Environmental Standards and International Trade in Automobiles and Copper: The Case for a Social Tariff
Natural Resources Journal
Environmental Standards and International Trade in Automobiles and Copper: The Case for a Social
Recommended Citation 0
0 Duane Chapman, Environmental Standards and International Trade in Automobiles and Copper: Th e Case for a Social Tariff , 31 Nat. Resources J. 449 (2020). Available at:
Environmental Standards and
International Trade in Automobiles
and Copper: The Case for a
Worldwide acid rain sulfur emissions are increasing, in spite of
reduced emissions in the United States. This increase is due to the
growth in manufacturing outside the United States without adequate
pollution control. Copperand automobilesmake useful case studies.
Copper is a pollution-intensive raw material, and automobiles are
pollution-intensiveproducts.Both are tradedon a global
basis.Production of each is less costly outside the United States because
environmental protection is frequently absent. Ultimately, global
agreementswill be reachedon industrialpollution control,analogous
to the MontrealProtocolon chlorofluorocarbons.Until such
agreements exist, a socialtariffon UnitedStates importscould be valuable.
The tariffcould be based upon the degree of avoidance of
environmentalprotection costs in the manufacture of imported goods.
Since 1970, global sulfur oxide emissions have increased 50 percent.'
As Figure 1 indicates, this global increase has occured despite the 25
percent decrease in United States emissions. National reduction of sulfur
emissions have been more than offset by increased emissions in the rest
of the world. The growth in world emissions, and subsequent pollution,
arises from the increased use of coal, oil, and metal ores without adequate
This paper examines the significance of pollution control costs, and
the avoidance of these costs, to trade and development. The general belief
among economists is that industrial country pollution control and
workplace safety are unimportant economic factors in influencing productivity,
*Professor of Resource Economics, Department of Agricultural Economics, Cornell University,
Ithaca NY 14853. Research supported by the New York State Agricultural Experiment Station at
Cornell University and the Department ofAgricultural Economics. The assistance of Wendy Bacon
is appreciated, as are the valuable comments of a Journalreviewer and several readers of earlier
versions of the paper.
1. Environmental Data Report, U. N. Environment Programme at 12, U.N. Doc. GC14/L.6
(million metric tons)
the location of manufacturing, or levels of global pollution. For example,
Edward Denison finds that United States productivity declined at a 0.22
percent rate from 1973 to 1981. He attributes this primarily to a reduction
in managerial and technical knowledge and efficiency and finds
environmental factors to be unimportant with respect to productivity. The OECD
Environment and Economics Conference, Walter, Smith and Ulph, and
Leonard agree. 2 This paper, however, develops evidence to the contrary.
TRADE IN AUTOMOBILES AND COPPER
The largest flow of cars, trucks, and buses in world trade is the
movement of four million vehicles from Japan and Korea to the United States.
2. Denison, Accounting for Slower Economic Growth:An Update, in International Comparisons
of Productivity and Causes of the Slowdown 46 (. Kendrick ed. 1984); Organisation for Economic
Cooperation and Development, Environment and Economics 61-6
); Walter, Environmentally
InducedIndustrialRelocation to Developing Countries, in Environment and Trade 22 (S. Rubin &
T. Graham eds. 1982); Smith & Ulph, The Impactof EnvironmentalPoliciesin DevelopedCountries
on the Trade of Developing Countries, in Environment and Development in Asia and the Pacific
112 (1982); and H. Leonard, Are Environmental Regulations Driving U.S. Industry Overseas? (1984).
This is nine percent of all of the 4
5 million vehicles produced in 1986
and 22 percent of all world trade in vehicles.' An average passenger
vehicle consists of 3200 pounds of material. Table 1 indicates the raw
materials in an automobile.
Japan almost wholly lacks the mineral and fuel resources needed to
produce those automobiles. Table 2 shows Japan's resource dependency
in basic resources. This dependency is matched by South Korea and the
other newly industrializing countries in East Asia. It seems, then, that
Japan's contribution to global pollution is connected to both the supply
of raw materials from resource-exporting, developing countries, and to
United States product demands. In other words, an American buyer of a
Japanese vehicle purchases a commodity which incorporates the
environmental protection or worker safety problems of the resource exporting
Copper trade is studied here because of its importance as a source of
acid rain sulfur emissions in the absence of sulfur control. In northern
Mexico and the southwestern United States, the ore types result in a
smelter feed which generally has one unit of sulfur for each unit of copper
produced. Consequently, in the absence of sulfur controls, for each ton
of copper produced by the smelter, one ton of sulfur is emitted into the
atmosphere, or passed into the water or land environment. Each ton of
sulfur, in turn, will be transformed into two tons of sulfur oxide, and
three tons of sulfuric acid. 6 Table 3 illustrates that the major exporting
countries average 32 percent control of sulfur emissions. Those industrial
importing countries that smelt copper average 89 percent control. (The
index of sulfur control in Table 3 is zero if the maximum of one ton of
sulfur is released into the atmosphere for each ton of copper produced.
The index of control would reach 100 percent if no sulfur became airborne,
and any sulfur in the ore was converted into sulfuric acid, or retained in
slag or a tailings pond.7 )
Sulfur control presumably increases production costs. If this is the
case, an automobile manufacturer that buys lower cost copper attains a
cost advantage due to the absence of pollution control in the refining
process. At this point, the relevant generalization is that Canadian and
West German manufacturers of automobiles exporting their product to
the United States obtain their raw materials from countries with generally
less pollution control than is practiced in the United States, Japan, or
Western Europe. This subject is addressed in greater detail in the fourth
IMPORTER COST ADVANTAGES FOR AUTOMOBILES
Table 4 summarizes five types of cost differences between Japan and
the United States for automobile manufacturing. The most important
6. Some fraction of contained sulfur, on the order of 3 percent, will be held in the smelter slag;
the remainder will be airborne.
7. The index is constructed to reflect the amount of atmospheric SO 2 emitted per ton of copper
produced, taking into account ore type and hydrometallurgical processing as well as sulfur control.
8. Table 3 includes all market-economy countries with at least 200,000 metric tons at one or
more stages in 1985. The sulfur control index in Table 3 is defined as I less the ratio of emitted
sulfur to copper production. (See supra notes 6 and 7). Stockpile changes are not shown. World
Bureau of Metal Statistics, World Flow of Unwrought Copper (1985); see also D. Maxim, The
International Competitiveness of the U.S. Non-Ferrous Smelting Industry and the Clean Air Act
(1982), and Air Pollution Requirements for Copper Smelters, United States compared to Chile, Peru,
Mexico, Zaire, and Zambia (May 1985); and author's visits to copper manufacturing plants in the
U.S., Mexico, Chile, South Africa, Zambia, and Zaire. Group totals are weighted averages.
difference is the estimated cost savings from avoided environmental
protection costs. These costs are estimated from 1-0 (input-output) data
summarized in Table 5. The 1982 data for eighty-one commodities were
aggregated into the eighteen categories in Table 5.o Since $15,000 is the
current average price of both domestic and imported cars,"' it is used as
the basis for Tables 4 and 5. It is assumed that smaller environmental
costs for developing country suppliers reduce the costs of Japanese
mining, chemicals, energy, and metals by ten percent. It is further assumed
that the same factor is five percent for other manufactured commodities.
The magnitude of these percentages is critical, and the validity of the
percentage assumptions is examined in the following section.
Summing up the entries in Table 4, the Japanese manufacturer has an
advantage of about $1300 per vehicle. As illustration, about $1550 is
before-tax capital income for the United States manufacturer,' 2 and the
figure may be about $2850 for the Japanese manufacturer. The overall
context, then, is that an estimate of $2100 in avoided environmental
protection costs is essential to providing for the profitability of the sale
of Japanese cars in the United States. It is critical to examine the empirical
basis for this estimate.
9. For Table 4 notes, see Appendix.
10. Table 5 in the text is developed from the coefficients in Table 4 of the 1982 1-0 set, provided
to the author by the Bureau of Economic Analysis, Department of Commerce.
11. 28 Survey of Current Business 20 (Nov. 1987).
12. Before-tax capital income is 0.104 of sales. 65 Survey of Current Business 17
and Table 3, 1982 1-0 analysis supra note 10. R. Cole and T. Yakushiji, The American and Japanese
Auto Industries in Transition 117 (1984), report that the United States government tax on labor and
capital is $650 higher than the Japanese tax.
As noted above, most authors have found pollution abatement costs to
be generally insignificant for most industries. However, a detailed review
of copper production costs indicates that environmental protection and
worker safety cost approximately 15 cents per pound.' 4 The major items
involved are shown in Tables 6 and 7. With current production costs in
the United States at about 65 to 70 cents per pound, 5 these costs are 20
to 25 percent of production costs.
The most comparable figures from other economic studies are on the
13. See supra note 8.
14. Based upon actual accounting records for equipment and personnel. Sources withheld to
protect confidentiality. It should be noted that an informed reviewer believes some of the items in
Tables 6 and 7 would be utilized by an efficient management in the absence of environmental
regulations. This seems logical. However, on-site visits to copper mines and smelters in South Africa,
Chile, Mexico, Zambia, and Zaire did not support this view because the Table 6 and 7 items were
generally absent. South Africa, in particular, used the most modem production technology without
these aspects of environmental and worker protection.
15. I U.S. Dep't of the Interior, Bureau of Mines, 1986 Minerals Yearbook 331 (1988).
order of one or two percent. These are the estimates of environmental
protection costs for the nonferrous mining and smelting industries. 6 The
other studies have used Federal survey data. 7
How can these differences be reconciled? There are several
possibilities. First, many environmental activities that are a part of a production
process may not be reported in surveys. For example, the labor, fuel,
and equipment costs of dust control in a pit mine by use of watering
trucks may not be reported. Similarly, collateral protection devices that
are a secondary part of production equipment, such as the cost of a dust
hood, or fans and hoods on a grinder, may not be reported.
Second, monitoring and planning activity may be excluded from
surveys. For example, surveys may omit the environmental protection
expense of: (a) professional time spent with visitors inspecting protection
systems; (b) meteorological monitoring of ambient air quality; (c)
Environmental Protection Activities and Equipment in Copper Production s
a. air and water pollution control for coal burned for power generation
b. bag house on crusher
c. berms for chemical storage
d. covered conveyor
e. primary convertor hoods
f. fugitive emission hoods
g. gas collection fans, electricity
h. hazardous waste control
i. meteorological data and forecasting for possible pollution emergencies
j. monitors for air and water quality
k. PCB control
1. storm catchment reservoir for ten-year storm
m. tailing reservoir and drain
n. tall stack
o. waste oil control and monitoring
p. water discharge plans and monitoring
q. water recycle zero discharge
r. water spray for dust control
s. wet scrubbers
t. acid plant
u. professional environmental protection personnel
v. Federal and State reports and meetings
16. I. Walter, Environmental Resource Costs and the Patterns of North-South Trade 16 (April
1986) (unpublished manuscript); C. Pasurka, EnvironmentalControl Costs and U.S. Effective Rates
of Protection, 13 Pub. Fin. Quarterly 167, 168 (April 1985); see also supra note 2 and text.
17. Historically, there were two Federal surveys. The Bureau of Economic Analysis surveyed
business investment in plant and equipment for pollution abatement. The Census Bureau surveyed
both operating and capital costs incurred by manufacturing industries. These surveys are being
reorganized. Comments here are based upon discussions with personnel in the Bureau of Economic
Analysis, the Bureau of the Census, and management personnel at mines and smelters.
18. Sources for Tables 6 and 7 are personal interviews and confidential accounting records at
mines and smelters.
ronmental planning; (d) time and expense in report preparation and
meetings with State and Federal regulatory personnel.
A third omission from survey data is the cost of protection of workers
from environmental hazards. All the items in Table 7 are excluded.20
A fourth excluded item is productivity loss. When production stops or
is slowed because of environmental problems, this is not counted as an
A fifth factor in under-reporting environmental costs in surveys may
be the perception of managment: current management may not perceive
practices that preceded them as protective; they may simply focus on
environmental practices introduced during their tenure. Examples here
may be respirators, or tall stacks.
Finally, interest expense or opportunity cost for investment in protection
equipment is not included in the survey data. This could be significant.
Whether or not these six factors are sufficient to account for the
difference between my estimate (20 to 25 percent) and previous economic
studies (two percent) is an open question. Certainly field research on
environmental protection, worker safety, and productivity should be
In 1973, only 11 percent of United States' primary copper consumption
was imported, refined copper. By 1986, this proportion had grown to 36
percent.2' Although copper demand has been constant for fifteen years at
about two million tons annually, we now import about one million tons
embodied in automobiles, consumer electronics, capital goods, plumbing
equipment, and so on.' Apparently, overall copper consumption has
increased in the United States while smelting has decreased.
20. Sources for this discussion are those for Tables 6 and 7; see supra note 18.
21. U. S. Dep't of the Interior, Bureau of Mines, supranote 15, at 303; Copper: A Chapter from
Minerals Facts and Problems 13 (1985) (unpublished manuscript).
22. Analysis of U.S. imports prepared by T. McClive (1988) using 1-0 data on copper content.
McClive estimates about 200,000 tons of copper are embodied in imported motor vehicles and
This, presumably, is part of the explanation for the growth in sulfur
emissions from copper production outside the United States. The rest of
the world has increased its SO2 emissions from copper production from
five million tons in 1970 to 12 million tons in 1983. The United States
reduced emissions from copper from three million tons to one million
over the same period. 23 This reduction was attained both by sulfur
scrubbing and the closing of noncontrolled smelters. Thus, the global acid rain
problem is increasing, as Figure I indicates, in spite of United States
In one important respect, the copper data do not support the hypothesis
that Japanese cars have lower production costs because of the absence
of pollution control in pollution-intensive raw materials. In Table 3, Japan
had smelter production equal to 75 percent of its copper consumption,
and 98 percent control of sulfur emissions. Consequently, it cannot be
argued that imported copper is a major component of potential Japanese
cost advantage with respect to pollution-control avoidance. A Japanese
cost advantage must be found in the importation of industrial resources
other than copper, or in the avoidance of environmental protection costs
at the mining stage before smelting.
However, the copper data support the concept for two other major
automobile exporters, Canada and West Germany. Canada is the second
leading exporter of vehicles to the United States, and Table 3 shows its
sulfur control at 26 percent. West Germany, the world's second leading
producer and exporter of vehicles after Japan, imports 99.9 percent of
its copper, two-thirds of which has been smelted prior to import, primarily
in developing countries.
The recent fluctuations in copper commodity markets help protect United
States copper producers from the competitive advantage held by
developing country producers. In early 1989, the price exceeded $1.50 per
pound, more than twice average production costs. As long as this
worldwide price is near this level, United States producers will continue to
prosper. If competitive markets return to the much lower price levels
present in the mid-1980s, it would seem unlikely that United States
production could be internationally competitive at its present production
A SOCIAL TARIFF
The illustrative subject matter here has been a raw material (copper)
and a complex final product (vehicles). Both are consumed and produced
on every inhabited continent, and widely traded. Forty percent of all
vehicles and 70 percent of all copper are sold in international trade.
Several other pollution-intensive commodities are also traded globally.
These include metals, petroleum, coal, forest and paper products, fruit
and vegetable produce, and ordinary metal products such as silverware.
The pollutants produced affect domestic workers and the public, and
have transnational and global effects. We have reasonably firm data on
growth in world sulfur, chlorofluorocarbon, and carbon dioxide
emissions. We can speculate with some confidence that global accumulations
of hydrocarbons, nitrogen oxides, and lower atmosphere ozone are also
increasing. There may be growing pollution of world oceans with metals
and chemicals, but there are no global data on these two pollutants to
confirm or deny the possibility. If current national policies continue, we
may confidently anticipate exponential growth in world emissions of
Economic theory offers three broad kinds of remedies: bribes or
payments from pollution recipients to polluters, taxes or fines on pollution,
and regulation of pollution emission.
Ultimately, we can anticipate global agreements on standards for
emissions of major pollutants. If these standards are defined on a per capita
basis, they will tend to encourage developing countries to "catch up" in
terms of higher pollution levels. If international agreements are defined
in terms of percentage reductions from current levels, they will tend to
hinder developing country growth by freezing their pollution-intensive
industry potential. One likely outcome is some balance between per capita
emissions and percentage rollbacks, as with the Montreal
chlorofluorocarbon treaty. Another approach would be framed on the basis of pollutant
emissions per unit of final product. Some future agreements may also
consider pollutant emissions per dollar of gross domestic product.
Ambient air standards are another potential basis for agreement.
Before global agreements are established, consideration should be given
to environmental or social tariffs based upon cost advantages obtained
by avoiding commonly available pollution control technologies. If, for
example, this tariff is set at 50 percent of the avoided cost, then given
the estimates in this paper, imported Canadian copper might have a tariff
of up to seven cents per pound. A Japanese or Korean car might be subject
to a tariff of 33 cents per pound.24
An environmental tariff could have several positive consequences. First,
the funds raised could be used to finance pollution control in developing
countries. Second, some part of the tariff funds could be allocated for
research on the general problems of transnational externalities and world
24. This is 50 percent ofthe $2100 per vehicle environmental cost differential in Table 4, divided
by the 3,178 pounds in Table 1.
pollution. Third, producer resistance to pollution control would be
diminished. With a 50 percent tariff, each dollar spent on pollution control
in the production of an exported car or refined copper would reduce net
income by fifty cents. In addition, a social tariff would enhance the
domestic political position of groups advocating environmental protection
in developing countries. Finally, the economic position of producers using
modem methods of pollution control would be enhanced rather than
There are three approaches to such a tariff: global, including the Soviet
Union and China; industrial-market economies, meaning Japan, Korea,
Australia, Western Europe, and North America.; or, finally,
implementation by the United States. The most practical near-term approach is
probably to build domestic United States support for a general trade and
tariff agreement covering environmental protection costs.
Should wage levels and worker safety be included in international social
tariffs or standard-setting agreements? At some future point in time, the
answer will probably be yes. But for the present, the need for policies
addressing global pollution arising from world trade is evident.
APPENDIX: WAGE AND PRODUCTIVITY CALCULATIONS
All figures in Table 4 are rounded to the nearest $25. With higher U.S.
taxes and benefits, the after-tax advantage is probably much greater. The
$1500 pricing differential for cars of the same quality is assumed to be
10 percent of retail price, and is based on the Cole and Yakushiji discussion
of marketing and pricing.' The higher Japanese transport cost of $450
is also based on their analysis.2 6 The labor cost differential of $1150 per
vehicle was based upon three different methods of estimating U.S. labor
cost per vehicle, each giving an estimate of approximately $3200 per
vehicle. In the first method, production worker pay is assumed to equal
the average for all employees. In other words, the production worker rate
of $20.53 per hour is assumed to be the average for all clerical, sales,
managerial, and production workers. This is multiplied by 2000 hours
per year, and divided by 13 vehicles per employee.27 The result is an
estimate of $3158 per vehicle. In the second method, an average 1985
motor vehicle industry pay of $31,559 was assumed to be the basis for
industry-wide benefits of 29.1 percent for all employees, the I-0 value
for the auto industry. This was increased by 1.7 percent, the reported
increase for production workers in total compensation per hour between
1985 and 1987.' Dividing again by 13 vehicles per employee, the result
is an estimate of $3187 in labor cost per vehicle, $29 higher than the
first method. In the third method, the value-added coefficient (0.331) is
multiplied by the ratio of wages and benefits to value added (0.685), and
by the $15,000 price.29 The result of this method is a labor cost per
vehicle estimate of $3401.
The estimate of 13 vehicles annually per employee is based upon the
U.S. production of 10.906 million vehicles in 1987 with 841,508
Cole and Yakushiji review several studies of Japanese and American
productivity, and cite Japanese studies finding a 10 percent differential,
or 14.3 vehicles per employee.3 Total Japanese compensation is 70
percent of the U.S. level for production workers.32
Taking the median U.S. estimate of $3187 total labor cost per vehicle,
and assuming 10 percent greater Japanese productivity and 30 percent
lower wages, the result is an estimate of $2028 labor cost per Japanese
The arithmetic difference of an $1159 wage differential is rounded to
$1150 in the text.
3. Japan and Korea each export more than one-half of their production . In 1986 , Japan produced 12 . 9 million vehicles and South Korea produced 600 thousand . Motor Vehicle Manufacturers Association, Facts and Figures 8 ( 1986 - 1988 ).
4. Data are estimated for U.S. cars . In 1976 , the average was a higher 3761 pounds. The reduction is equal to the smaller amount of steel used now . Motor Vehicle Manufacturers Association, Facts and Figures 74 ( 1988 ).
5. Statistical Yearbook for Asia and the Pacific,  Y.B. on Energy Statistics, U.N. Doec . ST/ ESCAP/427; III U.S. Dep't of the Interior , Bureau of Mines , 1986 Minerals Yearbook: Area Reports, International ( 1988 ).
23. See U.S. Envtl . Protection Agency, National Air Pollution Emission Estimates ( 1984 ), and Environmental Data Report, supra note 1.
25. R. Cole and T. Yakushiji , supra note 12, at 108.
26. Id . at 116.
27. Motor Vehicle Manufacturers Association ( 1988 ), supra note 4 , at 70.
28. Id ., and Motor Vehicle Manufacturers Association ( 1986 ), supra note 4 , at 69.
29. 65 Survey of Current Business 17 ( Nov . 1985 ) ; 1-0 Table 3 , supranote 10.
30. Motor Vehicle Manufacturers Association ( 1988 ), supra note 4, at 6 and 70.
31. R. Cole and T. Yakushiji , supra note 12, at 122.
32. Motor Vehicle Manufacturers Association ( 1988 ), supra note 4 , at 70.