A BARGAINING FRAMEWORK FOR THE GLOBAL COMMONS



By
Daniel W. Bromley
and
Jeffrey A. Cochrane



Working Paper No. 21, 18 pages, November, 1995



For more information and copies of this paper, contact:

Daniel W. Bromley
Department of Agricultural Economics
Taylor Hall, Room 331
University of Wisconsin-Madison
Madison, WI 53706-1562

Tel:  (608) 262-6184
Fax:  (609) 265-3061
Email: bromley@vms.macc.wisc.edu



Produced by:
  
Ellen A. Maurer
Communications Director
EPAT/MUCIA Research & Training
University of Wisconsin-Madison
1003 WARF Office Building
610 Walnut Street
Madison, WI 53705-2397

Tel:  (608) 263-4781
Fax:  (608) 265-2993
Email: eamaurer@facstaff.wisc.edu




Edited by Sharon Graham



* Some figures and/or tables included in the printed version of
this publication could not be included in this electronic
version.  If you need copies of these figures or tables, please
contact the author.  



PROJECT INFORMATION

A USAID-funded global program, the Environmental and Natural
Resources Policy and Training Project (EPAT), is implemented, in
part, by 15 universities and development organizations through
the Midwest Universities Consortium for International Activities,
Inc. (MUCIA). 

EPAT/MUCIA has research, training, and communication components
that offer the latest information about:

  * Energy, Industry & the Urban Environment
  * Forestry & Watershed Management
  * Macroeconomic Policy
  * Population & the Environment
  * Institutions & Policy Implementation
  * Environmental Policy Training 
  * Environmental Policy Communications

EPAT/MUCIA publications include:

  * Policy Briefs - short overviews of environmental policy  
    concerns
  * Case Studies - outlining specific in-country policy 
    challenges
  * Manuals - how-to-do-it environmental policy handbooks for 
    easy reference
  * Working Papers - of environmental policy research in progress
  * Monographs - detailing the state of knowledge in particular
    subject matter areas.

EPAT/MUCIA environmental policy partners have applied their
research to real problems, and they collaborated with researchers
throughout the world.

For more information about EPAT/MUCIA contact: 
  Chief of Party
  1611 North Kent Street, Suite 807
  Arlington, VA  USA 2209-2134   
  Tel: (703) 841-0026
  Fax: (703) 841-0699

EPAT/MUCIA-Research and Training partners include University of
Arizona; Cornell University; University of Illinois; Indiana
University; University of Iowa; University of Michigan; Michigan
State University; University of Minnesota; The Ohio State
University; Purdue University; University of Wisconsin; Abt
Associates; Development Alternatives, Inc.; Development
Associates, Inc.; and World Resources Institute.

Working Paper ISSN # 1072-9496



ABSTRACT



The global commons represents a class of environmental problems
that require international cooperation.  Global environmental
problems arise because the actions of some individuals or
governments in one location hold serious implications for
individuals and governments in other locations.  There are global
environmental policy problems because adversely-affected
individuals (and governments) attempt to alter the behaviors of
those responsible for global pollution.

We develop a framework with which to analyze such global
environmental problems.  Our goal is to craft a resource
management policy that will satisfy both those who seek a change
and those who prefer the current situation.  This environmental
incentive policy will align the interests of the two parties. 
Incentive alignment is the policy problem in the global commons. 
That is, we must find ways to align interests through realigning
incentives for individual and group behaviors.




FOREWORD



This Working Paper is a product of the Environmental and Natural
Resources Policy and Training (EPAT) Project funded by the United
States Agency for International Development (USAID).  EPAT is
part of USAID's effort to provide environmental policy
information to policymakers and practitioners in developing
countries.  The objective is to encourage the adoption of
economic policies for promoting sustainable use of natural
resources and enhancing environmental quality.

EPAT Working Papers are written for development professionals and
policymakers in developing countries who are responsible for
establishing and implementing policies on the sustainable use of
natural resources and for civil servants, project officers, and 
researchers who are directly involved in the implementation of
development activities.  

This Working Paper deals with a class of environmental problems,
called the global commons, that require international
cooperation.  They arise because the actions of individuals or
governments in one location hold serious implications of
individuals and governments in other locations.  This paper
develops a framework for analyzing such problems.  The goal is to
craft an environmental incentive policy that will align the
interests of both those who seek change and those who prefer the
current situation.  Policymakers may find incentive alignment to
be a useful strategy for modifying individual and group behaviors
in order to deal with the problems of the global commons.

The contribution of USAID toward writing, printing, and
distributing this document is estimated to be $8,000.  The
document is being distributed to more than 2,000 policymakers and
professionals in developing countries.  We will assess its
effectiveness by soliciting the views of recipients.  An
evaluation sheet is enclosed with each mailing of EPAT
publications for that purpose.




David Hales                         William Sugrue
Deputy Assistant Administrator      Acting Director
Center for the Environment          Office of Environment &      
USAID/G/ENV                         Natural Resources
Washington, D.C.  20523             USAID/G/ENV/ENR
                                    Washington, D.C.  20523




CONTENTS



A BARGAINING FRAMEWORK FOR THE GLOBAL COMMONS

THE EMPIRICAL PROBLEM

AGENCY THEORY AND BILATERAL NEGOTIATIONS

THE DOMESTIC POLICY PROBLEM

THE INTERNATIONAL POLICY PROBLEM
  Efficiency
  Feasibility

CONCLUSIONS

REFERENCES




A BARGAINING FRAMEWORK FOR THE GLOBAL COMMONS



The global commons represents an environmental problem area that
requires international cooperation (Bromley and Cochrane 1993). 
Individual and governmental actions in one location cause global
environmental problems and seriously impact individuals and
governments in other locations.  Certain natural resource use
threatens individual (and state) interests.  There are global
environmental policy problems because adversely-affected
individuals (and governments) attempt to alter the behaviors of
global polluters.

We develop a framework with which policymakers can analyze such
global environmental problems.  Our goal is to craft a resource
management policy that will satisfy both those who seek a change
and those who prefer the current situation.  This environmental
incentive policy will align the interests of the two parties. 
Incentive alignment is the policy problem in the global commons. 
That is, we must find ways to align interests by realigning
incentives for individual and group behaviors.




THE EMPIRICAL PROBLEM



We will focus on greenhouse gas emissions and their relationship
to the world's forests.  The link between the earth's atmosphere
and forests is direct and critical to sustaining life.  The
interests of one group is at odds with the interests of another. 
Those who harvest trees and clear land for settlement in the
Amazon Forest, for example, stand against those who seek to
protect the forests.

The problem is global.  For example, it has been estimated that:
"...South and Southeast Asia contribute about 25% of the carbon
dioxide emissions caused by burning wood, or about 6% of total
carbon dioxide emissions (Archer and Ichord 1989: 13)."

The industrial world--with its fossil-fuel driven factories and
automobiles--is a major contributor to the total annual
production of greenhouse gases.  It is reasonable to suggest that
the wealthy citizens of the industrial North wish to protect the
Amazonian "lungs of the earth" so it can process the carbon
dioxide that arises from our self-indulgent lifestyle.  Tropical
forests have become a free waste-processing facility for richer
nations--Japan and European and North American countries.

Human activities in the industrialized world generate large
quantities of greenhouse gases.  Tropical forests serve an
important function by processing much of those gases.  However,
land uses in the agricultural tropics threaten the sustainability
of that forest cover.  These land-use decisions link agricultural
nations with industrialized nations and their activities. 
Current energy consumption in the industrialized world imposes
demands on the biosphere's resources, and tropical land-use
activities threaten the sustainability of those resources. 
People in the industrialized North want to protect tropical
forests to maintain the forests' ability to process their
increasing greenhouse gas production. 

Figure 1 shows the problem in simplified terms.  (Graph will not
transfer on internet.  Please write to the author for a hard
copy.  See address at beginning of this paper.) In panel (a), Gs
represents total greenhouse gas production from agricultural
countries located in the southern hemisphere.  Gn represents
total greenhouse gas production from more industrialized
economies primarily in Europe, North America, and Japan.  For
simplicity we will refer to these two general zones as North and
South or N and S.

G* represents the total level of greenhouse gas production that
will not change the atmosphere's chemistry.   Call G* the
sustainable greenhouse gas level.  The figure shows gas
production from industrial nations holding constant into the
foreseeable future.  The figure also shows greenhouse gas
production in the agricultural South increasing as those nations
advance economically and as factories and automobiles become more
common.   Developing countries' increased reliance on fossil
fuels can seriously impact carbon gas emission levels unless they
also change their technology.  For instance,

"...China's ratio of CO2 emissions to gross national product
(GNP) is roughly five times that of Japan....  If China were to
achieve even 60% of Japanese efficiency and carbon intensity
levels in its new energy-producing and energy-consuming
infrastructure, it could improve this ratio substantially in a
relatively short time.  The one major hurdle is obtaining the
technical information, management assistance, and capital
required to promote more efficient and less polluting supply and
use of energy and other natural resources [Nitze 1990: 608]."

The line labeled G shows total greenhouse gas production over
time (where G = Gn + Gs).  The cross-hatched region in panel (a)
shows excess gas production that the earth cannot absorb (G >
G*).  Here one sees the threat to both the industrial North and
the agricultural South.

International efforts to impose a global production limit at G*
would meet opposition from citizens in industrial nations because
it would threaten their lifestyles.  The industrialized world
might argue that increased greenhouse gas production from the
newly industrialized agricultural South has "created the
problem."  Similarly, a limit on total production concerns those
in the agricultural South because it the industrial North might
receive a more favorable allocation.  Citizens in the
agricultural South might argue that the rich wish to limit their
(the southerner's) industrial ambitions.

There is something even more threatening in panel (a).  Line G~
traces the earth's absorbent capacity for greenhouse gases
assuming that industrializing southern tropical forests will
include clearing and burning.  That is, G* is no longer the long-
run sustainable capacity to process greenhouse gases.  Clearing
Amazonian forests drives the absorbent capacity down to G~ .  The
shaded region in panel (a) shows the processing deficit resulting
from land use changes in the southern agrarian nations. 
Together, the two marked areas between G and G~ show that excess
greenhouse gases may significantly change the global climate.

Panel (b) illustrates a policy that cuts excess production.  To
do this, the industrial North must reduce total greenhouse gas
production at the same rate as the newly industrial South
increases production.  Then total greenhouse gas production from
both regions matches the predicted absorbent capacity (G = G*). 
As the South industrializes, it reduces forest cover so the
actual processing capacity is less than predicted (G~ < G* = G). 
A processing deficit still exists.  As drawn, the industrial
North, because of stricter environmental policies, has cut its
total greenhouse gas production below that in the newly
industrial South.

In panel (c), we do not assume massive forest clearing in the
industrializing South or that compensating reforestation occurs
in the North.  So G* remains the absorbent greenhouse gas
capacity, and total global production matches that capacity.  The
hard part remains.  How do we restructure international resource
management policy to achieve the result in panel (c)? 

There are two aspects of this challenge.  First, there is the
problem of agreement between these two artificial regions,
"North" and "South."  That is difficult enough.  The second
problem will concern changing behaviors of individuals within the
two regions.  To avoid discussing differing interests among
countries, we will regard the "North" (N) as one state and the
"South" (S) as another.  This does not change the nature of the
basic bargaining problem but simplifies it by reducing the number
of policy units.

To simplify further, the policy problem of the global commons is
finding incentives that will align the interests and, thus,
individual behaviors in the North and South.  That alignment of
interests will become apparent when individuals agree to choices
that will reduce two tendencies.  The first tendency is one that
threatens the earth's capacity to process greenhouse gases.  In
this paper, we consider the world's forests essential to that
global absorbent capacity.  For simplicity, we will focus on the
Southern deforestation problem.

The second tendency is one that threatens to increase total
greenhouse gas production in both regions.  Notice that the
physical aspects of atmospheric chemistry link individuals in
both regions.  More  individual greenhouse gas production in the
North, or South, increases the economic value of the forests.  

Higher total greenhouse gas production in both hemispheres will
reduce absorbent capacity.  The resulting accelerated climate
change will impose greater costs on the world.  To avoid that
climate change, the two regions will need to undertake expensive
alternatives to the forests' processing capacity.  Continuing to
reduce the Amazonian forest and holding total greenhouse gas
production at its present level, without other compensating
reforestation, will carry serious costs.

We assume that tropical and other forests provide a resource
service, waste assimilation, at a lower cost than alternative
technologies.  The only "cost" of using the forests for this
purpose is preserving them.  The cost then is that people cannot
use the forest area in alternative ways and still maintain it. 
We are assuming that tropical forests are important assets in
their own right, besides acting as  greenhouse gas processors.

The policy problem requires introducing an action-forcing event. 
Without it, there is no reason the current resource management
policy cannot continue despite increasing evidence of higher
greenhouse gas concentrations.  We will begin by assuming that
one of the governments (either N or S) determines that it is
desirable to stabilize greenhouse gas production so total
production balances with the earth's processing capacity.  Panel
(c), figure 1 shows G = G*.  The government could make this
decision autonomously or be "forced" into the decision by a
rising internal political force.  In November 1990, in Geneva,
several industrial world governments agreed to stabilize total
greenhouse gas production.  The 1992 Rio Earth Summit carried
this commitment one step further.




AGENCY THEORY AND BILATERAL NEGOTIATIONS



One possible mechanism for addressing greenhouse gas problems is
to imagine that some international organization, for example the
United Nations, receives the authority to impose emissions
standards on all nations.  Because of the historical experience
with mandates from such organizations, we need to be cautious
about the prospects for successfully enforcing emission
standards.

Another method is to approach the greenhouse gas problem in a
bargaining framework.  We follow this argument and apply the
economic model of agency (explained below) to the global commons
problem.  The usual examples in agency theory concern the
problems of team production where monitoring costs of individual
efforts are high.  The standard employment contract embodies the
essence of the agency problem.  In this paper, we will apply
agency theory both to the relationship between individuals and
the state and to the relationship between one state and another.

The logic for this approach is straightforward.  The legal
foundations of an economy define a domain of individual choice
for each of us.  The presumption is that on the whole, these
millions of independent choices will be in the "public interest."

This insight of Adam Smith and a policy of non-interference still
requires that we recognize that markets cannot function without a
clear and precise definition of who owns what (property rights),
who may do what to whom (civil and criminal law), and who must
pay whom to protect their interests (contract law).  (So called
"free markets" are clearly not free of collective definition of
the choice range open to market participants (Bromley 1989).)

Assume that it is the government of N that seeks a new resource
management policy to achieve sustainable greenhouse gas
production as in figure 1 panel (c).  In agency theory, we talk
of the principal and of the agent.  It is the problem for the
principal to design an incentive policy that will align the
agent's interests with the principal's.

Consider a simple example.  A family may wish to have a
neighbor's child weed their garden.  However, the child will be
working while the family is away for three weeks.  The gardener
is the principal; the child is the agent.  The gardener problem
is to design an incentive scheme that will align the child's
interests with the gardener's own interests.  Put more directly,
the gardener wishes to design a compensation system so the child
will carry out that task exactly as the gardener wishes.

Imagine the gardener says to the child: "I will pay you $1.00 per
day to weed my garden.  When I return in three weeks I expect to
find the garden weed free.  I will then pay you for the total
days you worked."  It is obvious that this scheme has an inherent
incentive alignment problem.  The gardener wants to have the
garden weeded at the least possible cost.  The child wants to
earn as much money as possible.  A person can easily imagine that
the youngster has a strong incentive to inflate the actual time
involved.  It is also possible that this payment scheme will
influence how aggressively the child works.  The child may report
days worked honestly, but how hard did she actually work during
those days?  Or, she may pull the weeds on the final day of the
gardener's absence, meaning that the garden is weed free but that
the plants are stunted.  This payment scheme leaves the gardener
bearing all the uncertainty.

An alternative, but impractical, payment scheme would be to pay
the child a "piece rate," an amount per weed pulled.  The
monitoring costs of this approach are prohibitive.  The obvious
compromise is to pay her for the entire job.  Both the child and
the gardener will share the uncertainty.  The gardener must
estimate, from experience, how many hours the job might take. 
The child must also calculate the estimated hours.  To maximize
earnings per hour, the child has an incentive to work fast. 
However, the gardener has an incentive not to underestimate. 
Otherwise the child will not accept the job or may work so fast
as to be careless.  This compromise aligns the interests of both.

The gardener wants to have a weed-free garden and pay a
"reasonable price" for it.  The child wants to earn a fair wage
yet have enough time to do the job well.

Many such services--leaf raking, lawn mowing, and house painting-
-operate precisely on this basis.  The policy problem in garden
weeding is to align the interests of the principal (the gardener)
and the agent (the weeder).  Agency theory deals with precisely
this problem, whether applied to gardeners, home owners, business
owners, or state governments.

Returning to the global commons problem, consider the government
of N that faces two "agency" problems.  The first is to induce
its own citizens to alter their behavior in producing greenhouse
gases.  Call this the domestic policy problem for N.  The second
agency problem concerns the government of S.  That is, how can
the government of N induce the government of S to limit
individual citizens' choices (individual decisionmakers) in S? 
Call this the international policy problem.  This hierarchical
agency problem has two "agents"--the government of S and the
citizens in S.  That is, the government of N is the principal to
its own citizens as well as to the government of S.  Note that
the government of S, besides being an agent for the government of
N, is a principal for its own citizens (who are, in turn, agents
to the government of S).

Government of N
Principal to its citizens and to S

Citizens in N
Agents to N

                    Government of S
                    Agent to N; Principal to its citizens

                    Citizens in S
                    Agents to S


Within this hierarchical problem, we assign to the government of
N the role of ultimate principal, recognizing that wealth affords
N greater freedom in action.  Thus, the event that forces action
and  leads to a policy initiative will most likely be found in N.

This event may be that scientists in N have discovered that
trends in atmospheric greenhouse gas levels may seriously impact
life, particularly in the industrialized world.  This revelation
focuses public attention, in N, on the issue.  The result is that
the government of N declares that it seeks to change the resource
management policy that over produces greenhouse gases.

While there may be public concern about greenhouse gas
production, we assume it is unimportant to many individual
citizens of both N and S.  This is because an individual
contributes only minimally to the problem.  It is the cumulative
effect that concerns us.  There is no great incentive for an
individual to alter behavior.  Indeed, there may be strong
incentives in other directions, particularly in S.  The
government may want to increase economic development at the
expense of the tropical forests (and their processing services). 
That policy would also increase greenhouse gas production through
accelerated industrialization.

As the principal, the government of N is not only challenged to
modify the millions of individual decisions within N but to do
the same within S.  The difference lies in the inability of the
government of N to develop direct behavior modifying policies in
S.  Generally, without applying direct physical force against S,
the government of N may only develop policies that affect S at
the border.  The government of N must work through the government
of S to encourage policies that require internal enforcement or
other action.  Therefore, the government of N faces both a
domestic and an international problem.




THE DOMESTIC POLICY PROBLEM



Initially, certain individuals in N can ignore the interests of
others in N who care about total greenhouse gas production.  In
more formal language, we would say that those concerned about
high greenhouse gas production have no rights, while those well
served by the current resource management policy have privilege
(Bromley 1991: 17).  By privilege we mean individuals can act
without regard for others.  If those alarmed about greenhouse
gases were to take legal action to prevent continued emissions,
the court would say, "Sorry, there is no law against the
greenhouse gas production.  You have no right to seek relief."

Current resource management policy is that the air is an open
access resource.  Anyone who wishes to partake of its services
may do so freely.  The costs that arise from greenhouse gas
emissions are of little concern to those responsible.  Of course,
some citizens in N care about such emissions.  They have tried to
pressure their government to change the current resource policy. 
As noted, the court told them that they had no right to prevent
such emissions.

We can define the domestic policy problem concerning greenhouse
gas emissions in terms of a change in the air resource policy.  
Those now in a situation of no right gain a right, and those now
with privilege gain, instead, duty.  That is, the policy changes
the economy's legal structure so that those with air emission
privileges acquire a duty to consider the harm their emissions
are causing others.  The state may set limits on gas emission
limits or ban them altogether.  It may require compensation for
greenhouse gas emissions, perhaps in taxes.  The state may also
offer subsidies to encourage change to cleaner technologies.  In
the language of resource policies, the state can use various
actions to create rights and corresponding duties.  They include:


* a right to be free of emissions and a duty to refrain from
emitting them,
* a right to compensation and a duty to pay it,
* a right to receive funds to purchase cleaner equipment and a
duty to provide those funds by paying taxes.

In each case, the government's aim is to alter the resource
policy structure to realign incentives and encourage particular
actions.

Consider the air emissions policy example in southern California.

It incorporates a variety of incentive realignment measures.  The
state established a greenhouse gas emissions ceiling for firms
but permitted them to meet this ceiling several ways.  The Amoco
oil company delayed reducing its own emissions by purchasing
older, heavily polluting automobiles from individuals.  This
enabled the individuals to purchase newer and cleaner cars.  By
eliminating Amoco's privilege to emit greenhouse gases and by
replacing that privilege with a duty to respect newly established
rights to be free of such gases, the government redefined Amoco's
economic environment.  The government did not, however, simply
specify levels of gas particular Amoco factories could emit. 
Instead, it specified a broad emissions limit that Amoco was free
to meet in several ways.  Amoco found it cheaper to cut emissions
from older polluting automobiles than to refit its refineries. 
Amoco met the goal, of both citizens and the government, at a
presumably lower cost than they could have if the government
simply mandated that all emissions reductions take place in Amoco
factories.

In the preceding agency examples, the principal could manipulate
the agent's actions by offering or withholding various forms of
payment.  It was the payment form that was particularly
interesting.  By changing the form of payment the principal was
able to minimize monitoring and enforcement costs.  In domestic
policy, the government of N faces a similar problem.  While it
can enact a resource policy with new rights and duties, its form
will have important monitoring and enforcement implications. 
Consider a hypothetical example in which the government fails to
give adequate attention to monitoring and enforcement costs.

Suppose the government requires people to stop driving their
automobiles once they emit a certain amount of emissions that
year.  The advantages of cheating, particularly for the poor who
could not easily afford to purchase cleaner cars, would be great.

To enforce such a policy, the government would need to take three
steps.  It must require meters on all vehicles.  It must inspect
all vehicles to assure that people did not operate them beyond
the mandated limit.  And it must conduct additional inspections
to assure that vehicle operators did not tamper with emissions
meters.  The cost of such a policy would be prohibitive.

As an alternative to emissions management at the individual
vehicle owner level, government might alter incentives to vehicle
manufacturers.  The government might require manufacturers to
install catalytic convertors to reduce emissions.  And it could
easily monitor their presence before vehicles are sold.  Of
course, that does not completely eliminate monitoring and
enforcement problems.  Not only would the new policy require
frequent exhaust inspection, but the government may also levy
stiff fines on those who disconnect (or tamper with) their
vehicle's catalytic converter.  These inspections and litigations
have costs that the public must pay to maintain the resource
policy.




THE INTERNATIONAL POLICY PROBLEM



Two general strategies are available to N at the international
level.  One involves the creation of an authority that could
mandate and enforce policies in both N and S.  The authority of
this organization would supersede that of both the N and S
governments.  The policy actions available to this organization
would be little different from those previously mentioned.  This
approach would end the hierarchical nature of the agency problem
since this authority would not need to work through the
governments of N and S.  It could presumably implement policies
directly affecting individuals in both states.  The absence of
this organization, and the difficulties faced by international
bodies such as the United Nations, reveals the difficulties of
establishing an authority that presently sovereign states would
accept as a ruling body.

Without an international authority, the problem of the global
commons remains hierarchical.  To change individual behaviors in
S, the government of N needs to work through the government of S.

The government of N does not have the authority to change the
behavior of citizens in S.  So, policy methods appropriate to the
government of N in domestic policy are not available.  We must
find new methods to align incentives.  Decisionmakers must
coordinate the policies within N with policies between N and S. 
They must consider the problems S will face when implementing
such policies at its own citizens' level.  We consider two
aspects of this problem, the first relates to efficiency
concerns, and the second relates to a possible hierarchical
bargaining solution.



Efficiency


In figure 2, we consider the relationship between actions in N
and S.  (Graph will not transfer on internet.  Please write to
the author for a hard copy.  See address at beginning of this
paper.)  Consider panel (a) of figure 2.  Here we show, as in
figure 1, the earth's sustainable assimilative capacity for
greenhouse gases, G*.  The curve Gn shows emissions over time in
N assuming N begins effective domestic policies.  Notice that the
policy path in N is a function of the policy path in S, and vice
versa.  That is:

* if the total production of greenhouse gases (G*) has been
identified as the policy target,

* then it follows, by definition, that Gs*|Gn = G* - Gn,

* where Gs* is the policy target level of greenhouse gas
production in S, and

* where G* is fixed because the area of tropical and other
forests remains constant at its current level.

We will assume that an integral part of any policy agreement
between N and S entails maintaining the earth's assimilative
capacity at present levels by preserving existing forests.  We,
therefore, leave it to others to argue the questionable merits of
relying on "backstop" technologies that might replace the earth's
forests.  We also assume for the moment that S responds to the
emissions level of N to meet policy goal G*, leaving aside for
the moment why S might do so.

The distance (G* - Gn) between the two curves in panel (a) of
figure 2 represents the assimilative capacity for greenhouse
gases available for use by individuals in S.  Notice that should
the government of N have a more lax policy regarding greenhouse
gases, then the production path Gn' applies.  The obvious
implication is that total production of greenhouse gases from S
must be significantly less under Gn' than would be possible in N.

That is, total "allowable"greenhouse gas production from S is now
given by:
  Gs*|Gn' = G* - Gn'
  where Gn' > Gn

Panel (b) of figure 2 shows these two situations as Gs*|Gn and as
Gs*|Gn'.  These two functions suggest a policy domain for the two
governments.  If we think of:

* the function Gn' as depicting the most lax feasible policy
outcome in N,

* and the function Gn as depicting the most severe yet still
feasible policy outcome in N,

* then these two outcomes represent, in essence, the range of
bargaining open to the government of N in its discussions with
the government of S.

By similar logic, the government of S will regard the two
functions Gs*|Gn and Gs*|Gn' as defining its own bargaining
domain, given the two constraints that:

* the total area of tropical and other forests remains at its
current level, and

* S responds to N to meet the G* policy.

Another way to regard this bargaining domain between Gs*|Gn and
Gs*|Gn' is to consider the difference in costs that will fall
upon individuals in S if the governments of both N and S try to
set a resource management policy in S as a function of Gn' rather
than of Gn.  We show that extra cost in panel (c) of figure 2. 
We may expect that this amount represents the minimum willingness
to accept compensation on the part of the government of S for
adopting a sustainable resource management policy regarding
greenhouse gases.  That is, if the government of S is to preserve
its tropical forests and require its citizens to reduce
greenhouse gas production, it will insist that the government of
N recognize the costs of pursuing a more stringent policy that
fits N's domestic policy.

In panel (c) of figure 2 C(Gs*|Gn), the lower curve, represents
baseline costs within S of maintaining the tropical forests in
their rather pristine condition to process greenhouse gases.  We
base this lower-cost view on the assumption that a Gn emissions
policy exists in N.  Then S must absorb its forest preservation
costs but can pursue a lax greenhouse gas emission policy.  If,
however, N pursues a less stringent emissions policy of Gn', then
S must absorb higher emissions control costs to meet the overall
emissions policy of G*.  The upper curve in panel (c) of figure 2
represents higher costs in S.  The difference in costs is the
extra amount S is likely to demand from N.  This would happen if
the government of N fails to impose upon its own citizens a
strict emission reduction policy.  These higher costs in S result
from N's failure to do all it can to solve the greenhouse gas
problem.

Figure 3 shows the bargaining domain for the governments of both
N and S.  (Graph will not transfer on internet.  Please write to
the author for a hard copy.  See address at beginning of this
paper.)  In panel (a), we show the present-valued costs of
various reduction levels in the total greenhouse gas production
from S.  As greenhouse gas emissions increase along the
horizontal axis, the present-valued cost of emissions reduction
decreases along the vertical axis.  Figure 3 also shows the fixed
tropical forest maintenance cost to S at a level of Ts.  In panel
(b), we show the same relationship for N though we assume N is
not incurring its own forest maintenance costs.  We show the
sustainable emissions policy G* for both states.  It is a
function of the forest cover Ts.  Thus, if we alter the level of
forest cover in S, then the positioning of G* also would change.

The relationship between figures 2 and 3 is critical for policy
formulation between N and S.  We can see this by placing panels
(a) and (b) of figure 3 together.  Invert panel (a) and place it
on top of panel (b) so that the cost curves are tangent.  The
horizontal dimension then is equal to policy limit G*.  Along the
lower abscissa (NG*), the far right point (G*) shows the
situation if N could use all of the greenhouse gas assimilative
capacity.  Individuals in S would not have any access to that
assimilative capacity.  Similarly, along the upper abscissa (SG*)
we see the situation again at G* if S could use the entire
atmospheric assimilative capacity.  This assumes N will not emit
any greenhouse gas.  The left and right ordinates show,
respectively, the present-valued costs to N and S from reductions
in their citizens' greenhouse gas production.

We can next consider the two levels of greenhouse gas production
within N from before.  We continue to assume that S will maintain
its forests at a level Ts as in panel (a), now represented by the
dashed horizontal line in panel (c).  We also continue assuming
that S will respond to emissions from N to meet policy goal G*. 
Thus emissions from S are determined by emissions from N.  If N
pursues a rather lax greenhouse gas policy, earlier referred to
as Gn', it will be necessary for S to pursue a very restrictive
policy of Gs*|Gn' with implied costs for S of C(Gs*|Gn').  This
domestic policy is very expensive for S (reading down the right
ordinate) but cheap for N (reading up the left ordinate) since
N's costs will be only C(Gn').  Alternatively, if N pursues a
more restrictive domestic policy regime, earlier referred to as
Gn, then N's costs will be much higher, C(Gn), while costs to S
will be considerably less, C(Gs*|Gn).

The cooperative efficient policy to pursue is one in which
marginal (not total) costs of emission reduction are the same
between the two countries.  We define efficiency as that level of
emissions where neither state is willing to pay the amount
required to induce the other to reduce emissions further.  Since
the curves Cn and Cs are the total costs of emission abatement,
we know that their slopes are the marginal cost of greenhouse gas
abatement.  We also know there is a unique point at which the
slopes of Cn and Cs are identical when G* is satisfied (along a
vertical line through the box).  This is also the point of
tangency between Cn and Cs when we align panels (a) and (b) as in
panel (c).  That point yields the efficient level of reduced
greenhouse gas production for the two states.  For N the
efficient production level (or, conversely, abatement) is at Gn."

Hence the efficient level for emission reductions in S is at
Gs*|Gn."  Here C(Gn") shows the costs for N, while C(Gs*|Gn")
shows the costs to S.  There is no other possible cross-state
production and abatement distribution that will cost less.  Of
course "efficiency" as determined here is silent about each
region's capacity to incur these reduced production costs.



Feasibility


And this brings us back to our hierarchical principal-agent
problem.  We have repeatedly assumed that S will follow a policy
of maintaining its forests at a level Ts and respond to N's
emissions level to satisfy the policy G*.  But it may not be in
the interest of S to do so unless N takes further action.

We assume that N is the principal in seeking a new resource
management policy more to its liking and that S is the agent. 
But of course the government of S is also the principal regarding
its own citizens.  The early resource management policy, in which
N must persuade the citizens and government of S to incur costs
of C(Gs|Gn"), requires some sustainable agreement.  N must be
able to count on S ensuring that its new environmental policies
both sustain the tropical forests and persuade or require its
citizens to reduce their total greenhouse gas production.

As noted earlier, unlike domestic policy for N (or S),
international policy has no authority system that can force the
government of S to abide by the interests of the government of N.

But, the two governments have mutual interests.  The problem here
is to explore their nature and extent.  Let us assume that the
government of S has scant interest in preserving the tropical
biomass.  Preserving such forests may not allow the government to
earn large amounts of foreign exchange.  Biomass preservation
also may force the government of S to undertake other economic
development policies to address the problems of landless peasants
wanting new land.

Indeed, preserving tropical forests may require the government of
S to seize large estates of wealthy ranchers and redistribute
them to the landless.  The tropical frontier provides a "safety
valve" allowing the government of S to offer land to the poor
without confronting the landed gentry.

Preserving the present tropical forest cover is largely a goal of
the principal not the agent.  The problem is that the forest of
interest to the principal (N) lies within the sovereign territory
of the agent (S).  We must remember that this "problem" would not
be so pressing if the principal had not already cut most of its
own domestic forests.  We have assumed there is no motivation
within S to preserve the forest at the level Ts specified in
figure 3 panels (a) and (c).  Figure 3 panel (c) is the same as
figure 4 panel (a), with the addition of a new, lower level
tropical forest cover, Ts-, that S must maintain.  (Graph will
not transfer on internet.  Please write to the author for a hard
copy.  See address at beginning of this paper.)  The government
of N could persuade S to preserve its forests at the higher level
TS by paying C(Ts) - C(Ts-).  Further payment of C(Gs*|Gn") -
C(Ts) also would be necessary to persuade S to reduce emissions
to the efficient level.  We are assuming S otherwise would be
unwilling to take emissions-reduction measures.  Finally, N would
acquire costs C(Gn") to persuade or compel its own citizens to
reduce emissions to the efficient sustainable level.

Figure 4 panel (b) shows the outcome if N does not adopt a policy
of payments to S.  Without payments, S permits deforestation
until Ts-.  Because of the reduced forest cover, emissions can
only be sustained at level G*-, which results in a reduced domain
of choice.  To achieve its goal of emission sustainability
without paying S, N must now incur costs C(Gn-) to persuade its
own citizens to reduce emissions.  Note that if C(Gn-) is greater
than C(Gs*|Gn") - C(Ts-) + C(Gn"), then N must persuade S to
preserve its forests.  

We could have as easily drawn panel (b) so the cost curve for S
intersected the C(Ts-) line outside the box.  Without payments
from N, S alone would emit more than the sustainable greenhouse
gas quantity.  This latter situation presumably would have much
harsher consequences for N.  Perhaps N would need to begin a
reforestation policy to compensate.

The government of N need not rely exclusively on inducement
policies toward S, however.  Figure 4 panel (c) shows a punitive
policy, perhaps including trade sanctions, imposed by N if S
refuses to preserve its forests without inducements.  Assuming
the trade volume with S is inconsequential to N, but important to
S, the panel suggests high costs of C(Ps) to S from the sanctions
but only the smaller costs of C(Pn) to N.  Note the assumption
implicit in panel (c) that S would preserve even less forest once
sanctions are imposed.  This would reduce further the dimensions
of the policy arena.

We have abstracted from many additional costs that might fall
upon N from this further reduction in tropical forest cover. 
They would include the natural forest value to citizens of N or
any eventual pharmaceutical value from destroyed species.  These
additional costs weaken the credibility of N's threat of punitive
sanctions.  The costs of this sanction policy to N (if imposed)
are at least C(Pn) + C(Gn|P), assuming that S will not cooperate
with N while sanctions are in force.  

The hope of N, of course, must be that its threat is credible.  S
would then find it helpful to avoid punitive sanctions by
preserving its forests voluntarily, while N would incur little if
any cost.  If N may incur great costs in imposing sanctions on S,
S will probably not believe the threat.

Figure 3 suggests a policy that seems the most palatable.  It
preserves the forest and entails a politically acceptable
transfer of funds from wealthy N to the less-wealthy S.  This
path preserves the earth's diminishing forests at present levels.

It also takes into account a wide variety of benefits--intrinsic,
pharmaceutical, and others--that all the earth's citizens receive
when we maintain those forests.  It seems reasonable that the
government of S might want increased economic aid to promote
economic opportunities for its landless poor.  That is, foreign
aid may be useful in making the difficult choice between seizing
haciendas and savaging the forest.  Similarly, if preserving the
forests implies confronting the powerful timber concessionaires,
it is possible that S could use payments from N to redirect these
contractors into other work.

It may be that the area of mutual interests is too restricted to
accomplish what the principal (N) seeks.  Perhaps political
pressure on the government of S to continue its timber
concessions is too severe for payments (or policy concessions)
from N to overcome.  If the exported (S) timber is imported into
N, the solution is straightforward.  The government of N could
decide to ban exports from S.  This would be consistent with the
situation proposed in figure 4 panel (c).  If, however, S exports
timber to a third country, the government of N will need to
negotiate with that third region.  Such punitive policies are the
last resort because they create "winners" and "losers."  The
essence of international policy is seeking results that allow all
governments to view their new positions as "winners."




CONCLUSIONS



Agency theory provides a set of organizing concepts to define the
nature of the policy problem in the global commons.  We see the
important difference between domestic policy in the two states. 
We also see also how the domestic component of the two states
informs and defines the international sphere.  The interests of
the principal drive the bargaining process.  Yet, we see that the
domain for new rules on emissions really depends on the two
political entities (N and S) and their individual citizens'
interests and importance.  It is a mistake to imagine that
governments simply order their citizens around.  Indeed, recent
failure of the USSR's economy suggests that unless governments
get the incentives right, individual choices of the citizens can
be counterproductive.

The policy problem in the global commons is realigning interests
by redesigning domestic and international institutional
arrangements that result in new behaviors.  It is not a matter of
directing individual decisionmakers.  Such a policy requires huge
monitoring costs.  The powerful incentives for defecting from the
newly-imposed policy make these monitoring costs necessary.

Rather, the ideal policy is one that uses inducements to change
the individual choices of millions of people.  When people change
their behaviors in producing and processing greenhouse gas, we
will solve the problem of the global commons.  There is no magic
wand that will produce a new international policy.  Currently
there is no all-powerful supra-state to impose a new
institutional policy that will affect the daily lives of millions
who farm or work in the South's emerging industrial centers or
those who live in  the already industrialized North's cities. 
There is only the long process of changing incentives to alter
individuals' choices.

In the authoritative work on international policies, Young talks
of three possible origins of policy formation - spontaneous,
negotiated, and imposed (Young, 1989).  This discussion has
focused on the second and third origins.  It is possible that the
governments of N and S both will come to realize the seriousness
of greenhouse gases and begin to change their domestic policies
accordingly.  This would represent the spontaneous case.  That
case is not very interesting for the problems of the global
commons.  It suggests that governments will immediately recognize
the problem of greenhouse gases and independently solve it. 
While this would be a wonderful situation, it is also not very
realistic.  Independent states have different interests in how
their citizens' behaviors relate to the natural environment and,
hence, the global commons.  So the second two sources--
negotiation and imposition--are much more likely.  These two
classes of international policy origins have included the nature
of the approach followed here.

Young also points to the obvious result that negotiated and
imposed regimes usually avoid the temptations offered by changing
circumstances.  Spontaneous policies are only as durable as the
separate calculations of the interests of the respective parties.

In that sense, this approach, based on agency theory, offers a
conceptual guide to solve the problems of the global commons.




REFERENCES



Archer, R. and R. Ichord.  1989.  "Global Warming and Climate
Change: Role of the Asia Near East Region."  U.S. Agency for
International Development: Washington, D.C.

Bromley, Daniel W.  1991.  ENVIRONMENT AND ECONOMY: PROPERTY
RIGHTS AND PUBLIC POLICY.  Oxford, United Kingdom: Basil
Blackwell.

-----.  1989.  ECONOMIC INTERESTS AND INSTITUTIONS: THE
CONCEPTUAL FOUNDATIONS OF PUBLIC POLICY. Oxford, United Kingdom:
Basil Blackwell.

Bromley, Daniel W. and Jeffrey A. Cochrane.  1993.  UNDERSTANDING
THE GLOBAL COMMONS.  U.S. Agency for International Development,
Environmental and Natural Resources Policy and Training
Project/Midwest Universities Consortium for International
Activities Working Paper 13.  Arlington, Virginia.

Nitze, William A.  1990.  "A Proposed Structure for an
International Conventions on Climate Change."  SCIENCE 249(August
10, 1990):607-08.

Young, Oran R.  1989.  INTERNATIONAL COOPERATION: BUILDING
REGIMES FOR NATURAL RESOURCES AND THE ENVIRONMENT.  Ithaca, New
York: Cornell University Press.