Tietenberg Tom, Lewis Lynne. Environmental & Natural Resource Economics
Подождите немного. Документ загружается.
381Cost-Effective Policies for Nonuniformly Mixed Surface Pollutants
8
Emissions could rise with ambient allowance trades as well. This would occur whenever the transfer
coefficient of the seller was larger than that of the buyer.
allowing the same amount of emissions to each allowance holder. Using ΔK
R
to
represent this permitted rise and E to indicate the units of emissions allowed to the
ith source, we can see from Equation (1) that
(3)
The larger the transfer coefficient (i.e., the closer the source is to the receptor),
the smaller the amount of emissions legitimized by the allowances held by that
firm. Proximate sources must purchase more allowances than distant sources to
legitimize a given level of emissions. In this ambient allowance market, the sources
pay the same price for each allowance, but the amount of emissions allowed by each
allowance varies from location to location. The market automatically determines
this common price, and the resulting allocation of allowances is cost-effective.
With respect to Table 14.1, the ambient allowance price would be $6. This cost-
effective system is called an ambient allowance system to differentiate it from the
emissions allowance system, which is used to achieve a cost-effective allocation of
control responsibility for uniformly mixed pollutants.
We can reinforce our understanding of what is going on with the ambient
allowance system by examining a specific trade. Suppose our two sources in Table 14.1
want to trade permits with the first source buying from the second. To maintain the
same concentration level before and after the trade, we must ensure that
(4)
where the subscripts refer to the first source and second source and the ΔE
i
refers
to a change in emissions by the ith source. Solving this for the allowable increase in
emissions by the buyer yields
(5)
For a
2
⫽ 0.5 and a
1
⫽ 1.0, this equation suggests that for each allowance traded,
the buyer (the first source) is allowed to emit only one-half the amount of emissions
allowed by that same allowance to the seller. After this trade, the total amount of
emissions by both sources goes down.
8
This could not happen in an emissions
allowance system, since the design of those allowances causes all trades to leave
emissions (but not concentrations!) unchanged.
The Many-Receptors Case
This analysis generalizes easily to the many-receptors case. The cost-effective
ambient charge paid by any source would, in this case, be
T
i
=
a
J
j
= 1
a
ij
F
j
¢E
1
=
a
2
a
t
¢E
2
.
a
1
¢E
1
= a
2
¢E
2
¢K
R
a
i
=¢E
i
.
382 Economics of Pollution Control: An Overview
where
T
i
⫽ charge paid by the ith source for each unit of emissions
a
ij
⫽ transfer coefficient that translates emissions by source i into concentration
increases at the jth receptor
J ⫽ number of receptors
F
j
⫽ monetary fee associated with the jth receptor
Thus, the source has to pay a charge that incorporates its effect on all receptors.
The control authority could manipulate F
j
for each receptor location until the
desired concentration level is achieved at that receptor.
9
The extension of the ambient allowance system to the many-receptor case requires
that a separate allowance market be created for each receptor. The price prevailing in
each of these markets would reflect the difficulty of meeting the ambient standard at
that receptor. All other things being equal, ambient allowance markets associated with
receptors in heavily congested areas could be expected to sustain higher prices than
those affected by relatively few emitters.
Since both the ambient allowance and the ambient charge systems take location
into account, when these policies are chosen, the marginal cost of emissions control
varies from location to location. Sources located in heavily populated portions of
the region would pay higher marginal costs, since their emissions have a greater
impact on the receptors of interest. Having control costs depend on location
provides incentives for new sources to choose their location carefully. Since heavily
polluted areas have high control costs, new firms have some incentive to locate
elsewhere, even though pollution-control expenditures are only part of the costs a
firm considers when deciding where to locate. For nonuniformly mixed pollution
problems, where the emissions occur is important and relocation may be one way
to approach the least cost allocation. With the ambient allowance and charge
systems, this is precisely what occurs.
As a practical matter, however, ambient charge and allowance systems have
proved to be excessively complex to implement. As a result, control authorities
have developed a number of rule-of-thumb procedures designed to deal adequately
with spatial issues while promoting cost-effectiveness. One approach allows
unrestricted trading within predefined zones on a “one-for-one” basis, but allows
trading between zones only after adjusting the trading ratios to take location into
account. Another approach allows unrestricted trading, but restricts use, depending
on the conditions around the acquiring source. A review of these approaches,
where they have been applied, and the evidence on their success can be found in
Tietenberg (1995).
9
Because any higher F
j
reduces concentrations at several locations, not just at the jth receptor, not all
selections of F
j
that result in the ambient standards being met will result in cost-effective allocations. In
the single-receptor case, the charge equilibrium is unique and equal to the cost-effective one. This is a
further burden on the control authority of using an emission charge system as opposed to the allowance
system where the equilibrium is unique and cost-effective. The allowance system equilibrium is unique
because all equilibria other than the cost-effective one involve higher costs and, therefore, further
opportunities for trade.
383Other Policy Dimensions
10
The literature contains a third reason. It suggests that unless emitters cover all external costs via a
revenue-raising instrument, the cost of production will be artificially low, production will be artificially
high, and the industry will contain too many firms.
Other Policy Dimensions
Two main pollution-control policy instruments rely on economic incentives—
charges and cap-and-trade. Both of these allow the control authority to distribute
the responsibility for control in a cost-effective manner. The major difference
between them we have discussed so far is that the appropriate charge can be
determined only by an iterative trial-and-error process over time, whereas for
the cap-and-trade approach the allowance price can be determined immediately by
the market. Can other differences be identified?
The Revenue Effect
One of the differentiating characteristics of these instruments is their ability to
raise revenue. Environmental taxes and auctioned allowances raise revenue, but
cap-and-trade programs that gift the allowances to users free of charge do not.
Does this difference matter?
It does, for at least two reasons.
10
First, a number of authors (Bovenberg and
Goulder, 1996; Goulder, 1997; Parry, 1995) have noted that the revenue from
environmental taxes or transferable allowances could be substituted for the revenue
from distortionary taxes, thereby reducing those taxes and their associated
distortions. When this substitution is made, the calculations indicate that it allows
an increase in the present value of net benefits from the application of this
instrument, an effect that has been called the “double dividend.” This effect creates
a preference for instruments that can raise revenue as long as both the implementa-
tion of a revenue-raising instrument and the use of this revenue to reduce
distortionary taxes are politically feasible.
The second important consideration is that the revenue from taxes or auctions
could be used to reduce the burden on low-income households. The empirical
evidence suggests that gifting allowances produces a regressive distribution of the
control burden. (A regressive distribution is one that places a higher relative cost
burden on low-income households or individuals as a percentage of their income
than middle- or high-income households or individuals.) That same evidence has
also demonstrated that when the revenue from auctions or taxes is targeted
appropriately, the regressiveness of the policy can be eliminated.
A final consequence of raising revenues involves their political feasibility. It seems
quite clear that until 2008, using a free-distribution (“gifting”) approach for the ini-
tial allocation of allowances has been a necessary ingredient to build the necessary
political support for cap-and-trade legislation to be implemented (Raymond, 2003).
Existing users frequently have the power to block implementation, while potential
future users do not. This has made it politically expedient to allocate a substantial
part of the economic rent from these resources to existing users as the price of
securing their support, sometimes in creative ways (see Example 14.3).
384 Economics of Pollution Control: An Overview
The Swedish Nitrogen Charge
One of the dilemmas facing those who wish to use charges to control pollution is
that the amounts of revenue extracted from those subject to the tax can be con-
siderable and that additional expense can produce a lot of political resistance to
the policy. This resistance can be lowered if the revenue is rebated to those who
pay it, but if all firms know they are getting their money back, the economic incen-
tive to limit emissions is lost. Is it possible to design a system of rebates that will
promote political feasibility without undermining abatement incentives?
The Swedish Nitrogen Charge was designed specifically to resolve this
dilemma. Sweden’s nitrogen oxide emissions charge was first imposed in 1992 on
large energy sources. Some 120 heating plants and industrial facilities with about
180 boilers were subject to the tax.
It was intended from the beginning to have a significant incentive effect, not to
raise revenue. Although the charge rate is high by international standards (thereby
producing an effective economic incentive), the revenue from this tax is not
retained by the government, but rather is rebated to the emitting sources (thereby
promoting acceptance of the policy by the regulated sources). It is the form of this
rebate that makes this an interesting scheme. While the tax is collected on the
basis of
emissions,
it is rebated on the basis of
energy production.
In effect, this
system rewards plants that emit little per unit of energy and penalizes plants that
emit more per unit of energy, thereby providing incentives to reduce emissions
per unit of energy produced.
As expected, emissions per unit of energy produced fell rather dramatically. The
Swedish Ministry of the Environment and Natural Resources has estimated that
the benefits exceeded the costs by a factor of more than 3 to 1. Note, however,
that rebating the revenue means that this tax cannot produce a double dividend
and it provides no incentives to reduce energy consumption.
Sources
: Robert Anderson and Andrew Lohof. “Foreign Experience with Incentive Systems,” Section 11 in
The United States Experience with Economic Incentives in Environmental Pollution Control Policy
(Washington, DC: Environmental Law Institute, 1997); and Thomas Sterner.
Policy Instruments for
Environmental and Natural Resource Management
(Washington, DC: Resources for the Future, 2003):
286–288.
EXAMPLE
14.3
While this strategy reduces the adjustment costs to existing users, generally it
raises them for new users. Interestingly in the climate change case, the empirical
evidence suggests that only a small fraction of the total revenue would be needed to
assure that the profits of carbon suppliers would be unchanged by a switch to a
revenue raising approach (Bovenberg and Goulder, 2001). Gifting all allowances
therefore may not be inevitable in principle, even if political feasibility con-
siderations affect the design. As we shall, see some of the newest programs auction
the permits off.
On January 1, 2009, the historic tendency to “gift” allowances changed with the
implementation of the Regional Greenhouse Gas Initiative (RGGI) in ten
Northeastern states, from Maryland to Maine. This cap-and-trade program covers
CO
2
emissions from large fossil fuel–fired electricity-generating plants.
385Other Policy Dimensions
RGGI Revenue: The Maine Example
The revenue received by Maine from the quarterly RGGI auctions is received by
the Efficiency Maine Trust, a specially created, quasi-independent organization.
The enabling statute requires the Trustees to spend 85 percent of the effi-
ciency funds on measures that reduce electricity use (remember the funds
are raised from electricity generators) through the year 2011. The remaining
15 percent may be used for measures that directly reduce the consumption of
fossil fuels.
The statute further stipulates that the money is to be allocated on a competi-
tive bid basis, where the bids are compared on the basis of the amount of kilo-
watt-hours reduced (for electricity) or tons of CO
2
reduced (for fossil fuels)
per
trust dollar expended
. (Notice how focusing on public dollars, as opposed to the
sum of public and private dollars, provides an incentive for cost sharing on the part
of bidders—putting more of their own money and less public money into the
project raises the ratio of the savings per trust dollar and, hence, increases the
competitiveness of their bid.)
As of December 1, 2010, roughly $23.5 million had been raised by RGGI for the
Trust. According to economic studies, the resulting investments have generated
over $110 million in energy savings. By making local firms more cost competitive
despite the higher electricity prices (due to the much lower energy consumption
resulting from the investment in energy efficiency), these investments have saved
jobs and bolstered the local economy, while reducing emissions of one of the
gases that contributed to climate change. As of the end of 2009, CO
2
emissions
from RGGI plants were 33 percent lower than they were in 2005, the date in
which the plan was announced.
EXAMPLE
14.4
A number of RGGI states have chosen to auction nearly 100 percent of these
allowances, using a sealed-bid system, with the revenue returned to the states. Most
states have chosen to use the revenue to promote energy efficiency (see Example 14.4),
although three states—New York, New Jersey, and New Hampshire—have, in
retrospect, chosen to siphon off some of the money for budget relief.
Using the revenue from auctions to promote investment in energy efficiency
reduces the cost of meeting the carbon targets. Costs are reduced not only because
less energy is used (and hence less carbon emitted), but also because the lower
demand for energy lowers the price of the allowances. (Can you see why?)
It would be hard to overemphasize what a departure from the previous norm
this venture into auctioning represents. Allowing emitters to pollute up to the
emissions standard without paying for the right to pollute (the traditional
approach) implies that emitters have an implicit property right to pollute already;
they don’t have to buy it.
A cap-and-trade program with allowance auctions implies, in contrast, that the
atmosphere is held in trust for the community. Institutions that use the atmosphere
for emissions must therefore pay to acquire that scarce right. Notice that this
understanding of who actually holds the property right to the atmosphere completely
changes the lens through which this regulation is viewed.
386 Economics of Pollution Control: An Overview
Economic analysis reminds us that putting a price on carbon (as a cap-and-trade
policy does) is a necessary, but not sufficient, component of the efficient path to
stabilizing the climate. Pricing carbon is necessary to correct the externality associ-
ated with the damages from climate change, but the price is not sufficient because,
by itself, it does not promote the efficient level of investment in energy efficiency
(remember the discussion in Chapter 7?). While the cap-and-trade program
corrects the climate externality whether or not the allowances are auctioned,
auctioning the allowances and using the revenue to promote energy efficiency
provides a faster, more efficient transition.
Responses to Changes in the Regulatory
Environment
One major additional difference between charges and allowances concerns the
manner in which these two systems react to changes in external circumstances in
the absence of further decisions by the control authority. This is an important
consideration, because bureaucratic procedures are notoriously sluggish and changes
in policies are usually rendered slowly.
11
We consider three such circumstances:
growth in the number of sources, inflation, and technological progress.
If the number of sources were to increase in a cap-and-trade program, the
demand for allowances would shift to the right. Given a fixed supply of allowances,
the price would rise, as would the control costs, but the amount of emissions or
pollution concentrations (in the case of the ambient allowance system) would
remain the same. If charges were being used, in the absence of additional action by
the control authority, the charge level would remain the same. This implies that the
existing sources would control only what they would control in the absence of
growth. Therefore, the arrival of new sources would cause a deterioration of air or
water quality in the region. The costs of abatement would rise, since the costs of
control paid by the new sources must be considered, but by a lesser amount than
with cap-and-trade, because of the lower amount of pollution being controlled.
If the choice is between a fixed fee and a fixed number of allowances in a growing
economy, the dominance of the allowance system over the fixed-fee system
increases over time (Butler and Maher, 1982).
With cap-and-trade, inflation in the cost of control would automatically result in
higher allowance prices, but with a charge system it would result in lower control.
Essentially, the real charge (the nominal charge adjusted for inflation) declines with
inflation if the nominal charge remains the same.
We should not, however, conclude that over time, charges always result in less
control than allowances. Suppose, for example, technological progress in designing
pollution-control equipment were to cause the marginal cost of abatement to fall. With
cap-and-trade this would result in lower prices and lower abatement costs, but the same
aggregate degree of control. With a charge system, the amount controlled would
11
This is probably particularly true when the modification involves a change in the rate at which firms
are charged for their emissions.
387Other Policy Dimensions
actually increase (see Figure 14.5) and, therefore, would result in more control than a
cap-and-trade program that, prior to the fall in costs, controlled the same amount.
If the control authority were to adjust the charge in each of the above cases
appropriately, the outcome would be identical to that achieved by cap-and-trade.
The allowance market reacts automatically to these changes in circumstances, while
the charge system requires a conscious administrative act to achieve the same result.
Price Volatility
Future abatement costs depend on current investment decisions when the adoption
of new technologies are involved. The desirability of current abatement investment
depends not only on the level of the price associated with emitting (either the
allowance price or the emissions charge), but also on its volatility. Volatility can
inhibit investment incentives.
Do these two policies differ in terms of the likelihood that prices will exhibit
volatility? They do. Because emissions charges fix the price, price volatility is not
an issue with that approach unless the government keeps changing the price.
Allowances, however, fix the quantity and leave the price to the market. Large shifts
in the demand for allowances, coupled with the fixed supply, can cause prices to
vary a lot. In terms of price volatility, charges have the edge.
Instrument Choice under Uncertainty
Another major difference between allowances and charges involves the cost of being
wrong. Suppose that we have very imprecise information on damages caused and
avoidance costs incurred by various levels of pollution and yet we have to choose
either a charge level or an allowance level and live with it. What can be said about
the relative merits of allowances versus charges in the face of this uncertainty?
The answer depends on the circumstances. Allowances offer a greater amount of
certainty about the quantity of emissions, while charges confer more certainty about
the marginal cost of control. Allowance markets allow an ambient standard or an
aggregate emissions standard to be met with certainty, but they offer less certainty
about marginal costs. When the objective is to minimize total costs (the sum of
damage cost and control costs), allowances would be preferred when the costs of being
wrong are more sensitive to changes in the quantity of emissions than to changes in
the marginal cost of control. Charges would be preferred when control costs were
more important. What circumstances give rise to a preference for one of the other?
When the marginal damage curve is steeply sloped and the marginal cost curve
is rather flat, certainty about emissions is more important than certainty over
control costs. Smaller deviations of actual emissions from expected emissions can
cause a rather large deviation in damage costs, whereas control costs would be
relatively insensitive to the degree of control. Allowances would prevent large
fluctuations in these damage costs and therefore, would yield a lower cost of being
wrong than charges.
Suppose, however, that the marginal control cost curve was steeply sloped, but
the marginal damage curve was flat. Small changes in the degree of control would
388 Economics of Pollution Control: An Overview
have a large effect on abatement costs but would not affect damages very much. In
this case it makes sense to rely on charges to give more precise control over
control costs, accepting the less dire consequences from possible fluctuations in
damage costs.
These cases suggest that a preference either for allowances or for charges in the
face of uncertainty is not universal; it depends on the circumstances. Theory is not
strong enough to dictate a choice. Empirical studies are necessary to establish a
preference for particular situations.
One interesting current application of these insights involves the control of the
gases that intensify climate change. The conventional wisdom suggests that the
marginal benefits of reducing greenhouse gas (GHG) emissions are likely to be flat,
since the damages from climate change are driven by the accumulated concentra-
tion of GHGs (e.g., Nordhaus, 2008). The implication is that when marginal costs
are quite sensitive to the level of emissions reduction, but the damages from climate
change are not, a carbon tax is preferred on efficiency grounds.
The conventional wisdom, however, assumes that the effects of climate change
increase steadily as a function of atmospheric concentrations of GHGs. In fact,
growing scientific evidence suggests that climatic responses to temperature
increases may well be highly nonlinear, characterized by thresholds or abrupt
changes. This understanding of the science leads to a greater sensitivity of damages
to the level of emissions reduction, shifting the preference toward cap-and-trade
(Keohane, 2009). Given these conflicting views, the dominance of one approach
over the other in the presence of uncertainty is not clear.
Product Charges: An Indirect Form of
Environmental Taxation
The use of emissions charges presumes that it is possible to monitor and keep track
of the level of emissions so the appropriate tax can be levied. Sometimes that is
either impossible or impractical.
One strategy that has been employed in this circumstance is to tax the com-
modity that is most directly responsible for the emissions, rather than the emissions
themselves. For example, one might tax gasoline rather than attempt to measure
(and tax) the emissions from every gasoline-powered vehicle. And several counties
tax fertilizer rather than attempt to measure the amount of contamination of
groundwater sources from each bag sold. The Irish have even taxed plastic bags to
prevent littering (see Example 14.5).
While product charges frequently are simpler to administer, it is important to
keep in mind that they are not equivalent to emissions charges. Not every unit of
the taxed product may have the same impact on the environment. For example,
some purchased fertilizer may be used in sensitive areas (and therefore should
be heavily taxed to reflect its high marginal damage), while others may be used in
areas with lots of natural buffering (and therefore should not be taxed as heavily).
Since the normal product charge would be the same per bag, it would not be able to
389Other Policy Dimensions
make these kinds of distinctions. Product charges are most efficient when all
purchased units of that product cause exactly the same marginal damage. Although
full efficiency is probably rarely achieved by product charges, they may be better
(even much better) than doing nothing.
The Irish Bag Levy
Rapid economic growth in Ireland in the 1990s was marked by a significant
increase in the amount of solid-waste per capita. The lack of adequate landfill sites
resulted in escalating costs of waste disposal, which in turn led to more illegal
dumping and littering. It was feared that tourism, one of Ireland’s largest
industries, would be negatively affected as a consequence of the degradation of
the environment. The food industry, which based a significant amount of their
marketing strategies on a healthy, wholesome reputation, also suffered as a result
of the public perception of their role in the increased litter.
The most visible element of litter was plastic bags, so in 2002 the government
introduced the Plastic Bag Environmental Levy on all plastic shopping bags, with a
few exceptions that were sanctioned for health and safety reasons. Retailers
were charged a fee of 15¢ per plastic bag, which they were obliged, by the
government, to pass on to the consumer. This levy was designed to alter
consumer behavior by creating financial incentives for consumers to choose more
environmentally friendly alternatives to plastic, such as “bags-for-life.” (Bags-for-life
are heavy-duty, reusable cloth or woven bags, which were made available in all
supermarkets, at an average cost of €1.27.)
Expectations that this levy would bring about a 50 percent reduction in the
number of plastic bags used were exceeded when the estimated actual reduction
turned out to be 95 percent! In a single year, Irish consumers reduced their
consumption of plastic bags from 1.26 billion to 120,000, while concurrently
raising approximately €10 million in revenue for the government. Placed in the
Environmental Fund, this revenue finances environmental initiatives such as
recycling, waste management, and, most importantly, antilitter campaigns.
This levy has been viewed as a major success by the government and environ-
mental groups alike. It has also been enthusiastically embraced by Irish consumers,
thanks to an intensive environmental-awareness campaign that was launched in
conjunction with the levy. Irish retailers, although skeptical in the beginning, have
also recognized the huge benefits of this levy. Estimates suggest that their costs
were offset by the savings from no longer providing disposable bags to customers
free of charge, as well as the profit margin earned on the sale of “bags-for-life,”
whose sales have increased by 600–700 percent since the introduction of the levy.
The amount of plastic being sent to Irish landfills has been dramatically reduced,
bringing about a clear visual improvement. The success of this case has promoted
the diffusion of this idea. For example, in 2008 China banned superthin plastic bags
and imposed a fee on other plastic bags.
Source
: Linda Dungan. “What Were the Effects of the Plastic Bag Environmental Levy on the Litter
Problem in Ireland?” http://www.colby.edu/~thtieten/litter.htm/
EXAMPLE
14.5
390 Economics of Pollution Control: An Overview
Summary
In this chapter we developed the conceptual framework needed to evaluate current
approaches to pollution-control policy. We have explored many different types
of pollutants, and found that context matters. Different policy approaches are
appropriate for different circumstances.
Stock pollutants pose the most serious intertemporal problems. The efficient
production of a commodity that generates a stock pollutant could be expected to
decline over time. Eventually, a point would be reached when all of the pollutant
would be recycled. After this point, the amount of the pollutant in the environment
would not increase. The amount already accumulated, however, would continue to
cause damage perpetually unless some natural process could reduce the amount of
the pollutant over time.
The efficient amount of a fund pollutant was defined as the amount that
minimizes the sum of damage and control costs. Using this definition, we were able
to derive two propositions of interest: (1) the efficient level of pollution would vary
from region to region; and (2) the efficient level of pollution would not generally be
zero, although in some particular circumstances, it might.
Since pollution is a classic externality, markets will generally produce more than the
efficient amount of both fund pollutants and stock pollutants. For both types of pollu-
tants, this will imply higher-than-efficient damages and lower-than-efficient control
costs. For stock pollutants, an excessive amount of pollution would accumulate in the
environment, imposing a detrimental externality on future generations as well as on
current generations.
The market would not provide any automatic ameliorating response to the
accumulation of pollution as it would in the case of natural resource scarcity. Firms
attempting to unilaterally control their pollution could be placed at a competitive
disadvantage. Hence, the case for some sort of government intervention is
particularly strong for pollution control.
While policy instruments could, in principle, be defined to achieve an efficient
level of pollution for every emitter, it is very difficult in practice because the
amount of information required by the control authorities is unrealistically high.
Cost-effectiveness analysis provides a way out of this dilemma. In the case of
uniformly mixed-fund pollutants, uniform emissions charges or an allowance
system focused purely on emissions (not ambient effects) could be used to attain the
cost-effective allocation, even when the control authority has no information what-
soever on either control costs or damage costs. Uniform emissions standards would
not, except by coincidence, be cost-effective. In addition, either emissions trading
or charges would provide more incentives for technological progress in pollution
control than would emissions standards.
Policies to control nonuniformly mixed pollutants must take the location and
possibly the timing of the emissions into account as well as the amount. In principle,
this can be accomplished with either an appropriately designed ambient allowance
trading system or ambient charge; either one can result in a cost-effective allocation