the biological capacity
(BC = 1,178,166.3 gha) of Zanzibar. Thus, the scaling factor a of
the benet function b(Z) = a ln (Z + 1) is dened as follows
b EFBC EFBC
a
EFBC
ln EFBC 1
15; 239:47:
B:1
B.1. PAC Threshold Prices
Corporate and common threshold prices for buying PACs are calcu-
lated using equations (13), (14) and (24), respectively.
1. Z
A,B
= Y
A,B
e
with (s
A
,s
B
) = (C,F)
p
Z
A
p
P
p
Z

A

aln Z
A
1
Z
A
US$1283:70
2. Z
A,B
N Y
A,B
e
with (s
A
,s
B
) = (C,C)
p
Z
A
p
Z

A

a
Z
A
ZB1
US$151:44
3. Z
A,B
= Y
A,B
e
with (s
A
,s
B
) = (GC,GC)
p
Z;G
p
P
p
Z

G

aln Z
A
ZB1
Z
A
ZB
US$1326:26
B.2. Welfare Impacts
In stage 1, the change of the producer and consumer surpluses are
dened by equations (16) and (19). In stage 2, the increase of social
welfare can be determined by equations (27) and (28). In the grand co-
alition, it holds that PS
12
= T
max
G
.
1. Payoff increase of company A in stage 1 with (s
A

,s
B

) = (C,F)
PS
A,01
=(p
Z
p
P
)

Y
A
e
+p
Z

A
a

ln(Y
A
e
+ 1) =
US$690.310
3
2. Increase of social welfare in stage 1 with (s
A

,s
B

) = (C,F)
W
01
=(p
Z
p
P
)

Y
A
e
+ p
Z

a

ln(Y
A
e
+ 1) = US$19.6

10
6
3. Aggregated payoff increase of company A and B in stage 2 with
(s
A

,s
B

) = (GC,GC)
PS
12
p
Z
p
P
_ _
Y
e
B
p
Z

G
a lnY
e
A
Y
e
B
1ln Y
e
A
1
_ _

US$11:3 10
3
T
G
max
4. Increase of social welfare in stage 2 with (s
A

,s
B

) = (GC,GC)
W
12
=(p
Z
p
P
)

Y
B
e
+ p
Z

a

[ln(Y
A
e
+ Y
B
e
+ 1) ln(Y
A
e
+ 1)] =
US$655.9

10
3
B.3. Subsidy Scheme
Equation (35) estimates the maximum subsidy price company A is
willing to pay, while equation (36) denes the minimum subsidy
price company B is willing to accept.
1. Maximum subsidy price of company A
p
C;S
max
p
Z

A

a ln Y
e
A
Y
e
B
1 ln Y
e
A
1
Y
e
A
US$42:16
2. Minimum subsidy price of companies A and B, respectively
p
F;S
min
p
Z
p
P
p
Z

B

a ln Y
e
A
Y
e
B
1 ln Y
e
A
1
Y
e
B
US$97:77
p
C;S
min

p
F;S
min
Y
e
B
Y
e
A
US$24:4
Z
1 0
W
*
Z
Z
e
1
Y

Fig. A.1. Welfare change caused by PAC investment.

156 N. Meiner / Ecological Economics 95 (2013) 148158
References
Abbink, K., Moller, L.C., O'Hara, S., 2005. The Syr Darya River Conict: An Experimental
Case Study. CeDEx Discussion Paper No. 2005-14. University of Nottingham, UK.
Ambec, S., Ehlers, L., 2008. Cooperation and equity in the river-sharing problem. In:
Dinar, A., Albiac, J., Snchez-Soriano, J. (Eds.), Game Theory and Policymaking
in Natural Resources and the Environment. Routledge, London, New York,
pp. 112131.
Table B.1
PAC threshold prices, welfare impacts and subsidies.
Scenario 1:
market share 0.8/0.2
Scenario 2:
market share 0.5/0.5
Scenario 3:
market share 0.5/0.2/0.3
Zanzibar tourism
Tourists from Germany in 2007 10,843 10,843 10,843
Company A 8674 5422 5422
Company B 2169 5422 2169
Company C 3252
Eco-friendly customers (Y
e
)
Tourists from Germany in 2007 796 796 796
Company A 637 398 398
Company B 159 398 159
Company C 239
PAC threshold prices p
Z
; p
Z;G
_ _
in US$ for Z
A,B,C
= Y
A,B,C
e
Company A with (s
A
, s
B
) = (C, F) 1283.70 1195.12
Company A with (s
A
, s
B
) = (C,C) 364.24 191.14
Company B with (s
A
, s
B
) = (F, C) 1023.14 1195.12
Company B with (s
A
, s
B
) = (C, C) 102.23 191.14
Coalition A + B with (s
A
,s
B
) = (GC, GC) 1326.26 1326.26
Company A with (s
A
, s
B
, s
C
) = (C, F, F) 1195.12
Company A with (s
A
, s
B
, s
C
) = (C, C, F) 296.77
Company A with (s
A
, s
B
, s
C
) = (C, F, C) 245.31
Company A with (s
A
, s
B
, s
C
) = (C, C, C) 191.14
Company B with (s
A
, s
B
, s
C
) = (F, C, F) 1023.14
Company B with (s
A
, s
B
, s
C
) = (F, C, C) 156.47
Company B with (s
A
, s
B
, s
C
) = (C, C, F) 123.65
Company B with (s
A
, s
B
, s
C
) = (C, C, C) 102.23
Company C with (s
A
, s
B
, s
C
) = (F, F, C) 1097.91
Company C with (s
A
, s
B
, s
C
) = (F, C, C) 233.05
Company C with (s
A
, s
B
, s
C
) = (C, F, C) 148.89
Company C with (s
A
, s
B
, s
C
) = (C, C, C) 127.51
Coalition A + B with (s
A
, s
B
, s
C
) = (PC, PC, F) 1259.12
Coalition A + B with (s
A
, s
B
, s
C
) = (PC, PC, C) 287.57
Coalition A + C with (s
A
, s
B
, s
C
) = (PC, F, PC) 1283.70
Coalition A + C with (s
A
, s
B
, s
C
) = (PC, C, PC) 364.24
Coalition B + C with (s
A
, s
B
, s
C
) = (F, PC, PC) 1195.12
Coalition B + C with (s
A
, s
B
, s
C
) = (C, PC, PC) 191.14
Coalition A + B + C with (s
A
, s
B
, s
C
) = (GC,GC,GC) 1326.26
PAC threshold prices p
Z
_ _
in US$ for Z
A,B
N Y
A,B
e
Company A with (s
A
,s
B
) = (C,F) 189.18 189.06
Company A with (s
A
,s
B
) = (C,C) 151.44 94.65
Company B with (s
A
,s
B
) = (F,C) 188.59 189.06
Company B with (s
A
,s
B
) = (C,C) 37.86 94.65
Welfare impacts stage 0 1 in US$
Producer surplus (PS
01
) 885,194.14 847,838.92 847,838.92
Company A 690,319.31 396,059.46 396,059.46
Company B 194,874.83 451,779.46 180,711.78
Company C 271,067.67
Social welfare (W
01
) 19,595,146.10 18,197,995.48 18,197,995.48
(GC, GC) (GC, GC) (PC, F, PC) (GC, GC, GC)
Welfare impacts stage 1 2 in US$
Producer surplus (PS
12
) 11,311.21 48,666.43 37,355.23 48,666.43
Company A 26,856.97 52,193.22 35,407.61 52,193.22
Company B 15,545.76 3,526.78 14,163.05 1,382.71
Company C 12,215.43 2,144.07
Max transaction costs (T
max
G
) 11,311.21 48,666.43 23,192.18 48,666.43
Social welfare (W
12
) 655,946.23 2,053,096.85 1,397,150.61 2,053,096.85
Subsidy scheme (p
C,S
, p
F,S
) in US$
Company A: Max subsidy price 42.16 131.14 88.96 131.14
Min subsidy price 24.40 8.86 30.69 8.86
Min punishment (P
A
) 89,074.42 7,053.57 5,817.67 7,053.57
Company B: Min subsidy price 97.77 8.86 8.70
Min punishment (P
B
) 0.00 0.00 14,163.05
Company C: Min subsidy price 51.11 8.97
Min punishment (P
C
) 0.00 15,180.50
157 N. Meiner / Ecological Economics 95 (2013) 148158
Atmosfair, 2013. Compensate my ight. URL https://www.atmosfair.de/en/projekt2/
ugrechner/ (July).
Barrett, S., 1994. Self-enforcing international environmental agreements. Oxf. Econ. Pap.
46, 878894.
Barrett, S., 2003. Environment and Statecraft: The Strategy of Environmental Treaty-Making.
Oxford University Press, New York.
Barrett, S., 2007. Why Cooperate? The Incentive to Supply Global Public Goods. Oxford
University Press, New York.
Barrett, S., Stavins, R.N., 2003. Increasing participation and compliance in international
climate change agreements. Int. Environ. Agreements 3, 349376.
Becken, S., Frampton, C., Simmons, D.G., 2002. Analysing international tourist ows to
estimate energy use associated with air travel. J. Sustain. Tour. 10 (2), 114131.
Camerer, C.F., 2003. Behavioral Game Theory: Experiments in Strategic Interaction.
Princeton University Press, Princeton, New Jersey.
Carius, F. (Ed.), 2010. Payments for Ecosystem Services: Towards an Implementation
Strategy. Report of the International Expert Workshop 13th - 16th December 2010.
International Academy for Nature Conservation Isle of Vilm, Bundesamt fr
Naturschutz, Bonn, Germany.
Carraro, C., Siniscalco, D., 1993. Strategies for the international protection of the environ-
ment. J. Public Econ. 52, 309328.
CDP, 2012a. Business Resilience in an Uncertain, Resource-Constrained WorldCDP
Global 500 Climate Change Report 2012. Carbon Disclosure Project, London, UK.
CDP, 2012b. Collective Responses to Rising Water ChallengesCDP Global Water Report
2012. Carbon Disclosure Project, London, UK.
Chander, P., Tulkens, H., 1997. The core of an economy with multilateral environmental
externalities. Int. J. Game Theory 26 (3), 379401.
Chander, P., Tulkens, H., 2008. Cooperation, stability and self-enforcement in international
environmental agreements: a conceptual discussion. In: Guesnerie, R., Tulkens, H.
(Eds.), The Design of Climate Policy. MIT press, Cambridge, pp. 165186.
Drr, A.C., 2008. Economic analysis of certication in the Brazilian fruit chain. Ph.D. thesis,
Leibniz University of Hannover, Germany.
Dudley, N. (Ed.), 2008. Guidelines for Applying Protected Area Management Categories.
IUCN, Gland, Switzerland.
Dummett, K., 2006. Drivers for corporate environmental responsibility (CER). Environ.
Dev. Sustain. 8, 375389.
Duraiappah, A.K., Naeem, S., 2005. Ecosystems and Human Well-being: Biodiversity
Synthesis. MillenniumEcosystemAssessment. World Resources Institute, Washington,
D.C.
Emerton, L., Bishop, J., Thomas, L., 2006. Sustainable Financing of Protected Areas: A global
Review of Challenges and Options. IUCN, Gland, Switzerland.
Esteban, E., Dinar, A., 2013. Cooperative management of groundwater resources in the
presence of environmental externalities. Environ. Resour. Econ. 54, 443469.
Esty, D.C., Winston, A.S., 2009. Green to Gold: How Smart Companies Use Environmental
Strategy to Innovate, Create Value, and Build a Competitive Advantage. John Wiley &
Sons, Hoboken, New Jersey.
Ewing, B., Moore, D., Goldnger, S., Oursler, A., Reed, A., Wackernagel, M., 2010. The Eco-
logical Footprint Atlas 2010. Global Footprint Network, Oakland, California.
Finus, M., 2001. Game Theory and International Environmental Cooperation. Edward
Elgar Publishing, Cheltenham, UK.
Finus, M., 2003. Stability and design of international environmental agreements: the case
of transboundary pollution. In: Folmer, H., Tietenberg, T.H. (Eds.), The International
Yearbook of Environmental and Resource Economics 2003/2004: A Survey of Current
Issues. Edward Elgar Publishing, Cheltenham, UK, pp. 82158.
Finus, M., Rundshagen, B., 1998. Toward a positive theory of coalition formation and en-
dogenous instrumental choice in global pollution control. Public Choice 96, 145186.
Finus, M., Saz, M.E., Hendrix, E.M.T., 2009. An empirical test of new developments in co-
alition theory for the design of international environmental agreements. Environ.
Dev. Econ. 14 (1), 117137.
Gssling, S., 2000. Sustainable tourism development in developing countries: some
aspects of energy-use. J. Sustain. Tour. 8 (5), 410425.
Gssling, S., Borgstrm-Hansson, C., Hrstmeier, O., Saggel, S., 2002. Ecological footprint
analysis as a tool to assess tourism sustainability. Ecol. Econ. 42, 199211.
Grote, U., 2009. Environmental labeling, protected geographical indications and the inter-
ests of developing countries. The Estey Centre. J. Int. LawTrade Policy 10 (1), 94110.
Grote, U., Basu, A.K., Chau, N.H. (Eds.), 2007. New Frontiers in Environmental and Social
Labeling. Physica-Verlag, Heidelberg, Germany.
Hyer, K.G., 2000. Sustainable tourism or sustainable mobility? The Norwegian case.
J. Sustain. Tour. 8 (2), 147160.
Jobin, D., 2008. A transaction cost-based approach to partnership performance evaluation.
Evaluation 14 (4), 437465.
Juutinen, A., Mitani, Y., Mntymaa, E., Shoji, Y., Siikamki, P., Svento, R., 2011. Combining
ecological and recreational aspects in national park management: a choice experi-
ment application. Ecol. Econ. 70, 12311239.
Karousakis, K., Brooke, Q., 2010. Paying for Biodiversity: Enhancing the Cost-Effectiveness
of Payments for Ecosystem Services. OECD Publishing, Paris, France.
Koellner, T., Sell, J., Navarro, G., 2010. Why and how much are rms willing to invest in
ecosystem services from tropical forests? A comparison of international and Costa
Rican rms. Ecol. Econ. 69, 21272139.
Lange, G.-M., Jiddawi, N., 2009. Economic value of marine ecosystemservices in Zanzibar:
implications for marine conservation and sustainable development. Ocean Coast.
Manag. 52, 521532.
Laszlo, C., 2008. Sustainable Value: How the World's Leading Companies are Doing Well
by Doing Good. Stanford University Press, California.
Loh, J. (Ed.), 2000. Living Planet Report. WWF - World Wide Fund for Nature, Gland,
Switzerland.
McWilliams, A., Siegel, D.S., 2011. Creating and capturing value: strategic corporate
social responsibility, resource-based theory, and sustainable competitive advantage.
J. Manag. 37 (5), 14801495.
Munro, G.R., 2008. Game theory and the development of resource management policy:
the case of international sheries. In: Dinar, A., Albiac, J., Snchez-Soriano, J. (Eds.),
Game Theory and Policymaking in Natural Resources and the Environment.
Routledge, London, New York, pp. 1241.
OCGS, 2007. Zanzibar Statistical Abstract. Ofce of Chief Government Statistician,
Zanzibar, Tanzania.
Osborne, M.J., Rubinstein, A., 1994. A Course in Game Theory. MIT Press, Cambridge,
Massachusetts.
Pascual, U., Muradian, R., 2010. The economics of valuing ecosystem services and biodi-
versity. In: Kumar, P. (Ed.), TEEBEcological and Economic Foundation (Chapter 5).
Penner, J.E., Lister, D.H., Griggs, D.J., Dokken, D.J., McFarland, M. (Eds.), 1999. Aviation and
the Global Atmosphere. Intergovernmental Panel on Climate Change. Cambridge Uni-
versity Press, UK.
Pintassilgo, P., Finus, M., Lindroos, M., Munro, G., 2010. Stability and success of regional
sheries management organizations. Environ. Resour. Econ. 46, 377402.
Pintassilgo, P., Lindroos, M., 2008. Application of partition function games to the manage-
ment of straddling sh stocks. In: Dinar, A., Albiac, J., Snchez-Soriano, J. (Eds.), Game
Theory and Policymaking in Natural Resources and the Environment. Routledge,
London, New York, pp. 6584.
Porter, M.E., Kramer, M.R., 2006. Strategy and society: the link between competitive
advantage and corporate social responsibility. Harv. Bus. Rev. 84 (12), 7892.
Reinhardt, F.L., 1999. Market failure and the environmental policies of rms: economic
rationales for beyond compliance behavior. J. Ind. Ecol. 3 (1), 921.
Rockstrm, J., Steffen, W., Noone, K., Persson, ., Chapin, F.S. III, Lambin, E., Lenton, T.M.,
Scheffer, M., Folke, C., Schellnhuber, H.J., Nykvist, B., de Wit, C.A., Hughes, T., van der
Leeuw, S., Rodhe, H., Srlin, S., Snyder, P.K., Costanza, R., Svedin, U., Falkenmark, M.,
Karlberg, L., Corell, R.W., Fabry, V.J., Hansen, J., Walker, B., Liverman, D., Richardson, K.,
Crutzen, P., Foley, J., 2009. Planetary boundaries: exploring the safe operating space
for humanity. Ecol. Soc. 14 (2), 32 ([online], URL http://www.ecologyandsociety.org/
vol14/iss2/art32/).
Rotherham, T., 2004. The Trade and Environmental Effects of Ecolabels: Assessment and
Response. United Nations Environment Programme (UNEP), Genf, Switzerland.
Segerstedt, A., Grote, U., 2013. Ecosystem certicates in the tourism sector - gaining
ground for better ecosystem protection? Unpublished Article, Leibniz University of
Hannover, Germany.
Tietenberg, T., Lewis, L., 2012. Environmental & Natural Resource Economics, 9th edition.
Pearson, Boston, Massachusetts.
Trucost, 2013. Natural Capital at Risk: The Top 100 Externalities of Business. TEEB for
Business Coalition, Trucost Plc, London, UK.
Wackernagel, M., 1994. Ecological footprint and appropriated carrying capacity: a tool
for planning toward sustainability. Ph.D. thesis, University of British Columbia,
Canada.
Wackernagel, M., Rees, W.E., 1996. Our Ecological Footprint: Reducing Human Impact on
the Earth. New Society Publishers, Gabriola Island, Canada.
Weikard, H.-P., Finus, M., Altamirano-Cabrera, J.-C., 2006. The impact of surplus sharing
on the stability of international climate agreements. Oxf. Econ. Pap. 58, 209232.
Wiedmann, T., Minx, J., Barrett, J., Wackernagel, M., 2006. Allocating ecological foot-
prints to nal consumption categories with inputoutput analysis. Ecol. Econ.
56, 2848.
Wood, P.J., 2010. Climate Change and Game Theory. Environmental Economics Research
Hub, Research Report No. 62. Crawford School of Economics and Government,
Australian National University, Canberra.
158 N. Meiner / Ecological Economics 95 (2013) 148158