___es w 615
POLICY RESEARCH WORKING PAPER 1975
Acting Globally While Locally motivated air quality
programs have only minor
Thinking Locally collateral benefits for the
global climate. If agencies
Is the Global Environment Protected with global and local agendas
did business together, then
by Transport Emission Control individuals and firms-and
Programs? even cities - would act
globally when thinking
Gunnar S. Eskeland locally, and one would see
Jian Xie greater synergy.
The W7orld Bank
Development Research Group
Public Economics
and
Environment Department
Global Environment Unit U
September 1998
POLICY RESEARCH WORKING PAPER 1975
Summary findings
Eskeland and Xie find that locally motivated air quality Corntrol programs developed under joint stimulus to
programs for urban transport have limited collateral protect the global and local environment have not yet
benefits in terms of protecting the global climate. This been seen, and they may surprise us when they come.
could puzzle some, since these two public goods - one However, they will likely rely more on reducing demand,
global, one local - seem to be jointly produced. using instruments such as corrective (Pigovian) taxes on
However, air quality in Mexico City, Santiago, and fuels. The authors show how, if locally and globally
elsewhere is predominantly pursued by technical charged agencies can do business together, consumers,
improvements (making cars and fuels cleaner), and not producers, and cities will act globally when thinking
by reducing demand for polluting goods and services locally. Only then will we know the extent to which local
(though in Europe high fuel taxes help reduce demand). and global benefits are produced jointly.
This paper - a joint product of Public Economics, Development Research Group, and the Global Environment Unit,
Environment Department - is part of a larger effort in the Bank to analyze environmental problems and policies in
developing countries. Copies of the paper are available free from the World Bank, 1818 H Street NW, Washington, DC
20433. Please contact Cynthia Bernardo, room MC2-501, telephone 2T02-473 -1148, fax 202-522-1154, Internet address
cbernardo@worldbank.org. The authors maybe contacted atgeskeland@,2Dworldbank.orgor jxie@worldbank.org. September
1998. (19 pages)
The Policy Research Working Paper Series disseminates the findings of Diork in progress to encourage the exctange of edeas about
developmnent issues. An objectizve of the series is to get the findings out qluickly, euen if the presentations are less thaiz fully polisbed. The
papers carry the names of the autbors and sbou/d be cited accordingly. The fin&Fngs, interpretations, and conclusions expressed in this
paper are entirely those of the authors. They do not necessarily represent the zview of the World Bank, its Executizve Directors, or the
counrtnes they represent.
Produced by the Policy Research Dissemination Ccnter
Acting Globally while Thinking Locally:
Is the Global Environment Protected by Transport Emission Control Programs?
Gunnar S. Eskeland and Jian Xie*
The World Bank
Abstract: Locally motivated air quality programs have only minor collateral benefits for the
global climate. If agencies with global and local agendas did business together, then individuals and
finms - and even cities - would act globally when thinking locally, and one would see greater synergy.
Summary
Eskeland and Xie find that locally motivated air quality programs for urban transport have limited
collateral benefits in terms of protecting the global climate. This could puzzle some, since these
twio public goods-one global, one local-seem to be jointly produced. However, air quality in
Mexico City, Santiago and elsewhere is predominantly pursued by technical improvements
(making cars and fuels cleaner), and not by reducing demand for polluting goods and services
(though in Europe, high fuel taxes help reduce demand).
Control programs developed under joint stimulus to protect the global and the local environment
have not yet been seen, and they may surprise us when they come. However, they will likely rely
more on reducing demand, using instruments such as corrective (Pigovian) taxes on fuels. The
authors show how, if locally and globally charged agencies do business together, consumers,
producers and cities will act globally when thinking locally. Only then will one know the extent to
which local and global benefits are produced jointly.
* 'Valuable inputs from Noreen Beg and Charles Feinstein are gratefully acknowledged, as is financial
support from the Global Overlay Program, partially funded by the Governments of Denmark and Norway.
I. Introduction: A Local Public Good and a Global Public Good
Local air pollution tends to be awarded priority over global climate change
concerns, due to the adverse effects on health resulting from local pollutant emissions.
Such health damages have long been recognized in the industrialized world, and polluted
cities in industrialized countries-such as in the Ruhr area of Germany, in London, and in
Pittsburgh-have long ago moved to control dust emissions.' Even now, as particle and
lead concentrations have been reduced in many cities, and the main priority is reduction
in ground-level ozone, expected improvements in public health are the main driving
force. In many developing country cities, air quality is an emerging priority, and the main
motivator is again health effects (see, for instance, World Bank 1992; Ostro 1992; Ostro
et al. 1996, 1998; Working Group 1997, and Cropper et al. 1997). As examples, in
Jakarta, Indonesia, an estimated 1500 premature deaths are caused annually by air
pollution (Ostro 1992). A recent World Bank report (World Bank 1997) estimates that
each year 178,000 Chinese suffer premature deaths because of urban air pollution. Health
effects of air pollution are demonstrated in cities as diverse as Beijing, Krakow, New
Delhi, Sao Paulo, and Santiago, and are expected to be important in hundreds of cities in
developing as well as industrialized countries. In a study of Santiago, Chile, an air
pollution control strategy was found to be fully justified merely by modestly estimated
improvements in public health (Eskeland, 1997).
There is also increasing evidence that human activities (including the use of fossil
fuels) cause global climate change, with a range of adverse impacts on economy and
environment in many countries (see, for instance, IPPC 1996; Cline 1992). The
Intergovernmental Panel on Climate Change (IPCC) concluded that "the balance of
evidence suggests a discernible human influence on global climate."(Houghton et al.
19,96). The IPCC also estimates that unabated emissions of greenhouse gases will lead to
a rise of 1 to 3.5 degrees Centigrade in global temperature and a 15 to 95 cm rise in sea
level by 2100. This could result in the inundation of low-lying areas, changes in cropping
palterns, increased drought and flooding in some areas, and loss of biodiversity, among
other effects. Changes in the frequency and intensity of tropical monsoons could increase
flood deaths and other damages.
This study uses the term public goods for air quality and the global climate, not
because the control should be by the government-the government is anyway just a
conduit for private resources. Rather, the term public goods is used because enjoyment of
air quality and climate is non-exclusive-such goods can be enjoyed by one individual
without excluding the enjoyment by another. The contrast to public goods are private
goods-sometimes called rivalrous-such as bread, for which each slice is had by one
person or another. Air quality is called a local public good because its benefit domain is
confined to a city, or an airshed (e.g., a valley), while the climate is thought of as a global
IIn "The Economics of Welfare" (1920), A.C. Pigou used activities emitting "smoke" to illustrate that social returns
can differ from private returns, and cited studies giving quantitative estimates for damages from Manchester and
Pittsburgh. The term "Pigovian taxes," reflects his demonstration of a coordinating role for government, and taxes on
emissions as the recommended policy instrument.
1
public good for obvious reasons (the world we share is spherical). A long-standing theme
in public finance, originating with Pigou (1920), is that public goods such as air quality
and public safety create a need for coordination beyond what would be delivered by
voluntary trades and the market mechanism. Emission taxes and regulation are amongst
coordination mechanisms discussed in the literature.
The question we ask in the following study is to what extent there is synergy in
the solutions to urban air pollution problems and global warming problems. In other
words, we ask whether two public goods, one with a local domain, and one with a global
domain, are jointly produced, in the same way as are rmutton and wool. A number of
studies have addressed this problem in different ways. Fcr example, Burtraw and Toman
(1997) review studies that assess environmental ancillary effects of greenhouse gas
mitigation policies. Our approach is somewhat different, since we look at how two
differently directed policy agendas would interact through their strategies and whether
local and global benefits can be achieved simultaneously.
In our paper, we study some aspects of integration between local air pollution
problems-the way they are experienced in metropolitan areas such as Santiago, Chile
and Mexico City-and the problem of global climate change. The integration involves
aspects of environmental science and engineering, of economics and of institutional
mechanism design. The two problems have in common that they are caused in part by
emissions from human activities, and in particular by byproducts of combustion of fossil
fuels. On this basis alone, one could say that there is a strong synergy in the sense that a
complete shutdown of combustion activities would likely solve much of urban air
pollution problems as well as saving the global climate. In our study, however, we
examine more realistic (and indeed more attractive) strategies, such as modest reductions
in fuel combustion, and carefully designed urban air quality control programs. Along
these lines of collective action, we will show that the "jointness" in production of the two
public goods is not as obvious-not as firm-as one might expect.
The structure of the paper is as follows: In part [I, we review locally motivated
vehicular emission control programs for Mexico City and Santiago, and explain how the
goals pursued differ from those relating to the global climate change agenda; in part III,
we examine quantitatively the synergy between the two objectives, and check how local
programs would be modified if credited with 'collateral' global benefits; in part IV, we
show how fuel taxes can be used alone or together with other instruments to pursue both
goals; and in part V we envision a scenario wherein ,agencies with local and global
agendas do business together.
II. The Background Studies: Two Locally Motivated Emission Control Programs
While emissions from the same polluting sources contribute to both local and
global problems, the type of emissions that contribute to each are very different. Thus, an
urban air quality agency and a climate protection agency dislike different byproducts of
combustion, in the same way as wool merchants and butchers appreciate grazing sheep
for different reasons.
2
In this study, we use measures of six local air pollutants and three greenhouse
gases (GHGs). We have explained in Annex 1 our central assumptions and sources
regarding emission coefficients and how they are given priority weights according to a
local and a global agenda, respectively. A few reports (see Eskeland 1994 and World
Bank 1994) explain the basis for weights representing local priorities. These should not
be considered "final answer", but estimates based on present knowledge and plausible
assumptions. Central to our argument here is that emission components weighted locally
are not weighted globally and vice versa. This proposition is not likely in serious dispute.
Major local air pollutants emitted by motor vehicles are nitrogen oxides (NOx),
volatile organic compounds (VOCs), carbon monoxide (CO), sulfur oxides (SOx),
particulate matter less than 10 microns in diameter (PM10), and lead. Their relative
weighting is also associated with uncertainty and professional debate. Among these the
problems and priorities will-and should-differ from one city to another: Early phases
of urban air pollution control will typically emphasize dust and small particles (the now
mostly historical London smog, which alerted the profession to air pollution's effect on
premature mortality) and lead, but this emphasis fades as those pollutants are removed.
Present programs in the US and Western Europe emphasize the precursors of ozone:
volatile organic compounds and nitrogen oxides, and this emphasis also inadvertently
raises the emphasis of gasoline vehicles in a control program. Effects on human health
worill usually be key amongst the concerns in urban air pollution control programs. Such
programs will not pay any attention to carbon dioxide (CO2), and typically not to methane
and nitrous oxides either-the three most important pollutants in terms of global climate
change.
In an attempt to answer how urban air pollution control programs could be
modified by taking into consideration global climate change concerns, we revisit locally
motivated pollution control programs for transport in Mexico City and Santiago, Chile
(see World Bank 1992, 1996; Eskeland 1992, 1994, 1997). The analytical bases in the
two programs differ considerably, and we shall therefore use the studies differently. The
Mexico City analysis evaluates in detail a wide range of technical measures in which
vehicles and fuels can be made less polluting. 26 control measures remained-to be
represented on the curve-after others were excluded because they did not belong on a
cost-effective expansion path (Figure 1). The measures can be grouped as follows:
vehicle retrofitting, emission standards and inspection programs, fuel improvements and
alternative fuels. The study used a cost-effectiveness measure developed by giving
different pollutants different weights, but did not produce estimates of the benefits of
urban pollution reductions. The weights were dominated by accepted health
considerations, giving lead a high weight, CO a low weight, and with PM10, SOx, and
T4Ox intermediate weights (per emitted ton). The program ranked the 26 options
according to cost effectiveness, to show how emission reductions could be provided at a
lowest possible cost (Figure 1, adopted from Eskeland, 1994). However, since benefits
were not estimated, the study did not indicate a desirable level of emission reductions.
The Santiago study, in contrast, only analyzed three broadly defined emission
control strategies (emission standards for buses, cars and trucks), and thus does not have
the same amount of detail in terms of technical control alternatives. Rather, analytical
resources were geared towards obtaining estimates for the benefits of pollution
3
reductions, employing models of health effects, pollulant exposure and dispersion. The
Santiago study broke new ground in terms of completing a multi-pollutant cost-benefit
analysis of emission controls-including costing control alternatives, dispersion and
exposure modeling, dose response estimation for health effects, and valuation of these
health effects (see World Bank 1994; Eskeland 19,97). Ostro (1992) had laid the
groundwork for using transferred dose response functions for health effects. In the
Santiago study, a subset of the dose-response functions were estimated locally, lending
empirical support to the working assumptions that dose response functions can be applied
for transfers from other cities with similar conditions (World Bank 1994; Ostro et al.
1996, 1998).
Figure 1. Program to Reduce Local Air Pollution
Emissions in Mexico City, with and without a fuel tax
Marginal cost of emission reductions
(dollars per ton)
2600
Fuel
2100 improvements
Emission
standards
1600 Pas ger cars
axis..
1100 epla ment) 6.
Ga line Strengthen inspection
tru s s.......
600 -
oUO ~~~ Mini uses\
Inspection of p4ssenger cars
100 Inspection of high-use vehicles
0
, _00 Retrofitting (natural gas and LPG)
-400
Cumulative emission reductions Target Reduction
(million weighted tons) 1.2 million tons
Technical controls only
. Controls, matched with gasoline
Table 1 shows health benefits, in dollars per ton of emitted pollutants, resulting
from modestly estimated and modestly valued effects on human health in Santiago. To
facilitate a practitioner's understanding, we have also included what this would imply in
terms of dollars per liters of gasoline in the city, with alternative figures when we assume
controlled and uncontrolled vehicles. However, uncertainty and discussion prevail in
terms of what should be the priorities between emitted pollutants in locally motivated air
pollution control programs. The approach using transferred dose-response functions for
health effects has led to a high relative value on small dust particles (PM1 0), due to their
well-documented effects on a range of morbidity symLptoms as well as on premature
mortality. Present professional debate gives many reasons to believe that the relative
4
weighting of PM1O could be even higher than what is reflected in Table 1, due to
downward biases both in terms of quantity and value of mortality effects. The Santiago
study-in contrast to the Mexico City study, excluded effects of lead and carbon
monoxide. Lead was excluded because lead in gasoline was insignificant, since it was
already in the process of being phased out. Carbon monloxide was excluded because there
are as yet no quantified dose-response functions in the literature. Neither of these
omissions have any importance in a study of the interaction with global objectives: Lead
can be ignored because lead reduction strategies operate independently of GHG
emnissions, carbon monoxide because carbon is valued equivalently in a GHG program
whether it appears as CO or as CO2.
Table 1. Health Benefits of Emission Controls, Santiago, Chile
US$ per Illustration: Implied Cost:
emitted ton US cents per liter gasoline
Controlled vehicles Uncontrolled vehicles
Small particles (PM1O) $18,200 0.2 0.9
Nitrogen Oxides (NOx) $1,400 0.9 1.8
Volatile Organic Compounds (VOC) $500 0.2 1.3
Total 1.3 4.0
Note: Four cents per liter is roughly 20 to 25 percent ad valorem, assuming a world price of gasoline of
0.8 to $1 per gallon. Source: World Bank, 1994; authors' calculations.
Prioritization of emitted pollutants as greenhouse gases (GHG) is according to
their global warming potential (GWP), which puts a heavy weight on gases that play little
or no role in urban air pollution control programs. Table 2 shows the relative weights for
greenhouse gases used in this study, with the level illustrated by using a global benefit of
US$20 per ton of carbon equivalents (i.e., $5.4 per ton of CO2).2
Table 2. Global Benefits by Emitted Pollutant
(according to global warming potential, GHG)
US$ per Illustration: Implied Costs
emitted ton US cents per liter gasoline
controlled vehicles uncontrolled vehicles
Carbon Dioxide (CO2) $5.4 1.3 1.3
Mlethane (CH4) $134 0.0 0.01
Nitrous Oxide (N2O) $1,744 0.25 0.02
Total 1.5 1.3
Note: The global benefit of $20 per ton of carbon ($5.4/ton CO2) is from Fankhauser (1995). The
benefits of CH4 and N20 are derived by using their global warming potential factors recommended by
IPCC (1995). VOC and CO are not valued directly, only in terms of their 'terminal' status as CO2.
2 Since the global climate has not yet been given any scientifically based (or consensus) value, other values could also
be contemplated (our essential points are unchanged). $20 per ton is indicated by Fankhauser (1995).
5
Comparing Tables 1 and 2, it can be seen that the pollutants (gases and small dust
particles) that are valued in a locally motivated program (Table 1) are not valued in terms
of GWP (Table 2), and vice versa. The hydrocarbons that are targeted for emission
control in an air pollution control program, volatile organic compounds (or non-methane
hydrocarbons) are targeted precisely because of their reactivity. In terms of GWP, in
contrast, the only hydrocarbon that is given a value different from its terminal role as CO2
is methane, which has a high discounted global warming potential because it lives long in
the atmosphere in a form with much higher spontaneous global warming potential than
CO2. When emissions of VOC and carbon monoxide ('CO) are reduced in a locally
motivated air pollution control program, the result is merely to increase the share of
carbon atoms that are emitted directly as CO2 (more complete combustion). Such
technical controls, therefore, have no significant effect on global warming. When a
technical option contributes to GHG emissions reduction, it is typically because the
option makes vehicles more fuel efficient. Thus, there are less pollution emissions per
mile or kilometers driven, but typically not per liter of fuel consumed.3
III. Locally Motivated Programs and their 'Collateral' Global Benefits?
The local programs in Mexico City and Santiago are quite effective in reducing
air pollution locally-the Mexico City program can reduce 64 percent of the locally
weighted air pollutant emissions from motor vehicles (see Figure 1). What is the effect of
local programs on GHG emission reductions? The 26 measures identified in the Mexico
City study are technically oriented, and none of them deal with demand management or
alternative transportation modes.4 In part, this is why this program has a very limited
effect on the global environment. Figure 2 reveals that the Mexico program would reduce
only 6.5 percent of GHG emissions despite the success in local pollution reduction. In
fact, the 6.5 percent may well be an upwardly biased estimate, because no changes in
travel demand are assumed for these technical options, even though some of them deliver
a gain in fuel efficiency. More likely, travel and transport demand would increase when
higher fuel efficiency reduces short term marginal costs (Demand change and demand
management will be analyzed in section 4 of this paper). This rather unimpressive
synergy is found also in the case study of Santiago, Chile, for which the identified
measures in the locally motivated program reduces 65 percent of local pollution from
these sources but only 5.3 percent of GHGs.
3In fact, as the example showed in Table 2, the light-duty gasoline vehicle control option emits more GHGs per liter
of gasoline in part because of higher N20 emission per liter resulted from higher temperature and more complete
combustion.
4 The study (Eskeland, 1994) included arguments for and estimates of a 'mnatching gasoline tax', optimally to be
combined with other instruments (figure 1). In Mexico at the time, gasoline prices were increased (and other taxes
reduced to compensate), but whether pollution control was the dominant motivator is subject to interpretation.
6
Figure 2. Transport Air Pollution Abatement Supply Curves
in Mexico City: Local vs. Global
3500 -
GHGs abatement supply curve Local pollution abatement supply curve
3000 ($/ton of C02 equivalent) ($/ton of local toxicity-wtd. pollutant)
t 2500
' 2000
� 15000
a ooo t
E 500|-
fio is-oc_
-500 l
-1000
0% 20% 40% 60%
Percent of reduction in total emissions
Perhaps more interesting than this comparison in terms of total quantitative effects
is a more detailed look at what each individual measure on the local control cost curve
does to GWP emissions. One perspective is provided by reranking the 26 control
rneasures according to GWP cost-effectiveness (Annex Table 1). In Figure 3, we have
plotted the resulting cost-effectiveness ranking according to GWP against the cost-
effectiveness ranking for the locally motivated program.
Figure 3. Ranking of Pollution Control Measures:
Local vs. Global, Mexico City, Mexico
Gasoline truck replacement
Gasoline truck '93 stds.
8-
> t: Minibus '92 stds.
6 RcOC bus re-engine Passenger car U.S. Tier I stc s.
6 RI * lO bus re-engine
o AW SoM bus re-engine
,, 3 4 - * Passenger car'93 stds.
Taxi U.S. Tier I stds.
d 2 Taxi replacement
* Gasoline truck LPG retrofit
0
0 5 10 15 20 25
Ranking by cost-effectiveness of local toxicity-wtd. air pollution reduction
Retrofitting gasoline trucks for liquefied petroleum gas (LPG), the most cost
effective measure for controlling local air pollution is also the most cost effective for
climate control, but action number two and three in a locally motivated program (see
7
Figure 1) have no effect on the climate. In general, there are two important observations
to make from Figure 3. First, there i's a marked lack of correlation in this figure. The
technical measures that are cost effective from the perspective of controlling local air
pollution are not more likely (other than accidentally) to be cost effective from the
perspective of protecting the global environment (there is no tendency of a negative
correlation either). Second, 16 out of the 26 identified control measures in the locally
motivated program have no effect on the global climate. These are the measures that are
all on the upper frame of the figure, coming in last in terms of cost effectiveness (for
additional detail, see Annex Table 1, and Eskeland, [994). Practically speaking, only
strategies that make vehicles more fuel efficient or take them towards less carbon
intensive fuels have an effect on the global enviromnent (unless if they lead to less
driving, see below)
In Figure 4, we examine the impact on a locall) motivated program for Santiago
when it is credited with 'collateral' global benefits. Figure 4 displays the control cost
curve for three control options evaluated in the Chile study: emission standards for diesel
buses (Bus '91 stds), standards for diesel trucks (Truck '91 stds), and standards for light-
duty gasoline vehicles (LDGV '93). The upper curve includes local benefits and costs
only, which was the basis for evaluation in the original Santiago study. Using the benefit
weights of Table 1, all emission consequences are converted into PM 10 equivalents, so as
to arrive at a cost-effectiveness ranking reflecting local benefits. Given the estimated
benefit of $18200 per ton of PMIO, all of the three pollution control measures were
attractive even when credited only with local health effects.
Figure 4. The Pollution Abatement Curve in Santiago, Chile
Accounting for Global Benefi'ts
$20,000 T
Benefit of pollution reduction
(per ton of PMIO equivalent)
$16,000-
e $12,000
LDGV '93 stds
= o $8,000 -- - -
Truck'9lstds
w _ $4,000 Bus'91 stds ' w
' I With a subsidy for global benefits
($20 per ton of carbon)
$0 i
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Locally weighted pollution emission reductions
(percent of total emissions)
The lower, dotted curve represents the following thought-experiment: If these
three control strategies were credited with a certain value for the reductions they offer in
8
terms of GHG emissions, how would this modify the costs? We assume a subsidy
received by the policy planner from outside, but not by transport users, so no demand
response is included. The figure shows that a positive value for greenhouse gases ($20
per ton of carbon and equivalents in this case) would lower the cost curve. Among these
three measures, the car standards provide the higher amount of collateral GHG reductions
per dollar of control costs. It therefore is the one that displays the greatest 'jointness'
beiLween the local and the global agendas. If a subsidy (or a tax on GHG emissions) for
global benefits was applied at a higher rate, then climate-friendly measures (for example,
the LDGV standards measure) would be more cost-effective and attractive to decision
makers. Importantly, the quantitative impact of these subsidies is limited. If they were to
generate a re-ranking of measures for a locally motivated policy maker, they would have
to be in a radically different magnitude than $20 per ton of carbon. Furthermore, since the
background study only included measures found attractive within a narrow local
perspective, the study's limitations precluded that collateral global benefits could make a
control measure attractive.
Generally speaking, if there is agreement on a strategy between local and global
objectives in the transport sector, then it is typically because the strategy either alters total
fuel consumption or because it shifts consumption towards less carbon-intensive fuels. In
this case, according to the technical assumptions, the car standards improve fuel
efficiency of gasoline cars (from 7.8 to 9.7 km/liter) as one of its associated effects,
resulting in collateral global benefits. Under the assumption of constant travel distance,
this measure will reduce GHG emissions by 40,000 tons of carbon, which means a saving
of $0.8 million (being credited at $20 per ton of carbon) in total technical control costs.
Notice again the assumption that there is no change in travel demand. Without
complementing measures to control travel demand, this is a questionable assumption. In
the following, we will take a careful look at the demand change and demand
management.
IV. Demand Management: Using Fuel Taxes to Pursue both Agendas
The observed ineffectiveness of the locally motivated programs in reducing
greenhouse gas emissions raises the question of whether this really is all the available
synergy between local and global objectives. Before we turn to that question, we should
remember that the basic idea behind environmental taxes is that they can challenge
everybody-and polluters among them-to invent and reveal ways to reduce emissions.
Thus, the fact that locally motivated programs do not include great improvements for the
global environment could simply reflect that those programs are not developed with the
global environment in mind, and thus would only by accident include global benefits.
Mfeasures to deliver local and global emission reductions have not yet-to our
knowledge-received stimuli other than those conceived in desk studies such as this one.
Thus, we can only speculate what control measures would come out of the woodworks if
the policy environment contained stimuli for both the local and the global public good.
Turning to the practical task at hand, we know that a greater scope for jointness
cen be found when one also includes demand management in the emission control
strategy. Proposals for GHG reductions often come in the form of 'carbon taxes', whose
9
principal effects will be to turn users away from GHG-intensive consumption, and thus to
stimulate energy conservation as well as substitution towards less GHG-intensive energy
sources (from coal to petroleum products, and from petroleum products-and gas-
further towards hydro- and solar energy). Box 1 explains how a simple perspective from
welfare economics allows the development of a cost concept when demand is
manipulated to attain emission reductions. The perspective does not include the
environmental objectives in the welfare function, and thus represents costs in terms of
sacrificed consumption of other, nonenvironmental goods and services. In Eskeland
(1994) it is demonstrated that this cost concept is also valid when a 'matching tax' on
polluting goods and services is used in combination with measures such as emission
standards, which stimulate cars and fuels to be cleaner.5
Figure 5, with a locally motivated perspective (locally weighted pollutants, in
PMIO equivalents, along the x-axis), displays the three control strategies of Santiago as a
supply curve. The first, boldfaced curve represents the reduction in locally weighted
emissions provided by the identified technical controls. Fuel consumption and travel
demand are adjusted to reflect changes in fuel economy and resulting cost reduction per
kilometer. This is done by assuming that the demand function for fuel is derived from
demand for travel and transport. Then, an increase in fuel efficiency that is induced by
regulation will result in an increase in fuel demand corresponding to the reduction in user
cost for the vehicle.6 The second curve displays the additional locally weighted emission
reductions that would be obtained if one levied fuel taxes to protect the global
environment, increased in a stepwise fashion. As shown in the figure, local emission
reductions, then 'freeriding' on efforts to protect the global environment, would then be
increased to about 71 percent, up from the 61 percent lcical reductions, if carbon taxes
were levied at a rate of $150 per ton of carbon (equivalent to US 10 cents per liter of
gasoline, or about 30 percent ad valorem). The carbon tax rate at $20 per ton
($0.013/liter) would expand locally weighted emission reductions from 61 to 63
percent-an important but not stunning contribution to an urban air quality program.
5Eskeland and Feyzioglu (1997), "Is demand for polluting goods manageable? An econometric model of car
ownership and use in Mexico" aimed to see what are the costs to consumers of giving up gasoline consumption, when
price instruments were used to suppress the least essential trips first. Eskelarid and Feyzioglu (1997b) "Rationing can
backfire: The day without a car program in Mexico City" analyzed a one-day driving ban used to manage demand in
Mexico City. The ban was found to be counterproductive, leading to increased driving. This unfortunate outcome was
in part because the ban does not have a self-selection property, to suppress th.- least essential trips first, in part because
the regulation implicitly bundled cars with driving permits. Mexico City drivers responded in part by purchasing
additional used cars from the rest of the country, thus not only driving more than they otherwise would have, but also
with older, more polluting vehicles, and unnecessarily tying up hardware that could have been better used by others.
6 The price elasticity for fuel is assumed to be -0.80 for Santiago, Chile, based on a conservative estimate from the
study in Mexico City (Eskeland, 1994; Eskeland and Feyzioglu 1 997a).
10
Box 1. Emission Reductions with Fuel Tax Changes: Marginal Welfare Costs
Cost per liter
Demand curve
Tax change, dt I .
Initial tax, t .Marginal production cost for fuel
i dt. / Quantity of fuel, x
dx =dt - /-P
Quantity of pollutant emissions, E
dE = e dx
(E = e E x)
The top part of the figure is a traditional demand curve for a fuel - for example, gasoline. Under the
assumption that tax revenues are as useful to society as is income to consumers (and that other taxes are
zero), the welfare costs of a tax increase will be equal to the shaded area, approximated by the rectangle
dw z t * dx = t dt /p
where ac/7 is the slope of the demand curve.
The lower part is just an alternative x-axis, showing how emissions represent a constant times fuel
consumption, as long as the demand changes do not change emission factors (which for fuel tax changes is
a p ausible approximation, if not an accurate representation). Thus, the effect on emissions of a change in
the fuel tax rate is the emission factor times the demand change:
dE = e dx = e dt d .
It follows, if we divide the expression for the welfare costs by the expression for the emission reduction,
thalt the welfare costs of emission reductions, when delivered by fuel tax changes are:
dw/de = tle.
Thus, the part of the demand curve over the production cost for fuel appropriately can be seen as a supply
curve for emission reductions (reading it from right to left). The measure of emissions, E (say, tons per
year), and the corresponding emission factor e (say, grams per liter), can be chosen for the problem at
hand. It could be weighted by local toxicity if urban air quality is the objective or by global warming
potential if the objective is climate control, so this perspective is applicable in many ways.
11
It is noteworthy that a fuel tax of $20 per ton of carbon is more expensive-as a
means of acquiring local benefits-than any of the proposed locally motivated measures.
Thus, carbon taxes that would matter for the global environment cannot be justified by
saying that they would do something for the local environmnent: they would, but not
cheaply if one does not value the global goals. This does not mean that fuel taxes should
not play a part in a local program-indeed here they do-but adding a globally motivated
tax does not buy local improvements cheaply.7
Figure 5. Local Pollution Abatement Curves: Technical vs. Tax
$/ton PM10 equivalent Santiago, Chile
$150,000 - Technical options only Tech. options plus carbon txes
$125,000 $1 5OtC
$100,000
$75,000 t00C
$50,000 {ot_
$25,000 LDGVs'93 stds
1 Bus'91 stds Truck'91 stds Vs $2tC
$0 I i i . lll
0% 10% 20% 30% 40% 50% 60% 70%K 80% 90% 100%
Locally weighted pollution reduction
(percent in total emissions)
In Figure 6, we have turned the table, to ask: From the perspective of a body
charged with protecting the global environment, what does the pollution control program
in Santiago offer? When viewed from the narrow perspective of greenhouse gas
reductions, we can see that the three technical controls are costly and offer only tiny
emission reductions (only 5 percent of total emissions from. these vehicles at a high cost:
at $340 per ton of carbon for the gasoline vehicle standard, and at $1,550 per ton of
carbon for the diesel bus standard). The illustrative lvels of carbon taxes offer
greenhouse gas reductions much more cost-effectively. .A 5 percent greenhouse gas
reduction can be achieved by a fuel tax of $20 per ton of carbon. The carbon tax of $150
per ton can reduce GHG emissions by 29 percent. This figure illustrates amply how the
technical controls of a locally motivated program are not vvell geared towards providing
global climate protection and how demand management instruments such as fuel taxes
will play a role in GHG reduction.
Indeed, matching taxes on fuels are important for local programs dominated by standards. The Mexico City emission
reductions come at a welfare cost 30-45% higher if the fuel tax is excluded, since more expensive technical controls
must then be used (Eskeland and Feyzioglu 1997a).
12
Figure 6. GHG Reduction in Santiago, Chile
Technical Options and Fuel Taxes
$/ton of carbon
Tech. options only Tech. options and carbon taxes
$2,000
$1,500 Truck '91 stds
. 1 ~~~~~~~~~~~~~Bus '91 stds
$1,000
S500 t LDGV '93 stds
1 2t 5t $100tC $150tC
$ 0 1=
0% 10% 20% 30% 40% 50% 60%
GHG reductions (percent of total emissions)
V. A Globally Motivated Agency in Business with Air Quality Agencies
As shown above, a carbon tax can be effective in reducing GHG reduction:
Locally motivated urban air quality programs offer no serious competition, though they
do at some point offer a sensible supplement. We also showed, however, how a carbon
tax entails welfare costs locally-deterring but not killing incentives to levy such taxes.
Let us now introduce an experiment: What if an outside party-a firm,
government or agency who is interested in "buying" greenhouse gas emission reductions
were looking for potential providers. We could foresee such a demand for GHG
reductions coming about through international treaties, including binding quotas, but also
that the international agency would be authorized to purchase additional reductions from
anyone. Similarly, if quota agreements are tradable, then payments will be provided
between a party purchasing emission reductions from another party, not as a subsidy, but
in return for services rendered.
Of course, it is up to the seller (the city of Santiago, in this hypothetical case, or
perhaps the Chilean government) to determine how and whether it would offer emission
reductions, and it is up to the outside buyer how much he/she is willing to pay for certain
reductions or policy measures. The analysis here indicates, however, how to calculate the
welfare costs to the local party (population) of delivering GHG emission reductions. We
thus have an indication for the potential buyer of how much will be offered at various
price levels, or what he/she will have to pay. The maximum compensation necessary
would be the resulting welfare loss from the global-oriented policy. As illustrated in
Figure 7, the loss is represented by the area in triangle ABC (any incremental loss, from
additional purchases, is equal to a trapezoid like the one in Box 1: two such areas together
is also a triangle). The average cost at which the international agency would be buying
the emission reductions, however, would be considerably lower than the marginal costs,
and exactly half the marginal cost if the demand curve is linear.
13
Figure 7. Gasoline taxes and carbon emissions from light-duty gasoline vehicles
fuel cost per km. demand function
$0.025 .A $20 per ton of carbon
t tax revenue \
$12 million
$0.024 B .. C $0 per ton of carbon
6996 7395 million km
carbon taxes demand function
($/ton of carbon)
et welfare loss, $0.32
\ /g[ecost=$9.3/tC
$20tC I tax revenue
30tC | $12 million
0.593 0.627 carbon emissions, million tons
In Figure 7, a tax at $20 per ton of carbon is assumed to be levied on gasoline in
Santiago, Chile (similar analysis would apply to other fuels). With a simple log-linear
demand function, the potential impact of the tax on the operation of light-duty gasoline
vehicles, especially in terms of travel distance, fuel consumption, emission reduction, and
average cost of GHG reduction, is assessed.! The figure inclicates that a tax of $20 per ton
of carbon would reduce 34,000 tons of carbon emitted frorn those vehicles (5.4 percent if
total emissions from these vehicles) at the average cost of $9.3 per ton of carbon (about
half the marginal cost, or the tax rate). The net social welEare loss (assuming no loss, or
gain, from the transfer of tax revenue) is about $320,000.
Clearly, if the Chileans would do business with the globally motivated buyer at
this price, the carbon tax approach is cost-effective from the perspective of GHG
8 Again, we use the price elasticity for gasoline of -0.8
14
emission reduction. Local or national governments who adopt fuel taxes (or other similar
resl:raints) over and beyond possible treaties and obligations can be compensated for the
we]Lfare loss, not as a subsidy but as a business transaction.
Importantly, if local governments facing such a proposal apply the same
perspective, they would for local reasons apply fuel taxes representing local benefits of
emission reductions, and the carbon taxes levied after the international business is
cornpleted will come on top of those. This represents no double-counting, as each agency
pays for what it gets-not more, not less.
It should be noted that a framework including payment for emission reductions-
as we have proposed here-rather than payment of emission taxes to the internationally
charged body, involves some difficult issues of benchmarking, particularly in a long run
perspective. The problem basically is that the internationally charged body should pay
other parties only for measures taken over and above what they would be interested in
undertaking themselves for other reasons-in other words, the incremental cost . Such
problems prevail also in schemes with tradable quotas, for which the corresponding
problem is associated with the initial allocation of quotas. It is a deep problem, by many
found to be the cause of resistance to transferability of quotas. It does not preclude,
however, that trade can take place on a bilateral basis, but it does require the involved
bodies to be knowledgeable about their business and to be aware of what they are doing.
This basically means that an internationally charged body should be aware that a polluted
urban area has a self-interest in reducing local pollution, and that urban as well as rural
areas should know that they have something to offer.
15
References
Charles River Associates. 1997. World Economic Impacts of U'S Commitments to Medium Term
Carbon Emissions Limits. Paper prepared for American Petroleum Institute, CRA No.837-06,
Cambridge, Mass., U.S.A.
Cline, William R. 1992. "The economics of global warming." Washington: Institute for
International Economics.
Cropper, Maureen L. et al. 1997. "The Health Effects of Air pollution in Delhi, India." Policy
Research Working Paper Series, No. 1860, December. The World Bank, Washington, D.C.
Dower, R. and M. B. Zimmerman. 1992. The Right Climate for Carbon Taxes: Creating
Economic Incentives to Protect the Atmosphere. World Resources Institute, Washington,
D.C.
Eskeland, G. S. 1992. "Attacking Air Pollution in Mexico City."'Finance and Development, Vol.
29, No. 4, December: 28- 30.
Eskeland, G. S. 1994. "A Presumptive Pigovian Tax: Complementing Regulation to Mimic an
Emissions Fee." The World Bank Economic Review 8(3): 373-94.
Eskeland, G. S. 1997. "Air pollution requires multipollutant analysis: The Case of Santiago,
Chile" American Journal ofAgriculture and Resource Economics Vol. 79, No. 5, December.
Eskeland, G. S., and T. N. Feyzioglu. 1997a. "Is demand for polluting goods manageable? An
econometric study of car ownership and use in Mexico." Journal of Development Economics
53: 423-45.
Eskeland, G. S., and T. N. Feyzioglu. 1997b. "Rationing Can Backfire: The "Day without a Car"
in Mexico City." The World Bank Econometric Review 11(3): 33-408.
Eskeland, G. S., E. Jimenez, and L. Liu. 1998. "Prices that clear the air: Energy use and pollution
in Chile and Indonesia" Energy Journal, Vol. 19, No. 3.
Fankhauser, S. 1995 Valuing Climate Change: the Economic of the Greenhouse. Earthscan
Publications Ltd., London.
Houghton, J.T., L.G. Meira Filho, B.A. Callander, N. Harris, A. Kattenberg, and K. Maskell.
(eds.) 1996. Climate Change 1995: The Science of Climate C.hange, contribution of Working
Group I to the Second Assessment Report of the Intergovernmental Panel on Climate
Change. Cambridge, England: Cambridge University Press.
IPCC (Intergovernmental Panel on Climate Change). 1996. "The Revised IPCC Guidelines for
National Gas Inventories." Geneva, Switzerland.
Ostro, B., et al. 1996. "Air Pollution and mortality: results from a study of Santiago, Chile."
Journal of Exposure Analysis and Environmental Epidemiology 6: 97-114.
Ostro, B. et al. Forthcoming. "Air Pollution and Health Effects: A Study of Respiratory Illness
Among Children in Santiago, Chile." Policy Research Working Paper Series. The World
Bank, Washington, D.C.
OECD/IEA. 1993. Cars and Climate Change. OECD, Paris, France.
Parry, I. 1997. "Reducing Carbon Emissions: Interactions with thle Tax System Raise the Cost."
Resources (Summer)128. Resources for the Future, Washington, D.C.
Pearce, D. 1991, "The Role of Carbon Taxes in Adjusting to Global Warming. " The Economic
Journal 101 (July 1991): 938-48.
Pigou, A.C. 1920. The Economics of Welfare. London: Macmillan.
Repetto, R., R. Dower, R. Jenkins, and J. Geoghegan. 1992. "Green Fees: How a Tax Shift Can
Work for the Environment and the Economy." World Resources Institute, Washington, D.C.
16
Repetto, R. and D. Austin. 1997. "The Cost of Climate Protection: A Guide for the Perplexed."
World Resources Institute, Washington, D.C.
Turner, S., Michael Reale and Christopher S. Weaver. 1993. "Cost and Emissions Benefits of
Selected Air Pollution Control Measures for Santiago, Chile." Engine, Fuel, and Emissions
Engineering, Inc., Sacramento, California, U.S.A. Processed.
Working Group on Public Health and Fossil-Fuel Combustion. 1997. "Short-term Improvement
in Public Health from Global-climate Policies on Fossil-fuel Combustion: An Interim
Report." The Lancet 350(9088): 1341-48.
World Bank. 1992. World Development Report: Development and the Environment. Cambridge:
Oxford University Press.
World Bank. 1994. "Chile: Managing Environmental problems Economic Analysis of Selected
Issues." Chapter 3: The net benefits of an air pollution control scenario for Santiago. World
Bank, Washington, D.C., U.S.A. No.13061.
NVorld Bank. 1997. Clear Water, Blue Sky: China's Environment in the New Century. The World
Bank, Washington, D.C.
17
Annex 1
The Greenhouse Gases (GHGs) are carbon dioxide (CO2), methane (CH4) and
nitrous oxides (N20). C02 is harmless in terms of local pollution but a predominant
source of greenhouse gases (GHGs). The contribution of CH4 and N20 to global warming
is often measured by their global warming potential (GWP), a ratio of the global warming
effect from one kilogram of a GHG relative to that from one kilogram of CO2 over a
specified period of time. There are professional and learned disagreements about how to
measure GWP (it is a question, in part, of how one discounts the effect on the greenhouse
effect of the gas over time-since this effect differs for different gases) as well as the
environmental costs of global warming. The GWPs used in the study are 24.5 for CH4 and
320 for N20. These figures are suggested by the Intergovernmental Panel on Climate
Change (IPCC, 1996) for a timeframe of 100 years. Due to lack of GHG emission
information in study cities (this is a general case in most developing countries), we adopt
GHG emission factors in U.S., reported by IPCC (1996). This can be justified by the facts
that most vehicles operating in Mexico and Chile are mainly manufactured in U.S. and
other developed countries and that pollution emission standards adopted or going to be
adopted in these two countries, to certain degree, are based on U.S. emission standards. In
addition, we assume that the average cost of global warming is US$20 per ton of carbon,
i.e., $5.4 per ton of carbon dioxide equivalent.9
9This number is based on a study by Fankhauser (1995).
18
Annex Table 1: Ranking of Pollution Control Measures, Mexico City
Ranking by
Mgasure Toxic-wtd. GWP
LPG retrofit for gasoline trucks 1 1
Taxi replacement (Mexican '93 standards) 11 2
U.S. Tier 1 standards for taxi (incr.) 9 3
Passenger car '93 emi. standards (new veh.) 12 4
Re-engine light-diesel buses (U.S.' 91 stds.) 5 5
Re-engine R100 buses (CA'88 standards) 10 6
U.S. Tier I stds. for pass. cars (incr.) 22 7
Miinibus '92 emissions standards 6 8
Gasoline truck'93 emissions standards 8 9
Gas. truck replacement ('93 standards) 17 10
CNG retrofit for minibuses 2 11
CNG retrofit for gasoline trucks 3 11
Gasoline vapor recovery 4 11
Central I/M for high-use vehicles 7 11
Diesel especial (0.4 percent sulfur) 13 11
Lower Nova RVP (pre-'91 pass. cars) 14 11
Nova Sin (pre-'91 pass. cars only) 15 11
Decentralized l/M for passenger cars 16 11
5 percent MTBE Nova Sin (incr., pre-'91 pass. cars) 18 11
Lower Magna Sin RVP to 7.5 19 11
Road paving (1000 km.) 20 11
0.1 percent sulfur in diesel fuel (incr.) 21 11
Diesel meeting US 1993 specs. (incr.) 23 11
11 percent MTBE Nova Sin (incr., pre-'91 pass. cars) 24 11
5 percent MTBE in Magna Sin 25 11
1 1 percent MTBE in Magna Sin (incr.) 26 11
Note: The measures ranked at No. 11 by GWP reduction actually have no gain in GWP reduction.
19
Policy Research Working Paper Series
Contact
Title Author Date for paper
WPS1952 Enterprise Isolation Programs in Simeon Djankov August 1998 R. Vo
Transition Economies 33722
WPS19'i3 Trade Policies and Incentives in Garry Pursell August 1998 L. Tabada
Indian Agriculture: Methodology, 36896
Background Statistics and Protection,
and Incentive Indicators, 1965-95-
Background Paper 1, Sugar and
Sugarcane
WPS1954 Politicians and Firms in Seven Stijn Claessens August 1998 R. Vo
Central and Eastern European Simeon Djankov 33722
Countries
WPS1955 Appraising Workfare Programs Martin Ravallion August 1998 P. Sader
33902
WPS1S56 Benefit Incidence and the Timing Peter Lanjouw August 1998 P. Sader
of Program Capture Martin Ravallion 33902
WPS1 957 Bidding for Concessions Michael Klein August 1998 J. Dytang
37161
WPS1958 Total Strangers or Soul Mates? J. Luis Guasch August 1998 J. Troncoso
Antidumping and Competition Sarath Rajapatirana 37826
Policies in Latin America and the
Caribbean
WPS1 959 Is There a Credit Crunch in Wei Ding August 1998 F. Fernandes
East Asia? Ilker Domac 80453
Giovanni Ferri
WPS1 960 Calm After the Storms: Income Francisco H. G. Ferreira August 1998 G. Ilogon
Distribution in Chile, 1987-94 Julie A. Litchfield 33732
WPS1961 Protecting the Environment and the Gunnar S. Eskeland August 1998 C. Bernardo
Poor: A Public Goods Framework Chingying Kong 31148
Applied to Indonesia
WPS' 962 The Political Economy of George R. G. Clarke August 1998 P. Sintim-Aboagye
Privatization: An Empirical Analysis Robert Cull 38526
of Bank Privatization in Argentina
WPS1963 Commodity Risk Management and Donald F. Larson August 1998 P. Kokila
Development Panos Varangis 33716
Nanae Yabuki
WPS1964 Public Investment and Economic Ulrich Lachier August 1998 C. Lazcano
Growth in Mexico David Alan Aschauer 37776
Policy Research Working Paper Series
Contact
Title Author Date for paper
WPS1965Manufacturing Firms in Developing James Tybout August 1998 L. Tabada
Countries: How Well Do They Do, , 36869
and Why
WPS1966 Sulfur Dioxide Control by Electric Curtis Carlson August 1998 T. Tourougui
Utilities: What Are the Gains from Dallas Burtraw 87431
Trade? Maureen Cropper
Karen L. Palmer
WPS1967 Agriculture and the Macroeconomy Maurice Schiff August 1998 A. Valdes
Alberto Valdes 35491
WPS1968 The Economics and Law of Rent Kaushik Basu August 1998 M. Mason
Control Patrick Emerson 30809
WPS1969 Protecting the Poor in Vietnam's Dominique van de VValle September 1998 C. Bernardo
Emerging Market Economy 31148
WPS1970 Trade Liberalization and Endogenous Thomas F. Rutherford September 1998 L. Tabada
Growth in a Small Open Economy: David G. Tarr 36896
A Quantitative Assessment
WPS1971 Promoting Better Logging Practices Marco Boscolo September 1998 T. Tourougui
in Tropical Forests Jeffrey R. Vincent 87431
WPS1972 Why Privatize? The Case of Geroge R. G. Clarke September 1998 P. Sintim-Aboagye
Argentina's Public Provincial Banks Robert Cull 38526
WPS1973 The Economic Analysis of Sector Sethaput Suthiwart- September 1998 C. Bernardo
Investment Programs Narueput 31148
WPS1974 Volatility and the Welfare Costs Pierre-Richard Ag6nor September 1998 S. King-Watson
of Financial Market Integration Joshua Aizenman 33730