You are on page 1of 18

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/5147231

Externalities - A Market Model Failure

Article in Environmental and Resource Economics · February 1997


DOI: 10.1023/A:1026438001650 · Source: RePEc

CITATIONS READS
75 1,965

2 authors:

Arild Vatn Daniel Bromley


Norwegian University of Life Sciences (NMBU) University of Wisconsin–Madison
125 PUBLICATIONS 6,463 CITATIONS 155 PUBLICATIONS 5,222 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Arild Vatn on 07 September 2015.

The user has requested enhancement of the downloaded file.


Environmental and Resource Economics 9: 135–151, 1997. 135
c 1997 Kluwer Academic Publishers. Printed in the Netherlands.

Externalities – A Market Model Failure

ARILD VATN1 and DANIEL W. BROMLEY2


1
Department of Economics and Social Sciences, Agricultural University of Norway, Box 5033, 1432
Aas, Norway; 2 Department of Agricultural Economics, University of Wisconsin-Madison, Madison,
Wisconsin 53706, USA

Accepted 1 May 1996

Abstract. We focus here on a set of conceptual problems related to the accepted theory of externalities.
We are primarily concerned with difficulties that arise when a theoretical system is extended beyond
its logical domain. This is the practice in externality theory when the market model assuming
independent agents is used to analyze physical interdependency. The different kinds of dependencies
obscure the standard use of the Paretian analysis, as the issues of rights and efficiency are mixed
up. The creation of emissions and the creation of externalities are further not held apart producing
flows in the efficiency evaluations. Due to the interdependencies involved, actions of both emitter
and victim must be taken into account while searching for efficient policies. Finally, we analyze the
interrelationships between what is termed the internal structure of the market model and the annexed
sphere of externalities. We conclude that the accepted policy prescriptions both assume and demand
no interrelationships between these two spheres. We find the assumption unrealistic and inconsistent
as concerns the basic foundations of the market model. There are two main traditions addressing
externalities – the Coasean and the Pigovian. This paper shows that both are vulnerable to the above
critique. Thus the presumption that externality theory is now settled and coherent is seen to be without
theoretical support.

Key words: externalities, market model, rights, Pigovian taxes, Coase Theorem, social costs

1. Introduction

The unfortunate label ‘market failure’ has long been associated with externalities –
probably owing to the seminal paper by Francis Bator (1958). However, externali-
ties do not represent market failure. Given the market, the presence of externalities
can be interpreted as a rational result and thus cannot properly be called a ‘failure’
of the market. Dahlman (1979) claims that transaction costs are the sole source
of externalities, and Randall (1983) clarifies: ‘: : : it has long been clear that the
non-existence of certain markets is a rational market response to transaction costs
in excess of potential gains from trade’ (p. 137).
However, there is a problem of circularity embedded in the above observations
with important implications for efficiency evaluations of market outcomes. The
form and magnitude of transaction costs are a function of the institutional setting
– be it markets or other forms of transaction arrangements. Thus the amount of
what becomes ‘external’ simply reflects the structure of the ‘internal’ and cannot
be analyzed independent of that very same structure.

VICTORY: PIPS No.: 113316 MATHKAP


eare545.tex; 15/04/1997; 14:25; v.5; p.1
136 ARILD VATN AND DANIEL W. BROMLEY

Basically, the standard analysis of the efficiency of markets with external


(dis-)economies attached to them is formulated as a two-stage operation. First
economist show that markets, under certain conditions, fulfill the necessary and
sufficient conditions for efficiency according to the theorems of welfare economics.
The list of conditions states that goods must be costlessly demarcated objects with
clearly defined property rights attached to them, and they are costlessly exchanged
between independent and rational agents. Second, economists observe that not all
goods (or bads) conform to the given claims. Actions of one or more economic
agents may give rise to uncompensated physical and/or real economic effects for
others.1 A sphere of externalities is attached to the model to cover such inter-
dependencies and rules for internalizing these external effects are formulated to
reestablish optimality.
Such a two-stage analysis demands, however, no interrelationships between
what is defined internal, i.e. what is the internal structure of the system, and what
becomes regarded as external. The above observation about the role and character
of transaction costs is sufficient to show that this does not hold. It is therefore not
surprising that the externality debate has notoriously been hampered with disputes
over definitions and consistency since the concept first materialized in Marshall’s
work approximately 100 years ago (Marshall 1890; Papandreou 1994).
In contemporary externality theory, there are two major camps: the Pigovian
and the Coasean. Relating to the Pigovian stand and its tax rules, the above problem
arises in three interdependent ways:
First, the logic of the Paretian analysis on which the Pigovian conclusions rest
breaks down simply because it demands a priori and exclusively defined rights
to goods. This is, by definition, not the case for externalities. Hence, there are
persistent inconsistencies in the literature.
Second, to the extent that the problem is defined as one of physical interde-
pendency between a ‘polluter’ and a ‘victim,’ it is impossible to technically
isolate the action of one from the action of the other. Standard theory, how-
ever, is built on methodological individualism. Externality theory extends
this primary structural feature – independence – into a domain whose basic
attribute is interdependency, mixing up the relation between causation, rights
and efficiency.
Third, the standard Pigovian analysis disregards the fact that what is considered
necessary to make the market efficient – atomization and rational choice – will,
by definition, be a source of externalities. Again the model of the ‘internal’ fails
as a suitable metaphor for the dynamics of the ‘external’ and the interaction
of the two.
The Coasean position is vulnerable to the same type of critique. Still, its attack
on the Pigovian stand might be regarded as an effort to overcome some of the
above inconsistencies. By focusing exclusively on negotiating agents – thus claim-
ing that everything is or could become a market – Coase (1960) preserved an
apparent consistency with standard economic analysis, but at the steep price of lost

eare545.tex; 15/04/1997; 14:25; v.5; p.2


EXTERNALITIES – A MARKET MODEL FAILURE 137

coherence. By ignoring transaction costs, the propped-up post-Coasean externality


model became incapable of generating any class of Pareto-relevant externalities
(Buchanan and Stubblebine 1962; Dahlman 1979).
Still, the Coasean critique of the standard Pigovian tax solution is helpful in
detecting flaws in traditional thinking about externalities. We will thus utilize the
venerable Pigou–Coase controversy to elaborate and clarify the above arguments.

2. Externalities and Pareto Optimality


A fundamental and confusing observation is that both Coaseans and Pigovians use
Pareto optimality (or Pareto improvement) as the basis for their quite different
policy recommendations. It is clear that externalities arise when the actions of
one or more economic agents give rise to uncompensated physical/real economic
implications for others. Difficulties arise when we start searching for mechanisms
to determine the efficient level of intrusion into this setting.
It is important to recognize that even though resources used in the creation
of externalities may be owned, externalities appear outside the sphere of defined
property rights. In a system based upon private property, externalities appear largely
due to high costs of demarcation. Further, externalities are basically novelties.
They will mostly be recognized after they have been produced. Both elements
relate to the fact that externalities are – to a large extent – the result of interactions
between human activities and the integrated physical and biological processes of
the environment (Vatn and Bromley 1994). This invariably influences demarcation
costs as it often produces large time spans between when a physical act (e.g.
emission) takes place, and one becomes aware of the external effects it creates
(Perrings 1987). This inevitably means that fundamental questions about rights
and duties must be determined ex post. This reactive imperative creates at least
two serious problems for the economist. First, we have the problem of defining –
or redefining – what it means to talk of ‘efficient resource allocation.’ The second
problem concerns the design of efficient resource allocation mechanisms.
In a world where transaction costs govern resource allocation decisions, the
distribution of rights and duties is of paramount importance in determining what
shall be regarded as efficient resource allocation – i.e. an efficient level of external
economies and diseconomies. What is thought to be efficient cannot be defined
without a prior judgment about which party to a conflict has (or ought to have) the
protection afforded by a right and its correlated duty (Bromley 1991). The problem
of circularity relates to the fact that standard externality theory draws conclusions
about what is an efficient rights structures on the basis of reasoning that actually
presupposes this structure as given.
We know very well that the Pigovian economist would ‘solve’ the externality
problem by making the emitter – the source of the physical dimension – liable for
economic damages (harm). But how is it possible to give the Pareto-improvement
rule a retroactive force without destroying its basic feature? The distinction between

eare545.tex; 15/04/1997; 14:25; v.5; p.3


138 ARILD VATN AND DANIEL W. BROMLEY

ex ante and ex post definition is crucial. The character of the problem is such that
the emitter may not know – or will not be made liable for creating any harm – until
some considerable time after the activity is undertaken. And given the situation as it
then exists, it may be quite impossible to achieve a traditional Pareto improvement.
That is, it is almost certain that someone in such circumstance must lose, while
others will gain.
We find this dilemma in the classic work of Baumol and Oates (1988). After
discussing the properties of a Pigovian tax for solving an externality, they emphasize
that:
The price-tax conditions necessary to sustain the Pareto optimality of a compet-
itive market solution under the assumed convexity conditions are tantamount
to standard Pigovian rules, with neither taxes imposed upon, nor compensation
paid to, the victims of externalities (Baumol and Oates 1988, p. 45).
However, in their discussion of transboundary externalities, Baumol and Oates
lament the absence of a world government and the obvious problem that one
cannot expect a country emitting pollution to reduce it to some optimal level. In
considering this problem, they conclude:
We cannot simply rely on a program of pollution abatement in country A
[the polluting country] for this would impose costs on A with no offsetting
benefits to the polluting country. The OECD’s Polluter-Pays-Principle is thus
inconsistent with our insistence on a Pareto improvement. Mutual gains to the
countries necessarily require the victim country B to make some payments to
A (Baumol and Oates 1988, pp. 280–281).
If a Pigovian tax is necessary and sufficient to solve polluter-victim problems
efficiently within a country, why is it not good enough for transboundary pollution
and its victims? From the perspective of a Pareto improvement and physical inter-
actions, there is no substantive difference between two individuals, two firms, or
two nations. After the imposition of a Pigovian tax, it is clear that there will be
gainers and losers as measured against the status quo ante.
We see, therefore, a mixing of measures of optimality with the institution-
ally defined ability to define and enforce a certain rights structure addressing
external effects. The Baumol and Oates prescription for transboundary externali-
ties, a prescription that contradicts their earlier general policy rule, is derived from
the ordinary Pareto rule – not from the Potential Pareto Improvement (PPI) rule. Of
course, in the absence of the enforcement power of the state, international pollution
leaves little scope for alternative policies. Actually, a common authority like the
state – a power at a level above the atomistic agents – is a prerequisite for breaking
the circularity about rights and efficiency previously defined. Not only that the PPI
rule becomes a potential basis for action. As we now see, the PPI rule is the only
Paretian rule that can be used to defend the Pigovian solution with its consequences
for value neutral policy evaluations (Mishan 1980; Griffin 1995). Schmid (1987)
seems to be alone in recognizing this.

eare545.tex; 15/04/1997; 14:25; v.5; p.4


EXTERNALITIES – A MARKET MODEL FAILURE 139

3. Physical Interdependency, Rights and Efficiency


3.1. COASEAN CAUSALITY AND PIGOVIAN PERCEPTIONS
Consider the matter of who causes an externality to come into being. From the
Pigovan perspective this is a simple matter – the emitter causes the externality.
Making the emitter liable and imposing an emissions tax is thought to create
the necessary and sufficient conditions for restoring optimality. Thus there is an
inference from physical causation to efficiency determination. Coase (1960), on
the other hand, claimed that externalities are reciprocal. To Coase there is no reason
to blame one party as opposed to the other. Rather, Coase argued that efficiency
requires determining which party could change behavior most cheaply. On this
view, the responsible party should be the one whose modified situation is cheapest
for society to bear.
We suggest, therefore, that clarity in externality policy is purchased by recog-
nizing that the Coasean position differs from the Pigovian one in the fundamental
matter of the presumptive entitlement of one party versus another. That is, Pigou
saw externalities – and more specifically pollution – as an act undertaken by one
party that caused harm for another. In its simple Pigovian manifestation, factories
emit smoke and so laundries have to wash linens hung out to dry again. Coase
entered the discussion and caused us to ask whether the laundry might be able,
more cheaply, to find another way to dry its linens. Or, perhaps the laundry should
move as to avoid the smoke.2
In defending the efficiency aspects of Pigou’s solution, Baumol (1972) argued
that if it is cheaper for the laundry to move as opposed to suffering the unwanted
costs of smoke, it will certainly move of its own volition. Or, it might be ‘induced’
to move if the factory is liable. Thus there will be no victims as Baumol terms it,
and there is no need to tax the smoke emitter. The correct Pigovian tax would be
zero.
But to render this conclusion meaningful, Baumol has to accept the strange
Coasean world of zero transaction costs. By ignoring such costs, Baumol is forced
to discard the most basic argument for his own position. With zero transaction costs,
the Coase Theorem governs with no room for Pigovian policy. To make confusion
complete, Coase showed that even Pigou (1932) did not apply the ‘Pigovian rule’
consistently (Coase 1960, pp. 34–35).
To some extent, the problem arises because Pigou and his followers do not
seem to acknowledge that there are two processes involved when an externality
is created; the interdependence between the action of the emitter and that of the
receiver has two components. First, there is the emission of smoke. Second, this
smoke is transformed into an externality which happens first when a ‘victim’ is
– or has come – within the realm of the emission. With the case of the ‘moving
victim’ it is easy to see that there is no necessary coincidence between what causes
the emission (in a physical or technical sense) and what causes the externality to
come into being.

eare545.tex; 15/04/1997; 14:25; v.5; p.5


140 ARILD VATN AND DANIEL W. BROMLEY

It is therefore impossible to advance a rule that can generally define even the
quite narrow question of physical causality. What then is the basis for the Pigovian
perception and its alleged value neutral taxing rule? Coase’s arguments imply that
the Pigovian position is based foremost on moral grounds. While Randall (1974)
argues that Coase presumably cast the issue in reciprocal terms so as to purge
environmental policy of the moral dimension implicit in Pigou.
The Pigovian perception and its taxing rule may seem intuitively obvious
because the most common pollution cases are ones in which the emitter was
not there first, but came after established uses were underway. There have always
been laundries hanging out the linen, but smokey factories are the product of the
Industrial Revolution; a period of some significance when Pigou learned his eco-
nomics. But by altering the sequence of things, the apparent clarity disappears.
Coase (1960) offers an illustrative example – Bryant v. Lefever. Coase writes:
The plaintiff and the defendants were occupiers of adjoining houses which
were of about the same height. Before 1876 the plaintiff was able to light a
fire in any room of his house without the chimneys smoking: the two houses
had remained in the same condition some thirty or forty years. In 1876 the
defendants took down their house, and began to rebuild it. They carried up the
wall by the side of the plaintiff’s chimneys much beyond its original height,
and stacked timber on the roof of their house, and thereby caused the plaintiff’s
chimneys to smoke whenever he lighted fires (Coase 1960, p. 11).
In Pigovian terms it seems obvious that the plaintiff had a convincing argument;
the externality was indeed created by the defendant’s action. The plaintiff had done
nothing different from long-accepted practice and therefore could not be thought to
have ‘caused’ the problem. As the case worked its way through the courts, however,
things became somewhat less clear. In the first decision, the plaintiff was awarded
compensation for the defendant’s actions. However, on appeal the judgment was
reversed Coase cites the judge:
They (the defendants) have done nothing in causing the nuisance. Their house
and timber are harmless enough. It is the plaintiff who causes the nuisance
by lighting a coal fire in a place the chimney of which is placed so near the
defendants’ wall, that the smoke does not escape, but comes into the house.
Let the plaintiff cease to light his fire, let him move his chimney, let him carry
it higher, and there would be no nuisance. Who then causes it? It would be
very clear that the plaintiff did if he had built his house or chimney after the
defendants had put up the timber on theirs, and it is really the same though he
did so before the timber was there. But (what is in truth the same answer), if
the defendants cause the nuisance, they have a right to do so. If the plaintiff
has not the right to the passage of air, except subject to the defendants’ right
to build or put timber on their house, then his right is subject to their right and
although nuisance follows from the exercise of their right, they are not liable
(Coase 1960, p. 12).

eare545.tex; 15/04/1997; 14:25; v.5; p.6


EXTERNALITIES – A MARKET MODEL FAILURE 141

We see that the appellate court recognized the importance of rights to air move-
ment. We also see recognition of the priority of one type of rights as against another.
If further seems evident that the two courts placed very different weight on the
issue of the order in which the events occurred. The party being ‘harmed’ is the one
who simply wished to light a fire – an act almost as old as human history. Yet the
hidden Pigovian perception causes the analyst to start looking for the cause of the
‘externality.’ If the owner of the fireplace – the one hurt – ‘causes’ the externality
himself by lighting his own fire, then it would hardly qualify as an externality
in the eyes of received economic doctrine. The appellate judge saw this essential
point. Our point here is not to argue which party had the more compelling case.
Rather, it is to make clear the general impossibility of creating a rule that, from
mere technical logic, offers a consistent basis for determining the correct rights
structure. At bottom, coherent rights can only be established through a process that
starts not with physical acts, but with Kantian reason (Bromley 1991). All rights
are value based because they are willfully determined through collective action.
Interestingly, these arguments trap Coase as well. It was from cases such as
Bryant v. Lefever that Coase drew the conclusion that the only reasonable position
is that the actions of both parties are of singular importance in causing the problem.
Both parties were, in fact, the ‘cause’ of their new conflict. But arguing this way
deprives externality issues of their basic character – the moral issues related to the
intrusions we make into each other’s lives and space.
Let a butcher (B), throw the waste from his activities into the garden of N,
a nearby nursery school. Formally there is no difference to the above case in a
Coasean sense. Both parties – B and N – had to be undertaking their respective
activities in their particular locations for a problem to arise. But it does not seem
obvious that this is a case of reciprocal responsibility given the values involved.
Indeed some will argue that pollution itself is a culturally defined notion (Douglas
1966). Douglas argues that pollution is ‘matter out of place’ and so it should be
obvious that one must address both the substance of the ‘matter’ in dispute, and
what it means to be ‘out of place.’ Thus the endorsement of externalities as a
moral issue implicit in Pigou’s policy rule is appropriate, while the explicit claim
of advancing a value free and objective efficiency rule becomes contradictory.

3.2. EFFICIENCY AND THE PIGOVIAN RULE


Actually, the argument that a Pigovian tax is a least cost measure is incorrect. It
may be morally right to make the emitter liable, but it is not always ‘efficient’ to
put a tax on this side of the conflict. The Coasean allegation that the Pigovian tax
may impose costs on the ‘wrong’ party – that the net social dividend sometimes
would be enhanced if the ‘victim’ would have to take action – bears substantial
merit.
In the Coasean world of bargained solutions to externalities, the only relevant
cost in considering judgments about responsibility for action is the level and

eare545.tex; 15/04/1997; 14:25; v.5; p.7


142 ARILD VATN AND DANIEL W. BROMLEY

incidence of transaction costs. Liability for remedial action must lie with the party
best able to handle change with the minimum of such costs. Transaction costs are
either so high that no change in outcomes is deemed ‘efficient,’ or transaction costs
are low enough – at least for one side – to permit a bargained transaction. In this
case, the cheapest solution will be found through bargaining.
Actually, the above discussion is a generous interpretation of Coase (1960) since
he seems to mix up the reasoning. When making his point, Coase focuses only on
the direct abatement costs and ignores the distribution of transaction costs. But as
Coase turns to the Pigovian world, he offers an important challenge. The standard
Pigovian solution is that it will always be Pareto optimal to tax the emitter at the
level where marginal damage equals marginal abatement costs.
Baumol and Oates (1988) make, as we have seen, the Pigovian position very
clear. Except for the case of transnational pollution, they affirm that all arguments
about the need for additional measures toward the ‘victims’ are not appropriate.
They further clarify their position by noting ‘: : : the damages that victims suffer
from the detrimental externality provide precisely the correct incentive to induce
them to undertake the efficient levels of defensive activities’ (Baumol and Oates
1988, p. 22).
The literature shows unanimity on this matter and so the issue seems settled
(Burrows 1980; Fisher 1990; Hartwick and Oleweiler 1986; Pearce and Turner
1990). Despite the apparent consensus, the analysis is flawed. The problem arises
from the fact that Baumol and Oates fail to consider the full impact of moving the
responsibility between ‘polluter’ and ‘victim.’

3.2.1. The Effect of Moving Liability

Recall that the standard solution is to make the emitter(s) liable through the impo-
sition of a tax on emissions. However, there may be further defensive expenditures
by recipients even after abatement is undertaken by the emitter(s). If such defensive
expenditures are undertaken, it suggests that it might have been cheaper to let the
recipients carry out more (or all) of the remedial action under standard convexity
assumptions. This is not recognized in the Baumol and Oates analysis. Consider
Figure 1.
We depict pollution and costs along the axes. Curve I shows the marginal
damage costs of pollution, with OC as the current pollution level. Curve II shows
the marginal cost of abatement when undertaken by the emitter(s) only. Curve III
shows the marginal cost of defensive action that might similarly be undertaken
exclusively by the recipients of pollution. Marginal costs are assumed to be strictly
increasing. To simplify the exposition, we assume fixed abatement costs to be
equal in the two cases. Following standard Pigovian reasoning, a tax TO should be
imposed. Net social gain would be CEG with pollution reduced to OB. Given this
outcome, recipients will find it preferable to defend themselves. If we assume that

eare545.tex; 15/04/1997; 14:25; v.5; p.8


EXTERNALITIES – A MARKET MODEL FAILURE 143

Figure 1. Negative externality and cost relationships under different liability rules.

marginal costs are related only to the level of pollution confronting the recipient,
then pollution would be reduced further to OA, and the extra gain will be FDE.
But if it is economically sound for victims to take any defensive action, it is
most probably cheaper to make them (the victims) liable. With cost curves as in
Figure 1, victim’s responsibility creates the same reduction in effective pollution
(AC) with the area CFE comprising an extra net gain compared with the traditional
Pigovian solution. Victim’s liability may not seem fair, but it is ceteris paribus the
cheapest.
An immediate challenge to this approach is to claim that curve III will usually
lie above curve II, making Pigovian taxes (or other emitter oriented measures) the
cheapest solution. But this particular circumstance does not establish the universal
case in favor of polluter responsibility being cheapest.
Another argument relates to the sequence of action. If the victim acts before
a tax is imposed, the least costly solution in the given setting may be reached.
But if the Pigovian rule is generally accepted, victims will be motivated to await
state (collective) action. Some cases may be such that the costs of waiting are high
enough to bring forth (some) victims’ action. But again, this instance does not save
the Pigovian solution in general.
The policy reversal observed here arises from incompleteness in traditional
analyses. There are actually two different remedial strategies of relevance. Notice,
as above, that the defined responsibility determines which abatement strategy
appears to be efficient. For each institutional setting – or for each presumed rights
regime – there is an ‘efficient’ remediation strategy. This policy reversal will occur
when transaction costs are high enough to prevent bargaining between the parties.
That must be the case if the Pigovian rule is to be of interest. In his search for
efficiency, the Pigovian planner must bear the responsibility of comparing the two
settings in their totality.

eare545.tex; 15/04/1997; 14:25; v.5; p.9


144 ARILD VATN AND DANIEL W. BROMLEY

3.2.2. The Effect of Moving Victims

The Pigovian claim is that taxing the emitter gives the correct static as well as
dynamic results. It has long been clear that if the recipient must face all of the
remediation costs, ‘too many’ polluters will persist, thereby creating long-run inef-
ficiency. But given a clear understanding of the interdependency between the parties
– the reciprocal nature of externalities as physically defined – there is a similar
argument about what may happen on the side of the recipients. Certainly Coase
makes a relevant point when he recognizes that making the emitter responsible
may induce ‘too many’ victims to move into the area.
Consider a locality with a polluting firm and a set of residences living nearby.
Let us further think of another non-polluting firm that considers moving in. There
are two possible rules of rights: Rule R would change the tax level as soon as the
firm moves in to restore equivalence between the new marginal environmental cost
curve and the tax. Rule S would leave it unchanged. Unfortunately, neither rule
will – in general – induce a social optimum. This is most easily shown of the case
of a non-depletable externality (a general public ‘bad’). Consider Figure 2.
Curve I shows the marginal pollution costs from the status quo number of
emitters and victims in an area. Curve III depicts the marginal abatement costs.
The optimal tax is OT1 with pollution OF. Let the policy rule be R. Suppose our
non-polluting firm finds it profitable to move in on these grounds – the marginal
costs of pollution shift to curve II. The ex post optimal tax would then be OT2 . If
we look at the environmental costs given that the firm has already moved, this tax
increase induces a net gain equal to CDE consisting of an increased gain of the
old residents amounting to ABEF, a gain for the moving firm of BCDE, and an
increased abatement cost of ACEF.
But added to that we need to look at the other costs and gains of moving –
illustrated by the situation created if the tax was not changed (rule S). Let us
assume that the firm wants to move because this would lower transportation costs,
but that this gain is counteracted by increased costs due to higher pollution as
compared to the previous location. Let us say that the net relocation costs for the
firm with pollution at OF an positive. If they are less than CDE, it is still socially
optimal to motivate the firm to move. To create this we simply invoke rule R. On
the other hand, it might be that the relocation costs are greater than CDE (but less
than BCDE to make it a relevant case). If so, then rule S is appropriate.
Baumol and Oates (1988), as Mäler (1974) before them, consider this a case
of pecuniary externality and therefore irrelevant for policy. Baumol and Oates are
quite explicit in their treatment. First, they define a technological externality to be
a physical interdependency generating shifts in some production function, while
a pecuniary externality only affects financial circumstances. Second, they accept
that moving victims will create an increase in the tax level – implying that they
interpret R to conform with the Pigovian policy. They emphasize however that the
increased tax only creates pecuniary effects:

eare545.tex; 15/04/1997; 14:25; v.5; p.10


EXTERNALITIES – A MARKET MODEL FAILURE 145

Figure 2. External cost relationships with moving ‘victims’.

An increase in the laundry activity that increases the tax rate is precisely
analogous to an increase in shoe production that increases the cost of leather
to handbag manufacturers. In each case, a resource (in one case, leather, in the
other, clean air) has become more valuable and the price of the resource has
increased commensurately, as proper resource allocation requires (Baumol and
Oates 1988, p. 32).

In our mind the two cases are clearly different. The leather case reflects the
price effect of some occurring scarcity, which should not in any case be viewed
as an externality since this is the way markets are supposed to work as long as
rights are clearly defined. The clean air example on the other hand is about how
to define rights and measure scarcity in a case where physical interdependencies
are involved. From our example we see that the rule applied – R or S – defines
which costs enter the agents’ calculations and which resource allocation is thereby
induced. The question is thus not about a change in the scarcity of air. It is about
which costs the various agents are to face in a situation where (1) the action of one
influences the value of the resource for the other, and (2) the rights system defines
whether this is to be considered a cost – and for whom.
Again the misinterpretation occurs as an interdependency is treated as if it
consisted of only independent actions. For an optimal tax policy to be put into
practice, both the action of the polluting firm and all moving firms (even potentially
moving firms) must be taken into consideration. Both activities influence the net
social product and it would seem efficient to change the rule from case to case. This
is certainly not feasible, thus demonstrating the limits for choosing institutions on
efficiency grounds.

eare545.tex; 15/04/1997; 14:25; v.5; p.11


146 ARILD VATN AND DANIEL W. BROMLEY

4. Externalities as Systemic Phenomena


From the above we see that the institutional setting defines what is to be con-
sidered internal and what becomes external. The policy prescriptions that follow
from existing theory, however, demand no interdependence between the way the
economy is organized and the occurrence of externalities. If this does not hold, the
quality of standard policy advice diminishes, as we have seen.
Moving from the level of the individual to the level of the economic system,
this problem materializes in a lack of a coherent theory concerning relationships
between occurrence of externalities and the structure and dynamics of the economy.
As the market model is constructed, consistency demands some systemic relation
to exist between the internal and the external.

4.1. EXTERNALITIES AND MARKET STRUCTURE


First we mention the inconsistency related to assumptions about the structure
of the market. The accepted economic model demands individualization of con-
trol over resources. This then requires individualized private property rights. But
the imperative to divide control over resources among atomistic agents is, at the
same time, the mechanism responsible for creating some of the limitations of that
very same model. Through atomization, the number of borders among economic
agents increases, thereby amplifying transaction costs and hence contributing to
the generation of externalities:
The individualization of the world – its atomization really – is argued to be the
very best means of individuals to be made better off and, by simple aggregation,
for the collection of all individuals (call it society) to be better off. Now, if
externalities arise at the boundary of decision units, and if theory and policy
celebrate and sanctify atomization, then theory and policy would seem to
advocate the maximization of decision units and, ipso facto, the number of
boundaries across which costs might travel. Bluntly put, atomization ensures
potential externalities (Bromley 1991, p. 60).
The Pigovian answer – state-sponsored taxes on emitters – certainly reduces
transaction costs, thereby making possible some progress toward solving the prob-
lem. However, this two-stage approach, with the state as an exogenous force,
ignores important parts of the efficiency issue. This route is still less contradictory
than the Coasean version that claims that markets are necessary and sufficient to
foster all-around efficiency. Coase and his more ardent followers then undermine
this overly large claim by insisting that many externalities are Pareto irrelevant
due to the deus ex machina of high transaction costs. That is, transaction costs
demarcate Pareto-relevant from Pareto-irrelevant externalities. But since transac-
tion costs are a function of how the economy is organized (that is, its institutional
setup), the issue of efficiency is caught up in a severe circularity.

eare545.tex; 15/04/1997; 14:25; v.5; p.12


EXTERNALITIES – A MARKET MODEL FAILURE 147

4.2 EXTERNALITIES AND INDIVIDUAL RATIONALITY


In standard economic theory, externalities are regarded as unintended by-products
of economic activity. Mishan (1971) and Baumol and Oates (1988) are explicit
about this.3 Unfortunately, the presumption of externalities as being incidental and
unintended strips the problem of its essential economic character. To be blunt,
behavior that is unintended finds no role in the currently accepted world of rational
choice among self-interested maximizing agents. In economics as it is currently
taught and practiced, there can be no ‘unintended’ action by rational agents.
While it is not necessary to allege that all externalities are the clear result
of intended cost shifting, it is clearly incoherent and incorrect to assert that the
opposite is universally so. Certainly, if costs can be shifted – and done so without
violating any previously established and enforceable rights – then it will happen
under conventional assumptions about the objectives and motivation of the relevant
economic agents. Kapp writes:
Hence, a system of decision-making operating in accordance with the principle
of investment for profit cannot be expected to proceed in any way other than
by trying to reduce its costs whenever possible and by ignoring those losses
that can be shifted to third persons or to society at large (Kapp 1971, p. xiii).
While cost shifting is about getting rid of costs, it need not be driven by the desire
to hurt others. While the moral dimension of hurting others is somewhat clear,
the moral dimension of shifting costs – especially when doing so is not clearly
unlawful – is more ambiguous. Indeed, shifting costs in a permissive rights regime
can be equated with good business practices. Who, after all, wishes to pay more
than is required by law? This is especially the case when one’s stockholders are
sufficiently alert. And of course once one agent in a competitive economy starts to
shift costs, all competitors would be foolish not to follow suit.
Given the assumptions about behavior and rationality embedded in the conven-
tional model of the market and the economy, externalities will almost certainly
increase over time. Indeed, one will need to be on constant lookout for the creation
of new cost-shifting possibilities as competition intensifies. While the ‘invisible
hand’ clearly conduces to efficiency within the ‘internal’ economy, it must – with
the same ruthless logic – conduce to the shifting of costs throughout the ‘external’
sphere. This is not an instance of market failure, but of model failure.
From this recognition, one must inevitably draw an essential policy conclusion
about externalities. Unfortunately, the Coasean prescription, so well understood
in the conventional wisdom, misses the point we are raising here. Moreover, the
Pigovian prescription will find itself chasing a moving target. From the perspective
advanced in this section, taxing regimes are destined to address symptoms rather
than the fundamental problem that creates the externality in the first instance.
One may protest that the standard policy prescription is the best that economists
can hope to accomplish – to block or reduce some of the many possible ways
in which cost shifting can be successfully undertaken. On the other hand, since

eare545.tex; 15/04/1997; 14:25; v.5; p.13


148 ARILD VATN AND DANIEL W. BROMLEY

the basic problem relates to the double-sidedness of individual rationality, our


analysis points in directions other than those we as economists normally follow.
Observing the above inconsistencies and the limited role tax systems may play,
issues such as moral commitment, collective standards, social norms, and network
processes may attain a higher position in the understanding of externality policy.
Policy ‘instruments’ like moral suasion and individual modesty find scant place in
a model of individually optimizing agents. They are, however, observed especially
in areas of society where quality of life and personal integrity is at stake (Etzioni
1988; Sen 1977; Vatn and Bromley 1994).

4.3. EXTERNALITIES AND DYNAMIC MARKET PROCESSES


Finally, we address the matter of structural dynamics. Our proposition here is that
one must understand the strong relationship between the competitive dynamics
of a market economy and the evolving domain of externalities that goes beyond
the one of intended cost shifting. But in making this case, neither the reasoning
– nor the conclusion – are as straightforward as seen previously. Recall that the
willingness and the ability to develop new production methods is a desirable
outcome especially correlated with the normal incentives and structures of a market
economy. As a consequence, one sees a rapid growth in the development of new
chemical compounds (and other products) that may constitute a non-trivial threat
to the environment. In essence, we see here systemic externalities that are linked
to the size and dynamic aspects of the general economy.
A growing economy – with its increased throughput – induces externalities as
an effect of its very growth. To the extent that the magnitude of this growth in
material use is a systems feature, it follows the physical laws of nature. This is
certainly acknowledged in the literature – especially with relation to the effect of
different trade regimes on the rate of growth. But its importance for the externality
problem seems to have been overlooked.
The relations here are neither simple nor unidirectional. Growth that is induced
through competitive pressures does not necessarily result in increased throughput.
For resources that are priced, there are incentives that also will lead to reduced
resource use and thus less externalities of the kind proportional to throughput.
On the other hand, there are no mechanisms in the standard economic model that
will protect against the negative effects of growth on the free services of nature
– other than ex post charges (the Pigovian tax). Unfortunately, such charges will
have little effect on the ex ante choices of alternative technological trajectories
and investments. In this setting, what is regarded as ‘internally competitive’ – that
which yields the lowest expected market costs – will drive choice.
From the above discussion we see that the more expansive (extensive) the market
is, the more necessary it may be to ‘regulate’ it. The more stable its institutions – thus
securing inventors the fruits of their inventions, and securing investors the fruits
of their risks – the stronger will be the incentive for technological change. Given

eare545.tex; 15/04/1997; 14:25; v.5; p.14


EXTERNALITIES – A MARKET MODEL FAILURE 149

these circumstances, we suggest that the greater may be the need for institutional
change to address the undesirable side effects emanating therefrom.
An important policy question follows logically from such self-contradictions.
What will be most ‘efficient’ in the totality of things: To direct or constrain the
expansion of the market through certain ex ante restrictions on the constellation
of choice sets? Or to tax the externalities that market development and evolution
tend to create as we observe them ex post? If the ‘internal’ and the ‘external’ are
interrelated, there exists a tradeoff problem here. These are certainly wide and
complex issues. But given a model that abstracts from these interrelations, the
problem is neglected.

5. Conclusions
This paper has focussed on a set of conceptual flaws in externality theory. Our aim
has been to detect inconsistencies and offer explanations for their occurrence. The
problems observed arise because a theoretical model with its conceptual basis is
extended beyond its logical domain. The process of studying physical interdepen-
dence among actors from within a theoretical construct that in its basis disregards
such relationships is certain to offer limited return on one’s investment of time and
effort. As Krupp (1963) has shown, the combination of units in deductive systems
must obey the same laws as the basic or individual units.
We have identified several problems with the received wisdom of externality
theory and policy. When analysts start with the stylized model of the market –
which, perforce, assumes that rights are already defined and allocated – the basic
question to be addressed with respect to externalities is incapable of resolution.
Further, there is a curious twist in the standard Pigovian approach to the matter
of efficiency. From the assurance that the criterion is a standard Paretian rule, one
encounters the claim of value neutrality. But only the Potential Paretian Improve-
ment rule, with its extended problems of neutrality, fits the character of the problem.
Further, we see that the Pigovian rule does not necessarily secure the greatest net
social product. That is, it will sometimes be ‘efficient’ for victims to undertake
remedial action.
The basic problem here is related to the structure of the traditional externality
model. We have seen that the conventional approach is to graft an external sphere
on to the traditional model of the fully ‘internalized’ world. This two-stage model
confuses the central issue by concealing important dynamic questions. Indeed,
the contemporary method of analysis demands no interrelation between what is
regarded as the internal structure of the model of the economy, and what then
creates externalities and also determines their form and magnitude. When these
interrelations exist, the coherence and correctness of policy prescriptions deduced
from the model must be open to serious suspicion.

eare545.tex; 15/04/1997; 14:25; v.5; p.15


150 ARILD VATN AND DANIEL W. BROMLEY

Acknowledgement
The authors would like to acknowledge the valuable suggestions of Michael Farmer
and two anonymous reviewers on an earlier version.

Notes
1
This formulation is based on the standard definition of externalities as found in Mishan (1971)
and Baumol and Oates (1988). Here externalities are defined as situations where one economic actor’s
utility or production function includes variables whose values are chosen by others, and these others
do not compensate (for costs) or receive payments (for benefits) relative to the shifted values. Baumol
and Oates emphasize that even though payments are made, not all externalities are brought to zero.
This means that they accept an implicit distinction between Pareto relevant and Pareto irrelevant
externalities.
2
Such logic always invites incredulity and causes the reader to insist that the right should go to
the party who was ‘there first.’ But of course this time-dependent approach cannot be relied on for
coherent policy for the simple reason that ‘in the beginning’ there were probably no environmental
externalities at all. Policy requires more than freezing situations in time. More will be said on this
below.
3
There is a difference between Mishan (1971) and Baumol and Oates (1988) here. Mishan talks
about unintended or incidental acts. Baumol and Oates refer to Mishan, but use the phrase ‘without
particular attention to’ or ‘deliberately does something to affect A’s welfare’ (p. 17). This is not the
same, though Baumol and Oates do not seem to acknowledge any difference. Our point is not that the
economic actors deliberately produce the externality to harm someone. They do it to intentionally
reduce their own costs, perhaps acknowledging that it also may harm others, which is something
different indeed.

References
Bator, F. M. (1958), ‘The Anatomy of Market Failure’, Quarterly Journal of Economics 72, 351–379.
Baumol, W. J. (1972), ‘On Taxation and the Control of Externalities’, The American Economic Review
62, 307–322.
Baumol, W. J. and W. E. Oates (1988), The Theory of Environmental Policy. Cambridge: Cambridge
University Press.
Bromley, D. W. (1991), Environment and Economy: Property Rights and Public Policy. Oxford:
Blackwell.
Buchanan, J. M. and W. C. Stubblebine (1962), ‘Externality’, Economica 29(Nov), 371–384.
Burrows, P. (1980), The Economic Theory of Pollution Control. Cambridge, MA: The MIT Press.
Coase, R. H. (1960), ‘The Problem of Social Cost’, The Journal of Law and Economics 3, 1–44.
Dahlman, C. J. (1979), ‘The Problem of Externality’, The Journal of Law and Economics, 22,
141–162.
Douglas, M. (1966), Purity and Danger: An Analysis of Concepts of Pollution and Taboo. New York:
Praeger.
Etzioni, A. (1988), The Moral Dimension: Toward a New Economics. New York: The Free Press.
Fisher, A. C. (1990), Resource and Environmental Economics. Cambridge: Cambridge University
Press.
Griffin, R. C. (1995), ‘On the Meaning of Economic Efficiency in Policy Analysis’, Land Economics
71(1), 1–15.
Hartwick, J. M. and N. D. Oleweiler (1986), The Economics of Natural Resource Use. New York:
Harper Collins Publisher.
Kapp, K. W. (1971), The Social Costs of Private Enterprise. New York: Schoken Books.
Krupp, S. (1963), ‘Analytic Economics and the Logic of External Effects’, American Economic
Review 53, 220–226.

eare545.tex; 15/04/1997; 14:25; v.5; p.16


EXTERNALITIES – A MARKET MODEL FAILURE 151

Marshall, A. (1890), Principles of Economics, London: Macmillan.


Mishan, E. J. (1971), ‘The Postwar Literature on Externalities: An Interpretive Essay’, Journal of
Economic Literature 9, 1–28.
Mishan, E. J. (1980), ‘How Valid Are Economic Evaluations of Allocative Changes?’, Journal of
Economic Issues XIV(1), 143–161.
Mähler, K.-G. (1974), Environmental Economics: A Theoretical Inquiry. Baltimore and London: The
John Hopkins University Press.
Papandreou, A. A. (1994) Externality and Institutions. Oxford: Clarendon Press.
Pearce, D. W. and R. K. Turner (1990), Economics of Natural Resources and the Environment.
Baltimore: Johns Hopkins University Press.
Perrings, C. (1987), Economy and the Environment. Cambridge University Press.
Pigou, A. C. (1932), The Economics of Welfare. London: Macmillan and Co.
Randall, A. (1974), ‘Coasian Externality Theory in a Policy Context’, Natural Resources Journal 14,
35–54.
Randall, A. (1983), ‘The Problem of Market Failure’, Natural Resources Journal 23(1), 131–148.
Schmid, A. A. (1987), Property, Power, and Public Choice. An Inquiry into Law and Economics.
New York: Praeger.
Sen, A. (1977), ‘Rational Fools: A Critique of the Behavioral Foundations of Economic Theory’,
Philosophy and Public Affairs 6, 317–344.
Vatn, A. and D. W. Bromley (1994), ‘Choices Without Prices Without Apologies’, Journal of Envi-
ronmental Economics and Management 26, 129–148.

eare545.tex; 15/04/1997; 14:25; v.5; p.17

View publication stats

You might also like