Charles
460382012 Associates
Final Report
SCARCITY, RECYCLING AND SUBSTITUTION OF POTENTIALLY
CRITICAL MATERIALS USED FOR VEHICULAR EMISSIONS CONTROL
Prepared for -
U.S. Environmental Protection Agency
2565 Plymouth Road
Ann Arbor, Michigan 48105
CRA.Report No. 501
Prepared by
Charles River Associates
200 Clarendon Street
Boston, Massachusetts
February 1982
-------
Charles
River
Associates
TABLE OF CONTENTS
Page
Chapter 1: INTRODUCTION AND SUMMARY 1-1
Chapter 2: MATERIALS CRITICALITY AND EPA POLICYMAKING 2-1
Critical Materials and Strategic Materials 2-1
Market Contingencies That May Be the Basis for Considering
A Material Critical 2-2
Disrupted Factors of Production 2-3
Inadequacy of Reserves and Resources in the Long Run. . . . 2-4
Technological Shocks 2-4
Disrupted Foreign Sources of Supply 2-5
The Business Cycle 2-6
Defense 2-6
Principles of Measuring Materials Criticality and Implications
for EPA Pol icymaking 2-7
Materials Prices and Compliance Costs 2-7
Compliance Costs and the Critical ity of Materials Used
to Meet EPA Regulations 2-9
The Criticality of Materials from a National Perspective. . 2-14
Income Redistribution and Noneconomic Dimensions of
Materials Criticality 2-21
Role of Secondary Production and Inventories 2-22
Adequacy of Private Adaptations to the Threat of Supply
Disruptions and Other Contingencies 2-24
Implications for EPA Policymaking 2-27
Should EPA Sponsor R&D on Vehicular Emissions Control?. . . 2-31
A Major Qualification and A Possible Policy Prescription. . 2-32
Material Imports and Balance of Payment Problems 2-34
A Simple Economic Model for Estimating Criticality Due To
Foreign Supply Disruptions 2-36
Key Parameters and Formulas 2-37
Treatment of Stockpiling 2-43
Meaningful!ness of Results from the Model 2-46
Sample Estimates of the Critical ity of Platinum, Palladium,
Rhodium, Chromium, Manganese, Nickel, and Titanium Metal . . . 2-51
Normal Consumption, Production, and Prices 2-52
Contingency Threats and Price Elasticities 2-54
Price Elasticities of Consumption for Vehicular Emissions
Control 2-59
Illustrative Criticality Estimates and Conclusions 2-60
Bibliographic Note 2-64
Appendix 2-A: Availability of Materials from the U.S. National
Stockpile 2-67
Appendix 2-B: Computer Program to Calculate Average Annual
Economic Losses from Contingencies in Material Markets .... 2-70
Appendix 2-C: Earlier Approaches to Materials Criticality . . . 2-79
Chapter 2 References 2-117
-------
Charles
River
Associates
TABLE OF CONTENTS (Continued)
Page
Chapter 3: PROJECTIONS OF MATERIALS CONSUMPTION FOR CONTROL
OF VEHICULAR EMISSIONS 3-1
Appendix 3A: EPA ESTIMATES OF CONSUMPTION OF PLATINUM-GROUP
METALS FOR CONTROL OF VEHICULAR
EMISSIONS IN 1981 3A-1
Chapter 4: PLATINUM-GROUP METALS 4-1
Introduction 4-1
Stocks 4-1
U.S. National Stockpile 4-2
Refiner, Importer, and Dealer Stocks 4-2
Industry Shelf Stocks 4-7
Industry Stocks in Use 4-9
Private Speculative/Investment Stocks 4-11
Stocks and Increased Demand in the Short Run 4-11
Statistical Overview of Supply and Demand 4-12
Supply 4-12
Demand. 4-35
Consumption Trends in the 1970s 4-39
Prices 4-47
Primary Producers 4-54
South Africa 4-54
The Soviet Union 4-59
Canada 4-61
United States 4-62
Colombia 4-63
Other Countries 4-63
World Reserves 4-63
Supply Reliability 4-64
Consumption Elasticity and Secondary Recovery 4-65
Petroleum Reforming 4-66
Petroleum Cracking 4-66
Nitric Acid Production 4-66
Chemical Processes Other Than Nitric Acid 4-67
Telephone Switching Equipment 4-67
Dental and Medical Uses 4-68
Electrical -- Other Than Telephone Switches 4-68
Glass 4-68
Price Elasticities for Calculating the Critical ity of Platinum,
Palladium, and Rhodium 4-69
Speculation and Increased Demand for Platinum-Group Metals . . 4-69
Appendix: Annotated Bibliography and Guide to Sources of
Information on Platinum-Group Metals 4-72
Chapter 4 References 4-76
-------
Charles
River
Associates
TABLE OF CONTENTS (Continued)
Page
Chapter 5: RECYCLING OF PLATINUM-GROUP METALS FROM CATALYTIC
CONVERTERS 5-1
Background on the Function of the Catalytic Converter 5-2
The Characteristics of Catalyst Material 5-2
Cost Breakdown For New and Used Converters 5-3
Recycling Catalytic Converter 5-5
Phase I: Rejected Catalysts 5-5
Phase II: Replacement After 50,000 Miles 5-6
Phase III: Auto Catalyst From Salvaged Autos 5-6
Spent Catalyst PGM Refining 5-9
Future Availability of PGMs From Scrapped Converters. ... 5-11
Chapter 5 References 5-12
Chapter 6: SUBSTITUTES FOR PLATINUM-GROUP CATALYSTS IN VEHICULAR
EMISSIONS CONTROL 6-1
Background 6-1
Oxidizing Catalysts 6-2
NOX Removal Catalysts 6-4
Base-Metal Catalyst Research 6-5
Summary on Possible Replacement of Platinum-Group Metals . . . 6-6
Chapter 6 References 6-8
m
-------
Charles
River
Associates
H
INTRODUCTION AND SUMMARY
Technologies currently being employed in the United States to control
vehicular emissions require large expenditures on imported platinum-group
metals, and smaller expenditures on other materials sometimes characterized
as "critical." There is considerable concern about the reliability of
foreign supplies of many of these materials, and also about implications for
the U.S. balance of trade. In order to carry out its regulatory functions
efficiently, the U.S. Environmental Protection Agency must weigh these
concerns appropriately, which involves quantifying the important costs and
benefits to the United States that are associated with materials consumption.
It is important to recognize all the costs imposed by the potential
unreliability of foreign sources of materials supply, but it is also
important not to overestimate the importance of these costs (as some parties
may attempt to do when it is in their interest). In this study we explain
how EPA can quantitatively estimate the criticality of materials, and factor
those estimates directly into decisions about regulations controlling
vehicular emissions.
As groundwork for our analysis, subcontractor Rath and Strong projects the
quantities of platinum-group metals and other potentially critical materials
that will be used for vehicular emissions control in the United States during
the 1980s. In order to obtain a full assessment of the issue, we also
consider in some detail two further topics. The first topic is the current
extent of recycling of platinum-group metals and stainless steel scrap from
1-1
-------
Charles
River
Associates
catalytic converters; we also give particular attention to the extent of
recycling that is likely in the future. The second topic is the
technological possibilities for using emissions control technologies that
require smaller amounts of platinum-group metals.
The basic contents of this study are organized into five chapters following
this introductory chapter. Chapter 2 develops our main conclusions and
recommendations regarding how the Environmental Protection Agency can factor
into its policy decisions their effects on material markets. Our main
conclusion is that EPA must estimate compliance costs carefully, taking into
account likely increases in prices of materials due to EPA-induced demands,
and also include some adjustment for the probability of large (but usually
temporary) increases in prices of materials due to future market
contingencies such as foreign supply disruptions. By and large, however, it
is not reasonable to expect EPA to do more than estimate likely future costs
of compliance as these depend on likely future costs of materials.
Detrimental effects on the U.S. balance of trade from increased importation
of materials to satisfy EPA regulations is a minor consideration that can
usually be given low priority by EPA.
U.S. policymaking in response to the criticality of materials is more
efficiently undertaken at the national level, if it is justified at all.
National tariffs and stockpiles are the appropriate policy instruments, not
ad hoc decisionmaking by each individual government agency whose decisions
affect total U.S. consumption of materials.
Chapter 2 defines a "critical material" simply as one for which contingency
planning is worthwhile. If the contingency is a military conflict, then the
material is also "strategic." Chapter 2 then presents a comprehensive list
of nonmilitary contingencies that may justify preparatory planning, that is,
nonmilitary contingencies that may be the basis for considering a material
critical. Contingencies that are potentially important from the perspective
of consumers include mine disasters, labor strikes, equipment failure,
depletion of reserves, new competing demands for a material, or just
processing bottlenecks due to unexpectedly large total demand for the
material. However, the type of contingency that most often makes a material
highly critical is the threat of foreign supply disruptions.
Our analysis of the criticality of platinum-group metals and other materials
used for vehicular emissions control in fact concentrates upon the
unreliability of imported supplies. The seven materials for which we
quantitatively estimate criticality are:
1-2
-------
Charles
River
Associates
• Platinum;
• Palladium;
• Rhodium;
• Chromium;
• Manganese;
• Nickel; and
t Titanium metal.
Only in the case of titanium metal are we more concerned about problems other
than foreign supply disruptions.
The quantitative measure of a material's criticality is the expected future
cost per year due to the contingencies threatening the market, averaging out
years in which the contingencies do and do not occur. We provide
illustrative estimates for each of the materials listed above. In principle,
the expected cost of contingencies can be multidimensional, and include
noneconomic costs such as greater U.S. pollution from increased domestic
production of materials, or even from relaxation of vehicular emissions
standards, were that to be deemed likely. However, for this study we
concentrate on the strictly economic costs associated with contingencies in
material markets, usually disruptions in foreign supplies. Thus, we measure
criticality strictly in terms of expected dollar losses per year.
Criticality of a material can be measured from the perspective of a
particular end use, such as control of vehicular emissions, or from the
perspective of the nation as whole. When criticality is calculated from a
national perspective, it must be recognized that the same contingencies that
inflict costs on U.S. consumers of materials will often benefit U.S.
producers of those materials. This is generally the case where foreign
supply disruptions are the contingency of concern, so that some balancing of
criticality from the perspectives of consumers versus producers is required.
For purposes of illustrative calculations, we assume a dollar gained by U.S.
producers of a material compensates for a dollar lost by U.S. consumers.
For a contingency such as a foreign supply disruption, the key parameters of
the criticality estimation that we explicitly recognize in our calculations
are as follows:
1-3
-------
Charles
River
Associates
• The severity of contingencies that threaten U.S. consumers (or
producers) of a material, as measured by price increases (decreases)
that occur;
• The expected time between occurrences of these contingencies;
• The duration of these contingencies;
• The quantity of the material consumed in the United States under normal
conditions;
• The quantity of the material produced in the United States under normal
conditions, both from primary and secondary sources;
• The extent to which U.S. consumption can be reduced when prices rise
(the "elasticity" of short-run U.S. demand);
• The extent to which U.S. primary and secondary production can be
increased when prices rise (the "elasticity" of short-run U.S. supply);
and
• The size of normal U.S. inventories and stockpiles.
The most important determinants of the criticality that we estimate for the
seven materials listed earlier are subjective estimates of the severity,
frequency, and duration of a "representative" contingency for each market
(that is, the first three of the items listed above). Unlike the other
parameters of the critical ity estimates, it is unfortunately not possible to
estimate these parameters with any precision from historical data or
engineering analysis.
Not surprisingly, from the perspective of consumption for vehicular emissions
control it turns out that platinum is the most critical of the materials we
consider, by more than an order of magnitude. However, the criticality
penalty is very small relative to the apparent disadvantages of alternative
emissions control technologies, as described later in the report. Thus, the
criticality of platinum-group metals to the United States is not a strong
reason to discourage use of these materials for emissions control.
Moreover, our simplified methodology for calculating illustrative measures of
materials criticality does not take into account the fact that current
consumption of platinum-group metals for emissions control creates a "rolling
stockpile" of the material that will increasingly be available through
recycling in the future. If this consideration were factored into the
analysis, using a more sophisticated economic model, the criticality of
platinum-group metals used for emissions control would be considerably less.
(The same is true of the alloying metals used in the 409 stainless steel of
catalytic converters, since most of that material will be recycled in the
future as well.)
1-4
-------
Charles
River
Associates
The analysis in Chapter 2 summarized above draws on background information
presented in Chapters 3 through 6. Chapter 3 summarizes projections of U.S.
materials consumption for vehicular emissions control developed by
subcontractor Rath & Strong. Consumption of platinum and palladium is
projected to drop off somewhat from the high rates of 1980 and 1981, but
still represent a very sizable share of total U.S. consumption of those
metals. U.S. consumption of rhodium is projected to rise substantially by
the mid-1980s. Consumption for vehicular emissions control of other
materials is very small, both relative to total U.S. consumption of those
materials, and in terms of the share of compliance costs for which it
accounts (or might plausibly account in during future times of shortages and
high prices).
Chapter 4 presents extensive information on the markets for platinum-group
metals, particularly platinum, palladium, and rhodium. The most important
considerations for this study are U.S. reliance on potentially unreliable
exports from South Africa and the Soviet Union, and the very large price
increases necessary to induce most users of these materials to reduce
consumption to any significant extent. Ameliorating U.S. vulnerability to
foreign supply disruptions are substantial stockpiles maintained by most
consumers, and very high rates of secondary recovery, making most
applications of platinum-group metals interpretable as "stocks in use"
(rather than "consumption").
Chapter 5 assesses publicly available information about the young industry
recycling platinum-group metals from obsolete catalytic converters in the
United States. Taking into account the cost of gathering and processing used
converters, and the losses and contamination that occur during use, recovery
of platinum-group metals from this source appears to be only moderately
profitable at 1980 prices for platinum-group metals. Decreases in the prices
of platinum-group metals which would be large by historical standards, but
are conceivable, could make recovery at least temporarily uneconomical.
Finally, Chapter 6 discusses possible substitutes for platinum-group
catalysts in vehicular emissions control. There is no published evidence
that a catalyst system using only base metals could match the performance of
the present three-way catalyst system, leaving aside the question of
durability. It probably would be possible to design an oxidizing catalyst
system using only base metal catalysts that would meet 1980 standards for
emissions of carbon monoxide and hydrocarbons, but it would almost certainly
not meet 1981 standards. The unit would have to be somewhat larger than
present emissions converters using noble metals, and might well be more
costly at normal prices for materials. Most avenues for using base metal
catalysts in place of noble metal catalysts were investigated in the early
1970s and found (with a high degree of assurance) to be too unpromising to
justify further research. General Motors contin-ies to support research on
zeolite catalysts by Professor Hall at the University of Wisconsin, but
results are much too preliminary to warrant optimism.
1-5
-------
Charles
River
Associates
MATERIALS CRITICALITY AND EPA POLICYMAKING
This chapter contains the main conclusions of our study, drawing on
background information presented in later chapters. We first define what a
"critical" or "strategic" material is, and then describe market contingencies
that can justify such a designation. We explain how criticality can be
measured in terms of expected dollar losses per year, from the perspective of
consumption for vehicular emissions control or from the perspective of the
nation as a whole, and actually perform illustrative calculations for seven
elemental materials. We describe how the criticality measure can be factored
into EPA policymaking, particularly through its role in estimation of
compliance costs for EPA regulations. We also draw out implications for more
general EPA policy issues such as the likely adequacy of research on
emissions control undertaken by U.S. vehicle manufacturers, and possible
inadequacies in auto manufacturers' stockpiling of critical materials used
for emissions control.
CRITICAL MATERIALS AND STRATEGIC MATERIALS
The term "strategic and critical material" was institutionalized in 1939 by
the original legislation that established the U.S. National Defense
Stockpile. (See Appendix 2-A.) The currently effective version of that
legislation (as amended in 1979) defines strategic and critical materials as
those that "(A) would be needed to supply the military, industrial, and
essential civilian needs of the United States during a national emergency,
and (B) are not found or produced in the United States in sufficient
quantities to meet such a need."
2-1
-------
Charles
River
Associates
The more general term "critical material" has been used in a variety of
contexts by many different commentators and analysts. Webster's New
Collegiate Dictionary defines "critical" in this context to mean:
"indispensable for the weathering, solution, or overcoming of a crisis," and
in fact gives "the stockpiling of strategic and critical materials" as an
illustrative application of the term. Operationally, a material tends to be
described as "critical" for national policymaking if future events threaten
to inflict serious damage on the United States, and current planning and
policies can reduce the expected costs associated with the threat. If the
threatening event is a military conflict, the material is also described
as "strategic." Webster's New Collegiate Dictionary defines "strategic" in
this context to mean: "required for the conduct of war," and again gives
"strategic material" as an illustrative application of the term.
Our usage of the terms "critical" and "strategic" is entirely compatible with
general usage, as expressed in the above dictionary definitions. Since we
will mainly be concerned with nonmilitary contingencies in material markets,
we will focus on the "criticality" of materials. Our primary objective is to
provide an operational, quantitative definition of materials criticality,
that is as decisive as possible for determining what materials should be used
for vehicular emissions control, and also for assessing the likelihood that
the private sector will make the appropriate choices. (Alternative
definitions of materials criticality, and methods of measuring it, are
briefly surveyed in Appendix 2-C and in the Bibliographic Note at the end of
this chapter.)
The reader should be warned that terminology tends to be variable and
changing in this area. Currently, much publicity has been generated about
schemes through which private individuals can readily invest in stockpiling
of "strategic metals." In this context, a strategic metal tends to be an
imported metal with defense applications whose market volume is not large and
whose market price is volatile.
MARKET CONTINGENCIES THAT MAY BE THE BASIS
FOR CONSIDERING A MATERIAL CRITICAL
For the U.S. Department of the Interior, CRA is currently conducting a study
of methods for detecting and evaluating emerging problems in material
markets.* One early output of that study was a comprehensive categorization
of issues, problems, and contingencies in material markets that might inflict
losses on U.S. citizens, and thus be of concern to the Department of the
Interior or other federal agencies. The discussion that follows draws on
this related project to describe general types of market contingencies,
*See Charles River Associates (forthcoming).
2-2
-------
Charles
River
Associates
preparations for which may benefit from special planning by private firms and
government agencies. In other words, we now categorize and describe market
contingencies that can be the basis for considering a material "critical."
Despite the potentially broad applicability of the concept of materials
critical ity, there is one type of nonmilitary contingency that outstrips the
others in prominence, namely disruptions in foreign supplies of materials
that the United States imports. This preeminence of foreign supply
disruptions as the market contingency of concern is particularly true for the
materials of central interest in this study: platinum-group metals are
supplied to western markets predominantly by South Africa and the Soviet
Union, neither of whose reliability is unquestioned, though the nature of the
contingencies of concern is quite different for the two countries. Other
materials used in vehicular pollution control equipment, such as chromium for
stainless steel, are also obtained from potentially insecure foreign sources.
(For example, South Africa is the largest exporter of chromium as well as
platinum; the Soviet Union was an important exporter of chromium before 1975,
and Albania is now an important supplier to the West.)
We now systematically review all the major types of market contingencies, the
threat of which could conceivably be the basis for considering a material
"critical."
DISRUPTED FACTORS OF PRODUCTION
Production of minerals, like other goods and services, is interpreted by
economists to depend ultimately on the use of three main types of inputs or
"factors of production": land, labor, and capital. In addition, other
materials, transportation services, etc. are purchased from other firms (who
themselves use land, labor, and capital). Disruption of any essential factor
of production can stop production or delivery of a material, though
disruption of some factors are considerably more likely than others.
(Foreign producers, particularly those in less developed countries, are
considerably more prone to disruptions of this type than are domestic
producers; we distinguish problems with foreign production as a separate
category below.)
The most notable "land" used in the production of materials is of course the
mineral deposits from which metals and other materials are produced.
Deposits can be made suddenly unavailable through natural disaster, mine
accidents, or even sabotage. However, this type of disruption has not been
very important historically in the United States.
By far the most common contingency affecting the availability of labor for
mineral production is the deliberate strike, often organized by labor unions.
One can conceive of other contingencies, such as disease, having an effect,
but such events have not been important historically.
2-3
-------
Charles
River
Associates
Once capital equipment is installed, it can break down and disrupt
production. For major metals, this type of risk is spread so widely as to
render it unimportant. For materials produced from only a few sources, it
can be a problem from time to time, though usually one that is only temporary
and modest in proportions.
Materials and services purchased from other firms can also disrupt producers
of materials. Transportation routes can be severed by landslides or
breakdown of equipment. There may just be sudden competition for cargo space
from highly valued commodities that can afford to pay more for
transportation. For example, in South Africa, platinum is sufficiently
valuable as to justify transportation of refined material by airplane if
necessary. However, other minerals produced in South Africa, such as
chromium ore, are much more bulky, and shipments have at times been
significantly delayed by the priority shipment of other goods, such as
agricultural commodities in season.
Sharp increases in the prices of some inputs can be as disruptive to mineral
production as cutoffs in other inputs. The most prominent example in recent
years has been the sharp increase in the prices of petroleum and other forms
of energy. OPEC has claimed at times that it was embargoing the United
States and other regions of the world, but there is little evidence that this
strategy has ever been effective (in the sense of imposing much greater costs
on the embargoed regions compared to other importing countries). Energy
availability may also be subject to more localized disruptions. For example,
hydroelectric production of energy can be disrupted by low rainfall.
INADEQUACY OF RESERVES AND RESOURCES IN THE LONG RUN
For a few materials, published estimates of reserves and resources are
sufficiently low so there might appear to be some chance of "running out" in
a decade or two, before its use to meet EPA regulations has terminated. (A
"resource" qualifies as a "reserve" if production therefrom is economically
viable.) This fear is usually misplaced, because exploration can expand
reserves, and new technologies can make economical the exploitation of
previously uneconomical deposits. Because it generally emerges so slowly,
this problem is usually not included in the calculation of materials
criticality, though it should be considered when estimating the likely future
cost of complying with EPA regulations. Thinking in terms of "contingency
planning" is usually not that useful in this case.
TECHNOLOGICAL SHOCKS
The technological contingency of greatest concern to current consumers of a
material is usually the possibility that a very large new use will emerge,
resulting in an escalation of price. Fortunately, consumption of materials
2-4
-------
Charles
River
Associates
for new technologies tends to grow sufficiently slowly, or with sufficient
warning, so that major price increases can usually be avoided through
development of new productive capacity. The use of platinum as a catalyst
for pollution-control equipment is a case in point. On the other hand,
rhodium is produced as a byproduct, and has experienced relatively greater
price increases because supply is limited by the amount of platinum
production. Like inadequacy of reserves, this type of problem tends to
emerge sufficiently slowly so that it is usually not explicitly included in
the calculation of a material's criticality, though it should be factored
into the analysis of materials costs at some point, as we discuss further
below.
DISRUPTED FOREIGN SOURCES OF SUPPLY
Foreign producers of materials, particularly those in less developed regions,
are subject to all the contingencies described above, often to a considerably
greater extent than domestic producers. Localized military conflicts and
sabotage can cause major disruptions. Also, foreign producers of materials
are often not constrained by law from acting in a glaringly monopolistic
fashion, and their governments may even attempt to use their exports as a
political weapon.
There was concern during the 1970s that foreign exporters of minerals and
metals would use OPEC as a model and form effective cartels. Foreign
producers of metals such as aluminum and copper have attempted to increase
their joint monopoly power by forming producers' associations, but these
organizations have had very limited success. (For an extensive analysis of
the difference between OPEC and mineral producers, and the limited ability of
the latter to collude, see Charles River Associates (1976).) If foreign
producers of a mineral or metal organize effectively, there is a theoretical
possibility of "price gouging," where prices are raised very high, very
rapidly, in order to catch consumers before they can change their consumption
patterns. However, most foreign producers appreciate sufficiently the
detrimental long-term effects of such a policy, so that it has not been
common.
Most monopoly power in mineral markets is due to "natural" monopoly power,
stemming from the fact that one country has low production costs and a
dominant market share. This situation is often not too damaging to consumers
over time, since the dominant producer must keep prices low enough to
preclude entry by major competitors. Also, this situation tends to be quite
stable, as the low-cost dominant producer is greatly interested in
encouraging consumption of his material. Monopoly power can be exercised by
private companies or by foreign governments. One of the neatest ways for a
foreign government to accomplish this objective is simply to impose tarrifs
on exports of a mineral, which raises the world price, and funnels monopoly
profits directly into the national treasury. Caribbean producers of aluminum
ores have favored this technique. (The U.S. Constitution forbids export
tariffs in this country.)
2-5
-------
Charles
River
Associates
Moderate collusion among foreign producers is often tacit rather than overt.
This is probably the case in the platinum market. South African producers
are not explicitly organized, but each realizes that an attempt to expand its
market share may invite retaliatory expansion by its competitors. As a
group, they are better off if capacity expansions are planned
conservatively.
If, as is often the case, monopolistic aspects of foreign production are
stable over time, they may not give rise to sudden contingencies (such as
price gouging) and criticality planning of the type we develop below may not
be necessary. Monopolistic world prices will be continually somewhat higher
than they would be in a competitive environment, but consumers may rationally
come to accept that situation as a fact of life, rather than a contingency
subject to sudden reversal that requires planning.
Petroleum exports have in recent years been used with some success as a tool
of foreign policy, but there is no analogous example among mineral markets.
The value of world petroleum imports dwarfs trade in even major metals such
as iron, copper, and aluminum. The Soviet Union stopped exporting metals to
the United States prior to the Korean War, but no serious damage was
inflicted on this country. The United Nations attempted to impose an embargo
on Rhodesian chromium during the 1960s, but it was not enforceable.
THE BUSINESS CYCLE
The rate of economic activity in western market economies fluctuates
considerably, inducing large changes in the consumptions and prices of
materials. High prices (or even unavailability) of materials due to booming
competing demands can seriously affect some consumers in much the same way as
a foreign supply disruption. However, consumers with long-term contracts
with suppliers of materials may be somewhat protected. Materials whose
consumption is particularly sensitive to the rate of economic activity are
sometimes considered "critical," but this is far from a universal practice.
(We later consider the criticality of titanium metal from this viewpoint.)
DEFENSE
The premier contingency making a material "critical" is war. As discussed
above, the material is called "strategic" in that case. The costs imposed on
the United States by most of the contingencies described above are
predominantly economic, and so the resulting criticality can be analyzed
using economic theory. The costs of losing a war cannot be measured in
purely economic terms, so determining the extent to which a material is
strategic involves other considerations with which we do not attempt to deal.
Our analysis of materials used to meet EPA regulations will concentrate upon
peacetime contingencies.
2-5
-------
Charles
River
Associates
PRINCIPLES OF MEASURING MATERIALS CRITICALITY
AND IMPLICATIONS FOR EPA POLICYMAKING
Regulations implemented by the Environmental Protection Agency affect U.S.
material markets in a variety of ways. The most widely publicized effects on
material markets involve direct regulation, such as those telling producers
of materials the maximum amounts of specified air pollutants they are allowed
to emit. However, EPA rulemaking can also affect material producing
industries more indirectly, simply by increasing the demand for materials
which other industries need to meet EPA regulations. This study deals with
this latter issue, in particular increased demands for platinum-group metals
and other materials used by motor vehicle manufacturers to meet EPA emissions
standards.
Obviously, some important aspects of the availability of materials for EPA
regulations are normally factored into EPA's rulemaking processes, in
particular the prices of materials as they affect projected costs of
complying with the rules. Once EPA standards have been set, the regulated
industry has a continuing incentive to minimize the cost of compliance,
taking into account the prices of materials. The key question we consider in
this chapter is whether there is some aspect of the availability of
materials, beyond the inclusion of their price in estimated compliance costs
(such as their "criticality"), that EPA should take into account when making
rules and regulations. Our ultimate conclusion to this question in most
cases is "no." However, there can be exceptions, and in any case there is
often controversy surrounding decisions on this issue. Thus, in the
remainder of this chapter we construct a fairly elaborate rationale for our
conclusions, and also describe the possible exceptions.
MATERIALS PRICES AND COMPLIANCE COSTS
Predicting future compliance costs for a new or proposed EPA regulation, as
it depends on material prices, should not be a matter of simply referring to
the most recent price quote in Metals Week or the Chemical Marketing
Reporter. The current price of a material may be significantly above or
below the long-run market equilibrium, often because general economic
activity in the industrialized consuming countries (as measured by their
GNPs, for example) happens at the moment to be significantly above or below
the historical trend. Or there may be some transient supply problem, such as
a labor strike or a transportation bottleneck, causing current prices to be
significantly above prices that are likely in the future.
Usually, the most relevant basis for predicting future compliance costs is
the long-run equilibrium price of a material, where producers are earning an
adequate, but not excessive, rate of return. Predicting long-run equilibrium
2-7
-------
Charles
River
Associates
prices for materials under normal market conditions is not a trivial task if
the highest possible degree of accuracy is required. However, it is a
manageable problem. Here, we concern ourselves with whether these are
important further issues for EPA to consider, beyond simply accepting the
normal cost of materials implied by the methods of compliance chosen by the
regulated private sector.
Where EPA regulations will require relatively large increases in world
production of a material, one obvious complication is the effect of EPA
induced demands on the long-run equilibrium price of the material. For most
materials, long-run supply is very "elastic" in its response to price
incentives. In economists' jargon, that means that a very modest increase in
the market price of a material will eventually be a sufficient incentive for
its producers to increase greatly their output, allowing a number of years
for unrushed capacity expansions, and perhaps even additional time to explore
for new reserves. It may even be possible to increase greatly the production
of a material and eventually to have a lower market price than previously, if
expanded production allows increased economies of scale or induces advances
in processing technologies. The most notable exceptions to the above
generalization (that materials are available in very elastic supply in the
long run), occur when a material is produced as a byproduct of another
material. In that case, increased market prices may lead to very little
additional production. This consideration is particularly relevant for the
more minor of the platinum-group metals, most notably rhodium.
In addition to setting regulations to be met in the "long run," EPA must also
decide upon how quickly to impose standards of a given stringency. More
rapid imposition of a standard may lead to more rapid increases in the demand
for materials and short-run increase in their prices. These short-run price
increases will recede after capacity to produce the materials has expanded,
but they do imply temporarily higher compliance costs from faster
implementation of regulations. For simplicity in the following discussion,
we usually abstract away from the additional difficulties associated with
analyzing the speed of implementing EPA regulations, and consider only the
costs of regulations after they have been in effect for a while. However,
our conclusions about the adequacy of considering materials criticality only
to the extent that compliance costs are affected, also generally hold true
when evaluating the overall costs (and benefits) of different possible speeds
of implementation.
A final consideration relevant for predicting the costs of materials used
to comply with EPA regulations is the probability of contingencies such as
major foreign supply disruptions, that can greatly increase the price of a
material for a number of months, or even years. As discussed above,
significant susceptability to such contingencies qualifies a material to be
2-8
-------
Charles
River
Associates
considered "critical." We now consider how such considerations should in
principle be included in the normal course of predicting compliance costs for
EPA regulations, which requires analyzing how private firms include these
considerations in their choices among alternative methods for complying with
EPA regulations. This same analysis will be the basis for quantifying the
criticality of a material.
COMPLIANCE COSTS AND THE CRITICALITY OF MATERIALS
USL-U TO MELT EPA REGULATTUNF
Figure 2-1 presents a concrete example of the materials component of
compliance costs, in terms of the amount of an imaginary material
"catalystium" used to meet EPA vehicular emissions standards. The
illustrated catalystium "demand curve" assumes that when the market is in a
normal undisrupted state, the world market price is $100 per ounce, and
10,000 ounces per year are purchased to meet EPA regulations. However, we
suppose that there is a threat that the producers of catalystium located in
a foreign country will be disrupted by a localized military conflict. For
the sake of simplicity, we assume we know that the world market price rises
from $100 per ounce to $300 per ounce during such disruptions, and that the
disruptions last exactly one year. We further assume that the probability of
a disruption occurring in any future year is 0.1.
If producers of emissions control equipment continued to purchase 10,000
ounces of catalystium per year during disruptions, then it is very easy to
include the effect of the supply disruptions in the calculation of average
expected compliance costs in future years. The expected future price of
catalystium would be 0.9($100) + 0.1($300) = $120 per ounce, averaging out
years in which supply disruptions do and do not occur.
However, the example illustrated in Figure 2-1 is a bit more realistic. It
assumes that producers of emissions control equipment can cut back somewhat
on the use of catalystium when its price rises suddenly. A variety of
design changes may allow reductions in the use of catalystium, but for
simplicity we can assume here that simply using a greater proportion of
another material in the equipment allows EPA regulations to be met. Using
this greater proportion of the alternative material is not economical
(i.e., does not minimize compliance costs) when catalystium costs $100 per
ounce, but it is economical when catalystium costs $300 per ounce, and (we
suppose) the switch in technologies can be made rapidly after such a price
increase occurs.
How can the calculation of compliance costs take into account disruptions in
the supply of catalystium in this more complicated case? One straightforward
2-9
-------
Charles
River
Associates
Figure 2—1
CONSUMPTION OF "CATALYSTIUM" TO COMPLY WITH EPA STANDARDS,
AND ECONOMIC LOSSES FROM SUPPLY DISRUPTIONS
(A HYPOTHETICAL EXAMPLE)
PRICE ($/ounce)
300 -
200 -
100
5 8 10
QUANTITY CONSUMED (1000 ounces/year)
2-10
-------
Charles
River
Associates
way to proceed would be to calculate costs for each of the alternative
technologies separately, then weight the cost using less catalystium by 0.1,
and the cost of the normal technology by 0.9. (That is, weight by the
relative frequencies of a disruption occurring and not occurring.) However,
it is also possible to infer the effect of disruptions in the supply of
catalystium on compliance costs, just using the demand curve for catalystium
illustrated in Figure 2-1.
In order to appreciate how this is done, consider the more artificial demand
"curve" for catalystium illustrated in Figure 2-2, which assumes that at a
catalystium price of exactly $200 per ounce it is economical to switch
entirely from the normal technology to the alternative using less
catalystium. The resulting demand curve is a step function. (The more
realistic smooth demand curve in Figure 2-1 assumes that the switch away from
catalystium occurs gradually as prices rise from $100 per ounce to $300 per
ounce.) In Figure 2-2, $200 per ounce of catalystium represents a
"break-even" price at which it is equally economical to use either of the two
technologies to meet EPA standards. We can tell from the diagram that the
normal technology costs in total ($200-$100)(10,000) = $1,000,000 more per
year, when the catalystium price is $200 rather than $100. Thus, at the
catalystium price of $200 per ounce, the alternative technology using less
catalystium must also cost $1,000,000 more than the normal technology with
catalystium at $100 per ounce. Of this $1,000,000, additional costs for the
8,000 ounces of catalystium used with the alternative technology are
($200-$100)(8,000) = $800,000 per year. Thus, the cost of changing
technologies, apart from the effect of an increase in the price of
catalystium, is $1,000,000 - $800,000 = $200,000 per year.
This line of reasoning is probably clearer in the geometric terms of
Figure 2-2. When the catalystium price rises from $100 to just below $200,
the regulated industry continues to buy 10,000 ounces at an additional cost
for the year of almost $1,000,000. This loss is represented geometrically in
Figure 2-2 as the area of the rectangle made up of the two smaller rectangles
labeled 2a and 1. When the catalystium price rises from just below $200 to
just above $200, the total cost of compliance does not increase
significantly, but switching to the alternative technology causes the costs
to be broken down into
• the additional $200,000 cost of the alternative technology, represented
as Area 1 in Figure 2-2, plus
• the additional $800,000 cost of the remaining amount of catalystium
that is purchased, represented as Area 2a in Figure 2-2.
2-11
-------
Charles
River
Associates
Figure 2-2
CONSUMPTION OF "CATALYSTIUM" TO COMPLY WITH EPA STANDARDS:
ARTIFICIAL CASE WHERE DEMAND IS A STEP FUNCTION
PRICE ($/ounce)
300 -
200 -
100
I I
2b
2a
I
5 8 10
QUANTITY CONSUMED (1000 ounces/year)
2-12
-------
Charles
River
Associates
If the catalystium price then rises from $200 per ounce to $300 per ounce,
the remaining 8,000 ounces of catalystium that are purchased will cost an
additional ($300-$200)(8,000) = $800,000 per year, represented in Figure 2-2
as Area 2b. (We are assuming in Figure 2-2 that no further reductions in the
use of catalystium are economical in this price range.) Thus, the total
increase in the cost of compliance caused by the catalystium price rising
from $100 to $300 during supply disruptions is $200,000 in adjustment costs
(represented by Area 1), plus $1,600,000 in additional costs for 8,000 ounces
of catalystium (represented by Areas 2a and 2b). The cost savings achieved
by switching to the alternative technology (rather than just continuing to
buy 10,000 ounces of catalystium at $300 per ounce), is represented in
Figure 2-2 as Area 3, amounting to $200,000 for the year of the disruption.
The above line of reasoning generalizes readily to the case where
adjustment away from consumption of catalystium is gradual, as its price
rises from $100 per ounce to $300 per ounce during supply disruptions. This
more realistic case is illustrated in Figure 2-1. Again, Area 1 ($200,000)
represents adjustment costs and Area 2 ($1,600,000) represents the additional
cost of purchasing the remaining 8,000 ounces of catalystium during
disruptions. Area 3 ($200,000) represents the cost savings achieved by
adjusting to alternative technologies, rather than just continuing to consume
10,000 ounces per year during supply disruptions.
We now have sufficient information about the much simplified example of
catalystium to illustrate the preferred approach to measuring its criticality
for vehicular emissions control, due to the threat of foreign supply
disruptions. Remember our earlier assumption that the disruptions are
expected to occur on average in one year out of ten. Relative to the
situation where normal price $100 occurs with certainty, the expected
additional costs due to supply disruptions are (0.1) ($1,800,000) + (0.9)
($0) = $180,000 in each future year. (That is, additional costs of
$1,800,000 are borne on average in one year out of ten, and no additional
costs are borne in nine years out of ten.) This expected economic cost per
year is the quantitative measure of the criticality of catalystium
consumption for vehicular emissions control.
It is clear from the above catalystium example that there are at least five
basic determinants of the criticality of a material from the point of view of
consumption for vehicular emissions control. (We will discuss other
considerations in more general terms later.) The five determinants are:
• the severity of contingencies that threaten U.S. consumers, as measured
by price increases that occur;
t the probability of the contingencies occurring;
• the duration of the contingencies;
2-13
-------
Charles
River
Associates
• the quantity of calatystium consumed in the United States at normal
prices; and
• the extent to which U.S. consumption can be reduced when prices rise
(the "elasticity" of short-run U.S. demand).
It is interesting to note that the normal price of catalystium does not
directly enter the calculations, except as a base from which to calculate
plausible price increases during disruptions. Thus, for example, simply
examining the cost share of various materials used for vehicular emissions
control may not be a reliable guide to their respective critical!'ties, though
there is some relationship between cost shares and criticality, as we explain
further below.
The above calculation of criticality was simplified in a number of important
respects, notably in ignoring the effects of inventories and recycling. We
allowed the producers of emissions control equipment to switch to more
economical alternate technologies when the price of catalystium jumped, but
we did not allow them to accumulate an economical level of inventories in
preparation for disruptions. Also, we did not allow them to use recycled
scrap to a greater extent. These considerations deserve further discussion,
but it is convenient to discuss first criticality from a national
perspective, rather than just from the perspective of producers of emissions
control equipment.
THE CRITICALITY OF MATERIALS FROM A NATIONAL PERSPECTIVE
For other consumers of catalystium, criticality is measured in exactly the
same fashion as for producers of vehicular emissions control equipment. In
Figure 2-3, the demand curve for producers of vehicular emissions control
equipment is reproduced on the left, and another (also hypothetical) demand
curve for other consumers is given in the middle of the figure. The example
assumes that it is economical for other consumers of catalystium to cut back
on their consumption by 60 percent in response to tripled prices during
supply disruptions, in contrast with producers of vehicular emissions control
equipment, who find it economical to cut back consumption by only 20 percent.
As a result, adjustment costs for other consumers during supply disruptions,
represented as Area Ib in Figure 2-3, are larger than corresponding Area la
for producers of vehicular pollution control equipment. Triangular Area Ib
represents adjustment costs of (1/2)($300-$100)(10,000-4,000) = $600,000 per
year, while Area la represents adjustment costs of only $200,000 per year (as
previously calculated). But, of course, "other" consumers of catalystium
benefit from their greater flexibility by having to pay much less for
catalystium during supply disruptions: Area 2b represents additional costs
of only ($300-$100)(4,000) = $800,000 per year, as opposed to Area 2a for
producers of vehicular emissions control equipment, which was previously
calculated to be $1,600,000 for the year of a disruption.
2-14
-------
Figure 2—3
U. S. CONSUMPTION OF "CATALYSTIUM" FOR ALL END USES, AND
TOTAL ECONOMIC LOSSES FROM SUPPLY DISRUPTIONS
CONSUMPTION FOR
VEHICULAR EMISSIONS
CONTROL EQUIPMENT
CONSUMPTION FOR
OTHER END USES
TOTAL U. S.
CONSUMPTION
PRICE ($/ounce)
ro
en
300 -
100 -
0 12 20
QUANTITY CONSUMED (1000 ounces/year)
Charles
River
-------
Charles
River
Associates
On the right side of Figure 2-3, total U.S. consumption of catalystium is
obtained by summing (horizontally, at each price) consumption for vehicular
control equipment and consumption for other end users. (Note that the
quantity scale on the total consumption graph has been compressed.) The
reader can easily confirm that costs during disruptions are additive across
all end uses. That is, the national adjustment cost, represented by Area 1,
is the sum of Areas la and Ib ($800,000), and the remaining national
expenditure for catalystium consumption during disruptions, represented by
Area 2, is the sum of Areas 2a and 2b ($2,400,000). Total national costs
borne by U.S. consumers during disruption years are thus $800,000 +
$2,400,000 = $3,200,000.
If the United States produces no catalystium, and could not do so even when
the price of imports triples for a year, then sufficient information on the
above example has been given to calculate the criticality of catalystium
from a national perspective. Remembering that disruptions are assumed to
occur on average in one year out of ten, expected U.S. losses from
disruptions are (0.1)($3,200,000) = $320,000 per year.
However, suppose that the United States produces catalystium in normal times,
and could expand output somewhat during disruptions in foreign supplies. In
that case, the criticality of catalystium will be less from a national
perspective, though consumers will still face the same expected losses
(assuming the probability and severity of price increases from disruptions
are as before). The calculation of national criticality in terms of expected
losses yields this result, by recognizing that U.S. producers benfit greatly
from foreign supply disruptions, thus facing negative criticality from the
threat of this particular contingency. We now describe how the "criticality
calculus" described above can be extended to yield this result.
Figure 2-4 gives a hypothetical U.S. supply curve for catalystium, showing
production of 4,000 ounces per year at the normal price of $100 per ounce,
and production of 5,000 ounces during years in which supply disruptions
occur, when the price on the world market is assumed to be $300 per ounce.
How much additional benefit do U.S. producers receive as a result of
producing 5,000 ounces at $300, rather than 4,000 ounces at $100? It is
clear revenues rise from ($100)(4,000) = $400,000 to ($300)(5,000) =
$1,500,000, but the additional cost of producing 1,000 more ounces must be
netted out.
This additional cost to producers can be calculated by estimating how much
additional production would occur at prices between $100 and $300 (in a way
analogous to measuring additional costs to consumers in Figure 2-2).
Figure 2-5 gives a more artificial U.S. supply curve, specifying that the
additional U.S. production of 1,000 ounces per year all kicks in at $200 per
ounce. According to this supply curve, the additional labor, materials,
energy, and other factors of production required to produce an additional
1,000 ounces of catalystium cost $200 per ounce produced, so it is economical
(profitable) to produce the additional quantity when the market price is
above $200 but not when it is below $200. Thus, the total cost of producing
2-16
-------
Charles
River
Associates
Figure 2—4
U. S. PRODUCTION OF "CATALYSTIUM," AND ECONOMIC GAINS FROM
FOREIGN SUPPLY DISRUPTIONS
PRICE ($/ounce)
300 -
200 -
100 -
QUANTITY PRODUCED (1000 ounces/year)
2-17
-------
Charles
River
Associates
Figure 2—5
U. S. PRODUCTION OF "CATALYSTIUM": ARTIFICIAL CASE WHERE
SUPPLY IS A STEP FUNCTION
PRICE ($/ounce)
300-
200 -
100-
J
••••
1
I
1
0 45
QUANTITY PRODUCED (1000 ounces/year)
2-18
-------
Charles
River
Associates
the additional 1,000 ounces Is ($200)(1,000) = $200,000 for the year of the
disruption.
In general, even for the more realistic smooth supply curve in Figure 2-4,
the cost of additional production can be estimated as the area under the
supply curve, out to the point at which production occurs. Thus, in
Figure 2-4 the cost of additional production is represented as trapezoidal
Area 1, equal to (1/2)($100 + $300)(5,000 - 4,000) = $200,000.
U.S. producers thus gain Area 2 in Figure 2-4, equal to
($1,500,000 - $400,000) - $200,000 = $900,000, during years in which the
price of catalystium rises to $300 per ounce. (If producers just continued
to produce 4,000 ounces during a disruption year, they would only gain
$800,000.)
We are now in a position to recalculate U.S. losses during years in which
foreign supplies of catalystium are disrupted, recognizing the fact that U.S.
production will expand. Figure 2-6 combines the total U.S. demand curve of
Figure 2-3 and the U.S. supply curve of Figure 2-4. When the world market
price is at the normal level of $100 per ounce, U.S. consumption is 20,000
ounces and U.S. production is 4,000 ounces, requiring net imports of 16,000
ounces per year (the horizontal distance between the supply curve and the
demand curve). At the disruption price of $300 per ounce, U.S. consumption
is 12,000 ounces and.U.S. production is 5,000 ounces, requiring net imports
of 7,000 ounces per year.
The U.S. loss areas described in earlier figures are renumbered in
Figure 2-6. Area 1 ($800,000) is adjustment costs suffered by U.S. consumers
in order to reduce consumption from 20,000 ounces to 12,000 ounces. Areas 2,
3, and 4 together ($2,400,000) represent extra payments by U.S. consumers for
the remaining 12,000 ounces that are purchased at the high $300 price rather
than at the normal $100 price. Of that total additional transfer to (all)
suppliers by U.S. consumers, Area 4 ($900,000) accrues to U.S. producers as
increased profits and Area 3 ($100,000) represents additional revenues of
U.S. producers used to cover increased production costs (beyond the $100 per
ounce that consumers normally pay). Loss Area 2 ($1,4000,000) accrues to
foreign producers as extra payment for the 7,000 ounces of catalystium that
are still imported.
We should mention that the method described above for measuring the cost
imposed by a large increase in the price of a material may require
supplementation where adjustment to the disruption involves dismissal of
workers. The standard assumption implicit in the above methodology is that
these workers can find alternative employment at comparable wages. Where
this assumption is significantly overoptimistic, the additional cost of
unemployed labor should be added when calculating the criticality of a
material from a national perspective.
2-19
-------
Charles
River
Associates
Figure 2-6
TOTAL U. S. CONSUMPTION AND PRODUCTION OF "CATALYSTIUM," AND
NET ECONOMIC LOSSES FROM FOREIGN SUPPLY DISRUPTIONS
PRICE ($/ounce)
300 1
200 -
100
QUANTITY (1000 ounces/year)
2-20
-------
Charles
River
Associates
It may have already occurred to the reader that if the United States started
off as a net exporter of catalystium, rather than a net importer, then the
gains accruing to U.S producers from a foreign supply disruption will be
larger than the losses suffered by U.S. consumers, as U.S. producers export
more onto the world market at much higher prices. Thus, a material can be
quite critical to U.S. consumers, and yet have "negative" criticality from a
national perspective. This balancing of gains and losses by various groups
within the United Sates requires closer examination, to which we now turn.
INCOME REDISTRIBUTION AND NONECONOMIC DIMENSIONS OF MATERIALS CRITICALITY
Netting gains of U.S. producers against losses of U.S. consumers, each
calculated in dollar terms, is a common procedure, but it may be preferable
for purposes of national policymaking to keep separate accounts of these
gains and losses. For example, as a value judgment, legislators may feel
that a dollar gained by U.S. producers is not as important as a dollar lost
by U.S. consumers. The implied value judgment is that a redistribution of
income from U.S. consumers to U.S. producers is undesirable, rather than
being the neutral consideration that the netting procedure would require.
If the analyst keeps separate accounts of losses and gains by U.S. consumers
and producers, then a "multi-dimensional" measure of criticality results.
The user of the multi-dimensional measure can then apply his or her own
weights to losses suffered by various groups of consumers and producers, in
order to calculate a single summary measure of the criticality of various
materials (as is generally required to make final policy decisions). But,
of course, this summary measure will generally be somewhat different from
that which results from applying another person's "weights" (value
judgments).
Other, noneconomic effects of disruptions in material markets may make it
desirable to measure the criticality of materials in additional dimensions
that are not even denominated in dollar terms. Continuing with our earlier
example, suppose that increases in the world price of catalystium from $100
to $300 causes Congress to relax vehicular emissions standards. In that
case, consumption of catalystium would decrease during disruptions more than
previously, and, as our earlier diagrams indicate, the direct economic losses
from disruptions measured in dollar terms would be less. However, the
noneconomic effects of increased vehicular emissions due to relaxed standards
would be considered a cost of the disruption by most of the U.S. populace.
Some estimate of increased air pollution would then be an appropriate
additional dimension for a criticality measure used for national
policymaking. The prime example of a noneconomic dimension of materials
criticality concerns its usefulness for military contingencies (that is, the
extent to which it is "strategic"). Obviously, the cost to the United States
of being less well prepared for war cannot be measured entirely in dollar
terms.
2-21
-------
Charles
River
Associates
It is necessary to understand that a measure of materials criticality can be
extended to include noneconomic dimensions, but we cannot explore all such
possibly useful generalizations here. Hereafter we will simply assume that
all costs imposed on U.S. citizens by contingencies of concern in the
catalystium market are strictly economic, and that a dollar gained by a U.S.
producer compensates for a dollar lost by a U.S. consumer. Under these
assumptions, we can conclude that the criticality of catalystium in the case
allowing for U.S. production (as illustrated in Figure 2-6) equals expected
annual loss (0.1)($3,200,000 -$900,000) = $230,000 per year (that is,
averaging out years in which disruptions do and do not occur).
ROLE OF SECONDARY PRODUCTION AND INVENTORIES
Two important activities occurring in U.S. material markets that we chose,
for simplicity, not to include explicitly in the above illustrative
calculations for the catalystium market are inventory adjustments and
secondary recovery (recycling). It is straightforward to include secondary
recovery in the analysis in a roughly appropriate way, by simply including
secondary production with primary production (from mines) in the supply
curves illustrated in Figures 2-4 and 2-6. Just as for primary U.S.
production, the criticality of a material from a national perspective is
reduced the greater the amount of recycling in normal times, and the greater
the extent to which recycling can be expanded when prices suddenly rise.
(This description of the role of secondary recovery is qualitatively correct,
but it ignores the linkage between past consumption and the pool of
scrappable items from which secondary production can come during disruptions.
More sophisticated market models that explicitly recognize this linkage
should ideally be used to calculate the criticality of materials.)
Business firms faced with the threat of disruptions in the supply of an input
such as catalystium normally maintain inventories or stockpiles to be used
when supply disruptions occur. It is clear from the above analysis how the
existence of normal business inventories and stocks can decrease expected
costs from supply disruptions (that is, decrease the criticality of a
material), by reducing the amount of material that must be purchased on the
world market at very high prices during disruptions. On the other hand,
holding inventories imposes costs of its own that should also be attributed
to the disruption threat, and included in the measure of criticality.
Administering and maintaining a stockpile requires the time of a firm's
managers and employees, and involves other out-of-pocket expenses as well,
even after the inventories have been acquired.
The original cost of the stockpiled material is not counted by economists as
a cost to the firm (or nation) at the time of acquisition, since one kind of
asset (money) has just been transformed into another kind (stockpiled
2-22
-------
Charles
River
Associates
catalystium, in our example). The transformation could be reversed, if
desired (except for transaction costs, such as transportation and brokerage
charges). However, holding assets in the form of catalystium over time,
rather than in plant and equipment, in other immediately productive assets,
or even in a financial instrument earning interest, does impose costs on the
firm (and the nation). These costs are usually approximated by the estimated
interest costs of financing the stockpile (even if the stockpiler did not
actually borrow to finance his inventories). Many metals and other materials
do not cost much (out of pocket) to store, relative to their market value, so
the dominant cost of stockpiling is in fact the interest expense. As a
general rule, U.S. firms increase their stockpiles, held in anticipation of
market contingencies, until the next unit stockpiled is not expected to be
salable during the next disruption for enough to cover its expected holding
costs (recognizing that the dates, severities, and duration of disruptions
are uncertain events).
In summary, U.S. consumers of a material generally make two major types of
adaptations to the threat of supply disruptions (and other contingencies as
well). Before the disruption they acquire inventories, and after it occurs
they switch to alternative technologies, as summarized in market demand
curves for the material (such as Figure 2-3). (Having the capability to
switch quickly to alternate technologies during disruptions may, of course,
also require advance planning.)
For purposes of measuring the economic criticality of a material from a
national perspective, we can now lengthen the list of determinants (developed
earlier for particular consumers) as follows:
t the severity of contingencies that threaten U.S. consumers or producers,
as measured by price increases that occur;
• the expected time between these contingencies;
t the duration of these contingencies;
• the quantity of the material consumed in the United States under normal
conditions;
• the quantity of the material produced in the United States under normal
conditions, both from primary and secondary sources;
t the extent to which U.S. consumption can be reduced when prices rise
(the "elasticity" of short-run U.S. demand);
• the extent to which U.S. primary and secondary production can be
expanded when prices rise (the "elasticity" of short-run U.S. supply);
and
t the size of normal U.S. inventories and stockpiles.
2-23
-------
Charles
River
Associates
In addition to privately held inventories, the U.S. National Defense
Stockpile contains platinum-group metals and other materials that could
conceivably be used for vehicular emissions control during a nonmilitary
emergency. Appendix 2-A considers this possibility, and concludes that an
act of Congress would probably be required to authorize such releases. We do
not assume these stockpiles will necessarily be available during the
nonmilitary contingencies upon which we base the illustrative criticality
estimates given at the end of this chapter.
O PRIVATE ADAPTATIONS TO THE THREAT
OF SUPPLY DISRUPTIONS AND OTHER CONTTNGtHCTES
Recognizing the adaptations that private firms make in response to the threat
of contingencies in material markets, key questions are whether the U.S.
government should make additional preparations and adaptations with respect
to critical materials, and, more particularly, whether the Environmental
Protection Agency should weigh the critical ity of the material in its
decisionmaking, beyond including the effects of the underlying contingencies
on the usual calculation of expected future compliance costs (as we described
that process earlier). Under certain circumstances, which can be
approximately satisfied in some material markets, general government policies
(such as stockpiling, tariffs, quotas, or subsidies to domestic producers)
are not needed. In these circumstances, private firms can be expected to do
the amount of stockpiling, and choose the production technologies, that are
efficient from a national perspective. In these cases EPA only must consider
the effect of materials prices on expected compliance costs when choosing
among policy options.
We cannot analyze in detail here the conditions under which private
adaptations to market contingencies are efficient from a national
perspective. (Charles River Associates has filled many volumes analyzing
these issues, particularly for materials criticality stemming from the
threat of foreign supply disruptions. See the Bibliographic Note and
References.) However, we can state the most important of these conditions
and give some indication of their relevance.
The first of the conditions under which private firms would prepare
sufficiently for market contingencies is that there be no expectation of
price controls, material allocation, or other government interference with
the market, even during serious disruptions. For example, if firms expect
inventories to be reallocated from "have" firms to " have not" firms (as
actually occurred in the post-OPEC U.S. petroleum market), then they will
have reduced incentive to accumulate contingency stocks, and private holdings
will be less than is justified by benefits (and costs) measured from a
national perspective. Expectation of price controls (perhaps instituted with
the rationale of "moderating the inflationary impact of a supply disruption")
can have the same unfortunate consequences.
2-24
-------
Charles
River
Associates
A second condition is as follows: The preparations for contingencies by
individual U.S. firms must not reduce the likelihood or severity of the
contingency faced by other U.S. firms. This condition is likely to fail to
some extent. For example, when an individual U.S. firm holds an additional
unit of inventories, it must bear the entire cost of holding that unit, but
some of the benefit generally accrues to other firms. This "external"
benefit to other U.S. firms occurs during disruptions, when the firm with the
additional stockpiled unit needs to buy one unit less on the world market,
tending to decrease the price at which other U.S. firms obtain their imports.
One can imagine an artificial situation where this mechanism would not
operate, for example where a cartel of foreign producers threatens to
suddenly form and double the existing market price, no matter how much the
resulting rate of demand is reduced due to U.S. buyers consuming out of
stocks. However, it is much more likely that the cartel would set a lower
price, at least initially, if U.S. stocks are higher. This mechanism is
referred to in CRA studies as the "price deterrence" benefit of U.S.
stockpiling. Reductions in U.S. consumption during disruptions due to use of
alternative technologies can also have price deterrence benefits that are
external to the individual U.S. firm actually adopting the alternative.
In general, whenever a U.S. firm bears all the cost of some action preparing
for a disruption, or all of the cost of an adjustment made during the
disruption, but other firms reap some "external" benefit, private
preparations and adjustments tend to be less extensive than is desirable for
the nation as a whole. All of the costs of an additional increment of
preparation or adjustment are recognized by the private decisionmaking unit,
but all of the benefits to the nation are not. The private firm stops
preparing or adjusting when the incremental private cost equals the
incremental private benefit, whereas, from a national viewpoint, the firm
should continue preparing and adjusting until the incremental private cost
equals the incremental national benefit.
Another mechanism of the same nature can be relevant when a material market
is threatened by a disruption that is deliberately decided upon by foreign
producers of a material, or by their governments. In that case, greater U.S.
preparations, particularly larger U.S. stockpiles, can decrease the
probability of a disruption occurring in the first place. In CRA studies
this mechanism is called the "probability deterrence" benefit of U.S.
stockpiling. (See for example the appendix to Klass, Burrows, and Beggs
(1980).)
Depending on the types of threats facing a particular market, price
deterrence and probability deterrence can make private preparations and
adjustments much less than would be desirable from a national perspective.
In these situations, a strong case can be made for government policies, such
as tariffs or stockpiling, that will manipulate or augment private
preparations and adjustments.
2-25
-------
Charles
River
Associates
Other conditions can cause preparations for market contingencies by U.S.
firms to fall short of the national optimum. One such condition involves the
way private firms compare present costs and benefits with future costs and
benefits. Generally, decisionmakers "discount" costs and benefits occurring
in the future relative to current costs and benefits; that is, a dollar
gained tomorrow justifies the expenditure of something less than a dollar
today. But makers of public policy may conclude that private firms discount
future costs of disruptions too much, thus incurring insufficient costs for
preparations today. (Economic theory does not offer a definitive answer to
the question of what rate of time discount is most appropriate for national
policymaking.)
The above discussion assumed, as is often the case, that U.S. firms consuming
a critical imported material are individually small relative to the world
market, but together account for a sizable proportion of world consumption.
On the other hand, if a single U.S. firm accounts for most of U.S.
consumption, then the importance of price deterrence effects and probability
deterrence effects, as discussed above, may be considerably less. Most of
the benefits of preparation for, and adjustments to, disruption will accrue
to that single U.S. firm, rather than being "external," so the extent of
private preparations and adjustments will tend to be much closer to the
national optimum. In Japan, it is common for firms to coordinate decisions
about critical imported materials, thus gaining benefits that otherwise would
be possible only in the case of a single importing firm. However, in the
United States such coordination would run afoul of antitrust regulations.
(There are many competing considerations in deciding the desirability of the
Japanese institutional arrangement versus the U.S. arrangement, and this
consideration is very probably not the deciding one.)
One circumstance that can lead to more private preparations for disruption
(relative to the national optimum) rather than less, is "risk aversion" on
the part of private firms. This situation can be explained with reference to
our earlier sample calculation of criticality for consumers of the imaginary
material catalystium. We calculated the criticality of catalystium to
consumers by averaging losses that could be expected to occur over many
years. Since disruptions were assumed to occur in one year out of ten, the
average expected loss to consumers was (0.1)($3,200,000) = $320,000 per year.
Contrast this expected loss with that which would occur if a disruption loss
of $640,000 occurred in one year out of two. In that case, expected losses
are again $320,000 per year, and the criticality of the material is as
before.
This procedure for evaluating losses of different sizes occurring with
different frequencies is entirely appropriate for policymaking at the
national level, where losses in a given year of $640,000 or $3,200,000 are
miniscule relative to the size of the entire U.S. economy. However, consider
2-26
-------
Charles
River
Associates
a small U.S. firm for which catalystium accounts for a major proportion of
production costs, and suppose there is stiff competition from other firms
whose product does not require catalystium. For that small firm, a tripling
of the price of catalystium, sufficient to inflict costs of $3,200,000 on
U.S. consumers over one year, might be sufficiently severe to cause
bankruptcy. The fact that losses will "average out," over many succeeding
years, is small comfort if the firm has gone bankrupt in the meantime. In
this case, where the possible loss is catastrophic for the decisionmaking
unit, it is rational to be "risk averse," and undertake more stockpiling and
other preparations for disruptions than would be justified by average
expected costs alone.
In addition to stockpiling on its own, a small firm can also reduce the risk
of an increase in the price of catalystium by buying "forward" contracts for
future delivery of catalystium on a commodity exchange. In that way the
risk is spread among a great many speculators, who can individually diversify
their speculations so that they are not catastrophically affected by a
disruption in the catalystium market. In this way private risk aversion can
be reduced, which is beneficial to the individual firm, but may reduce
preparations for disruptions that they would otherwise undertake. Another
strategy for reducing risks of supply disruptions is to enter into long-term,
fixed-price contracts with reliable suppliers who are unlikely to be
disrupted.
IMPLICATIONS FOR EPA POLICYMAKING
We have described above a number of conditions in material markets that can
cause private firms' preparation and adjustments for market contingencies,
particularly supply disruptions, to be different from (usually less
extensive) those that would be justified by the costs and expected benefits
measured from a national perspective. In those circumstances, a case can be
made for government policies such as tariffs and national stockpiles, which
modify or augment private preparations or adjustments made in response to the
threat or occurrence of a contingency such as a supply disruption. A
material must be "critical," that is, threatened with serious contingencies
such as a major foreign supply disruption, to justify government actions in
addition to the private actions that profit-maximizing U.S. firms undertake
naturally. However, government actions such as tariffs and national
stockpiles are not necessarily justified for all critical materials. That
is, criticality is a necessary condition, but not a sufficient condition, to
justify general government policies such as tariffs and stockpiles.
With this analysis of criticality and general government policymaking in
material markets as a background, we can now address directly the principle
question that this chapter asks: how should the Environmental Protection
2-27
-------
Charles
River
Associates
Agency factor the criticality of materials into its decisionmaking, beyond
recognizing the effects of contingencies on prices of materials when
calculating average future compliance costs? We presume that any effect of
EPA decisions on the long-run equilibrium price of a material (roughly
speaking, its "normal" price) is taken into account. (The usual example is
where increased demands induced by EPA regulations raise the normal price of
the material somewhat.) We also presume that any operative, or likely,
general government policies in the material market (such as tariffs) are
taken into account by EPA when calculating compliance costs.
Our answer to the question of appropriate EPA policymaking procedures is
particularly easy to justify if general government policies in the material
market, such as stockpiles and tariffs, are presumed to adequately recognize
the material's critical ity. In that case, the price paid for the material by
producers of vehicular emissions control equipment respresents the material's
true cost to society, and no special EPA policymaking, beyond careful
forecasting of compliance costs, is required. This conclusion can be
justified in great detail, using elaborate versions of the type of economic
cost-benefit analysis that we described above as a basis for measuring the
criticality of materials. However, the conclusion is sufficiently plausible
on its face (and we have sufficiently burdened noneconomists reading this
chapter with unfamiliar concepts) so that we forgo its full development. It
is basically just one application of very general economic theories showing
how the price system can efficiently allocate resources in a free market
economy.
It is, of course, true that EPA decisions can greatly change the quantities
of a material that are consumed in the United States, as well as other
conditions affecting its criticality, thus changing the general government
policies that are appropriate. To take an obvious example, increased U.S.
consumption to meet EPA regulations would presumably increase the optimal
size of government stockpiles. In order to facilitate better and more timely
government policymaking for materials markets, it certainly could be
worthwhile in principle for the EPA to inform other government agencies,
particularly those with direct responsibility for policymaking in material
markets, concerning EPA decisions that will significantly affect material
markets. (We hope circulation of this study outside EPA will serve this
purpose to some degree.) In return, EPA might learn of possible changes in
general government policies that would affect the forecasting of compliance
costs and hence potentially affect EPA decisions.
The above line of reasoning might sound a trifle artificial to those familiar
with U.S. policies toward material markets, because in fact such policies
have been designed almost exclusively in response to the threat of military
contingencies. As we discuss further in Appendix 2-A, the U.S. National
Defense Stockpile is currently reserved exclusively for defense related
2-28
-------
Charles
River
Associates
applications, unless an act of Congress authorizes release for other purposes
(which has never happened in the forty-plus years of the stockpile's
existence). The United States has modest tariffs on the importation of a
number of raw and processed materials. The policy debate on such tariffs
often centers around the desirability of having secure domestic sources of
supply during wartime (though the political power of special interest groups
sometimes seems the more important consideration). Certainly, U.S. tariffs
have not been adjusted as the types of nonmilitary contingencies that we
identified earlier threaten and recede, in the way in which economic theory
suggests would be optimal.
On the other hand, there are a number of government policies aimed at the
structural causes of certain nonmilitary contingencies in material markets.
Government sponsored R&D is often aimed at reducing U.S. dependence on
imports of materials, for example, by making it profitable to exploit
previously uneconomical U.S. deposits. The U.S. Department of the Interior
is the lead agency in this area. U.S. government agencies such as the
National Labor Relations Board are concerned with settling domestic labor
disputes, including those affecting material markets. The U.S. Department of
State and other U.S. agencies are concerned with international relations that
may affect the conditions under which this country imports materials.
Nevertheless, it is still relevant to note that the United States in most
cases has not generally employed market-specific policy instruments, such as
tariffs and "economic" stockpiles, to counteract the threat of nonmilitary
contingencies such as foreign supply disruptions.
Is this lack of fine-tuned U.S. policies aimed at nonmilitary contingencies
a serious problem? In most cases, probably not. In many material markets,
it can plausibly be argued that the private sector undertakes sufficient
preparations for, and adjustments to, nonmilitary contingencies in material
markets so that the benefit of even theoretically optimal government policies
on tariffs and stockpiling would diminish expected national losses only
moderately. Moreover, when decisions are finally made in the real world
about such national policy instruments as tariffs and public stockpiles, it
must be recognized that they are often in practice more costly than analysis
of their theoretical optimality would indicate. Administering tariffs and
stockpiles can be much more costly than originally estimated, particularly
when vested economic interests and political realities intrude into the
management process. Recognizing these facts of political life reduce
considerably the potential scope for beneficial government policymaking, and
makes historical practice in the United States more understandable.
Suppose then that the EPA is contemplating policy options that will greatly
affect the market for certain materials (such as increasing demand for the
platinum-group metals), and it suspects that national policies (such as
2-29
-------
Charles
River
Associates
tariffs and economic stockpiles) do not adequately recognize the criticality
of the materials. What should EPA do? The first observation we make, which
is easy to understand in the light of discussions above, is that it is often
difficult to determine whether national policies "adequately" recognize the
criticality of materials, given private efforts to ameliorate that
criticality. To say the least, it would be very ambitious for EPA to analyze
conditions in material markets in sufficient depth to make such a
determination, particularly where government agencies with direct
responsibility for these issues have not done so. In fact, EPA can hardly be
held responsible for such policy analysis, and doubtless it is a
responsibility whose lack EPA does not regret.
Furthermore, any actions EPA could take to reduce consumption of a critical
material -- for example by requiring compliance through one technology rather
than another -- would fall into a category that economists call "second-best
solutions." It is more efficient from a national perspective to have
national policies that discourage all consumers from using a critical
material, if that is indeed called for because some condition exists in the
market that makes private policies otherwise inadequate.
The two main conclusions we reach regarding the role of materials criticality
in EPA decisionmaking are thus as follows: first, in almost all cases the
EPA need only calculate compliance costs in a comprehensive manner that
recognizes the likelihood of market contingencies temporarily raising market
prices in the future. (The effect of EPA induced demands for materials on
compliance costs should also be recognized, both during possibly rapid
implementation, and during the "long run" thereafter.) Second, where the EPA
anticipates its policies will have major impacts on material markets, it
should coordinate with the Department of the Interior and other federal
agencies with more direct responsibility for policies affecting material
markets.
In order to implement these suggestions, it can still be useful for EPA to
estimate roughly the criticality of materials required for compliance with
its regulations. Estimating criticality from the perspective of consumption
for vehicular pollution control equipment is very closely related to
estimating how much expected future compliance costs will be raised due to
the contingencies that threaten the market, which is something EPA should do
anyway. Estimating criticality from a national perspective suggests the
possible importance of EPA coordinating policies affecting materials markets
with other government agencies. The last part of this chapter pursues both
of these conclusions by developing rough estimates of the criticality of
platinum-group metals and four other materials required for vehicular
emissions control equipment.
2-30
-------
Charles
River
Associates
SHOULD EPA SPONSOR R&D ON VEHICULAR EMISSIONS CONTROL?
Our discussion above considered decisions by private firms about amounts of
critical materials to use, when there exists a range of different
technologies for complying with EPA standards on vehicular emissions. A
further, related question is whether private research and development on new
compliance technologies is sufficiently extensive, or whether government
sponsored research of some sort is called for.
Research is of course an uncertain process, and there is always the chance
that a government sponsored program will discover a new technology that
private investigators overlooked. However, there appears to be no general
evidence that government sponsored research tends to be more productive per
dollar spent. If anything, conventional wisdom asserts just the opposite,
that private R&D tends to be more efficient. There are horror stories about
ill advised and unsuccessful government supported research, but there are
other cases as well, where government research was successful where private
research was not successful, or was deemed too unpromising even to pursue. A
classic case in the history of material markets is the development by the
U.S. Bureau of Mines of technologies to process deposits of the low-grade
iron ore taconite.
We reach no conclusions here concerning the comparative cost effectiveness of
private versus government research and development. Rather, we simply
examine the general circumstances under which private decisions about R&D
expenditures in this area are made, to see if there are strong reasons to
suspect that the total amount spent would be inadequate.
The usual cause of inadequate private research (and hence the usual
justification for government sponsorship of research) involves reasoning much
like that we described above to analyze the adequacy of private stockpiling
and other preparations for contingencies in material markets. An individual
private firm bears all of the expenses of research it undertakes on its own.
If there are large "external" benefits to other firms from successful
research by one firm, which that one firm cannot substantially capture
through licensing fees and other arrangements, then the firm contemplating
research will tend to spend less than is justified by the expected benefits
of its research to the nation as a whole. So-called "basic" research tends
to have the most extensive external benefits, which cannot be appropriated by
the successful researcher. Thus, government sponsorship of research is
typically most justified for basic research.
How do the circumstances of private research and development on vehicular
emissions control stack up against the usual justification for government
sponsorship? By and large it appears that the private level of effort in
this area should be roughly appropriate. Most of the research is specific to
2-31
-------
Charles
River
Associates
emissions control, and has limited usefulness outside the motor vehicle
industry. To be more specific, much of the benefit of research in this area
undertaken by General Motors will accrue to General Motors. Moreover,
licensing fees for technologies patented at General Motors should allow GM to
appropriate some of the benefits received by other motor vehicle
manufacturers. The general case for government sponsorship of research in
this area appears quite weak.
In Chapter 6 we review research on alternatives to conventional catalytic
converters (containing platinum-group metals) that has taken place to date.
Certainly there is no direct indication that promising avenues of research to
develop more cost effective compliance technologies have been left
unexplored.
A MAJOR QUALIFICATION AND A POSSIBLE POLICY PRESCRIPTION
The relatively optimistic conclusions reached above about the adequacy of
private R&D and private preparations for contingencies such as supply
disruptions, are applicable to all industries, whether their consumption of a
critical material stems from the nature of consumers' demands in the
marketplace or from the need to satisfy government regulations. However,
there is one further major consideration where consumption of a material is
based predominantly on the need to satisfy a government regulation.
The main qualification we would make to our general case for the adequacy of
private R&D on vehicular emissions control technologies concerns private
expectations about the stringency of future emission standards. If U.S.
firms do not expect a future standard to be actually in effect and enforced,
then they may well not undertake adequate research on ways to meet that
standard in the manner least costly to themselves and the nation. EPA is
undoubtedly in a better position than we to assess the credibility of
scheduled future vehicular emissions standards.
This same type of issue can arise when assessing the adequacy of the
preparations that the U.S. auto industry makes for contingencies such as
foreign supply disruptions. If the U. S. auto industry believes that it will
be able to arrange relaxation of emissions standards whenever the price of
platinum, palladium, and rhodium rises greatly during a supply disruption,
then the industry will probably undertake less extensive preparations than
they would otherwise. In particular, they would stockpile much less of these
metals for such contingencies.
2-32
-------
Charles
River
Associates
It is easy to see why a politically persuasive case for relaxation of
emissions standards could be made during a major supply disruption in the
platinum-group markets. It would probably be virtually impossible for U. S.
vehicle manufacturers to obtain enough of these metals on the spot market if
they had no inventories and supplies from South Africa and the Soviet Union
were cut off. As discussed further below, the prices of the required
platinum-group metals could become "astronomical" in this case.
It would be very interesting in this regard to know the size of the
inventories of platinum, palladium, and rhodium that the U. S. auto industry
currently maintains. However, as discussed further in Chapter 4, consumers'
stocks of these metals are not officially estimated in the United States. If
stocks in the auto industry are quite low in comparison with other industries
(such as petroleum refining or glass manufacturing) where continuing use of
platinum-group metals is mandated by technology rather than by government
regulation, then a case could be made that the auto industry has some
expectation of throwing itself on the mercy of the regulatory and legislative
processes in the event of a major disruption in platinum-group supplies from
South Africa. (In fact, stocks of these critical materials in the auto
industry should be considerably greater than those "on the shelf" in the
petroleum and glass industries, because those other industries have control
over material obtained by secondary recovery (after a few years' use),
whereas the auto industry has no special access to material obtained from
obsolete catalytic converters.)
The obvious solution to this problem, if it is indeed a problem, would be to
require documentation from U. S. auto manufacturers that they have a
specified minimum level of inventories on hand at all times, unless given an
explicit exemption by EPA. We have not investigated the legal or practical
aspects of implementing such a new regulation. In setting the appropriate
level for such inventories, it would be important for EPA to decide whether
very large increases in the price of platinum-group metals could in fact
eventually justify some relaxation of emissions standards. (One interesting
aspect of such a policy is that it would probably be about as effective with
foreign vehicle manufacturers as with U.S. manufacturers, even though they
would not be directly subject to the stockholding requirement; the reason is
that they would fully expect EPA vehicular emissions standards to be
maintained as long as U.S. vehicle manufacturers could consume platinum,
palladium, and rhodium out of required inventories.)
2-33
-------
Charles
River
Associates
MATERIAL IMPORTS AND BALANCE OF PAYMENT PROBLEMS
The reader may have noticed in the preceding discussions that, despite its
references to general policymaking in materials markets, no explicit mention
was made of the effect of increased material imports on the U.S. balance of
trade or balance of payments. This omission was purposeful, since the issue
tends to be largely a red herring for efficient government policymaking in
markets for particular goods such as materials. It is relatively
straightforward to estimate the increased dollar value of U.S. material
imports that would result from increased consumption for vehicular emissions
control. However, in themselves, increased imports usually do not signal
significant national losses --and certainly not losses of the same magnitude
as the dollar value of increased imports.
It would be too time consuming to develop fully the rationale for the above
conclusion here, but the general idea is as follows. International trade
benefits all countries that participate, by allowing them to import goods in
whose production they are relatively inefficient, and export goods in whose
production they are relatively efficient. Imports are a necessary part of
this process, and as such they benefit the United States rather than harming
it.
Nevertheless, though it may at first seem paradoxical in view of the above
general truths, it is possible in theory to benefit the United States by
reducing U. S. imports. This can be done for all imports most easily by
imposing a general tariff, or for imports of a particular good (such as a
material) by imposing a specific tariff or quota. Where reduced imports of a
particular material are planned, the rationale is generally that the United
States is a large importer on the world market, and reduced U.S. imports can
reduce the price at which the remaining imports are obtained. In effect,
implementing such policies allows the United States to act monopolistically
("monopsonistically," to be more precise) with respect to its foreign
suppliers. The United States does benefit, but foreign exporters lose even
more than the United States gains.
The biggest problem with policies such as tariffs, particularly a general
tariff on all imports, is that they invite retaliation. If other trading
countries also reduce their imports, then all countries will typically be
worse off than with no tariffs at all, simply because the greater
international productive efficiency allowed by trade has been diminished.
Recognizing these facts, the United States has often been a world leader in
attempting to reduce international trade barriers, so that all countries
benefit from more free trade. U. S. tariffs on material imports can be a
useful response to the threat of particular market contingencies (as
discussed earlier in this chapter), particularly where retaliation by foreign
suppliers is not a problem. However, simply imposing trade barriers to
2-34
-------
Charles
River
Associates
improve the U. S. balance of trade largely represents a "beggar-thy-neighbor"
policy that runs counter to the U. S. tradition of supporting free trade
(continually threatened by special interest groups though the tradition may
be).
Some of the same effect as a tariff on imports of materials could be achieved
by EPA requiring that compliance with its regulations be achieved with less
of an imported material, such as the platinum-group metals. It is unlikely
such actions would invite retaliation in the way a tariff can, so, from a
narrow perspective, some small economic gain to the United States could
accrue. However, in order to obtain such monopolistic (monopsom'stic)
benefits, foreign prices for the remaining amount of U. S. imports must be
driven down. The first requirement is that EPA decisions affect a large
proportion of U. S. imports of the material. As we show later, EPA has this
"leverage" only in the market for platinum-group metals. Moreover, in most
major material markets, world supply is usually very price elastic in the
long run, which means that a reduction in U. S. consumption (or in the growth
of U.S. consumption) will cause only a very small decrease in the world price
of the material, after suppliers have had a chance to ajust to the new
situation. Thus, for example, the monopsonistic benefit to the United States
of reducing platinum consumption would be very small because of the
elasticity of supplies from South Africa.
There are exceptions to the above generalizations, however, particularly by
product materials where demand reductions can lead to sizable price
decreases. Among materials used for vehicular pollution control in the
United States, the outstanding example is rhodium, which is a byproduct of
platinum. Rhodium production is essentially at the limit imposed by current
world platinum production, and rhodium is usually the most expensive
commercial metal (per unit weight). If EPA eliminated U.S. rhodium
consumption for vehicular emissions control, there would be a significant
decrease in its price on the world market. It would be possible to do a
careful calculation of the optimal extent to reduce U.S. rhodium consumption
in order to realize the maximum monopsonistic gain for the United States.
(If EPA decisions determined all of U.S. rhodium consumption, then the
maximal monopsonistic gain to the United States would be obtained simply as a
result of EPA recognizing the effect of its decisions on the price of rhodium
when calculating compliance costs, as we have recommended.) However the gain
to the United States would still be very modest, and it seems almost certain
that EPA has more important policy concerns to occupy its attention.
Moreover, regulating rhodium consumption downward would indirectly violate
2-35
-------
Charles
River
Associates
the U.S. tradition advocating free trade. (An interesting further aspect of
the case of rhodium is that the parties most harmed by a reduction in U.S.
imports would be platinum producers in South Africa.)
We have attempted above to make the most reasonable case we can for EPA being
concerned with the implications of its policies for U.S. material imports and
the balance of trade. However, in the final analysis we do not think it is
of sufficient importance to warrant inclusion as a separate consideration in
EPA policymaking. It will be hard enough for EPA simply to take into account
such vagaries of international trade as changes in exchange rates among the
U.S. dollar and other world currencies, as these affect the dollar cost of
U.S. material imports, and hence compliance costs for EPA regulations. It is
not really reasonable to expect EPA to design optimal monopsonistic importing
policies as part of its decisionmaking.
A SIMPLE ECONOMIC MODEL FOR ESTIMATING CRITICALITY DUE TO FOREIGN
SUPPLY DISRUPTIONS
The economic model we use in this chapter to estimate the criticality of
materials used in vehicular emissions control equipment, is a slightly
generalized version of the simple analysis of U.S. supply and demand curves
that we described earlier in numerical terms for the imaginary material
"catalystium." As in that example, we concentrate on materials criticality
due to the threat of foreign supply disruptions, and calculate criticality
from both the national perspective and the perspective of consumption for
vehicular emmissions control.
2-36
-------
Charles
River
Associates
KEY PARAMETERS AND FORMULAS
Figure 2-7 displays, with general algebraic nictation, the linear supply and
demand curves that are assumed to be relevant to each of the materials we
consider. Key parameters of the criticality measurement, whose roles are
explicitly indicated in Figure 2-7, are as follows:
P0 = the normal world market price of the material
XP0 = the world market price during disruption
Qd = U.S. consumption at normal price P0
Qp = U.S. primary production at normal price P0
Qr = U.S. secondary production at normal price P0
ed = the price elasticity of U.S. consumption
ep = the price elasticity of U.S. primary production
es = the price elasticity of U.S. secondary production.
Other key parameters in the criticality measurement are the following:
D = the duration of disruptions, and
T = the time interval between (starts of) disruptions
When we consider later the possible role of stockpiling, we must also
specify
r = the "real" rate of interest (difference between observed
nominal rates and the rate of inflation), and
e = the out-of-pocket expenses of stockholding, measured
relative to the value of the material stockpiled.
2-37
-------
Charles
River
Associates
Figure 2—7
MARKET MODEL FOR ESTIMATING CRITICALITY OF IMPORTED
MATERIALS THREATENED WITH FOREIGN SUPPLY DISRUPTIONS
PRICE
XP
2P
SECONDARY,
PRODUCTION
0 Q
/ * PRIMARY PLUS
1 /< SECONDARY
PRODUCTION
CONSUMPTION
QUANTITY
2-33
-------
Charles
River
Associates
The price elasticity of U.S. consumption is the relative decrease in
U.S. consumption that results from a doubling of the price"! Thus, if a
doubling of price causes a 10 percent decrease in consumption (as was the
case for producers of emissions control equipment in the "catalystium"
example dicussed earlier), ej = 0.1. Since we are assuming all demand (and
supply) curves are linear, decreases in consumption resulting from larger
price increases are proportionately greater. Thus, an X-fold increase in
price during disruptions causes U.S. consumption to decrease to
Qd [l-(X-l)ed].
The price elasticity of U.S. primary or secondary production is the
relative increase in U.S. primary or secondary production that results from a
doubling of price. Along linear supply curves, an X-fold increase in
price causes U.S. primary production to increase to Qs [l-(X-l)es].
Figure 2-8 reproduces the total U.S. demand and supply curves from Figure 2-7
and labels the loss areas that were explained in numerical terms for the
catalystium example. Triangle 1 is the net cost consumers incur from
reducing consumption by fraction (X-Den. The rectangle made up of Areas 2,
3, and 4 is the additional payment (or transfer") consumers make to domestic
and foreign suppliers due to the price increase. Rectangle 2 is the
additional transfer to foreign suppliers. The rectangle made up of areas 3
and 4 is the additional transfer to domestic producers, of which Triangle 3
represents the additional cost of production (beyond the cost of importing
the material at normal price P0).
The first column of Table 2-1 translates the loss areas in Figure 2-8 into
algebraic formulas convenient for performing the actual calculations. These
formulas take into account the duration (D) and frequency (1/T) of
disruptions, to give average expected losses per year. All losses are
expressed as multiples of market values observable in normal times. For
example, U.S. consumers' losses are a multiple of the value of their
consumption in normal times (P0Qd)- "Losses" of primary and secondary
producers are generally negative (that is, "gains ), because the transfer (a
negative cost) is bigger than the adjustment cost by an amount represented in
Figure 2-8 as Trapezoid 4. The losses of primary and secondary producers are
expressed in Table 2-1 as multiples of the values of their production in
normal tmes (P0Qp and P0Qs> respectively).
Average expected U.S. losses per year, netting gains by U.S. producers
against the larger losses by U.S. consumers, are obtained by adding up the
losses in the first column of Table 2-1. The result is our measure of the
criticality of the material from a national perspective. We do not present
2-39
-------
Figure 2-8
U. S. ECONOMIC LOSSES (AND GAINS) FROM
FOREIGN SUPPLY DISRUPTIONS
PRICE
XP
Charles
River 4
Associates
Qs [1 + (X-1)6,1
QUANTITY
2-40
-------
River
Table 2-1 Associates
FORMULAS FOR AVERAGE EXPECTED ANNUAL U0S, ECONOMIC LOSSES FROM FOREIGN SUPPLY DISRUPTIONS*
Import-Eliminating
Economic Groups No Stockpiling Stockpile** Comprehensive Stockpile
Consumers
Adjustment (1)
Transfer (2,3 a
Primary Producers
Adjustment (3,
Transfer (3 and
Secondary Producer
Adjustment (3,
Transfer (3 and
Stock Holders
Holding Cost
Acquisition Cos
Less Revenues
2
HP 0 • •• vy - - C 11 D/T ^ 0
p
nd 4) {PQQd[tX-l)-(X-l) ed]} D/T ** '" '' 0
, .2
part; i_r LJ - — ?j — — t j u/ 1 ~ u
0
<1 Dart} f-P 0 r( X-l ^ hf X-1 ) c 11 n/T -^ - 0
i , (Jai u; x -i -VpL VA-iyT^A-i/ n '
S
2
part) [P Q -^ii- £ 1 D/T -* °
vj j C- o
4, part) {-P0QSC(X-1)+(X-1)2 es:> D/T -* °
0 (r+e) Po -Qp [1+(X-1) ep] D (r+e) Po (Qd-Qp-Qs)
l-Qs [1+(X-1) esl
tf o -(X-1J Po l-l^Hx-ij^ilD/T o
-o: n+fx-n e:i!
-------
ro
ro
Table 2-1
FORMULAS FOR AVERAGE EXPECTED ANNUAL U.S. ECONOMIC LOSSES FROM FOREIGN SUPPLY DISRUPTIONS*
(Continued)
SOURCE: Charles River Associates, 1981.
*Appendix 2-B describes a simple computer program that performs these calculations.
**Arrows indicate that formulas are unchanged from the column to the left.
Charles
River
Associates
-------
Charles
River
Associates
separate formulas for calculating economic losses from the persective of U.S.
consumption for vehicular emissions control, because they are a special case
of the formulas in Table 2-1. All that is required is setting U.S. primary
and secondary production (Qp and Qs) equal to zero, and setting U.S.
consumption (Q^) equal to consumption for vehicular emissions control
equipment, with the appropriate price elasticity
TREATMENT OF STOCKPILING
The economic model we are using is too simple to provide an adequate
explanation of the holding of inventories and stockpiles. As discussed
earlier, U.S. stockpiles can reduce U.S. losses during a disruption by
reducing the amount of imports required. Stockholding of a material can be
undertaken by U.S. producers, consumers, dealers, brokers, commodity
exchanges; or in fact any private party that wishes to bear the expense of
doing it, generally with the expectation of selling at some time in the
future at a significantly higher price that covers holding costs. In Table
2-1 we treat U.S. stockpiling as a separate activity, even though it may be
undertaken by consumers, primary producers, or secondary producers.
(Consumers holding stocks can be considered to sell to themselves during
disruptions.)
Table 2-1 allows the calculation of U.S. economic losses using three
different assumptions about stocks. None of these three oversimplified
assumptions about stocks will be exactly appropriate, but they provide useful
perspective in the area where our analysis of materials criticality would
otherwise be weakest. The formulas in the first column of Table 2-1, which
we discussed above, calculate average expected U.S. economic losses per year
assuming there is no U.S. stockpiling at all.
The formulas in the second column of Table 2-1 assume that the U.S. stockpile
is sufficiently large so that the United States need not import any of the
material during disruptions, but the domestic price is still equal to the
world market price (XPp) during disruptions. Thus, U.S. consumers and
producers experience the same losses and gains as they did with no stockpile.
(The arrows in the second column of Table 2-1 indicate where loss formulas
are unchanged from the first column.) The difference between the two cases
is that additional revenues that accrued to foreign suppliers during
disruptions when there was no U.S. stockpile now accrue to U.S. stockpilers.
As explicitly indicated in the labeling of the quantity axis in Figure 2-8,
the amount of U.S. imports during disruptions, measured at annual rates, is
2-43
-------
Charles
River
Associates
Annual Rate of Imports
During Disruptions = Qd [l-(X-l)ed] - Qp [l+(X-l)ep] - Qs [l+(X-l)es]
If disruption duration D is a half year, for example, the actual quantity of
imports during the disruption will be half of the amount determined by the
above formula, which we assume to be the quantity stockpiled in the second
column of Table 2-1.
The "real" cost of holding a dollar of assets in the form of stockpiles
(rather than in a form earning interest) is the "real" interest rate r, which
we specify to be 6 percent (r = 0.06) for results reported in this chapter.
The real interest rate is estimated as the difference between the "nominal"
interest rate actually observed in money and capital markets, and the rate of
inflation. Our use of an estimated real interest rate, to calculate the
costs of holding stockpiles for a year, implies the reasonable assumption
that the'value of stockpiled material will rise at the same rate as the
general rate of inflation in the U.S. economy, even in the absense of any
supply disruption. All prices, values, and economic losses that we report in
this chapter are in terms of constant 1981 dollars, that is, dollars deflated
to adjust for future inflation (which raises the "nominal," but not the
"real," prices of materials and most other goods.) (A real interest rate of
6 percent is approximate by historical standards. However, in 1981 real
interest rates were unusually high in the United States. Nominal interest
rates were nearly 20 percent per year, while general inflation was running at
less than 10 percent per year, implying a real interest rate closer to 10
percent. Since the analysis we are doing here bears on policies over many
future years, estimating future costs of holding stocks on a more normal
historical basis seems appropriate.)
In addition to the real interest cost of holding stocks, there are
out-of-pocket costs for management, warehousing, etc., that should in
principle be included. For most materials, the interest cost is considerably
larger, but we also allow for an annual out-of-pocket cost (e), per dollar
value of stockpiled material. We generally just assume e = 0.005 for results
reported in this chapter. The annual cost of holding the
"import-eliminating" stockpile is thus
Annual Rate of Imports
(r+e) P0 During Disruptions D
as entered under "Stock Holding Costs" near the bottom of the second column
in Table 2-1.
2-44
-------
Charles
River
Associates
In addition to holding costs, stockpilers of course also acquire the material
during normal times at price P0, and sell it during disruptions at price XP0.
The cost of acquiring stocks, net of revenues received from sale during
disruptions, is
Annual Rate of Imports
-(X-l) P0 During Disruptions D
which will generally be negative because prices rise during supply
disruptions (X>1). This negative cost, adjusted for the fact that it only
occurs every T years, is entered under "Acquisition Cost, Less Revenues" at
the bottom of the second column in Table 2-1. This amount would accrue to
foreign suppliers if there were no U.S. stockpile, as assumed in the first
column of Table 2-1, but with an import-eliminating stockpile it accrues to
U.S. stockholders instead. If the acquisition costs, less revenues from
sales during disruptions (on an annual average basis), is more negative than
the holding cost per year is positive, then holding stocks in anticipation of
the contingency is profitable, and U.S. economic losses are reduced by
stockpiling. (If disruptions are sufficiently mild or infrequent, it may not
be profitable to stockpile, in which case total U.S. economic losses would
not be reduced by stockpiling.)
A U.S. stockpile that would be sufficiently large to replace completely
normal imports during foreign supply discriptions, and hence prevent any
domestic price increases, clearly could not be profitable for private
stockpilers. It would also be larger than the optimum stockpile that
minimizes national losses (assuming it is not possible for the United States
to export stockpiled material). Nevertheless, such a stockpile may be closer
to the appropriate size than the "import-elminating" stockpile considered in
the second in second column of Table 2-1. Thus, in the third column of Table
2-1 we consider a "comprehensive" stockpile, sufficient to make up for normal
U.S. imports of (Q(j-Qp-Qs)D during the disruption. The cost of holding the
stockpile is indicated in the third column of Table 2-1. All other entries
in the third column are zero because there is no price increase during
disruptions in the U.S. market. (U.S. consumers suffer no losses, U.S.
producers experience no gains, and the revenue from sale of U.S. stockpiles
just equals the acquisition cost.)
(There would be some justification for estimating U.S. materials criticality
using estimates of actual U.S. stocks, rather than the hypothetical stock
sizes considered in Table 2-1. However, available data on U.S. stocks are
often incomplete, notably for the platinum-group metals, as discussed in
Chapter 4. There are also theoretical reasons for considering the
hypothetical stock sizes in a simple model that does not recognize any
relationships between stock sizes and the severity of price increases during
the specified foreign supply disruptions.)
2-45
-------
Charles
River
Associates
A computer program to perform the simple calculations described in Table 2-1
is presented in Appendix 2-B.
MEANINGFULLNESS OF RESULTS FROM THE MODEL
It has undoubtedly occurred to the reader, before reaching this point in the
analysis, that estimating the criticality of materials is a highly inexact
science, particularly as regards specifying the severity, duration and
frequency of disruptions threatening material markets. It can be instructive
to review the history of supply disruptions and other contingencies in the
markets of interest. But situations and threats change too rapidly for
historical analysis to provide definitive results. For example, historically
South Africa has been a quite reliable source of platinum-group metals,
chromium, and other materials often identified as critical. And yet its
reliability into the 1990s and beyond has been a source of concern to U.S.
policymakers. Clearly we must often rely on quite subjective estimates of
the likely severity, duration, and frequency of contingencies threatening
particular markets in order to obtain the most relevant measures of materials
critical ity.
The results we report below are of course also limited by the simplicity of
our economic model, and the fact that we are only specifying one type of
market contingency. Clearly there is in fact some probability of any of a
wide range of disruptions, characterized by different severities and
durations. Nevertheless, our specification of a single, representative
disruption (severity, duration, and frequency) for each market can still be
a roughly valid basis for comparisons among markets. These specifications
distill considerable CRA experience in analyzing contingencies in material
markets using more sophisticated models. The results we report here are
probably in the same ballpark as criticality estimates that would be obtained
with much more extensive applications of much more sophisticated models.
Moreover, the results reported here have the advantage that they clearly
indicate the reasons that the criticality of materials used for vehicular
emissions control equipment differ so markedly.
For the U.S. Department of the Interior, CRA has been working on a quite
sophisticated model specifically designed to simulate reliably much more
severe disruptions in the platinum and palladium markets than have occurred
historically. Elements of this model are discussed in Chapter 4. The model
is dynamic, so that, for example, the longer a disruption lasts the more
consumers can adjust away from materials that have become more costly. The
linkage between past consumption of plantium and palladium and current
secondary recovery is explicitly recognized. Supply and demand conditions
abroad, as well as in the United States, are recognized so that, for example,
it is possible to estimate the extent to which greater U.S. stockpiles will
2-46
-------
Table 2-2
MATERIALS USED FOR VEHICULAR EMISSION CONTROL:
CRITICALITY FROM A NATIONAL PERSPECTIVE.
Material (form)
Platinum - Group
Platinum (metal)
Palladium (metal)
Rhodium (metal)
Value of Normal U.S.
Price Elasticity of U.S.
Consumption
($ million)
729.0
123.0
76.0
Primary
Production
PoQp
($ million)
1.4
1.3
0.2
Secondary
Production
PO°S
($ million)
315.0
47.5
23.8
Consumption
ed
0.03
0.045
0.03
Primary
Production
eP
1.3
0.3
0.4
Secondary
Production
0.015
0.015
0.015
Other Materials
Chromium (ferro)
Manganese (ferro)
Nickel (metal)
Titanium Metal (sponge)
686.0
792.0
1,590.0
253.0
0
14.4
200.0
0
56.4
0
100.0
11.0
0.07
0.04
0.08
0*
0
0.1
0.08
0*
0.05
0
0.02
0.01
*For titanium, the price elasticities of demand and primary supply are set equal to zero, as an (Imperfect) adjustment for the fact that
the specified contingency of concern Involves an Increase 1n demand creating a processing bottleneck. See the discussion 1n the text.
Table continued on following page.
Charles
River
Associates
-------
ro
CO
Table 2-2 (continued)
MATERIALS USED FOR VEHICULAR EMISSION CONTROL:
CRITICALITY FROM A NATIONAL PERSPECTIVE.
Material (form)
Platinum - Group
Platinum (metal)
Palladium (metal)
Rhodium (metal)
Other Materials
Chromium (ferro)
Manganese (ferro)
Nickel (metal)
Titanium Metal (sponge)
Supply Disruption
Expected annual U.S. losses ($ million/year) with
Severity
(Proportional Increase
In World Price)
X
7
6
7
5
4
3
3
Duration
D
2.5
2.5
2.5
2
1.5
1
1
Expected Time
Between
Disruptions
T
20
20
20
20
20
15
15
No Stockpiling
246.0
36.0
32.9'
211.0
164.0
152.0
32.3
Import-Eliminating
Stockpile
103.0
17.0
11.9
96.1
77.3
84.4
15.7
Comprehensive
Stockpile
67.0
12.1
8.5
'81.8
75.8
83.9
15.7
SOURCE: Charles River Associates, 1981.
Charles
River
Associates
-------
ro
i
Table 2-3
MATERIALS USED FOR VEHICULAR EMISSION CONTROL:
CRITICALITY FROM THE PERSPECTIVE OF THAT END USE
Material (form)
Platinum - Grown
Platinum (metal)
Palladium (metal)
Rhodium (metal)
Other Materials
Chromium (ferro)
Manganese (ferro)
Nickel (metal)
Titanium Metal (sponge)
Vehicular Emission Control
Supply Disruption
Value of
U.S. Consumption
PoQd*
($ million)
280.0
18.7
18.0
10.0
0.3
2.0
3.8
Price Elasticity
of Consumption
*d*
.03
.03
.03
0.25
0.33
0.5
0.5
Severity
(Proportional Increase
1n World Price)
X
7
6
7
5
4
3
3
Duration
D
(years)
2.5
2.5
2.5
2
1.5
1
1
Expected Time
Between Disruptions
T
(years)
20
20
20
*
20
20
15
15
TOTAL
332.8
Table continued on following page.
Charles
River
Assnrintec
-------
Table 2-3 (continued)
MATERIALS USED FOR VEHICULAR EMISSION CONTROL:
CRITICALITY FROM THE PERSPECTIVE OF THAT END USE
ro
en
o
Material (form)
Platinum - Grown
Platinum (metal)
Palladium (metal)
Rhodium (metal)
Other Materials
Chromium (ferro)
Manganese (ferro)
Nickel (metal)
Titanium Metal (sponge)
Expected Annual Consumers Losses ($ million/year) With
No Stockpile
191.0
10.8
12.3
2.00
0.03
0.13
0.25
Import-ElIminatlng
Stockpile
56.2
3.5
3.6
2.00
0.03
0.13
0.25
Comprehensive
Stockpile
45.5
3.0
2.9
1.30
0.03
0.13
0.25
TOTAL
216.51
65.71
53.11
SOURCE: Charles River Associates, 1981.
Charles
River
Associates
-------
Charles
River
Associates
diminish the severity of price increases on the world market during
disruptions. We had hoped to be able to report results from these
sophisticated platinum and palladium models in this study, but unfortunately
appropriate results are not yet available.
On the other hand, there are definitive advantages to applying the same
simple model to all markets. The model used here, though simple, recognizes
all of the potentially crucial determinants of materials criticality.
Moreover, these determinants can be tabulated to indicate readily why
expected economic losses are much different in one case versus another. For
present purposes, clarity and organized analysis of the issues are
undoubtedly more important than refined estimates of the second digit of the
criticality measure.
SAMPLE ESTIMATES OF THE CRITICALITY OF PLATINUM, PALLADIUM,
RHODIUM, CHROMIUM, MANGANESE, NICKEL, AND TITANIUM METAL
As explained in some detail above, the most useful measure of the criticality
of a material is expected losses (usually predominantly economic) associated
with the threat of various market contingencies, the most important of which
often involves disruptions of foreign sources of supply. Expected economic
losses can be calculated from the perspective of the nation as a whole,
netting gains and losses of domestic producers and consumers, or from the
perspective of a particular end use, such as control of vehicular emissions.
Table 2-2 presents sample estimates of the criticality -- from a national
perspective — of seven elemental materials used for vehicular emissions
control in the United States, using the economic model described in Table
2-1. The required parameter estimates are presented on the first page of
Table 2-2, and expected economic losses (calculated under three different
assumptions about stockpiles that will be available during the contingency)
are presented on the second page.
Table 2-3 presents the corresponding criticality estimates calculated from
the perspective of consumption for vehicular emissions control. The severity
(X), duration (D), and frequency (1/T) of the market contingency are repeated
from Table 2-2.
In the following discussion, we first explain our projections of the values
of consumption and production for the seven raw materials. Then we specify
price elasticities of consumption and production, and the severity, duration,
and frequency of a representative contingency threatening each market. Given
these parameters of the criticality calculation, we finally interpret the
resulting estimates of expected economic losses given in the last three
columns of Tables 2-2 and 2-3, and draw out the policy implications for EPA.
2-51
-------
Charles
River
Associates
NORMAL CONSUMPTION, PRODUCTION, AND PRICES
The parameter estimates in Tables 2-2 and 2-3 are designed to be roughly
appropriate for the mid-1980s in the United States. Trend projections of
total U.S. consumption and production under normal conditions were obtained
from Mineral Trends and Forecasts, U.S. Bureau of Mines (1979). Appropriate
estimates of other parameters are much more uncertain than these quantities,
so any reasonable projections of total U.S. consumption and production are
adequate for our purposes.
Total U.S. primary production (Qp) includes only production from domestic raw
materials. Total U.S. secondary production (Qs) includes only production
from oT_d or obsolete scrap. Production from new scrap, generated by
processing before materials are sold in final goods (to consumers, investors,
or the government), is not explicitly considered because its supply can
usually not be expanded significantly, even under the incentive of much
higher prices for the material. Correspondingly, total U.S. consumption (Qj)
includes consumption out of primary production and old scrap, but consumption
of new scrap is netted out. (For platinum, palladium, and rhodium, the
estimates of secondary production (Qs) and total consumption (Qj) from U.S.
Bureau of Mines (1979) have been roughly adjusted upward by CRA to include
"toll refining" of the metal, where secondary refining of used metal is
undertaken on a fee-for-service basis, with the consuming industry
retaining ownership.)
The values of materials reported in Tables 2-2 and 2-3 are based upon the
following market prices, projected to be roughly appropriate for normal
market conditions in the mid-1980s:
• Platinum
• Palladium
• Rhodium
• Chromium
• Manganese
t Nickel
• Titanium
Metal
$450 per troy ounce of metal
$95 per troy ounce of metal
$475 per troy ounce of metal
$940 per short ton of elemental chromium contained in
ferrochromium
$480 per short ton of elemental manganese contained in
ferromanganese
$5,700 per short ton of metal
$11,000 per short ton of metal in the form of sponge.
2-52
-------
Charles
River
Associates
These prices (labeled P0) are measured in constant 1981 dollars, so that
inflation has no direct effect on the values reported in Tables 2-2 and 2-3.
(For simplicity, these values (that is, price P0 times quantity produced or
consumed) are calculated as though all of U.S. production or consumption is
processed through the indicated form of the material; this assumption is
least appropriate for chromium, where a majority of U.S. consumption indeed
requires production of ferrochromium or similar ferroalloys, but a
substantial proportion of U.S. consumption does not.
The first column of Table 2-3 estimates the value of U.S. consumption of each
of the seven elemental materials for vehicular emissions control in 1985 to
1987. The prices specified above (P0) are applied, while quantities consumed
(Q
-------
Charles
River
Associates
Clearly leverage is greatest in the market for platinum, and is negligible
for the four alloying elements.
CONTINGENCY THREATS AND PRICE ELASTICITIES
We now consider briefly major contingencies threatening the markets for the
seven elemental materials used for vehicular emissions control, and specify
the severity, duration,.and frequency of a representative contingency
(usually a foreign supply disruption) threatening each. The parameter
estimates are reported in Tables 2-2 and 2-3. As discussed earlier,
specification of a single contingency can only begin to characterize the
range of possible events in various material markets. Our main concern is
that the relative sizes of the single disruption threats be roughly
appropriate when comparing one market with another, and that the resulting
measure of expected U.S. economic losses be in the right ballpark. In the
following discussion of each market we also specify all elasticities that
characterize the responsiveness of U.S. consumption and production to much
higher market prices.
PLATINUM, PALLADIUM, AND RHODIUM
As discussed at greater length in Chapter 4, western consumption of the
platinum-group metals is supplied predominantly by South Africa and the
Soviet Union. There is clearly the potential for a very severe cutoff in
primary world supplies. Almost all platinum-group mining outside South
Africa is a byproduct of nickel and copper which implies that output from
alternative sources would not expand significantly in response to larger
increases in the price of platinum-group metals.
Moreover, the demand for platinum-group metals tends to be very unresponsive
to price increases. We estimate in Chapter 4 that a five-fold price increase
would only cut platinum and rhodium consumption by 12 percent, and palladium
consumption by 18 percent. That is,
EC! = (0.12/4) = 0.03 for platinum and rhodium
and
ed = (0.18/4) = 0.0045 for palladium
2-54
-------
Charles
River
Associates
The major consideration mitigating the severity of disruptions in world
supplies of platinum, palladium, and rhodium is secondary recovery. There
are very large quantities of these metals currently in use as catalysts, and
in other applications where a very high percentage can be recycled. During
disruptions in primary supplies, the natural operation of the price system
would redirect some of this recycled material toward the most critical
end uses (that is, to those uses that are willing to pay the most for the
material). However, because secondary recovery is already so extensive in
the platinum-group markets, there are only very modest opportunities to
expand recyling during emergencies. We specify secondary recovery elasticity
es= 0.015 for all three platinum-group metals under consideration.
For purposes of the sample criticality measurements in Tables 2-2 and 2-3, we
specify a seven-fold price increase (X=7) for platinum and rhodium during
disruptions. We specify the probability of a disruption to be sufficiently
high so that the expected time between disruptions is T = 20 years. Because
South Africa or the Soviet Union could be cut off from the United States for
a long time, we specify disruption duration D = 2.5 years. For palladium we
specify a somewhat less severe price increase during disruptions (X=6)
because, worldwide, price responsiveness is somewhat greater on the demand
side (and perhaps even on the supply side) of the market.
There may be significant mining of platinum-group deposits at the Stillwater
complex in the United States by the 1990s, but we do not factor that
possibility into our analysis. It would not be possible to create entirely
new underground U.S. capacity quickly enough after a supply disruption has
started to produce much during the first two or three years of the
disruption, but existing or abandoned sources, especially placer deposits and
the old mine at Goodnews Bay, Alaska, could be expanded or activated fairly
quickly. The Goodnews Bay deposit yields almost entirely platinum, and its
reactivation justifies a large primary supply elasticity for platinum in the
United States. We specify ep = 1.3 for platinum. There is actually more
rhodium than palladium in the Goodnews Bay deposits, justifying a slightly
higher overall U.S. supply elasticity for rhodium. We specify ep = 0.3 for
palladium and ep = 0.4 for rhodium. These price elasticities of primary
supply are quite large, but they have relatively little effect on the
calculations because normal U.S. primary production is so low.
CHROMIUM
South Africa is the dominant producer of chromium for western markets.
Neighboring Rhodesia produces substantial quantities of high-grade ores
suitable for metallurgical applications. Turkey and the Philippines also
produce substantial quantities. The Soviet Union formerly exported large
amounts of chromium to the West, but in recent years Albania has become the
important Communist exporter to the West. A complete disruption in southern
2-55
-------
Charles
River
Associates
Africa (South Africa and Rhodesia) would eliminate a large proportion of
western supplies and cause substantial, sustained price increases on the
world market. We specify disruption severity X = 5 for common grades
of ferrochromium, and disruption duration D = 2 years. We specify the time
between disruptions to be T = 20 years, the same as for the platinum-group
metals, which are also obtained from South Africa and Communist countries.
The United States does not produce significant amounts of primary chromium
under normal market conditions. A severe disruption such as that specified
in Tables 2-2 and 2-3 would induce some production from small U.S. deposits
in California and elsewhere, but our results would not be changed
significantly by recognizing this complication, so we simply specify £p = 0.
(A slightly generalized version of the economic model would be required to
recognize zero production at normal prices, with significant production
beginning at some higher price.)
Chromium is used predominantly for stainless steel and other steel alloys.
It is also used for refractories and in chemicals without first being
processed into ferrochromium. Consumption for chemicals is most responsive
to price changes, while consumption for stainless steel is least responsive.
There are very limited opportunities for reducing the chromium content of
stainless steels while still retaining the corrosion resistance at higher
temperatures required for the most demanding applications of stainless steel.
However, less than 20 per cent of stainless steel consumption is for
demanding applications such as turbines, while many other end uses could
substitute coated steels, plastics, and other materials. A five-fold
increase in the price of ferrochromium corresponds to roughly an 80 per cent
increase in the price of stainless steel, which would lead to substantial
conservation of stainless steels. Such a large price increase would reduce
chemical consumption of chromium by considerably more than half. We specify
the overall U.S. price elasticity of demand for chromium (valued in the form
of ferrochromium) to be e^ = 0.07.
Approximately eight per cent of U.S. chromium consumption is from secondary
sources, predominantly scrapped stainless steel used to make new stainless
steel. There are modest opportunities for increasing stainless scrap
recovery, for example by more carefully sorting stainless scrap from the
more common types. (A CRA pundit once remarked that there might also be
significantly increased recovery of stainless steel hubcaps through illicit
channels.) We specify the price elasticity of secondary supply to be
£s = 0.05.
MANGANESE
The Soviet Union produces more manganese than any other country by a
considerable margin, but uses the material intensively in its domestic steel
industry, and exports little to the West. South Africa is by far the
2-56
-------
Charles
River
Associates
largest exporter to the industrialized western countries, followed by Brazil,
Gabon, Australia, and India. Southern Africa is not as important in the
manganese market as in the chromium market, and there are a larger number of
alternative suppliers with more extensive possibilities for expanding
production. Thus, we specify the disruption severity to be less (X = 4) and
the disruption duration less (D = 1.5 years). We keep the same disruption
frequency as for the other materials obtained from South Africa (time
between disruptions T = 20 years).
Well over 90 percent of U.S. consumption of manganese is for steelmaking,
where it is an indispensable constituent of virtually all steels. There is
a small amount of flexibility in the quantity of manganese used per ton of
steel, and there could be a small amount more if manganese specifications for
steels were modified to reflect technologically minimal needs under the
current state of the art of steelmaking. Also, manganese that occurs
naturally in many iron ores and in steel scrap, and is normally lost in
processing, can be conserved by such procedures as slag recycling. We
specify the overall U.S. demand elasticity to be ej = 0.04.
Projected 1985 U.S. primary production of manganese is very modest, though it
could probably be expanded substantially during a sustained serious
disruption of foreign supplies. (The Bureau of Mines may be assuming some
production from ocean nodules.) We specify the domestic price elasticity of
supply to be es = 0.1.
NICKEL
The most common cause of supply disruptions in the world nickel market has
been labor strikes at Canadian mines, particularly at the dominant Canadian
producer INCO. The Canadian market share has dropped greatly over the last
two decades, to less than 40 per cent, so such labor unrest is less critical
to the United States and other nickel importing nations than was earlier the
case. Newer producers, such as Australia and the island of New Caledonia (an
"overseas department" of France, in the southwestern Pacific Ocean) have been
more stable, and in any case represent alternative sources of supply .
(Canadian production of nickel is discussed briefly in Chapter 4, because
Canadian production of platinum-group metals is predominantly as a byproduct
of nickel.)
Strikes tend not to last for very long, in comparison with other sorts of
contingencies, such as civil wars, so we specify disruption duration D = 1
year. Price increases during strikes will affect purchasers forced to go to
the spot market more than those under undisrupted long-term contracts, so we
specify the relative price increase to U.S. consumers to be only X = 3.
However, we specify somewhat more frequent disruptions occurring on the
average of every T = 15 years.
2-57
-------
Charles
River
Associates
Nickel is used in stainless steels, alloy steels, nonferrous alloys, and is
also used for electroplating. Limited substitution is possible, either
through use of different alloys, or use of different types of materials
altogether. We specify the U.S. price elasticity of demand to be e^ = 0.08.
(The appropriate elasticity would be somewhat greater if the disruption
lasted longer than one year.)
U.S. primary production of nickel is modest, normally accounting for between
10 and 15 percent of U.S. consumption. There has been a fair amount of
excess capacity in recent years, and production could probably be expanded
significantly over the course of a year. We specify the U.S. price
elasticity of supply to be es = 0.08.
There is modest secondary production of nickel from obsolete scrap, with
limited possibilities for expanding this recovery during periods of high
nickel prices. We specify the price elasticity of secondary U.S. supply to
be es = 0.02
TITANIUM
The critical ity of titanium is unlike that for the other metals we have
considered, in that disruptions in supplies of the elemental raw material are
not a serious threat, but "processing bottlenecks can be. Most titanium is
mined from deposits of the ore ilmenite, and is used in very mundane
applications as a white pigment. Less than five percent of titanium is
processed into a metallic "sponge" and then metal, the majority for
applications in aerospace hardware. Some is also used in steel alloys.
(Most metal is produced from the ore rutile, rather than ilmenite, though it
is possible at slightly higher cost to make metal from ilmenite if the
appropriate processing capacity has been constructed.)
The most likely bottleneck in the production of titanium metal is in the
capacity to make sponge from rutile. In fact, world production was seriously
constrained by sponge capacity in 1980 and 1981, causing a substantial
increase in prices over long-run equilibrium levels. Under incentives of
very high prices, new sponge plants can be built in a year or a year and a
half. Thus, disruptions in the supply of titanium metal are not likely to
last a long time. We specify duration D = 1 year. Consumers under long-run
contracts with producers may not suffer a great deal, which affects our
specification of a relatively low disruption severity X = 3.
By far the most common cause of insufficient sponge capacity is difficulty in
predicting the demand for titanium metal, as determined by such factors as
the rate of macroeconomic activity in industrialized countries and decisions
about military spending. As discussed earlier, there is some question
2-58
-------
Charles
River
Associates
whether the threat of such common commercial contingencies truly qualifies a
material to be considered "critical." Most of the same issues arise, but our
very simple economic model is not really set up to analyze situations where
the U.S. demand curve will likely be shifting outward when the contingency
occurs. We ignore this analytical complication for the present analysis,
except (as an imperfect adjustment) we specify negligible price elasticities
of domestic demand and supply, ej = 0 = es. In recent years the United
States has imported increasing amounts of titanium sponge from Japan and the
Soviet Union. Disruptions in these imports could also impose costs on the
United States.
Only modest amounts of titanium metal scrap are generated from obsolete
aircraft and chemical processing equipment. In most cases, scrappage of
obsolete titanium equipment would not be accelerated simply because of higher
scrap values, since titanium values are typically very small relative to the
total value of the equipment. Thus, we specify a very small price elasticity
of secondary production, es = 0.01.
PRICE ELASTICITIES OF CONSUMPTION FOR VEHICULAR EMISSIONS CONTROL
We have not investigated in any detail the efficiency of substituting for 409
stainless steel and other alloys in emissions control systems. Substitution
of materials such as aluminized steel has been proposed. The main cost
imposed by such a substitution would apparently be reduced durability,
necessitating replacement costs borne by car owners. EPA durability
requirements might have to be relaxed to allow such a substitution. It is
said that aluminized steels tested to date lasted less than half as long as
409 stainless steel in catalytic converters.
Under the price incentives that we assume to exist during disruptions (see
Table 2-2), we suppose for purposes of calculations that use of any of these
alloying materials can be eliminated (not necessarily simultaneously,
however). The implied elasticities of demand for vehicular emissions control
Ud) are as follows:
• Chromium: 0.25
• Manganese: 0.33
• Nickel: 0.5
t Titanium metal: 0.5
The above elasticity estimates should be regarded as illustrative rather than
definitive.
2-59
-------
Charles
River
Associates
For platinum, pallodium, and rhodium, we assume only very modest reductions
in usage would be possible in response to six- and seven-fold price
increases. For purposes of illustrative calculations, we specify price
elasticity ed = 0.03.
ILLUSTRATIVE CRITICALITY ESTIMATES AND CONCLUSIONS
Having explained all the parameters required for the calculations, we can now
compare the "criticality" (average annual ecoomic losses) for the seven
elemental materials described in Table 2-2 (national perspective) and Table
2-3 (vehicular emissions control only). From a national perspective,
platinum is the most critical of the materials considered, but chromium,
manganese, and nickel have average annual losses that are of the same order
of magnitude. The severity of the supply disruption threatening the platinum
market is ameliorated considerably by the extensive secondary recovery of old
material that takes place in the United States.
The value of normal U.S. nickel consumption is more than double that of any
of the other materials, but the severity and duration of disruptions is not
as great as for the other materials considered, and U.S. production, both
from primary and secondary sources, is significant. Also, the price
elasticity of demand is greater for nickel than for the other materials
considered.
The value of U.S. manganese consumption is somewhat greater than the value of
U.S. chromium consumption, but the greater severity of the disruption threat
for chromium makes it the more critical material.
The criticality of palladium, rhodium, and titanium metal from a national
perspective, is less than that of the other materials, largely because the
value of U.S. consumption in normal times is less. The criticality of
palladium and rhodium is substantially reduced by secondary recovery, while
the representative contingency threatening the market for titanium metal ties
with nickel for being the least severe of those specified.
The above conclusions about the relative criticality of the various materials
hold whether losses are calculated assuming no stockpiling or very large
stockpiles. (A comprehensive stockpile is usually only moderately larger
than an "import-eliminating" stockpile, and both are large relative to the
disruption threat.)
From the perspective of consumption for vehicular emissions control, platinum
is the most critical material by more than an order of magnitude relative to
any other material considered here. The value of platinum consumption dwarfs
2-60
-------
Charles
River
Associates
that for the other materials, and no other material faces a more severe
disruption threat. The criticality of palladium, rhodium, and chromium is
potentially significant but relatively low. The criticality of manganese,
nickel, and titanium is negligible. The criticality of the alloying
materials (chromium, manganese, nickel, and titanium metal) is lessened
considerably by large price elasticities of consumption (ej), reflecting our
assumption that alternative alloys (such as aluminized steels) could be used
during such a truly severe supply disruption.
As noted earlier, the above calculations of the criticality of platinum,
palladium, and rhodium, do not explicitly recognize the linkage between
current consumption of the material and possibilities for future secondary
recovery. In the case of platinum-group metals used for vehicular emissions
control, consumption taking place in the late 1970s and 1980s will develop
into a "rolling stockpile" of material that could considerably reduce U.S.
vulnerability to foreign supply disruptions in the 1990s. When this
consideration is fully taken into account, the criticality of platinum could
be considerably less than estimated above. We have not projected the size of
the U.S. rolling stockpile of platinum-group metals in this study. It will
be a simple exercise with the model mentioned earlier that CRA has designed
for the U.S. Department of the Interior.
The average annual losses from contingencies calculated in Tables 2-2 and 2-3
are much more illustrative than definitive, for reasons indicated throughout
the discussions above. However, these calculations clearly provide the right
types of information to guide policy decisions by the Environmental
Protection Agency. We finish our discussion in this chapter assuming the
criticality estimates in Table 2-3 are appropriate. It will be clear how to
use any revised estimates that become available.
Consider in particular the menus of materials required for the currently
projected emissions control technology as specified in Table 2-3. The total
annual cost for that menu, where all the indicated material markets are in a
normal state, is approximately $330 million per year. Expected annual losses
from contingencies in the various material markets are approximately
$50 million per year, assuming large stockpiles. (Private inventories are in
fact typically substantial for the more critical materials; Chapter 4
describes stockpiles for the platinum-group metals.) Total expected annual
costs for the menu of materials, including disruption costs, is thus
approximately $380 million per year.
Suppose that an alternative emissions control technology presented itself,
which was equally effective and otherwise like that described in Table 2-3,
2-61
-------
Charles
River
Associates
except that total expected annual costs for a substitute menu of materials
(including expected losses from market contingencies) was less than $380
million per year. That alternative technology would have a lower overall
expected cost than the currently projected technology using platinum-group
catalytic converters.
We confirm in Chapter 6 that no new technology currently on the horizon has
much promise of being more cost effective than platinum-group catalytic
converters. The critical materials penalty of approximately $50 million per
year constitutes only a very modest additional expected cost over the
normal cost of $330 million for the indicated menu of materials, particularly
when all the disadvantages of alternative control technologies are
considered.
If a new, more cost-effective technology should be developed, we have
reasonable confidence that the U.S. auto industry would choose that
technology in a way that appropriately weighs the criticality of the required
materials. The reasons for this conclusion were discussed earlier in the
chapter. They can be summarized by saying that, in this case, there appears
to be little reason for a significant discrepancy between the cost to private
firms of complying with EPA regulations, and the cost to society. (This
fortunate situation is not always the case for other EPA regulations.)
Moreover, assuming the U.S. auto industry believes that EPA emissions
standards of a given stringency will be in effect for many years into the
future, the industry should have sufficient incentive to undertake the amount
of research and development that is appropriate to that stringency. The
reason is analogous to that made above concerning appropriate private
decisions about which materials to use: The cost reductions made possible by
a new, lower cost emissions control technology would benefit the firm
developing the new technology to an extent comparable to the total national
benefit from reduced control costs. This optimistic conclusion, that U.S.
firms will undertake the amount of research and development approximately
appropriate for a given EPA standard, will, however, not hold if U.S. firms
seriously doubt that the standard will be maintained into the future.
The preparations that the U.S. auto industry has made for disruptions in
foreign supplies of platinum-group metals, particularly the size of
stockpiles they maintain, may be inadequate if they expect that relaxation of
EPA emissions standards can be arranged during those disruptions. As
protection against this possibility, EPA might consider requiring U.S.
vehicle manufacturers to hold specified minimum levels of inventories of
platinum, palladium, and rhodium.
The fact that we can directly use our definition of materials criticality to
choose compliance technologies having the lowest expected costs in the future
shows that we are using the appropriate concept. The concept of materials
criticality is designed to facilitate decisions about what materials to use
2-62
-------
Charles
River
Associates
and what policies (such as stockpiling and tariffs) to pursue, to protect
that rate of consumption of critical materials which is deemed efficient. In
this chapter we have emphasized the choice of materials to consume, but
extensions of the same methodology are appropriate for choosing efficient
stockpiles and tariffs. Earlier investigations have summarized appropriate
variables for determining materials criticality, but few have fashioned a
quantitative measure of criticality appropriate for direct inclusion in the
decisionmaker's balancing of economic (and noneconomic) costs and benefits.
Appendix 2-C, which is extracted from a 1978 CRA study, briefly surveys some
of these alternative analyses. Other work in this area, particularly more
recent work, is summarized briefly in the following Bibliographic Note.
2-63
-------
Charles
River
Associates
BIBLIOGRAPHIC NOTE
To our knowledge, no other study has specifically addressed in any detail the
issue of how a line agency of the federal government, without direct
responsibility for materials policymaking, should factor the criticality of
materials into its policymaking, for example where consumption of those
materials could be greatly increased by the agencies' decisions. However,
many other studies have treated in great detail more general aspects of the
problem, such as what contingencies threaten what markets, and what national
policies are most effective in reducing expected costs from contingencies.
We briefly survey here (and in Appendix 2-C) studies that are particularly
prominent, recent, unusual, or closely related to the analysis we have
presented in this chapter.
The Office of Minerals Policy and Research Analysis at the U. S. Department
of the Interior has taken important steps in recent years toward
implementating the type of methodology that we recommend for measuring the
criticality of materials. See Adams, White, and Grichar (1979), who consider
the criticality of bauxite, cobalt, copper, iron, and nickel. The analysis
by Adams, White, and Grichar is notable for considering a range of disruption
severities and for formally surveying (by "modified Delphi" techniques)
market experts to estimate disruption probabilities. An earlier
methodological and empirical study by the same group at the Department of the
Interior estimated expected economic costs from potential disruptions in the
markets for aluminum, chromium, platinum, and palladium. See U.S. Department
of the Interior (1975). This work was reviewed and revised in Charles River
Associates (1977). Two shorter articles summarizing basic methodologies and
conclusions are Adams (1977), and Burrows and Beggs (1977).
The CRA study that treats methodologies for estimating the criticality of
materials in most general terms is Charles River Associates (1978). The
chapter from that report which summarizes alternative methodologies for
measuring the criticality of materials is reproduced as Appendix 2-C to this
chapter. The most comprehensive CRA study in this area is the multi-volume
Charles River Associates (1976), which estimated economic costs from problems
in the markets for platinum, palladium, chromium, manganese, and three other
non-energy materials, as well as the market for petroleum. Other CRA studies
treating the criticality of materials which the United States imports, and
appropriate policy responses at the national level, include: Charles River
Associates (1977a), Charles River Associates (1976b), Charles River
Associates (1976a), and Charles River Associates (1975). Much of the CRA
work referred to above is summarized in the book Klass, Burrows, and Beggs
(1980), which specifically treats the same material markets as Charles River
Associates (1976).
2-64
-------
Charles
River
Associates
Since World War II, a number of landmark studies of problems and
contingencies affecting U.S. material markets have been conducted by
government-sponsored commissions. Among the most prominent were the U.S.
President's Materials Policy Commission (1952) and the National Commission on
Materials Policy (1973). Most studies by major government commissions
focused on general issues in mineral markets (such as resource depletion,
possibilities for more recycling, appropriate amounts of research and
development, and control of pollution from domestic mineral production),
rather than specifically on the types of contingencies that make a material
"critical." In the 1970s, several major government studies appeared that
focused more particularly on those types of contingencies, particularly
disruptions of foreign supplies. A short but useful example is the
Special Report on Critical Imported Materials by the Council on International
Economic Policy (1974).
The most sophisticated of the government studies undertaken in the 1970s was
that by the National Commission on Supplies and Shortages (1976), which went
into considerable depth on ways of improving federal policymaking on material
markets at the national level. Also notable for breadth in covering issues
affecting material markets, and appropriate policymaking at the national
level, are drafts produced by the recent Domestic Policy Review of Nonfuel
Minerals. See U.S. Department of the Interior (1979).
The third volume of U.S. Department of the Interior (1979) is entitled _A
Compendium of Issues, Options, and Recommendations Contained in Major
Post-war Nonfuel Mineral Policy Studies^It is available through the
Superintendent of Documents of the U.S. Government Printing Office. This
large volume provides a comprehensive bibliography and a collection of key
quotes from most of the significant studies of national policymaking for
nonfuel minerals undertaken since 1945. It includes many references that we
do not mention here because they are less directly relevant, less prominent,
less incisive, or representative only of special interests. We recommend
this source if more extensive bibliographic information than we provide here
is desired.
The Brookings study by Til ton (1977) is to be recommended for sensible
economic analysis of a broad range of contingencies and problems currently
facing minerals industries.
A recent study conducted at Resources for the Future, Inc. evaluates
contingencies in the markets for cobalt, chromium, manganese, aluminum,
copper, lead, and zinc, concluding that only for chromium does the United
States clearly face "undue vulnerability ... to contingencies that might
either seriously disrupt supplies or cause sharp upward movement of prices,
with consequent serious economic impact." No quantitative estimate of
economic costs or the criticality of the various materials is developed,
nowever. See Fischman (1979), p. 1, and also an abridged version issued in
book form, Fischman (1980).
2-65
-------
Charles
River
Associates
A study by Zabrowski and Lyle (1978) at the old Federal Preparedness Agency
(now the Federal Emergency Management Agency) investigates certain
characteristics of the input-output matrix for an economy that bear on the
criticality of its material industries, as we would recommend measuring it.
However, the analysis appears incomplete as a basis for choosing materials to
utilize, or policies to ameliorate supply disruptions, in part because it
does not consider the likelihood of contingencies of various severities.
A recent report by Kern 0. Kymn (1980) at the U.S. Federal Emergency
Management Agency is notable for assessing criticality of a material (steel
is used as an example) stemming from labor strikes, rather than foreign
supply disruptions.
It is worth noting that planning for contingencies affecting the price and
availability of materials can be considerably more unwieldy than the
situation faced by EPA in planning for vehicular emissions control. The
principal investigator of this study recently participated in a panel for the
National Materials Advisory Board, leading to the short report entitled
Identification of Critical and Strategic Materials for Naval Combat Systems,
NMAB (1981).Navy systems are so complex that it is inordinately costly to
trace all the materials used back through third and fourth level vendors to
the suppliers of the original raw materials, despite the fact that military
purchases of components are much better documented than most commercial
purchases. As a result, production bottlenecks can spring up in unexpected
places in the chain of materials supply.
There are, of course, a multitude of sources of information and data on the
markets for individual materials. Those for the platinum-group markets are
reviewed in Chapter 4. Similar sources of information are available for
chromium, manganese, nickel, titanium, and other potentially critical
materials. Publications of the U.S. Department of the Interior provide
particularly useful summaries of relevant information. The Annual Report of
the Secretary of the Interior Under the Mining and Minerals Policy Act of
1970 usually provides a useful summary of policy issues in minerals markets
that are considered most pressing at the national level. The U.S. Federal
Emergency Management Agency (FEMA) makes quarterly (National Defense) Stock-
pile Reports to Congress, describing the operative legislation, the current
status of the stockpile, goals for the future size of the stockpile, and
recent activities (including a separately bound statistical supplement).
2-66
-------
Charles
River
Associates
APPENDIX 2-A
AVAILABILITY OF MATERIALS FROM THE
U.S. NATIONAL DEFENSE STOCKPILE
2-67
-------
Charles
River
Associates
APPENDIX 2-A
AVAILABILITY OF MATERIALS FROM THE
U.S. NATIONAL DEFENSE STOCKPILE
In this appendix we consider possible avenues by which materials can be
released from the National Defense Stockpile for non-military contingencies.
Our general conclusion is that an act of Congress would almost certainly be
required to authorize release of materials required for vehicular emissions
control during a non-military emergency.
The U.S. National Defense Stockpile was established just prior to World War
II. Its sole objective is to "serve the interest of national defense." It
is "not to be used for economic or budgetary purposes." The enabling
legislation for the stockpiling, amended in 1979, further requires that the
quantities of material stockpiled are to be "sufficient to sustain the United
States for a period of not less than three years in the event of a national
emergency." Stockpile goals are estimated by the Federal Emergency
Management Agency (FEMA) as the difference between requirements for materials
in wartime and the amount likely to be available, assuming some austerity in
non-military consumption.
The Strategic and Critical Materials Stockpiling Act (as amended July 30,
1979) allows disposal of materials only under specified authorities, the most
general of which is an act of Congress. The President can authorize
disposals under more narrowly defined conditions. Perishable materials are
to be "rotated" out of the stockpile to prevent deterioration. With prior
notification of the Committees on Armed Services of the Senate and House of
Representatives, the President can dispose of materials that are in excess of
stockpile requirements, or that may deteriorate in value if not sold.
Beyond the above described types of stockpile releases, which are primarily
designed for routine management of the stockpile, Section 7 of the Strategic
and Critical Materials Stockpiling Act also grants the President more
discretionary authority. In time of war declared by Congress, or in time of
a declared national emergency, the President or his delegate may release
materials specifically required for the national defense. The President also
has the power during times other than declared emergencies or wars to release
material specifically required for purposes of the national defense.
Historically, there have been 28 releases authorized by the President under
Section 7; all but three of these were made during World War II, the Korean
Conflict, or the Viet Nam Conflict. All releases under Section 7, including
those made during wartime, have been made directly by the President, rather
than through delegated authority. The three releases made at other times
consisted of mercury (in 1956 and 1959) and asbestos (in 1979). The critical
2-68
-------
Charles
River
Associates
consideration in all these releases was that the material would be used
directly for defense purposes. Since use of material for control of
vehicular emissions ostensibly does not constitute application to national
defense, it is highly doubtful that releases under Section 7 could be made by
the President for that purpose.
It is conceivable that platinum-group metals or other material required for
vehicular emissions control, might be available from the stockpile, because
amounts held in excess of requirements are due to be sold. (There is no
excess currently.) However, in this case the President is required to use
competitive negotiation and formal advertising, unless prior explanation is
made to Congress. In the usual case of competitive bidding, producers of
emissions control equipment would have to vie with other consumers for the
materials they wish to obtain.
There is one further avenue under which the President has in the past
authorized releases from the National Defense Stockpile. Section 101 (b) of
the Defense Production Act of 1950 (as amended August 20, 1980), gives the
President general authority to allocate materials in the civilian market when
he finds:
(1) that such material is a scarce and critical material essential to
the national defense, and (2) that the requirements of the national
defense for such material cannot otherwise be met without creating a
significant dislocation of the normal distribution of such material in
the civilian market to such a degree as to create appreciable
hardships.
In two cases, historically, it has been found that operation of the defense
priorities system has created sufficient "hardships" in the civilian market
that releases from the national stockpile were justified. The materials
released were argon gas and titanium. Since contingencies other than
operation of the defense priorities system seem most likely to affect
materials required for vehicular emissions control, this avenue for arranging
releases from the National Defense Stockpile also does not promise to be very
useful for meeting EPA related materials consumption in a non-military
emergency.
The general conclusion that emerges from the above review of peacetime
releases from the national stockpile is that almost certainly an act of
Congress would be required to respond to the types of non-military
contingencies that appear most likely in the market for platinum-group
metals. This conclusion is particularly clear if EPA were to wish that the
material be specifically allocated for use in vehicular emissions control
equipment, rather than being put up for competitive bids among all consumers
in the market. A FEMA official with whom we discussed the matter concurred
with this conclusion.
2-69
-------
Charles
River
Associates
APPENDIX 2-B
COMPUTER PROGRAM TO CALCULATE AVERAGE ANNUAL ECONOMIC
LOSSES FROM CONTINGENCIES IN MATERIAL MARKETS
2-70
-------
Charles
River
Associates
APPENDIX 2-B
COMPUTER PROGRAM TO CALCULATE AVERAGE ANNUAL ECONOMIC LOSSES
FROM CONTINGENCIES IN MATERIAL MARKETS
This appendix lists a simple computer program designed to calculate the
economic losses determining the criticality of a material. The very simple
economic model upon which the program is based is described in the text of
the chapter. The formulas applied are described fully in Table 2-1. The
required criticality calculations can be performed by hand without
difficulty, but the program is a convenience, particularly where one is
interested in disaggregating losses according to whether they are suffered by
U.S. consumers, primary producers, secondary producers, or stockholders. The
numbers displayed in the following listing of the program and its output are
for the "catalystium" example explained in the early part of the chapter.
The program is written in the BASIC language for the Tektronix 4052 computer,
but should be readily adaptable to most other minicomputers. The program
requires less than 8,000 bytes of core.
2-71
-------
LIST
108 PEM—PROGRAM TO CALCULATE MATERIALS CRITICALITY (EXPECTED ANNUAL
118 PEM—ECONOMIC LOSSES FROM CONTINGENCIES)
128 REM—SAMPLE CALCULATIONS FOR IMAGINARY "CATALYSTIUM" MARKET
138 PEM—UERSION 1, AUGUST 15, 1981, S. BEGGS
140 REM--PARAMETERS
158 REH—VALUE OF NORMAL U.S. CONSUMPTION FOR VEHICULAR EMISSION
160 REM--CONTROL <$ MILLION)
170 Ui=l
ISO REM—PRICE ELASTICITY OF U.S. CONSUMPTION FOR VEHICULAR EMISSION
196 REM—CONTROL
260 Ei=8.1
210 REM—SUPPLY DISRUPTION SEVERITY (PROPORTIONAL INCREASE IN WORLD
228 REM—PRICE)
239 X=3
249 Z=X-1
258 REM—EXPECTED TIME BETWEEN (STARTS OF) DISRUPTIONS (YEARS)
266 T=10
278 REM—EXPECTED DURATION OF DISRUPTIONS (YEARS)
288 D=l
290 REM—VALUE OF TOTAL NORMAL U.S. CONSUMPTION <* MILLION)
300 1*2=2
310 REM—PRICE ELASTICITY OF TOTAL U.S. CONSUMPTION
328 E2=0.2
338 REM—VALUE OF NORMAL U.S. PRIMARY PRODUCTION ($ MILLION)
340 V3=0.1
350 REM—PRICE ELASTICITY OF U.S. PRIMARY PRODUCTION
360 E3=0.125
378 REM—VALUE OF NORMAL U.S. SECONDARY PRODUCTION ($ MILLION)
380 V4=6.3
390 REM--PRICE ELASTICITY OF U.S. SECONDARY PRODUCTION
400 E4=6.125
410 REM—STOCKHOLDING COST, PER $1 MILLION OF MATERIAL PER YEAR
420 REM—•<* MILLION, AT NORMAL PRICES)
430 C=6.065
-------
i
^j
CO
448
45Q
468
478
488
49Q
580
510
528
538
548
558
560
578
580
598
688
618
628
638
648
658
668
678
688
698
788
718
28
REM—CALCULATIONS FOR CONSUMERS PRODUCING UEHICULAR EMISSION
REM—CONTROL EQUIPMENT
Il=Ul*El*Zt2/2/T*D
I2=im-'T*D
18=11+12
REH—CALCULATIONS FOR THE UNITED STATES
Jl=U2*E2*2t2/2--'T*D
J2 = U2t.''T*D
J8=J1+J2
Ki=U3*E3*Zt2-''2/T*D
738
748
750
768
770
788
K8=K1+K2
Ll=U4*E4*ZT2/2/T*D
L2=-U4* < Z+Zt2*E4 > /
L8=LH-L2
Hl=8
H2=8
H8=8
PAGE
PRINT "EXPECTED U.S. ECONOMIC LOSSES FROM MARKET CONTINGENCIES"
PRINT " i* MILLION, AVERAGE PER YEAR;"
PRINT " "
P Ij T L I T II II
PRINT "ADDITIONAL COSTS FOR PRODUCERS OF UEHICULAR EMISSION"
PRINT "CONTROL EQUIPMENT (NO STOCKHOLDING)"
PRINT " TOTAL = "!10
PRINT " ADJUSTMENT = "511
PRINT " TRANSFER = "512
PRINT " "
PRINT " "
PRINT "U.S. LOSSES WITH NO STOCKPILE"
GOSUB 800
-------
no
i
798
800
810
820
838
848
GO TO 1010
REM—SUBROUTINE
TO PRINT OUT LOSS DISAGGREGATIOH
860
870
969
918
920
930
948
950
968
978
988
950
1800
1818
1828
1830
1040
1050
I860
1870
1088
1090
1180
1110
1120
1136
PRINT "
PRINT "
PRINT "
PRINT "
PRINT "
PRINT "
PRINT "
PRINT "
PRINT "
PRINT "
PRINT "
PRINT "
PRINT "
PRINT !I
PRINT "
PRINT "
PRINT "
PRINT "
PRINT "
RETURN
STOP
PAGE
PRINT "U.S
H2=-M2
M2=0
CONSUMERS"
SUBTOTAL = "JJO
ADJUSTMENT COST
TRANSFER COST =
PR I MAR'
PRODUCER?"
:.O
II
ST
ST =
SUBTOTAL = "5KO
ADJUSTMENT CO
TRANSFER CO
SECONDARY PRODUCERS"
SUBTOTAL = "!LO
ADJUSTMENT COST
TRANSFER COST =
STOCKHOLDERS"
SUBTOTAL = "5H0
HOLDING CO
3TS
J2
K2
";LI
L2
HI
u.s,
ACQUISITION COST, LESS REUENUES =
H2
TOTAL = "?M0
ADJUSTMENT &
TRANSFER (TO
STOCKHOLDING
FOREIGNERS) =
Ml
M2
LOSS
JES
WITH IMPORT-ELIMINATING STOCKPILE"
HO=H1+H2
M1=J1+K1+L1+H1
880
GOSUB
STOP
PAGE
PRINT
"U.S. LOSSES WITH COMPREHENSIVE STOCKPILE"
-------
OH3 00£T
098 anSQD 06cl
Q=SW QlZ
TH=TW 89c
tH=0H RSc
0=cH 0tc
0=01 0cc
0=0>l 06 1 1
0=c!M 08TI
0=1 X 0^H
09U
-------
EXPECTED U.S. ECONOMIC LOSSES FROM MARKET CONTINGENCIES
U MILLION, AVERAGE PER YEAR)
ADDITIONAL COSTS FOR PRODUCERS OF VEHICULAR EMISSION
CONTROL EQUIPMENT (NO STOCKHOLDING)
TOTAL = U.18
ADJUSTMENT =0.02
TRANSFER = 0.16
U.S. LOSSES WITH NO STOCKPILE
CONSUMERS
SUBTOTAL = 0.32
ADJUSTMENT COST = 0.08
TRANSFER COST = 0.24
PRIMARY PRODUCERS
-lu SUBTOTAL = -0.0225
ADJUSTMENT COST = 0.0025
TRANSFER COST = -0.025
SECONDARY PRODUCERS
SUBTOTAL = -0.0675
ADJUSTMENT COST = 0.0075
TRANSFER COST = -0.075
STOCKHOLDERS-
SUBTOTAL = 0
HOLDING COSTS = 0
ACQUISITION COST, LESS REVENUES = 0
U.S. TOTAL = 0.23
ADJUSTMENT & STOCKHOLDING COSTS = 0.09
TRANSFER (TO FOREIGNERS) = 0.14
STOP IN LINE 1010 PRIOR TO LINE 1020
-------
U.S. LOSSES WITH IMPORT-ELIMINATING STOCKPILE
CONSUMERS
SUBTOTAL = 0.32
ADJUSTMENT COST = 8.88
TRANSFER COST = 8.24
PRIMARY PRODUCERS
SUBTOTAL = -8.8225
ADJUSTMENT COST = 0.8825
TRANSFER COST = -8.825
SECONDARY PRODUCERS
SUBTOTAL = -8.0675
ADJUSTMENT COST = 8.8075
TRANSFER COST = -8.075
STOCKHOLDERS
SUBTOTAL = -0.0945
HOLDING COSTS = 8.0455
ACQUISITION COST, LESS REMENUES = -8.14
U.S. TOTAL = 0.1355
ADJUSTMENT fe STOCKHOLDING COSTS = 0.1355
TRANSFER
-------
U.S. LOSSES WITH COMPREHENSIUE STOCKPILE
CONSUMERS
SUBTOTAL = 0
ADJUSTMENT COST = 0
TRANSFER COST = 6
PRIMARY PRODUCERS
SUBTOTAL - 6
ADJUSTMENT
TRANSFER CO
SECONDARY PRODUCERS
SUBTOTAL = 0
ro
oo
COS
T
•T =
it
= 0
3
ADJUSTMENT COST = 0
TRANSFER COST = 0
STOCKHOLDERS
SUBTOTAL = 0.104
HOLDING COSTS = 0.104
3T
U.S
ACQUISITION
TOTAL - 0.104
ADJUSTMENT fe STOCKHOLDIHG
TRANSFER
-------
Charles
River
Associates
APPENDIX 2-C
EARLIER APPROACHES TO MATERIALS CRITICALITY
2-79
-------
Charles
River
Associates
APPENDIX 2-C
EARLIER APPROACHES TO MATERIALS CRITICALITY
The following survey of earlier analysis of materials criticality is taken
from Charles River Associates (1978). A few more recent studies are
discussed briefly in the Bibliographic Note. The chapters preceding the
following discussion (in its original report), discussed materials
criticality in more general terms, including the application of more
sophisticated economic models to the determination of materials criticality.
However, the text of the present chapter is sufficient introduction to
understand all important issues that are raised in this appendix.
Earlier studies of materials criticality have in the main
had the same concerns which motivated the analysis we have pre-
sented above. In this chapter we discuss some of the more prom-
inent of these earlier studies in light of the basic principles
and economic modeling recommended in the previous two chapters.
Examining the strengths and weaknesses of these earlier efforts
from such a perspective also allows us to elaborate upon our
earlier discussions and provide further examples.
Import Dependence
The most prominent area of policy analysis for which criti-
cality measurements have been developed is undoubtedly that of
foreign supply disruptions. This emphasis has been particularly
pronounced since the formation of OPEC and the subsequent Arab
oil embargo.
The typical first step toward investigating the criticality
of imported materials is to rank the materials by the percentage
of U.S. import dependence which each accounts for. A bar chart
2-80
-------
CHARLES RIVER
ASSOCIATES i
INCORPORATED
of such a ranking for the year 1972 is reproduced below as Figure
4-1.l U.S. import dependence for platinum group metals, chromium
and cobalt was nearly complete, whereas only a moderate percentage
of U.S. copper consumption had to be imported.
Although such a ranking of materials is useful, it obviously
has limitations as a measure of criticality. The most obvious
limitation of Figure 4-1 is the omission of information on the
absolute size of the market. For example, although U.S. import
dependence was slightly greater for tantalum than it was for
aluminum (ores and metal), the absolute value of the aluminum
ore imports is so much greater than the absolute value of the
tantalum imports that there can be little doubt that aluminum
is the more critical material.2
Chapter 3 of the present study is a good general guide to
the many further relevant considerations which need to be inclu-
ded in a measurement of criticality, such as the probability of
a disruption occurring and the existence of substitutes. It is
worth reiterating a point previously made about the use of pol-
icy models for measuring materials criticality. If materials
markets to be compared are assumed to be identical in all re-
spects except the value of imports — the same normal inventory/
consumption ratio, the same price elasticities, the same dis-
ruption severity, the same disruption probabilities and so on —
then the resulting criticality loss measures are simply propor-
tional to import dependence in absolute dollar terms.
JThis particular diagram is from National Commission on
Materials Policy, Material Needs and the Environment Today and
Tomorrow, Final Report (Washington, D.C.: U.S. Government Print-
ing Office, June 1973), pp. 2-25. Similar charts have been
produced by other investigators.
On the other hand, measuring import dependence as a percent-
age of consumption does have the great advantage of indicating
whether there are domestic producers who might expand production
when foreign supplies are cut off.
2-81
-------
Figure 4—1
PERCENTAGE OF U.S. MINERAL REQUIREMENTS
IMPORTED DURING 1972
CHARLES RIVER
ASSOCIATES
INCORPORATED
PERCENTAGE IMPORTED
MAJOR
FOREIGN SOURCES
100%
60%
2SX
i
0%
I
PLATINUM GROUP METALS
MICA »—;
CHROMIUM
STRONTIUM
COBALT
TANTALUM
ALUMINUM to—*—il
MANGANESE
FLUORINE
TITANIUM taM
UK. USSR, SOUTH AFRICA. CANADA, JAPAN. NORWAY
MMA. IRA2IL. MALAGASY
USSR. SOUTH AFRICA. TURKEY
MEXICO, VAIN
ZAIRE, ICLGIUM. LUXEMBOURG. FINLAND, CANADA. MOMMA Y
NIGERI A. CANADA, ZAIRE
l JAMAICA, (URIHAM, CANADA. AUSTRALIA
«RAZIL.GASO«, SOUTH AFRICA. ZAIRE
a MEXICO. SPAIN. ITALY.SOUTH AFRICA
^ AUSTRALIA
-'J CANADA. lOUTX AFRICA
MAl>rSlA. THAIUkNO. «OUVtA
H MEXICO. JAPAN. PERU. UK. KOREA
a CANADA. NORWAY
.^ ««A21L. NIGERIA. MAlAGAnr. THAI LAND
J IOUTH AFRICA. MtXICO. UK. *OUVIA
t m.l.l.l.1.1.1.' 11111' 11 WW] CANADA. IWITZERLANO.UBR
I Tftffi?*fPPfrf!rJJJJ'UMf* CANADA
I agiiiJAM^JMMMaM CANADA. MEXICO
' P,,. • ""• LU.IJJ.I. .IJU CANADA. MEXICO. FCRU
I IUJAIJJUAULIJUIUUUI CANADA. PERU. MEXICO. HONOURAt. AUVTRAUA
I tMHUWAfMHyorfm PERU. IREUkNO. MEXICO. ORECCE
I BA».".I.I.I.'.IJ.I.UUW< CANADA, MEXICO. JAMAICA
I KtfJ*fMf.VMm CANADA, JAP AN. MEXICO. UK
I KJJJJWAMJAAM PCRU, CANADA
f muuuuuuumaM •OUTH AFRICA, CHIUE. UBR
I ifff^fffBoaaa CENTRAL » SOUTH AMERICA .CANADA.MIDDLE EACT
1 IBPIIIflBflPflm CANADA. VENEZUELA, JAPAN. COMMON MARKET KECI
I HJLU.UJJUUU CANADA. AUTTRAUA.PERU. MEXICO
I •IIIIIM« MEXICO, AUSTRALIA. itLCIUM. LUIKHO URC. CANADA. FCRU
I BWWVT CANADA. PERU. CHILE
| JUUUWM CANADA. AUCTRAUA
1 FAMJUI AUSTRALIA. MALAYSIA. INDIA
I1 IUUW» GREECE. ITALY
| m CANADA,MEXICO. (AMAMAS
GREECE. IRELAND
CANADA
WEST GERMANY. FRANCE
CAMAOA. MEXICO. ITALY, PORTUGAL
*0% 2S% OK
SOURCE: National Commission on Materials Policy, Material Needs and the Environment Today and Tommorrow
(Washington D.C.: U.S. Government Printing Office, June 1973), p. 2—25.
2-02
-------
CHARLES RIVER
ASSOCIATES
INCORPORATED
Measuring Criticality Without Economic Models
Even if one's basic definition of materials criticality
is much less explicit and detailed than that developed above in
Chapter 2, it is not particularly difficult to identify many of
the factors which determine such criticality. A number of stud-
ies have compiled comprehensive lists of determinants on a sub-
jective basis, without constructing a formal economic model.
Typically, cross tabulations of characteristics for various ma-
terials have been prepared, with summary index values calculated
for each material by arbitrarily weighting the characteristics.
We will now review a few of the more important studies which
have developed measures of materials criticality without the use
of an economic model
Study by King and Cameron
One of the more detailed and formal efforts along these
lines has been done by Alwyn King and John Cameron at the U.S.
Army War College.1 The list of observable factors which they
considered relevant to determining U.S. vulnerability to for-
eign supply disruptions is reproduced below as Table 4-1.
In order to compute a summary measure of criticality, King
and Cameron first assigned arbitrary numerical weights to large
(L), medium (M) and small (S) effects on the materials' vulner-
ability for the factors listed in Table 4-1. Then the direction
and magnitude of the factor's effect on the vulnerability of par-
ticular materials were also characterized by arbitrary weights.
The two weights were multiplied in each cell of Table 4-1 and then
added, first for "economic, political and military" considerations
separately, and then for all three together. Each material inves-
tigated was treated in this fashion in turn.
1Alwyn H. King and John R. Cameron, "Materials and the New
Dimensions of Conflict" (Carlisle Barracks, Pa.: U.S. Army War
College, Strategic Studies Institute, May 15, 1974); Alwyn H.
King, "Materials Vulnerability of the United States — An Update"
(Carlisle Barracks, Pa.: U.S. Army War College, Strategic Stu-
dies Institute, April 30, 1977).
2-83
-------
CHARLES RIVER
ASSOCIATES
INCORPORATED
Table 4-1
FACTORS AFFECTING COMMODITY RELATIVE VULNERABILITY INDEX,
AS COMPILED BY KING AND CAMERON
L = Large
Factor Effect on Vulnerability M = Medium
S = Small
Economic Political Military
Domestic reserves:
Avai labiIity L L L
Cost of developing L L S
Domestic production industry:
Present capability L L L
Cost of augmenting L L S
Substitute materials:
Present availability L L L
Cost of research to develop L L S
Time required to develop ILL
Additional domestic resources:
Present availability ILL
Cost to develop suitable processes L M M
Time to develop suitable processes M MM
Probability of discovery if not available M M S
Cost of additional exploration M M S
Foreign suppIiers:
Number of controlling companies L S M
Number of supplier countries M L M
Political stability of suppIier countries M L M
Ideology of supplier countries L L L
Productive capacity of supplier countries L L L
Economic sufficiency of supplier countries L L S
History of political relations with US S 'M S
US dollar involvement in supplier country M M S
Accessibility of supplier countries (supply
routes) SSL
US Stockp Me:
Present US stockpile objective L L L
Actual quantity in US stockpile M M M
Customary industry stockpile M M M
Table continued on following page.
2-84
-------
CHARLES RIVER
ASSOCIATES
INCORPORATED
Table 4-1 (Continued)
FACTORS AFFECTING COMMODITY RELATIVE VULNERABILITY INDEX,
AS COMPILED BY KING AND CAMERON
L = Large
Effect on Vulnerability M = Medium
S = Small
Economic Political Military
Trend in usage of critical material M M S
Proportion of national consumption directly
related to military requirements SSL
Importance of secondary sources (recycling) M M M
SOURCE: Alwyn H. King, Materials Vulnerability of the United States
An Update (Carlisle Barracks, Pa.: U.S. Army War College,
Strategic Studies Institute, April 30, 1977), p. A-2.
2-85
-------
CHARLES RIVER
ASSOCIATES
INCORPORATED
King and Cameron have only reported results for a truncated
version of the above methodology based on a consideration of the
following five factors: availability of domestic reserves; avail-
ability of substitutes; number of foreign suppliers; ideology of
foreign suppliers; and U.S. stockpile objective. Also, they con-
sidered only eleven of the materials most likely to have high
criticality ratings. Results, in rank order with their "vulnera-
bility index ratings," were chromium (34), platinum group (32),
tungsten (27), manganese (23), aluminum (22), titanium (20), co-
balt (20), tantalum (16), nickel (14), mercury (11) and tin (6).1
These results are intuitively plausible. However, applying the
methodology to many more materials using all 27 factors in Table
4-1 could easily lead to questionable results since there is no
overall economic theory underlying the analysis.
Relative Inclusiveness of the CRA Policy Model
It is interesting to compare King and Cameron's list of de-
termining factors and subjective approach to materials criticality
with the use of the CRA economic policy model described in Chap-
ter 3. Both approaches are concerned primarily with foreign sup-
ply disruptions. Use of an economic model makes much heavier
demands on the analyst when new factors are to be formally in-
cluded in the methodology. However, the economic modeling ap-
proach described in Chapter 3 in fact includes explicitly or im-
plicitly virtually all of the observable factors listed in Table
4-1. Furthermore, application of the economic model greatly
enhances the value of this basic information by utilizing as
accurate a quantitative specification as possible, and by trans-
lating this information into dollar loss figures, broken down by
who gains and who loses. Thus, groups with differing interests
and values have information in the most useful possible form for
:King and Cameron, op. ait., p. 17.
2-86
-------
CHARLES RIVER
ASSOCIATES |
INCORPORATED
ordering their own priorities and for participating in the
process of determining a national consensus on priorities and
policies. The complexity and scope of the materials criti-
cality problem is reduced in the most meaningful possible way
by applications of an economic model.
In the remainder of this section we will-consider-the- -
factors listed in Table 4-1 and evaluate how adequately they
can be incorporated in the economic model described in Chap-
ter 3. In .some-cases, a generalization of the model in Chap-
ter 3 would allow somewhat fuller recognition of relevant cir-
cumstances (e.g., the depletion of reserves), but in general
the available policy models can be very comprehensive if care-
ful analyses precede specification of their parameters.
Domestic Reserves
U.S. reserves of the commodity under study should be
important determinants of the domestic supply curve in the
CRA. policy model; however, it is not possible to infer directly
the flow of domestic production from the stock of reserves.
If reserve estimates were accurate and inclusive, it might
be desirable to allow explicitly for backward shifts in the
U.S. supply curve as the domestic resource base is depleted.
However, reserve estimates only reflect the state of knowledge
at a particular point in time, and historically depletion has
been more than offset by discovery and technological change
in most cases. Thus, the fact that the existing CRA policy
model ignores depletion is usually not a serious omission,
though it may be so in cases such as petroleum where the size
of domestic reserves is relatively well known.
2-87
-------
CHARLES RIVER
ASSOCIATES
INCORPORATED!
Domestic Production Industry
The present capability of the domestic production industry
is indicated by the position of the short-run domestic supply
curve. The cost of (gradually) augmenting domestic production
capacity is indicated by the position of the long-run domestic
supply curve relative to the short-run curve.
The intersection of the short-run and long-run domestic
supply curve occurs at a single rate of production which can
be referred to as the "capacity" of the domestic industry,
though capacity in this sense represents an optimal adaption
of industrial capital to the existing rate of production, not
an upper bound on production.
Substitute Materials
The availability of substitutes under current technology
is usually the primary determinant of the short-run and long-
run price elasticities of demand in the CRA policy model. In
many cases important "substitutes" for a material whose criti-
cality is being investigated will be factors of production
other than alternative materials. For example, labor may be
substituted for a material such as manganese by increasing
the attentiveness of workers controlling its consumption in
steelmaking.1
Substitutions may also be made by consumers of final
goods and services; when the price of a potentially critical
material rises, consumers may decide to decrease their pur-
chases of goods requiring its use. All these considerations
When the price of a material input shifts upward, an
economist considers any factor of production which is conse-
quently used in greater amounts to be a "substitute" for the
material input. The engineer's definition of a "substitute"
is much less general.
2-88
-------
CHARLES RIVER
ASSOCIATES
INCORPORATED
affect the "derived demand curve" for a raw material which
enters the policy model. The demand curve for the raw material
is characterized as "derived" because it can be determined
from the demand curve for the final good and information about
the substitutions which producers can make.
In the standard market model underlying the CRA policy
model, changes in U.S. technology, such as development of new
substitute materials, are theoretically represented as shifts
in the demand curves. In practice, movements along a demand
curve are often interpreted to include some minor predictable
technological innovations and some efficiencies due to "learn-
ing by doing." The dynamic implications of such processes are
captured in their relevant form in the CRA node! by the short-
run and long-run supply and demand curves, and by solving for the
intersections of short-run and long-run curves which maximize
firms' profits over time.
Domestic Resources
U.S. "resources," as opposed to "reserves," would become
economical to exploit only at prices higher than normal. Thus,
the shape of the domestic long-run supply curve is strongly
affected by the existence of resources. However, unless a
market disruption (such as the formation of OPEC) is expected
to endure, it may not be efficient to exploit such resources
because of the long lags between substantial investments and
actual domestic production. If exploitation requires signi-
ficant research and development, the same distinctions between
dynamic adjustments in short-run curves and shifts of under-
lying long-run curves must be made on the supply side of the
domestic market as were discussed above for the demand side
of the market.
2-89
-------
CHARLES RIVER
ASSOCIATES
INCORPORATED
Foreign Suppliers
The number and nature of foreign suppliers of an imported
material determine the probability of supply disruptions of
various lengths and severities. There is in practice no reli-
able relationship between readily discerned characteristics
of supplier countries and the probability of disruptions. The
characteristics listed in Table 4-1 are suggestive in particular
cases, but expert assessment of the overall disruption proba-
bility is undoubtedly much more reliable than subjective weight-
ing of the tabled characteristics.1
Plausible disruption scenarios involving various producer
countries can often be translated into effects on the world
market price using econometric market models prior to appli-
cation of a policy model; the effect of stock releases on the
world market price during the disruption should also be esti-
mated. The particular approach chosen depends on the market
structure both in normal times and during the disruption. For
example, the structure of the chromite market was fairly com-
petitive in the 1950s and early 1960s, but some collusion among
producing countries has occurred during the 1970s. A sharp
cut-off of supplies from southern Africa would today encourage
much greater collusion than that which has occurred since 1974.
Thus, predicting world prices during such a disruption involved
both constructing a cartel pricing model utilizing historical
behavior patterns for consumers and suppliers who were assumed
!0ne characteristic not explicitly mentioned in Table 4-1
involves distinguishing sources of supply as to whether or
not they are controlled by less developed countries. Some
investigators have made this distinction an important element
in their analysis of materials criticality, but, as with the
other characteristics, it is only part of the broader situ-
ation which determines the probability of disruptions of
various severities and durations.
2-90
-------
CHARLES RIVER
ASSOCIATES
INCORPORATED
to remain competitive, and utilizing hypothetical beh'avior
patterns leased on profit maximization for the suppliers who
were assumed to collude.1
The situation described above for the chromite market
in which there is a competitive market structure in normal
times and monopolistic pricing during the supply disruption
is typical. However, in analyzing a cutoff of cobalt supplies
from Zaire due to the recent military conflict in Shaba Prov-
ince, it was appropriate to reverse this pattern. Zaire nor-
mally exercises considerable monopolistic power as the price
setter in the cobalt market. However, during a complete dis-
ruption of Zaire the "fringe" of competitive suppliers would
become the only sources of supply and the market structure
would hence become fully competitive. Determining a competi-
tive price which clears the disrupted market by utilizing
historical behavior patterns captured by the CRA econometric
model then became appropriate.2
Before applying the policy model discussed in Chapter 3,
it is appropriate to utilize formal market models in the man-
ner illustrated above to determine the severity of plausible
disruption scenarios in terms of world price increases. This
formal modeling approach contrasts with the approach recommended
earlier for determining the probability of a disruption begin-
ning and continuing, when this probability must generally rely
on a subjective evaluation of a multitude of relevant factors,
:See Charles River Associates Incorporated, The Report of
the U.S. Department of the Interior on the Critical Materials
Aluminum, Chromium, T?latinum and 'Palladium: A Review and Revision
(Cambridge, Mass.: CRA, July 1977), Chapter 6.
2See Charles River Associates Incorporated, Implications of
the War in Zaire for the Cobalt Market (Cambridge, Mass.: CRA,
June 1977).
2-91
-------
CHARLES RIVER
ASSOCIATES
INCORPORATED
like those characteristics of foreign suppliers which are
listed in Table 4-1. Expert opinions on probabilities are
likely to differ significantly. However, the probabilities
are readily interpreted intuitively and discussion of whether
the chance of a disruption is one chance in four or one chance
in forty over the next decade is much more incisive than an
arbitrary weighting of a mixed bag of relevant considerations.
U.S. Stockpile
The size of industry and government stocks which are
available during an economic disruption directly enters the
determination of materials criticality when the CRA policy
model is applied. Moreover, the specified size of stocks
held by other consuming nations can be of crucial importance,
though it is not listed in Table 4-1. In general, the pre-
paredness of foreign consumers determines how aggressively
they bid for remaining world production during a disruption,
and hence affects the price at which imports can be obtained
by the United States.
One of the significant advantages of applying a formal
economic model to the determination of materials criticality
is simply the identification of relevant considerations which
may be overlooked if the structure of the problem is delineated
more informally. Remembering to include the stocks of foreign
consumers is a case in point.
In general, there is a vast economic literature on such
problems as estimating supply and demand curves which can be
drawn upon to improve criticality measurements. For example,
the smaller the proportion of the cost of final goods which
is accounted for by a potentially critical raw material, the
less price elastic its demand will tend to be and hence the
2-92
-------
CHARLES RIVER <
ASSOCIATES I
INCORPORATED I
greater its criticality; formulas precisely specifying such
relationships are often very useful for studies such as those
we are recommending for determining materials criticality.
Trend in Usage
As importation of a material increases, the expected
dollar loss which measures its criticality will generally
increase as well. If the domestic quantities demanded,
supplied, and stocked at the normal price increase propor-
tionately, and if the price elasticities and other parameters
remain constant, then expected dollar losses calculated with
the CRA policy model will also increase proportionately. In
general, of course, trends in various categories of end users
can affect the price elasticity of demand for all categories
combined, which appropriately summarizes the information
required by the policy model; criticality loss measures would
then not increase precisely proportionately with the quantity
of total consumption.
At any time, the fact that there has recently been
an unexpected change in the trend of usage can make an impor-
tant difference in criticality ratings. A sharp increase in
usage can leave world producers above the capacity point (where
their short-run supply curve intersects their long-run supply
curve), thereby limiting their ability to respond if other
segments of the market are disrupted. On the other hand, an
unexpected trend decrease in consumption can result in world-
wide excess capacity and considerably lower estimates for
criticality loss measures. As circumstances change, naturally
criticality estimates should be periodically updated.
2-93
-------
CHARLES RIVER
ASSOCIATES
INCORPORATED!
Military Consumption
The U.S. military does not consume large amounts of raw
materials directly. However, its purchases of fabricated
items do indirectly require sizable quantities of raw materials
If the military demand is to be considered a fixed requirement,
e.g., because there is no suitable substitute for a critical
application, then all of the elasticity in the U.S. demand
curve will be due to decreases in civilian demand. Such a
demand curve can be utilized in the usual fashion when the
CRA policy model is applied to analyzing foreign supply dis-
ruptions. If a foreign supply disruption of concern is in
fact a major military conflict involving the United States,
then further issues are involved, as discussed elsewhere in
this study.
Secondary Recovery
Secondary recovery decreases the amount of U.S. primary
demand at each price, which is the appropriate input for the
CRA policy model. In fact there are further ramifications
to secondary recovery which can be of great importance in a
few cases where rates of secondary recovery are very high and
the lags between consumption and recovery are short. The
prime example is platinum, which is largely used as a catalyst
in petroleum refining and in chemical production.: In this
case the amount of platinum in use is essentially a sizable
stock which can be reallocated during disruptions.
:The recovery rate for platinum used in catalytic con-
verters for automobiles and in electrical equipment is also
very high, but recovery lags are longer for this end use than
for the other uses.
2-94
-------
CHARLES RIVER
ASSOCIATES
INCORPORATED
A special policy model has been designed at CRA for
application to the platinum and palladium markets, where the
most important precaution in the face of threatened foreign
supply disruptions need be only some extra metal held "in use."1
This platinum or palladium held "in use" is "extra" in the
sense that using more of the metal as catalysts has benefits;
however, these benefits would normally not justify the cost
of purchasing the metal unless there was also a significant
probability that the market price would increase after pur-
chase (due to a foreign supply disruption in the present
instance).
A. BOM Tabulation
The study by King and Cameron is similar in format to
a number of other cross tabulations of materials' character-
istics relevant for measuring criticality. Not all such
tabulations attempt to push the analysis so far toward subjec-
tive quantification of characteristics and computation of
summary indices as do King and Cameron. One of the more graphic
and authoritative of such cross comparisons has been constructed
by John Morgan of the U.S. Bureau of Mines. It is reproduced
below as Figure 4-2.2
:See Charles River Associates Incorporated, The Report of
the U.S. Department of the Interior, op. cit., Chapters 9
and 10.
2See John D. Morgan, "Mineral Data Improvements and Critical
Materials R&D at the U.S. Bureau of Mines," Proceedings of the Work-
shop on Government Foliates and Programs Affecting Materials Availability
(Columbus, Ohio: Metals and Ceramics Information Center, Feb-
ruary, 1976), pp. 319-344. This same paper was presented before
the National Symposium on Ceramics in the Service of Man in June
1976, and was reproduced as a Bureau of Mines Publication
entitled National Considerations of Strategic and Critical Materials.
2-95
-------
Figure 4—2
BOM CROSS TABULATION OF MINERAL PROBLEMS
CHARLES RIVER
ASSOCIATES
INCORPORATED
MINERAL PROBLEMS
1975 AND BEYOND
PftUIMINAIT DATA, AUOUST, 1975
BUREAU Of MINIS
AAAjXXX Ma|0« P'oblom Iran lha national
JOJSSXS? viewpoint
'-•••i^af* Moderate problem Iroia tka national
•:.•:".'.." Minor or more lacaliiad problem
_ Large itockpile eicenei prevent
' a currant probleia
i. U.i exparff contribute lubitanlialry
to our balance ol trade
ABRASIVES, NATURAL
ABRASIVES, 1ANUFACTURES,
ALUMINUM (Incl. BAUXITE t, ALUMIHA)
ANTIMONY
ARGON
ARSENIC (byproduct of Copper)
•»
>.
u
T
i
«
u
M
V
W
T9
w
3
a
"S = 0 >
u 3 > C o
• « g • o e
3 * 5 = J ^
A U 44 C
w 3 o ••* • •
i X « fi U * J — i
s -s s "• , i .
- S x ? 3
X *l U «* W*J J
STJ c '** — * u u g
« > « ^ • i a
3 ^ -• 3 • O °- ^
w^us »- ^3
« V « 6 »» 1- C J -- —
C C -O -- V 2
"*-xi- MO« c e
SS1^"0 • •=" 5 «• | |
h 4 U 4) « S CSC
3 3 *t a* ^ B X -- O O
OVO.C M-.UU
U « u u w
5 e 5 S 2 55
3 <• —
u 4 41 w
r i e s
s i s i
— o « t*
.s — u a
b e
!«
1 .
. -: ,
c c 2
1 J
=. =
S3 ^88SSS888i
%' -^ 'f.'i
L
ASBESTOS, CHRYSOTILE 8££8
SARITE
BERYLLIUM
BISMUTH (byproduct of Le*d)
BO RON
BROMINE
CADMIUM (byproduct of Zinc)
CALCIUM CHLORIDE
CEMENT
CES IUM
CHLORINE1, MANUFACTURED
CHROMIUM (tncl. FERSOCHROMIUM)
CLAYS
COAL
COBALT (byproduct of Cooper & Nickel)
COLUMBIUM (NI08IUM)
COPPER
CORUNDUM S EMERY
DIAMONDS, GEM STONES
DIAMONDS, INDUSTRIAL, STONES, NATUHAL
DIAMONDS, INDUSTRIAL, SORT, NATURAL
DIAMONDS, INDUSTRIAL, SORT, SYNTHETIC
DLATCMITE
FELDSPAR
FLUORSPAR
GALLIUM 'byproduct of Aluminum & Zinc)
GARNET
i
W«
oc^oddddf
?WVVW9
88 -S
35-'' SSSSS!
^SSSS^S^S:
SS
W" ^^^
SS «^5
r •
c -•»
^
w
j^J;
\'-^
rf '•<
•';*
'
O3 (A Wl
1 1 « 1 n
?• :
-
"'•"." r
'-•!. KS ^X
I *x"2v5
"35"-' 1VAVS
> :
?SS4««
N<
KXJ •'
sssss^ssss
Csx '-—•"-
r
•
*i» ^ 'i1*
H c
d
«
!SS
ss
t-' •
?5
r- i- S
'j&
y/i
j&&
SS
. . ?,
SS8S88S
•-.•"
Ft~ • -• /
--.
~-^
•V-.
-a-
WJ
.- ;
',•£ ^^' "
f*-*
*'^- ^ '^'"
•»
'" -•-•;•:
i- T-' ."
• ' -"V ^ i
t'Z
h| t
J k
, -.-«
v**
1 | « | n
VI
3
e
0
3
o
^
•$
W
=«
O
2-96
-------
Figure 4—2 (cent.)
BOM CROSS TABULATION OF MINERAL PROBLEMS
CHARLES RIVER
ASSOCIATES
INCORPORATED!
MINERAL PROBLEMS
1975 AND BEYOND
PIUIMINARY DATA, AUGUST. 1975
• U«EAU OF MINES
f Moior problems ttom sWe notiono!
> viewpoint
Moderate problems Iron the oolio
viewpoint
Minor or more locolned problems
S Lara* stockpile eacesiet prevent
o current problem
E US etporli contribute substantially
to our balance of trade
MAGNESIUM
HANGASESE (incl
MERCURY
MICA, SCRAP i FLAKE
MICA, SHEET
MOLYBDENUM
NATURAL GAS
KICKEL
NITROGEN, ELEMENTAL
•
1
•
J
&
-5
> c
s ^
*s • U
-• 0-
rj -O
"O /
a 4,
^
w «
3 r
*j W
* *• V
4« ^
U PI
\ Z ~
i : i
3f 1-
- 0 i
; — u
n w tn r.
?
c
!
3
C
—
*
•a
W
j
5
— •*• r
r: = =
i
3
£
V.
in
£ I
« 4.
2 '
O C
2 E
a. >
s s i
o *
s- ; »
1 » ^
00 —
>s •- U
U t X
w C U
3 b
1 S „
•- * • *J
U -* C
c ac «
— -. c -
i
•
1 C
=
c
o g
n v
b | c | d K. I i | j
K1TROGEN, FIXED
OIL SHALE
OWGEd
PEA1
PERLITE
PETROLEUM
PHOSPHATES
PLATINUM GROUP PLATINUM, PALLADIUM, OSMIUM,
RHODIUM, RUTHENIUM, 1R1DIUM
JS
POTASH
PUMICE 4 VOLCAS1C CINDER
QUARTZ CRYSTALS (LASCA for »ynthetlc)
RADIUM
(AXE EARTHS
HHEK1W1 (byproduct of Copptr-Molybdtnun
RUBIDIUM (byproduct of Lithium 4 Ce«tum)
GRAVEL
_
SCANDIUM (byproduct of Uranium & Pho«ph«te)
SELENIUM (byproduct of Copper)
SILICON (METAL i FERRDSILICON)
SILVER
SLAG - IRON & STEEL
SODIUM CARBONATE & SULFATE
STAUSOLITE (byproduct of Tlt»nium Biaerall)
STONE
STRONTIUM
SULFUR
TALC
TJUflALUM
TELLURIUM (byproduct of Copper)
THALLIUM (byproduct of Zinc)
THORIUM
TIN
T1TAK1UM Itocl. tLKENITS 4 RBT1LE)
TUNGSTEN
URANIUM
VANADIUM
VMMICULITE
UOLLASTONITE
TTTRIUM (byproduct of tort Earth«)
ZEOLITES
ZINC
ZIRCONIUM (incl. ZIRCON!
«5
p SSS
$$•
J8S
to w
m J
56666aSi
k 1 ~ n o
SOURCE: Dr. John D. Morgan, "Mineral Data Improvements and Critical Materials R & D at U.S. Bureau of Mines'
Proceedings of the Workshop on Government Policies and Programs Affecting Material Availability
(Columbus Ohio: Metals and Ceramics Information Center, February 1976), pp. 342-343.
2-97 '
-------
CHARLES RIVER
ASSOCIATES
INCORPORATED
Many of the problem areas identified in Figure 4-2 have
already been discussed in terms of our theoretical framework
and the application of policy models to determination of
materials criticality. However, other of these problem areas
deserve a brief comment.
A. mineral is apparently considered a "U.S. foreign ex-
change drain" (problem area f) if the net imports of the United
States have a high value. If there is an opportunity for
this country to develop domestic resources of a particular
mineral, say through investment in developing new recovery
technologies, then such a mineral does present an opportunity
for beneficial government R&D programs. In cases such as
petroleum, large expenditures on imports may be partially the
result of collusion among foreign suppliers. However, in
general international trade greatly benefits the United States,
and the theory of comparative advantage explains why it is good
that we import some goods and services in exchange for those
we export. Thus, large U.S. expenditures for imports are not
a sufficient condition for criticality; such expenditures may
or may not be a symptom of some problem or indicate the need
for revised government policy.
Health, safety, and environmental problems (areas h, 1,
m, and n) are included in Figure 4-2, in contrast to studies
restricted to criticality stemming from actual or potential
foreign supply disruptions. As we noted in the general the-
oretical discussion of Chapter 2, a restricted viewpoint is
quite workable as long as the national policies designed to
treat each problem area are relatively distinct. For example,
stockpiles counteract the threat of foreign supply disruptions
but have little direct impact on health, safety, and environ-
mental problems. The reverse is true of pollutant taxes or
job safety regulations.
2-98
-------
CHARLES RIVER
ASSOCIATES
INCORPORATED
Another set of problem areas identified in Figure 4-2
relates to a fairly distinct set of national policies, and
thus can often be usefully treated apart from criticality
due to foreign supply disruptions. Manpower, energy, and
transport (problem areas i, j, and o) are inputs into "the U.S.
minerals extraction and processing industries which are par-
ticularly prone to problems requiring national attention.
Examples of relevant policies would be forced arbitration of
strikes or regulation of the price of natural gas. Materials
which are considered a "load on the U.S. transport system"
are apparently simply voluminous and heavy, which may or may
not indicate criticality.
"Inadequate recycling" (problem area h) can be an impor-
tant element in the analysis of many policy problems, from
foreign supply disruptions to occupational health regulations,
However, it can be an area of concern simply on the grounds
of economic efficiency. Depletion allowances in the tax system,
discriminatory freight rates, and other institutional char-
acteristics of the U.S. economy appear to bias consumption
toward primary sources of materials and away from secondary
scrap. Economic losses due to these inefficiencies could in
principle be included in a materials criticality rating if
the perspective were to be broader than a concern only with
foreign supply disruptions.
It is clear that cross tabulations of market character-
istics like that in Figure 4-2 are a useful way of organizing
information. When the CRA policy model is applied to materials
criticality, cross tabulations of the inputs into the model
should be constructed as well as cross tabulations of outputs.
Such an input tabulation would be in many ways a particularly
incisive substitute for the type of presentation we have been
discussing.
2-99
-------
CHARLES RIVER j
ASSOCIATES !
INCORPORATED!
Measuring Criticality With Input-Output Models
There have been a number of studies of the effects of
shortages on the U.S. economy based on input-output models.
Such models explicitly consider all sectors of the economy
simultaneously; this inclusiveness is an advantage when one
wishes to consider simultaneous shortages of most materials, as
would occur for example during a war. The Federal Preparedness
Agency routinely uses such models in planning stockpile objec-
tives. Unfortunately, adding to an input-output model all the
features which the CRA policy model incorporates for measuring
materials criticality would be an extremely complex project
which is not likely to be undertaken soon. In any case, a
policy model based on a model of a single market implicitly
recognizes the interrelationships among markets to a greater
extent than may be immediately apparent.
A recent study done at the Stanford Research Institute is
representative of the results which an input-output model
generates when shortages are analyzed.1 Moreover, the study
explicitly addressed the issue of determining criticality of
materials in terms of prescreening those which were worth
careful consideration using the formal model. We will use this
report as the basis for most of our discussion of the applica-
tion of input-output models to materials criticality.
lSee the following publications: Mark D. Levine and Irving
W. Yabroff, Department of Defense Materials Consumption and the Impact of
Material and Energy Resource Shortages , prepared for the Defense
Advanced Research Projects Agency by the Stanford Research
Institute, November 1975; Evan E. Hughes, et al. , Strategic Resources
and National Security: An Initial Assessment, prepared for the Defense
Advanced Research Projects Agency by the Stanford Research
Institute, April 1975; Mark D. Levine and Irving W. Yabroff,
"Strategic Resources and National Security," Paper I in Proceed-
ings of the Department of Defense Materials Shortage Workshop, Metals and
Ceramics Information Center, January 1975.
2-100
-------
CHARLES RIVER
ASSOCIATES
INCORPORATED
Prescreening of Materials
The prescreening of materials for the SRI study was done
on a formal basis which is similar to other subjective weight-
ing schemes we have already discussed.l Seven criteria were
used to rate commodities on a scale of 1 to 10:
(1) Percent of U.S. consumption for defense purposes.2
(2) U.S. reserves.
(3) Percent of U.S. consumption imported.
(4) Vulnerability of sources.
(5) Difficulty of substitution.
(6) Value of consumption ("economic importance").
(7) Leverage of industry (high value of final good
output per dollar of raw material input).3
Jesuits for 74 potentially critical raw materials are
reproduced in Table 4-2. A rating of 10 for a particular
criterion is the maximum contribution to criticality which
is allowed; a rating of zero indicates a lack of data. A
:The methodology is described below only in enough detail
that the summary table of results can be roughly interpreted.
For a complete discussion, see Hughes and others, pp. 191-212
Data utilized for criteria (1) , (2), (3) and (6) were
explicitly for the year 1972. Criteria (4), (5) and (7) were
evaluated subjectively.
2The first criterion reflects the fact that the study was
done for the Department of Defense; in terms introduced in
Chapter 2, the perspective of this study is somewhat less
general than "national concensus," though introducing
specialized interests in such an ad hoc fashion is open to
criticism.
3The "leverage" of a raw material is directly taken into
account in the CRA policy model when the J .S. price elasti-
city of demand for the raw material is derived from the price
elasticity of demand for finished goods.
2-101
-------
PRIORITIES OF MATERIALS:
(1) (2) (3)
o
ro
Name of Material
DoD Use
L61*
C73»
[5]»
m*
Cio]»
ra»
DP
[8]"
M*
I.
2.
3.
4.
5.
6.
7.
8.
9.
10.
II.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
Aluminum
Iron
Manganese
Graphite
Copper
Yttrium
Chromium
Platl num-Group
Tungsten
Mica-Sheet
Nickel
Antimony
Cobalt
Fluorine
Mercury
Silver
Tantalum
Tin
Lithium
Asbestos
Co 1 umb 1 urn
Cesium
Bismuth
Potassium
Cadmium
5
8
7
5
10
4
7
5
8
10
7
8
10
6
7
5
6
8
7
3
6
0
4
1
9
U.S.
Reserves
10
7
10
10
5
9
10
10
9
10
10
10
4
10
9
9
10
10
5
9
6
0
8
7
6
Table 4-2
RANKING ON NATIONAL SECURITY AND ECONOMIC CRITERIA
(4) (5) (6)
Difficulty
of
Substitution
(7)
10
2
8
9
2
4
7
10
5
10
7
5
7
4
8
9
0
9
7
8
5
6
4
Vulnerability
of Sources
3
3
6
6
4
0
7
6
7
7
4
6
6
5
7
6
6
4
3
5
6
5
3
4
3
6
8
7
5
6
0
5
8
7
4
4
7
7
6
7
8
6
3
6
5
4
5
7
Economic
Importance
6
to
1
0
6
0
1
1
1
1
2
1
1
1
1
1
1
1
0
1
1
0
1
1
1
Industrial
Leverage
5
9
8
2
7
0
5
3
4
2
4
3
4
3
2
4
2
3
2
2
2
1
3
6
3
Geometric
Mean
Columns 2-8
6.0
5.9
5.7
5.5
5.2
5.2
5.1
4.9
4.9
4.8
4.8
4.7
4.7
4.6
4.5
4.5
4.5
4.3
4.2
3.8
3.8
3.8
3.7
3.6
3.6
Table continued on following page.
CHARLES RIVER
ASSOCIATES
INCORPORATED
-------
Table 4-2 (Continued)
PRIORITIES OF MATERIALS: RANKING ON NATIONAL SECURITY AND ECONOMIC CRITERIA
(1)
DID*
ro
O
GO
[I]"
Name of Material DoD Use
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
40.
49.
50.
Tha 1 1 1 urn
Indium
Zinc
Beryl 1 lum
Thorium
Gold
Lead
Vanadium
German 1 urn
Sul fur
Selenl urn
Arsenic
Hafnium
Strontium
Barium
Rubidium
Z 1 rcon 1 urn
Corundum
Gypsum
Tl tan lum
Chlorine
Sand and Gravel
Iodine
Stone -Crush
Te 1 1 u r 1 urn
10
0
8
10
10
4
8
5
10
5
6
1
0
5
6
0
5
4
1
4
4
2
3
3
3
(2)
U.S.
Reserves
I
7
6
I
I
9
2
9
6
8
I
3
I
I
I
10
I
10
6
I
I
I
4
I
I
(3)
(4)
(5)
(6)
(7)
Vulnerability
Imports of Sources
10
0
5
7
2
2
2
I
4
0
10
9
4
I
0
0
4
5
10
I
4
0
4
3
6
7
2
4
3
5
I
3
6
3
4
3
5
3
0
2
3
Difficulty
of
Substitution
5
5
2
6
4
5
5
4
3
7
7
4
5
4
4
5
6
2
3
4
7
6
3
6
6
Economic
Importance
0
0
1
1
0
1
1
1
1
1
1
0
0
1
1
1
0
1
1
1
2
3
1
3
1
Industrial
Leverage
1
1
4
3
1
1
4
2
1
5
2
2
1
1
3
1
1
1
3
2
8
7
1
7
1
Geometric
Means
Columns 2-8
3.5
3.4
3.4
3.3
3.2
3.1
3.1
3.0
2.9
2.8
2.7
2.7
2.7
2.6
2.6
2.5
2.5
2.4
2.4
2.4
2.4
2.4
2.3
2.3
2.2
Table continued on following page.
CHARLES RIVER
ASSOCIATES
INCORPORATED
-------
Table 4-2 (Continued)
PRIORITIES OF MATERIALS: RANKING ON NATIONAL SECURITY AND ECONOMIC CRITERIA
Name of Material
51. RhenI urn
52. Boron
53. Molybdenum
54. Nitrogen
55. Phosphorus
56. Clays
57. Silicon
58. Garnet
59. Rare Earths
60. Talc
(1)
DoD Use
0
0
9
3
2
3
9
10
2
4
(2)
(3)
(4)
(5)
ro
i
o
4^
61.
62.
63.
64.
65.
Gal Hum
Magnesium
Sodium
Ca 1 c 1 urn
Bromine
1
5
5
Q
4
U.S.
Reserves Imports
.
:
5
1
1
1
1
1
1
5 1
1
1
10
1
1
1
1
1
1
0
1
1 I
1
1
1
0
GNP. See text.
Vulnerability
of Sources
4
6
3
1
2
2
1
2
8
3
2
2
1
1
2
3
2
3
2
2
1
2
2
0
Difficulty
of
Substitution
6
7
3
8
B
7
6
2
7
5
3
4
6
7
5
3
5
3
3
3
3
3
2
0
(6)
Economic
Importance
I
I
I
2
I
(7)
Industrial
Leverage
I
3
3
6
6
4
3
I
I
2
I
3
5
9
2
2
2
66. Dlatomlte 4
67. Feldspar 3
68. Kyanlte 0
69. Mica-Scrap 4
70. Pumice 3
71. Stone-Dlmen 4
72. Vermlcullte 3
73. Perlite 3
74. Scandium 0
"Ranking by effect of shortage on GNP.
SOURCE: Evan E. Hughes, et al. , Strategic Resources and National Security: An Initial Assessment, prepared for
the Defense Advanced Research Projects Agency by the Stanford Research Institute, April 1975, pp. 202-205.
Geometric
Means
Columns 2-8
2.2
2.2
2.2
2.2
2.1
2.1
2.1
2.0
2.0
2.0
2.0
2.0
2.0
2.0
1.9
.8
.8
.7
.7
.7
.7
.5
.4
.0
CHARLES RIVER
ASSOCIATES
INCORPORATED
-------
CHARLES RIVER
ASSOCIATES
INCORPORATED
summary measure for the seven criteria is calculated simply
by taking a geometric average of all positive ratings. The
materials are listed in Table 4-2 by their rank according to
this summary measure.1
The 12 nonfuel materials which were studied intensively by
Levine and Yabroff with their input-output model were titanium,
platinum, cobalt, tin, chromium, aluminum, copper, silver,
nickel, tungsten, zinc and lead. These materials are indicated
on the left side of Table 4-2 by an asterisk and by an alter-
native ranking which will be explained below. The lack of
close correspondence between the formal prescreening rank and
the materials which were eventually deemed worthy of further
study is a pointed commentary on the weakness of subjective
weighting schemes. Considerations such as vulnerability of
supply sources and substitution possibilities are undeniably
important in determining materials criticality, but their
relative importance differs in complicated ways from material
to material; a formal economic model is generally required to
clarify such issues.
The point is not that prescreening of materials before
detailed construction of criticality ratings is infeasible or
undesirable. I&ther, given that there are only roughly a
hundred candidate raw materials, it is relatively easy to
select the most useful ones to analyze in depth according
to the principles presented in Chapter 2 of this study. The
same materials are the major candidates in most studies of
criticality, assuming the results from formal weighting
1 There is no weighting of the relative importance of the
various criteria, as was done by Cameron and King.
2-105
-------
CHARLES RIVER
ASSOCIATES
INCORPORATED
procedures are applied with good judgment. The criticality of
less major materials should also be analyzed in time, but just
using roughly conservative assumptions will often show quickly
that the criticality of a material such as dimensional stone is
not as great as for materials of more immediate interest.
Consider the case of sheet mica, which Levine and Yabroff
did not study intensively in spite of its having a high
rating (10) in Table 4-2. Despite the high rating, good
substitutes exist, consumption is declining steadily, and use
of natural sheet mica is expected to be very low by the year
2000. The United States has very few reserves of sheet mica,
but resources which could be exploited at considerably higher
prices are sizable.1 Moreover, the U.S. strategic stockpile
contains a very large quantity of the material, some of which
would very probably be available to U.S. industry during any
major disruption of imports. The criticality of sheet mica
would almost certainly be much lower than indicated in Table
4-2 if it were analyzed according to the principles we have
recommended in this study.
The Cost of Disruptions
The last report from the SRI study applied a 150-sector
input-output model to gas, petroleum, coal and the twelve
nonenergy materials listed above. Although this final stage
of the study does not explicitly consider the relative criti-
cality of materials, some basic results are generated which
indicate what one can expect from utilizing an input-output
model to determine materials criticality.
Producing sheet mica is labor-intensive, so the cost of
U.S. labor (relative to that in less developed countries where
production occurs) makes it uneconomical to exploit U.S.
resources.
2-106
-------
CHARLES RIVER
ASSOCIATES
INCORPORATED
The ultimate results reported in the SRI study are repro-
duced in their graphical form in Figure 4-3. Shortages of
individual raw materials are measured on the horizontal axis
as a percent of normal consumption, and the percent reduction
in attainable GNP is measured on the vertical axis.1 induc-
tions in gas, petroleum and coal availability quickly cause
catastrophic reductions in GNP according to results
generated by the input-output model. However, quite con-
ceivable shortages of the 12 nonenergy materials are also
represented as having disastrous effects on GNP.
It is important to put the results in Figure 4-3 into
perspective before critiquing them. First, input-output models
have intrinsic limitations which are very difficult to avoid
and which were quite familiar to the authors of the study.
Thus, results for shortages beyond certain points are con-
sidered unrealistic and are so indicated by using dashed lines
in the figure; even the more reliable results represented with
solid lines are artificial enough that they probably should
not be interpreted literally (as we discuss further below).
Second, the results in Figure 4-3 are not intended by the
authors to indicate complete criticality ratings, since they do
not consider such obviously relevant factors as the probabil-
ity of a disruption occurring.2
1A linear programming algorithm was used to maximize GNP
given the resource shortage. Because of unrealistic fixed
coefficients production functions and unrealistic constant
values of unit outputs, artificial constraints on industrial
capacities and on final demands had to be imposed in order to
obtain results which were not obviously unreasonable.
2Based partially on their formal analysis, the authors
do venture the tentative recommendation that chromium,
aluminum, tin and platinum may be the most critical commodi-
ties for economic stockpiling by the United States. See
Levine and Yabroff, Department of Defense Materials Consump-
tion, op. ait., p. 12. Additional study is also recommended.
2-107
-------
Figure 4—3
EFFECTS OF ENERGY AND MATERIAL SHORTAGES ON GNP
ESTIMATED BY SRI
PO
o
00
a.
2
2
O
H
O
a
ct
UJ
o
a:
UJ
a.
30
28
26
24
22
20
18
16
14
\2.
10
8
6
4
2
0
0
/*' ' Ki i /
A// i/ / /
I SI I 8 I
'$/' I
& !
15
20 25 30
PERCENT SHORTAGE
35
40
45
50
SOURCE: Mark D. Levjne and Irving W. Yabroff, Department of Defense Materials Consumption and the Impact
of Matmial and Energy Resources Shortages (Menlo Park, Ca.: Stanford Research Institute, November
1975), p. 78. Report prepared for the Defense Advanced Research Projects Agency.
CHARLES RIVCR
ASSOCIATES
INCORPORATED
-------
CHARLES RIVER
ASSOCIATES
INCORPORATED
Nevertheless, the ranking of materials from left to right
in Figure 4-3 is highly suggestive of the type of results which
a full input-output analysis of criticality would generate.
This ranking of nonenergy materials was indicated above in
Table 4-2 by the bracketed numbers and asterisks at the left
side.*
Comparing the rankings for nonmetallic titanium and
metallic titanium in Figure 4-3 is a revealing way to indicate
some of the most important limitations of input-output models.
It is suggested that a given percentage reduction in GNP can
be caused by a relatively small shortage of the nonmetallic
titanium but that the same percentage reduction requires a
much larger shortage of metallic titanium. According to our
interpretation of Figure 4-3, nonmetallic titanium ranks first
among the nonenergy materials in what we might call poten-
tial criticality, while metallic titanium ranks last.
Metallic titanium popularly has a high-technology image
because of its use in aerospace applications. However, non-
metallic uses account for the bulk of titanium consumption;
most goes into white pigments for very mundane usage in paints,
paper and plastics. The wide usage of titanium pigments as an
intermediate good throughout industry accounts for its apparent
importance in Figure 4-3. As the availability of titanium
decreases, buildings would not be built and appliances would
not be manufactured because of a shortage of white paint, at
least according to an input-output model which allows for no
substitutions.
*Four materials are omitted from Figure 4-3 for clarity,
but are included in Table 4-2 : tin and chromium fall between
cobalt and aluminum; copper and silver fall between aluminum
and nickel.
2-109
-------
CHARLES RIVER
ASSOCIATES
INCORPORATED
It is possible at considerable expense to modify input-
output models to take substitution possibilities into account
in a rudimentary fashion. However, adding to an input-output
model all the features which make the policy model in Chapter
3 very well suited to determining materials criticality
would be an unrealistically complicated job at the present
time. In particular, the input-output model is static and
does not take into account the successive adaptions which
the economy can make over time. A related difficulty is that
the GNP loss measures generated by an input-output model are
generally less satisfactory than the economic surplus
measures used in the policy model of Chapter 3; the surplus
measures recognize that successive unit reductions in the
availability of a good inflict larger and larger economic
costs on society.
Analysis of Two or More Markets
With a Policy Model
As mentioned earlier, the sectoral inclusiveness of
input-output models makes them an important tool when analyz-
ing simultaneous shortages in many markets, as would occur
during a war. However, even in planning mobilization for a
war, models which are sophisticated in other ways can be
useful for analyzing certain problems.
For analyzing disruptions in one market or several
markets, the advantages of models like those discussed in
Chapter 3 of this study are compelling. Most importantly,
such models allow a realistic dynamic treatment of the adjust-
ments the economy can make daring a disruption, including
substitution of alternative materials and the optimal alloca-
tion of available inventories. The simpler structure of the
2-110
-------
CHARLES RIVER '.
ASSOCIATES i
INCORPORATED!
underlying microeconomic models also means that other aspects
of materials criticality, such as the probability of disrup-
tions of various lengths and severities, can be treated in a
more satisfactory manner.
The policy model discussed in Chapter 3 does assume in
its basic form that prices for inputs other than the poten-
tially critical material are relatively stable during supply
disruptions. Such an assumption may be significantly
unrealistic for certain closely related markets. However,
generalizing the policy model to take account of several
closely related markets, such as iron and manganese (comple-
ments) or bauxite and copper (substitutes) would only be
moderately costly. More ambitious generalizations of the
policy model would allow treating simultaneous disruptions in
multiple markets, as long as most of the markets in the economy
were relatively unaffected. Fortunately for ease of analysis
as well as for peace of mind, it is very unlikely that dis-
ruptions of nonfuel minerals would massively disrupt the U.S.
economy to the extent that such microeconomic analysis would
be an insufficient tool for analysis. In any case, the
microeconomic policy model could be nested inside a macro-
economic model if this were considered necessary for analyzing
cataclysmic cases.
Because it ignores, or at best oversimplifies the substi-
tution possibilities and other dynamic adjustments which the
economic system can make, an input-output model can easily
consider simultaneous disruptions of many materials. How-
ever, in most cases a microeconomic approach to analyzing
widespread disruptions offers more realistic and reliable
results.
2-111
-------
CHARLES RIVER
ASSOCIATES
INCORPORATED
Other Studies of Criticality
In the remainder of this chapter we briefly discuss
interesting aspects of other studies bearing on the criticality
of materials.
NMAB Study
The National Materials Advisory Board (NMAB) recently
completed a study of critical materials which utilized defini-
tions different from those which we have proposed: "Critical
materials are those that are necessary to manufacture the pro-
ducts required for a national emergency and its accompanying
essential civilian needs."1 "Criticality" thus does not neces-
sarily imply that a serious disruption of supplies is probable;
however, "strategicness" is used to connote unreliable foreign
sources. In terms of the basic concepts developed in Chapter 2
of the present study, the definition of criticality used in
the NMAB study appears to be more closely related to the total
utility gained from a material than to expected losses of util-
ity.
In any case, the actual selection of materials in the
NMAB study is not closely tied to the formal definition of
criticality. In fact, the selection process is only described
in general terms as a subjective weighting of many relevant
factors, with the details omitted. The materials selected for
detailed study were chromium, germanium, iridium, rhenium,
National Materials Advisory Board, Committee on the
Technical Aspects of Critical and Strategic Materials, A
Screening for Potentially Critical Materials for the National
Stockpile, Publication NMAB-329 (Washington, D.C.: National
Academy of Sciences, 1977), p. iv. Critical and strategic
materials are discussed in this study specifically in the
context of decisionmaking for the U.S. national stockpile.
2-112
-------
CHARLES RIVER
ASSOCIATES j
INCORPORATED i
zirconium, hafnium and vanadium. Chromium was identified as
the most critical of these materials.
C.O.I.E.P. Study
A task force of federal agencies under the direction of
the Council on International Economic Policy and the National
Security Council produced a useful overview of critical mate-
rials in 1974. l The basic explicit criterion of criticality
employed was U.S. import dependence, but other considerations
relating to the threat of foreign supply disruptions were
weighted qualitatively; there was no attempt to rank materials
by their criticality. This study illustrates the type of pre-
screening of materials which can be usefully done before the
more formal and ambitious approach we have proposed is imple-
mented .
OTA S tudy
A recent study of stockpiling policies by contractors
for the Office of Technology Assessment has a short section on
criteria for selecting materials.2 The analysis is somewhat
unwieldy, involving consideration of separate stockpiles for
each of five objectives: to discourage or counteract cartel or
unilateral political actions affecting price or supply; to
cushion the impact of nonpolitical import disruptions; to
assist in international materials market stabilization; to
conserve scarce domestic materials; and to provide a market for
temporary surpluses and to ease temporary shortages. In a
1 Council on International Economic Policy, Critical
Imported Materials (Washington, D.C., December 1974).
2U.S. Congress, Office of Technology Assessment, An
Assessment of Alternative Economic Stockpiling Policies
(Washington, B.C., August 1976), pp. 52-57.
2-113
-------
CHARLES RIVER |
ASSOCIATES |
INCORPORATED!
rough way this approach does recognize that materials critical-
ity depends on the type of conditions which exist and the type
of contingencies which threaten. Selection of materials re-
lated to the above problem areas was finally done by a survey
and consensus of expert opinion, formalized as a "modified
Delphi technique."
NCSS Study
The National Commission on Supplies and Shortages did
not specifically recommend implementation of a system for
measuring criticality of materials. However, the approach to
criticality which we have proposed here is logically a prelim-
inary part of the process of policy analysis, most obviously
for determination of stockpile levels. The Commission's dis-
cussion of stockpiling in fact recognizes the usefulness of the
type of policy model discussed in Chapter 3, and earlier ver-
sions of the CRA policy model are explicitly examined.1 The
Commission's report also recognizes the importance of the cost
measurements which we have emphasized and it delves in depth
into many closely related issues, such as improving data
collection for policy analysis.
NSF Study
The National Science Foundation sponsored a methodolog-
ical study of materials criticality by International Research
and Technology Corporation in 1974.2 This study alludes to
many of the ideas which we have earlier identified as important,
1 National Commission on Supplies and Shortages, Govern-
ment and the Nation's Resources (Washington, B.C., December
1976), pp. 131-140.
2International Research and Technology Corporation,
Critical Materials: A Problem Assessment, prepared for the
National Science Foundation (Arlington, Va., May 1974) .
2-1H
-------
CHARLES RIVER
ASSOCIATES |
INCORPORATED l
but it has no unifying theory of criticality and in particular
fails to take full advantage of economic concepts and theory.
In the final analysis it simply lists many relevant considera-
tions without relating them to an explicit economic model
explaining losses from disruptions.
AFL-CIO Study
The AFL-CIO commissioned a study of imported raw mate-
rials and their importance for U.S. workers and consumers.
Nine commodities were considered of major signifi-
cance in terms of use by U.S. industry, import
dependency, vulnerability to price and supply
manipulation, and impact on U.S. employment . . .
aluminum, copper, lead, zinc, tin, nickel, manga-
nese, iron ore and chromium.1
This study did not consider the measurement of materials criti-
cality in any depth. It is interesting for current purposes
only because the materials of greatest concern to the most
prominent U.S. labor organization are no different from those
selected by other investigators, despite a perspective arguably
narrower than a national concensus.
POD Workshops
The proceedings of two workshops sponsored by the
Department of Defense in 1975 and 1976 are interesting for
expressing a broad range of concerns relating to and merging
with the issue of measuring materials criticality. Shortages
of fabricated items due to inadequate capacity, OSHA and EPA
regulations, and even scheduling difficulties are discussed
Gutchess and Stanley Ruttenberg, Rend Mate-rials
for America: A Program to Assure Meeting Future Needs, prepared for
the Industrial Union Department of the AFL-CIO (July 1975), p.10
2-115
-------
CHARLES RIVER
ASSOCIATES
INCORPORATED
in the various papers.1 The most frequent complaint is a lack
of timely and reliable forecasts of materials availability.
We have considerably narrowed the focus of our present inquiry
so that these important issues must be treated elsewhere.
Admittedly, we have had relatively little to say direct-
ly about how to determine materials which are most critical for
defense purposes, beyond showing in Chapter 2 how such consid-
erations can be made conceptually compatible with criticality
stemming from nonmilitary considerations. For some materials,
such as germanium (which is used in infrared optical technol-
ogy) , nonmilitary criticality may be small relative to other
materials, while military criticality is great.2
following Proceedings volumes were published by
the Metals and Ceramics Information Center, a Department of
Defense Information Analysis Center: Workshop on Government
Policies and Programs Affecting Materials Availability (Feb-
ruary 1976) ; Materials Shortage Workshop (January 1975) ,
Edward Dyckman, "Review of Government and Industry Shortages,"
Item A in the 1975 volume, is a useful overview of some earlier
work on materials criticality which we have also discussed.
2See "Demand of New Technology on DOD Material Supply —
Initial Findings," Item C in the 1975 DOD Workshop Proceedings
volume.
2-116
-------
Charles
River
Associates
CHAPTER 2 REFERENCES
1. Adams, Robert L. 1977. "Evaluative Methods -Some Basic Concepts and
Tools of Analysis." In conference proceedings on Contingency Planning
for Materials Resources. Sponsored by The Engineering Foundation and
The National science Foundation, Edward B. Berman, ed. New York:
United Engineering Center.
2. Adams, Robert L., Barbara A. White and James S. Grichan. 1979.
"Developing a Critical Minerals Index: A Pilot Study." Office of
Minerals Policy and Research Analysis, U.S. Department of the Interior.
July.
3. Burrows, James C. and Steven D. Beggs. 1977. "Policy Implications of
Producer Country Supply Restrictions: Approach and Conclusions." In
conference proceedings of Contingency Planning for Materials Resources.
Sponsored by The Engineering Foundation and The National Science
Foundation, Edward B. Berman, ed. New York: United Engineering Center.
4. Charles River Associates. 1978. "Measuring Materials Criticality for
National Policy Analysis." Prepared for the General Accounting Office.
March.
5. . 1977a. "Implications of the War in Zaire for the Cobalt
Market."[Revised). Prepared for the Office of Minerals Policy and
Research Analysis, U.S. Department of the Interior. June.
6. . 1977b. "The Report of the U.S. Department of the
Interior on the Critical Materials Aluminum, Chromium, Platinum and
Palladium: A Review and Revision." Prepared for the Office of Minerals
Policy and Research Analysis, U.S. Department of the Interior. July.
7. . 1976a. "Policy Implications of Producer Country Supply
Restrictions." Prepared for Experimental Technology Incentives Program,
National Bureau of Standards, U.S. Department of Commerce. Vol. I-V.
August-December.
8. . 1976b. "Public Policy in the Chromium Market." Prepared
for the Office of Minerals Policy and Research Analysis, U.S. Department
of the Interior. October.
9. . 1976c. "Public and Private Stockpiling for Future
Shortages." Prepared for National Commission on Supplies and Shortages.
August.
2-117
-------
Charles
River
Associates
CHAPTER 2 REFERENCES (continued)
10. . 1975. "Economic Issues Underlying Supply Access
Agreements: A General Analysis and Prospects in 10 Mineral Markets."
Prepared for the Bureau of International Labor Affairs, U.S. Department
of Labor.
11. Council on International Economic Policy. 1974. "Critical Imported
Materials." Special Report, Washington, D.C.: December.
12. Dyckman, Edward. 1975. "Review of Government and Industry Shortages."
13. Fischman, Leonard L. 1980. "World Mineral Trends and U.S. Supply
Problems." Prepared by Resources for the Future. Washington, D.C..
14. . 1979. "Major Mineral Supply Problems." Prepared by
Resources for the Future. Prepared for the Office of Science and
Technology Policy, Executive Office of the President. Washington, D.C.:
September.
15. Gutchess, Joceyln and Stanley Ruttenberg. 1975. "Raw Materials for
America: A Program to Assure Meeting Future Needs." Prepared for the
Industrial Union Department of the AFL-CIO, July.
16. Hughes, Evan E. et al. 1975. "Strategic Resources and National
Security: An Initial Assessment." Prepared for the Defense Advanced
Research Projects Agency by the Stanford Research Institute, April.
17. International Research and Technology Corporation. 1974. "Critical
Materials: A Problem Assessment." Prepared for the National Science
Foundation, Arlington, Va., May.
18. King, Alwyn H. and John R. Cameron. 1974. "Materials and the New
Dimensions of Conflict." Carlisle Barracks, Pa.: U.S. Army War
College, Strategic Studies Institute, May 15.
19. King, Alwyn H. 1977. "Materials Vulnerability of the United States --
An Update." Carlisle Barracks, Pa.: U.S. Army War College, Strategic
Studies Institute, April 30.
20. Klass, Michael W., James C. Burrows, and Steven D. Beggs. 1980.
"International Minerals Cartels and Embargoes: Policy Implications for
the United States." Praeger: N.Y.
21. Kymm, Kern 0. 1980. "Development of an IDEAS (Industrial Disruptions
Economic Analysis System) Critical Industries Strike Impact Assessing
System." June.
2-118
-------
Charles
River
Associates
CHAPTER 2 REFERENCES (continued)
22. Levine, Mark D. and Irving W. Yabroff. 1975. "Department of Defense
Materials Consumption and the Impact of Material and Energy Resource
Shortages." Prepared for the Defense Advanced Research Projects Agency
by the Stanford Research Institute, November.
23. ^ . 1975. "Strategic Resources and National Security: Paper I
in Proceedings of the Department of Defense Materials Shortage Workshop,
Metals and Ceramics Information Center, January.
24. Materials Shortage Workshop. 1975. Metals and Ceramics Information
Center, Department of Defense Information and Analysis Center.
January.
25. Morgan, John D. 1976. "Mineral Data Improvements and Critical
Materials R&D at the U.S. Bureau of Mines." Proceedings of the Workshop
on Government Policies and Programs Affecting Materials Availability.
Columbus, Ohio!Metals and Ceramics Information Center, February:
319-344.
26. National Commission on Materials Policy. 1973. "Materials Needs and
the Environment Today and Tomorrow." U.S. Government Printing Office,
June.
27. National Commission on Supplies and Shortages. 1976. "Government and
the Nation's Resources." December.
28. National Materials Advisory Board (NMAB). 1981. "Identification of
Critical and Strategic Materials for Naval Combat Systems." Commission
on Sociotechnical Systems, National Research Council. April.
29. . 1977. "A Screening for Potentially Critical Materials for
the National Stockpile." Committee on the Technical Aspects of Critical
and Strategic Materials. Washington, D.C.: National Academy of
Sciences.
30. Tilton, John E. 1977. "The Future of Nonfuel Minerals." The Brookings
Institution, Washington, D.C.
31. U.S. Congress. 1976. "An Assessment of Alternatives Economics
Stockpiling Policies." Office of Technology Assessment, Washington,
D.C.: 52-57.
32. U.S. Department of the Interior. 1979a. "Domestic Policy Review of
Nonfuel Minerals." Draft report, Vol. I-II. Volume III prepared by the
Department of Mineral Economics, Pennsylvania State University for the
U.S. Bureau of Mines.
2-119
-------
Charles
River
Associates
CHAPTER 2 REFERENCES (continued)
33. U.S. Department of the Interior. 1979b. "Mineral Trends and
Forecasts."
34. U.S. Department of the Interior. 1975. "Critical Materials: Commodity
Action Analyses (Aluminum, Chromium, Platinum and Palladium). Prepared
by the Office of Minerals Policy Development. March.
35. U.S. Department of the Interior. Mining and Minerals Policy. Issued
annually.
36. U.S. Federal Emergency Management Agency. Stockpile Report to the
Congress. Issued quarterly.
37. Webster's New Collegiate Dictionary. 1976. G. & C. Merriam Co.
38. Workshop on Government Policies and Programs Affecting Materials
Availability. 1976. The Metals and Ceramics Information Center,
Department of Defense Information Analysis Center. January.
39. Workshop Proceedings, Department of Defense. 1975. "Demand of New
Technology of DOD Material Supply — Initial Findings." Item C.
2-120
-------
Charles
River
Associates
PROJECTIONS OF MATERIALS CONSUMPTION FOR CONTROL OF VEHICULAR EMISSIONS
Rath & Strong projected consumption of materials for U.S. vehicular emissions
control for this study. Only projected consumption for the platinum-group
metals is reported in detail here, because, as confirmed in Chapter 2, these
are by far the most critical materials from the perspective of vehicular
emissions control. However, order of magnitude estimates for other
potentially critical materials, as used in Chapter 2, are also presented
below.
Tables 3-1 through 3-3 project consumption of platinum, palladium, and
rhodium by the four significant U.S. auto manufacturers, for emissions
control on cars and light trucks. These projections are based upon assuming
a constant volume of U.S. production over the reported time horizon from 1980
to 1987. Assumed production by U.S. manufacturers, in thousands of vehicles
per year, is as follows:
General Motors
Ford
Chrysler
American Motors
Total
Cars
5,700
2,090
950
285
9,025
Li ght
Trucks
1,382
916
272
154
2,724
Total
7,082
3,006
1,222
439
11,749
The projections in Tables 3-1 through 3-3 recognize differences in the
technologies used for emissions control by the four U.S. manufacturers, and
the effect of future mixes of engine types. One important reason for
3-1
-------
Charles
River
Associates
Table 3-1
U.S. CONSUMPTION OF PLATINUM FOR VEHICULAR EMISSIONS CONTROL, 1980-1987
(1000 grams per year"!
Year
1980
1981
1982
1983
1984
1985
1986
1987
Cars
20,096
19,060
15,651
14,780
13,980
14,244
15,595
14,468
Light
Trucks
6.
5,
4.
301
976
936
4,723
4,
4.
507
561
4,857
4,535
Total
26,397
25,036
20,587
19,503
18,487
18,805
20,452
19,003
SOURCE: Rath & Strong, 1981.
3-2
-------
Charles
River
Associates
Table 3-2
U.S. CONSUMPTION OF PALLADIUM FOR VEHICULAR EMISSIONS CONTROL, 1980-1987
(1000 grams per year)
Year
1980
1981
1982
1983
1984
1985
1986
1987
Cars
8,534
8,115
6,025
4,661
4,414
4,493
4,871
4,553
Light
Trucks
676
543
847
506
434
1,449
1,537
1,435
Total
11,210
10,658
7,872
6,167
5.
5.
6,
848
942
408
5,988
SOURCE: Rath & Strong, 1981.
3-3
-------
Charles
River
Associates
Table 3-3
U.S. CONSUMPTION OF RHODIUM FOR VEHICULAR EMISSIONS CONTROL, 1980-1987
(1000 grams per
Year
1980
1981
1982
1983
1984
1985
1986
1987
Cars
25.
13.
339.
947.
851.
858.
964.
Light
Trucks
7.2
3.9
140.
291,
266,
266,
291,
.5
.7
.1
.2
.2
877.7
268.1
Total
32.4
17.4
479.6
238.8
118.0
124.6
1,255.7
1,145.8
SOURCE: Rath & Strong, 1981.
3-4
-------
Charles
River
Associates
projected declines in platinum and palladium consumption between 1980 and
1987, is assumed reductions in the size of gasoline engines and greater usage
of diesel engines. Rath & Strong assumed that diesel engines will be
equipped with a monolithic substrate contained in a clamshell particulate
trap largely made of 409 stainless steel, in which particles are burned using
hot exhaust gases. No use of noble metals in the particulate traps is
assumed.
The Rath & Strong subproject report for this contract contains a detailed
description of the assumptions underlying their projections, and a wealth of
data beyond that required for the immediate purposes of this study (such as a
disegregation of platinum-group consumption by vehicle manufacturer). Thus,
their report is included with this study as a separately bound appendix.
Appendix 3-A presents alternative estimates, by the Environmental Protection
Agency, of consumption of platinum-group metals for control of vehicular
emissions in 1981. The EPA estimates differ considerably from those
presented for 1981 in Tables 3-1 through 3-3, particularly for rhodium.
The detailed projections of platinum-group consumption provided in Tables 3-1
through 3-4 are of some interest in their own right. However, such detail
and accuracy is not really required for the rough estimation of materials
criticality performed in Chapter 2. For that purpose, we simply round off
projected annual consumption for the period 1985-1987 roughly as follows (in
thousands of grams per year):
• Platinum: 20,000
• Palladium: 6,000
• Rhodium: 1,200
Other potentially critical materials used in vehicular systems for emissions
control, fuel management, and related purposes, are chromium, manganese,
nickel, and titanium. These materials are mostly contained in the stainless
steel (type 409) used in the catalytic converter. The following very rough
estimates of annual consumption of these materials by U.S. vehicle
manufacturers, for emission control and related purposes on cars and light
trucks, were provided by Rath and Strong (in short tons per year):
• Chromium: 10,600
t Manganese: 730
• Nickel: 360
• Titanium: 350
As explained in the preceding chapter, these quantities are small proportions
of total U.S. consumption of these metals, and their criticality from the
perspective of U.S. vehicle manufacturers and EPA is much less than the
criticality of the platinum-group metals. (We confirmed for this project
that consumption for emissions control of several other metals, including
molybdenum and vanadium, is nonexistent or entirely inconsequential.)
3-5
-------
Charles
River
Associates
EPA ESTIMATES OF CONSUMPTION OF PLATINUM-GROUP
METALS FOR CONTROL OF VEHICULAR EMISSIONS IN 1981
Tables 3-1 through 3-3 in the text provide Rath & Strong estimates of U.S.
consumption of platinum, palladium and rhodium for vehicular emissions
control in 1980 and 1981, as well as for the 1985-1987 time frame considered
by Charles River Associates in Chapter 2. The Rath & Strong estimates for
1981 differ substantially from best estimates produced internally by the
Control Technology Assessment and Characterization Branch of EPA's Office of
Air, Noise and Radiation. The EPA estimates are given in Table 3A-1.
Time did not allow us to investigate the reason for the large discrepancies
between Tables 3-1 through 3-3, and the EPA estimates for 1981. In any case,
the Rath & Strong estimates for 1985-1987 are much more compatible with the
EPA estimates for 1981, and the 1985-1987 projections were the basis for the
work in Chapter 2.
3A-1
-------
Charles
River
Associates
Table 3A-1
U.S. CONSUMPTION OF PLATINUM, PALLADIUM AND RHODIUM FOR VEHICULAR
EMISSIONS CONTROL, EPA ESTIMATES FOR 1981
Number of Light-Duty Gasoline Vehicles
Produced in the United States: 8,864,444
Grams of Platinum-Group Metals Consumed per Vehicle:
Platinum: 1.957
Palladium: 0.653
Rhodium: 0.184
Total: 2.794
Implied Total U.S. Consumption of Platinum-Group Metals,
in Thousands of Grams:
Platinum: 17,350
Palladium: 5,790
Rhodium: 1,630
Equivalent Total U.S. Consumption
of Platinum-Group Metals, in Troy Ounces
Platinum: 557,700
Palladium: 186,000
Rhodium: 52,400
SOURCE: Environmental Protection Agency, Office of Air, Noise and Radiation,
Control Technology Assessment and Characterization Branch.
August 5, 1981.
3A-2
-------
Charles
River
Associates
PLATINUM-GROUP METALS
INTRODUCTION
This chapter focuses on the platinum-group metals (PGMs) used in vehicular
emissions control systems. As discussed in another chapter, there is
currently no economical alternative to the noble-metal catalytic converter
which uses platinum, palladium, and rhodium. Supplies of these materials
thus become an important issue when contemplating the costs of auto emission
regulations.
The supply elasticity of PGMs is discussed in this chapter by time-frame of
response. Stockpiles of PGMs are first analyzed in terms of their ability to
bridge a short-term demand/supply gap. Following the stocks discussion is an
analysis of world primary production elasticity by country, a discussion of
total world PGM reserves, and an analysis of world supply reliability.
Following the primary production section, other PGM-using industries are
examined, focusing on the response of consumption and secondary recovery in
each industry to PGM price rises. The chapter concludes with a discussion of
the role of speculation in PGM markets.
STOCKS
The most immediate possible supply response to any increase in demand for
platinum group metals would come from stocks. The key questions then become
4-1
-------
Charles
River
Associates
what government and private stocks of PGMs exist, how large they are, and
what the immediate supply elasticity of the stocks is. The secretive nature
of the platinum industry makes these questions difficult to answer precisely;
nevertheless, much information is public and discussions with industry
personnel have revealed additional information. The industry secrecy often
has good cause; one firm recently told the U.S. Bureau of Mines that it would
not be reporting any PGM stocks during the current period due to forced
removal by armed robbery.
There are several different types of PGM stocks. Table 4-1 categorizes PGM
stocks in the United States and indicates whether reliable data are regularly
reported on each of them.
The following sections discuss each kind of stock and available information
about their size.
U.S. NATIONAL STOCKPILE
Currently, the U.S. government maintains stockpiles of three platinum-group
metals -- platinum, palladium, and iridium. As Table 4-2 indicates, current
inventories of each metal are well below stated goals of the General Services
Administration (GSA), which is in charge of maintaining stockpiles of U.S.
critical materials. The current total PGM inventory of 1,725,000 troy ounces
has not changed since 1971.
There is industry speculation that the GSA may soon purchase PGMs. One
mechanism that would facilitate this purchase is the proposed National
Defense Stockpile Transaction Fund. Monies entering the fund through sale of
commodities in excess of their goals can be used to purchase other
commmodities in deficit. Silver and tin are two excess commodities, although
recent GSA efforts to sell silver have been stopped in Congress.
The anticipated military buildup under the Reagan administration, according
to industry sources, is expected to lead to primary focus on platinum-group
metals and cobalt for stockpile acquisition. (See American Metal Market,
1981.) It is too early to tell how much of the PGM stockpile deficit will be
eliminated and, of course, planning of GSA activities in commodity markets is
kept secret in order to avoid speculative reaction.
REFINER, IMPORTER, AND DEALER STOCKS
U.S. Bureau of Mines data on stocks of PGMs held by refiners, importers, and
dealers from 1975 to 1980 are reproduced in Table 4-3. The figures show a
squeezing down of inventories starting in 1978 due to higher metals prices
and strong demand. Figure 4-1 plots the changing composition of refiner,
importer, and dealer stocks among the PGMs during the 1970s.
4-2
-------
Charles
River
Associates
Table 4-1
PLATINUM GROUP METAL STOCKS
Type Regular Data Reported?
U.S. National Stockpile Yes
Refiner, Dealer, and Importer Stocks Yes
Industry Shelf Stocks No
Industry Stocks in Use No
Private Speculative/Investment Stocks No
SOURCE: Charles River Associates, 1981.
4-3
-------
Charles
River
Associates
Table 4-2
CURRENT STATUS OF U.S. NATIONAL DEFENSE STOCKPILE
(Thousands of Troy Ounces)
Material Goal Current Inventory
Platinum 1,310 453
Palladium 3,000 1,255
Iridium 98 17
Total 3~TO T772T
SOURCE: U.S. Bureau of Mines, 1981, Mineral Commodity Summaries, Washington,
D.C.: U.S. Government Printing Office.
4-4
-------
Table 4-3
REFINER, IMPORTER, AND DEALER STOCKS
OF PLATINUM GROUP METALS, 1975-1980*
(Troy Ounces)
Charles
River
Associates
Year
Platinum
Palladium
Rhodium
1975
1976
1977
1978
1979
1980**
420,770
536,318
438,045
369,823
305,605
493,000
335,621
459,765
475,358
369,937
323,865
293,831
53,847
47,769
48,392
51,322
47,678
46,421
*Includes metal in depositories of the New York Mercantile Exchange.
**As of December 31, 1980.
SOURCE: U.S. Bureau of Mines, various issues, Minerals Yearbook and Mineral
Industry Surveys, various issues, Washington, D.C.: UTS. Government
Printing Office.
4-5
-------
t
r
o
o
u
n
c
Figure 4—1
Platinum Group Metal Stocks held by Refiners, Importers, & Dealers
CPrepared By: Char lee River Associates)
1 250038
1000000 -
750000 -
500000 ~
250000 ~
1970 1971 1972 1973 1974 1^75 1976 1977 1978 1979
Rhod i urn
Platinum
Pa I I adI urn
Other
years
SOURCE* U.S. Bureau of Mines, Minerals Yearbook,
Mineral Industry Surveys, various Issues, Washington, D.C.
Charles
River
Associates
-------
Charles
River
Associates
The U.S. Bureau of Mines (U.S. BOM) collects these data by sending out a
questionnaire to all U.S. domestic refiners, dealers, and importers. U.S.
BOM sources indicate that the returned questionnaires capture about 90
percent of the firms. A U.S. BOM telephone survey revealed that the less
than ten percent of firms that did not respond dealt with minute quantities
of PGMs. Hence, it can be safely concluded that the U.S. BOM reports capture
well over 95 percent of refiner, importer, and dealer PGM stocks.
INDUSTRY SHELF STOCKS
There exist no published data on the amount of platinum-group metals held in
inventory for future use by industries in the United States or abroad. Firms
generally will not disclose information of this nature, so knowledgeable
industry observers willing to talk are one of the few readily available
sources of information on industry PGM shelf stocks.
The National Materials Advisory Board (NMAB) in their recent report on
government PGM stockpiling strategies stated that the chemical, petroleum,
and glass industries hold large inventories of PGMs. (See NMAB, 1980.)
However, the NMAB did not provide any order of magnitude estimates of
industry shelf stocks. The only bit of quantitative information in their
report on shelf stocks is that from 1971 to 1977 the petroleum industry added
an estimated 350,000 to 450,000 troy ounces of platinum to its shelf
stockpiles.
Discussions with knowledgeable industry observers have tended to corroborate
the NMAB claim that the petroleum, chemical, and glass industries maintain
large PGM inventories. One source stated that the chemical industry probably
has over one year's supply of replacement needs, and the source "firmly
believed" that the glass industry has one year's supply as well. Another
source stated that petroleum companies sometimes lease their PGM stocks and
usually have one year's supply as shelf stocks.
Applying this one-year estimate to data on sales to consuming industries
data, one can arrive at lower bound estimates for PGM shelf stocks for the
chemical, petroleum, and glass industries (Table 4-4). (See Tables 4-21 and
4-23 for basic data.)
The U.S. automobile industry, the largest domestic user of platinum and
rhodium, and the third largest user of palladium (in 1979), is completely
secretive about its PGM inventories. It is reported, however, that the
automobile companies purchase PGMs directly from primary producers, and in
turn sell them back to catalyst manufacturers such as Engelhard or
Johnson-Matthey. Auto company personnel interviewed would not divulge the
size of their PGM inventories.
4-7
-------
Charles
River
Associates
Table 4-4
LOWER-BOUND ESTIMATES OF PLATINUM GROUP METAL SHELF STOCKS,
BASED ON 1979 RATES OF CONSUMPTION
(Troy Ounces)
Industry
Platinum
Palladium
Rhodium
Chemical
Petroleum
Glass
98,600
170,013
88,594
199,743
24,588
1,729
11,684
1,223
15,276
SOURCE: Order of magnitude estimates by Charles River Associates, 1981,
4-8
-------
Charles
River
Associates
INDUSTRY STOCKS IN USE
Platinum-group metals actually being used in production processes are
referred to as PGM "stocks in use," because in many industries (e.g.,
chemical, glass, and petroleum) PGMs are used as catalysts or otherwise
indirectly, and loss rates are quite low. In petroleum refining, for
example, platinum is used as a catalyst and thus not consumed by the
production process, but it must be periodically recycled because of high
temperatures and contamination; the loss rate is around 3 percent during
recycling. For the chemical industry the recycling loss rate is about 18
percent, and for the glass industry about 3 percent.
In a 1976 report, (CRA, 1976) CRA used these loss rates, data on industry
growth, and current annual industry PGM consumption data to arrive at order
of magnitude estimates of stocks in use. Very rough estimates of 1979 stocks
in use can be generated from these original estimates and the ratio of
current to previous annual PGM consumption by industry. These estimates are
provided in Table 4-5.
There are dangers inherent in this estimation methodology, however. First,
Bureau of Mines consumption data are not comprehensive, but are based on
reported sales to consuming industries. In its annual Mineral Commodity
Summaries, the Bureau of Mines calculates "apparent consumption" with an
accounting formula based on imports, and this figure is always well above
reported sales to consuming industries.* A second difficulty with an
estimation method of this nature is that industries purchase PGMs not only to
replenish stocks in use, but also to adjust shelf stocks. If in a given year
an industry's shelf stocks are being increased, consumption would be high and
so would the subsequent stock-in-use estimates.
In late 1980 CRA designed, for the U.S. Department of the Interior, a new
model of platinum and palladium consumption and production that focused on
engineering estimates of the speed and extensiveness with which consumption
of the two metals could be reduced after supply disruptions so severe that
their price would rise to five or ten times normal levels (which is
considerably beyond the range of historical experience).
Part of this modeling effort involved estimating the typical holding period
prior to recycling of stocks in use, and estimating recovery rates. This
*The accounting formula is:
Mine Production + Secondary Refining Production + (Imports - Exports)
+ (Beginning Stocks - Ending Stocks).
4-9
-------
Table 4-5
ESTIMATES OF PLATINUM AND PALLADIUM STOCKS IN USE, 1979
(Troy Ounces)
Charles
River
Associates
Industry
Petroleum
Chemical
Glass
Electrical
Other*
Platinum
1,787,500
1,158,168
1,098,900
1,653,106
1,895,528
Palladium
849,940
1,874,896
24,040
5,027,132
3,924,000
*Includes jewelry, medical, dentistry, and miscellaneous.
SOURCE: Charles River Associates estimates, based on U.S. Bureau of Mines,
Minerals Yearbook, various issues, Washington, D.C.:
U.S. Government Printing Office.
4-10
-------
Charles
River
Associates
model will provide more sophisticated estimates of platinum and palladium
stocks in use, which would largely overcome the two difficulties described
above. Unfortunately, this platinum-palladium model has not yet been fully
implemented for computation.
The stock-in-use estimates shown in Table 4-5 are rough order of magnitude
numbers, and the confidence intervals are unknown. They serve as a useful
reference point, however.
PRIVATE SPECULATIVE/INVESTMENT STOCKS
Platinum-group metals have recently become an object of private speculative
investment, probably due in large measure to the realization that the United
States is extremely import dependent and industrial use is rising. The New
York Mercantile Exchange sponsors trade in platinum futures contracts, in
response to this interest by investors outside the chain of producers and
consumers.
Absolutely no estimates are published on the amount of PGMs in the hands of
private investors. The Bureau of Mines omits sales to private investors from
the "miscellaneous sales" category in its publications, because of
confidentiality problems due to the small number of buyers. Bureau personnel
have indicated that these undisclosed numbers have been quite small to date,
however.
In commodities futures trading, only a small percent of the contracts are
actually consummated by delivery, because an investor's long or short
position is usually nullified with an offsetting contract before the
specified delivery date. At that point, profits or losses are counted and
the trader is out of the market, for richer or poorer. Hence, the amount of
existing valid platinum futures contracts or "open interest" is not the
proper indication of how much metal is actually trading hands through the
futures market. In general, it seems unlikely that much platinum is being
held by individuals who fear economic or political chaos, and consequent
depreciation of paper currency. Such individuals seem much more likely to
hold traditional media of exchange, such as gold or silver. Indications are,
therefore, that the stock of PGMs in the hands of private investors is quite
small relative to other kinds of stocks discussed in this section.
STOCKS AND INCREASED DEMAND IN THE SHORT RUN
The above order of magnitude estimates of various PGM stocks allow us to
address the question of how large a gap between demand and supply they could
bridge, given a one or two year lag in any response of primary production.
4-11
-------
Charles
River
Associates
U.S. auto manufacturers could perhaps buy a small percentage of shelf
stocks held by other industries at a moderate premium. However, astronomical
prices probably would be required to induce sale of anything like a majority
of shelf stocks held by other industries, particularly at times when market
conditions were perceived to be more uncertain than usual.
Simple arithmetic demonstrates that, except for rhodium, U.S. government
stockpiles could help meet increased PGM demand for the auto industry
(Table 4-6). However, such action is highly uncertain due to the fact that
government stocks are reserved for emergencies, unless compelling reasons
lead to special Congressional action. No shipments from U.S. government PGM
stockpiles have occurred in recent years.
Table 4-7 presents current refiner, importer, and dealer stocks as a percent
of 1979 auto industry consumption. It is physically possible for these
stocks to meet substantial percentage increases in auto industry PGM demands.
However, evidence indicates that flows from this source can be
price-inelastic. For example, while the dealer price for platinum increased
70.5 percent in 1978, refiner, importer, and dealer stocks of platinum fell
by only 17.4 percent during that year. Similar price-inelastic behavior
occurred for palladium and rhodium. These figures suggest that PGM prices
might have to increase dramatically to encourage flows from stocks held by
refiners, importers, and dealers in the United States. The price required to
call forth these private stocks could be more moderate than these numbers
suggest if the demand/supply gap is expected to be only temporary.
STATISTICAL OVERVIEW OF SUPPLY AND DEMAND
In this section, a brief outline of PGM supply and demand is presented to set
the stage for subsequent analysis. First, data on world primary production
are presented, and then industrial use of PGMs in the United States is
described.
SUPPLY
Due to lags in data gathering and publication, country-specific estimates of
world primary PGM production have been published by private sources only
through 1977. More current production data, where available, are discussed
in producer country profiles below. Tables 4-8 through 4-10 present world
estimates by country of primary production for platinum, palladium, and
rhodium, respectively. These tables reflect the central fact that over 98
percent of the world's primary production of platinum-group metals is
4-12
-------
Charles
River
Associates
Table 4-6
U.S. NATIONAL STOCKPILES AS A PERCENT
OF 1979 AUTO INDUSTRY PLATINUM METAL GROUP CONSUMPTION
(Nominal Data in Troy Ounces)
1979 Auto
Industry Consumption
U.S.
Stockpiles
November 30,
1980*
Stockpiles as a
Percent of Consumption
(Column 2/Column 1)
Platinum
Palladium
Rhodium
803,229
222,156
26,136
453,000
1,255,000
0
56.4
565.00
0
*As of November 30, 1980.
SOURCE: U.S. Bureau of Mines, 1979, Minerals Yearbook, and 1981, Mineral
Commodities Profiles, Washington, D.CT:U.S. Government Printing
Office.
4-13
-------
Charles
River
Associates
Table 4-7
RECENT U.S. REFINER, IMPORTER, AND DEALER STOCKS
OF PLATINUM GROUP METALS AS A PERCENT
OF 1979 AUTO INDUSTRY CONSUMPTION
(Troy Ounces)
Refiner,
Importer, and
1979 Auto Dealer Stockpiles
Industry Consumption September 30, 1980
Stockpiles as
a Percent of
Consumption
(Column I/Column 2)
Platinum
Palladium
Rhodium
803,229
222,156
26,136
402,310
321,065
50,021
50.1
144.5
191.4
*As of August 30, 1980.
SOURCE: Charles River Associates Incorporated, 1981.
4-14
-------
Table 4-8
ESTIMATED WORLD PRODUCTION OF PRIMARY
PLATINUM BY COUNTRY, 1971-1977
(Thousands of Troy Ounces)
Charles
River
Associates
1971
Australia
Canada 219
Colombia 26
Ethiopia o
Japan* 3
Philippines , 1
South Africa 750
USSR 690
United States* 10_
Total 1,701
1972
-
187
24
0
4
3
870
705
4
1,799
1973
0
163
26
0
4
2
1,416
735
6
2,354
1974
0
177
21
0
4
1
1,700
750
4
2,658
1975
0
184
23
0
5
1
1,559
795
5
2,572
1976
o
198
26
o
9
1,680
840
1977
o
216
25
o
10
1,740
840
2,753 2,831
*Production of refined metal, some from imported ores and crude palladium.
SOURCE: U.S. Bureau of Mines, Minerals Yearbook, various issues, Washington,
D.C.: U.S. Government Printing Office; and Roskill Information
Services, Ltd., 1979, The Economics of Platinum Group Metals,
London.
4-15
-------
Table 4-9
ESTIMATED WORLD PRODUCTION OF PRIMARY
PALLADIUM BY COUNTRY, 1971-1977
(Thousands of Troy Ounces)
Charles
River
Associates
Australia
Canada
Japan*
Philippines
South Africa
United States*
USSR
Total
1971
-
190
5
2
438
10
1,380
2,026
1972
-
162
6
5
508
11
1,410
2,103
1973
1
142
6
4
826
13
1,470
2,463
1974
1
154
11
2
991
9
1,500
2,669
1975
1
160
14
1
909
11
1,590
2,687
1976
2
172
18
-
980
6
1,680
2,859
1977
2
188
20
-
1,015
5
1,680
2,910
*Production of refined metal, some from imported ores and crude palladium.
SOURCE: U.S. Bureau of Mines, various issues, Minerals Yearbook, Washington,
D.C.: U.S. Government Printing Office; and Roskill Information
Services, Ltd., 1979, The Economics of Platinum Group Metals,
London.
4-16
-------
Table 4-10
ESTIMATED WORLD PRODUCTION OF PRIMARY
RHODIUM BY COUNTRY, 1971-1977 (Thousands of Troy Ounces)
Charles
River
Associates
Canada
South Africa
U.S.A.
U.S.S.R.
Total (countries
listed)
1971
1972 1973
1974
1975
1976
1977
14
44
0
46
104
12
51
0
47
110
11
83
0
49
143
11
99
0
50
160
12
91
0
53
156
14
98
0
56
168
14
101
0
56
171
SOURCE: Estimates by Roskill Information Services, Ltd.
4-17
-------
Charles
River
Associates
obtained from only three countries -- South Africa, the Soviet Union, and
Canada. Canada is a junior member of the "big three"; clearly the Soviet
Union and South Africa dwarf other PGM producing countries. South Africa
predominates in platinum with 61.5 percent of world production in 1977, while
the Soviet Union dominates in palladium with 57.7 percent of world production
in 1977. South Africa accounted for 59.1 percent of 1977 world rhodium
production, as reported by Roskill; the corresponding percentages for the
Soviet Union and Canada were 32.7 and 8.2, respectively.
These tables also indicate that U.S. production of PGMs is inconsequential;
we are almost completely import-dependent. This situation is expected to be
mitigated in the future, however, with the development of U.S. PGM deposits
discussed below.
Tables 4-11 through 4-14 indicate the countries upon which the United States
is directly import dependent. Countries that are important exporters to the
United States differ somewhat from the primary producing countries, because
ores and concentrates are often exported to other countries for refining. In
the long run, the distribution of output among the primary producing
countries is by far the more important consideration. However, in the short
run, trade patterns can be difficult to redirect due to long-term contracts
between producers, refiners, and users, the location of processing
facilities, and similar considerations. Thus, existing trade patterns are of
some importance for analyzing the short-run effects of supply disruptions.
Table 4-11 contains data for all platinum-group metals, while Tables 4-12,
4-13, and 4-14 pertain to platinum, palladium, and rhodium, respectively.
The importance of the United Kingdom as an exporter of PGMs to the United
States is clear from the tables. The United Kingdom is important because of
refining done there by companies like Johnson-Matthey, which owns part of
Rustenburg Platinum Mines, South Africa's largest platinum mine. Other
countries that export sizable amounts of PGMs to the United States include
Japan and Switzerland. Figures 4-2 through 4-5 aid in the interpretation of
the import data for platinum, palladium, and rhodium. These graphs trace
major countries' changing shares of the U.S. import market throughout the
1970s. Two related trends emerge clearly: the increasing importance of
South Africa and the decreasing importance of the Soviet Union in the 1970s.
U.S. refining of platinum-group metals throughout the 1970s is outlined in
Tables 4-15 through 4-18. Data are presented on primary and secondary toll
and nontoll refining for all platinum group metals. Data are presented
separately for platinum, palladium, and rhodium.
U.S. primary refining has sharply fallen over the 1970s. These data in part
reflect South African efforts to refine more metal within that country,
rather than exporting ores or concentrates.
4-18
-------
Table 4-11
U.S. IMPORTS FOR CONSUMPTION OF
PLATINUM METALS BY COUNTRY,1 1970-1980
(Troy Ounces)
Country
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
Australia
Belgium-Luxembourg
Canada
Chile
Colombia
France
Germany, Federal
Republic of
Italy
Japan
Mexico
Netherlands
Norway
Romania
South Africa,
Republic of
Sweden
Switzerland
USSR
United Kingdom
i Yugoslavia
— • Other
6,093
14,184
24,745
--
27,547
--
10,320
--
25,628
6,411
10,021
11,676
--
115,500
--
1,671
494,978
651,895
--
131,138
5,099
24,466
51,365
--
26,856
5
9,684
1,240
24,170
4,624
781
12,553
--
165,672
4,067
470
407,628
558,009
--
91.3543
8,352
15,917
39,605
--
25,535
--
19,676
--
111,875
15,669
1,428
23,361
--
237,697
2,807
2,099
736,264
589,711
-_
62,188''
4,904
21,807
33,281
1,761
27,210
953
10,957
--
164,468
12,935
49,649
38,287
--
245,411
4,253
29,149
882,267
806,423
_-
168,918s
37,056
57,972
--
20,165
2,205
12,731
--
208,119
25,980
6,206
74,974
--
1,016,458
8,077
9,381
1,012,321
734,458
--
14.5026
6,476
30,418
57,962
--
21,588
2,244
22,490
350
51,567
12,132
6,000
36,137
--
837,081
9,531
11,547
331,267
362,168
1,225
20.1017
6,280
31,416
93,648
--
18,201
11,985
36,014
5,409
19,864
20,424
6,130
28,103
--
1,241,669
7,679
13,641
652,112
455,193
4,862
14.4298
1,600
26,249
88,510
--
7,989
7,074
25,370
2,800
18,581
152,402
10,602
19,480
2,680
1,267,191
4,067
20,440
617,215
226,657
3,876
7,591
..
49,472
90,305
1,608
14,550
--
--
33,408
29,597
106,780
29,014
13,798
--
1,591,925
--
23,178
552,666
343,503
--
41,607
„
34,783
80,668
--
15,707
--
--
7,723
35,002
31,867
33,068
32,085
--
2,083,209
--
40,324
693,215
305,522
--
85,955
10,396
102,246
98,330
--
3,988
5,735
54,140
15,194
22,388
41,997
59,304
17,629
--
1,908,325
8,929
31,096
376,747
503,321
3,247
242.9309
Total
1,531,807 1,388,043 1,892,184 2,502,633 3,240,605 1,820,284 2,667,059 2,510,374 2,921,411 3,479,128 3,505,042
Table continued on following page.
Charles
River
Associates
-------
Table 4-11 (Continued)
U.S. IMPORTS FOR CONSUMPTION OF
PLATINUM METALS BY COUNTRY1, 1970-1980
(Troy Ounces)
'Includes unwrought and semimanufactured platinum-group metals, unspecified combinations, platinum-group metals from precious metal
ores, sweeping, waste, scrap, and materials not elsewhere specified.
2January-December 1980.
3Includes Argentina, Austria, Brazil, Denmark, Finland, Panama, Peru, Surinam, and New Zealand.
"•Includes Botswana, Brazil, Costa Rica, Finland, Ghana, Malawi, Netherlands-Antilles, New Zealand, Panama, Turkey, Venezuela.
5Includes Brazil, Costa R1ca, El Salvador, Finland, Ireland, Panama, New Zealand, Uruguay, and Venezuela.
6Includes Republic of Korea.
7Includes Costa R1ca, Finland, and Panama.
8Includes Costa R1ca, Finland, Peru, and Portugal.
'Includes Argentina, Costa R1ca, Finland, Hong Kong, and Namibia.
-P»
ro SOURCE: U.S. Bureau of Mines, Minerals Yearbook and Mineral Industry Survey, various Issues, Washington, D.C.: U.S. Government Printing
o Office. ""~
Charles
River
Associates
-------
Table 4-12
PLATINUM IMPORTS FOR CONSUMPTION
IN THE UNITED STATES BY COUNTRY,1 1970-1980
(Troy Ounces)
Country
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
19802
Australia
Belgium-Luxembourg
Canada
Colombia
France
Germany, Federal
Republic of
Italy
Japan
Mexico
Netherlands
Norway
Romania
South Africa,
Republic of
Sweden
Switzerland
USSR
United Kingdom
Yugoslavia
Other
Total
1,004
2,957
22,163
--
3,140
—
12,840
37
5,309
4,707
--
97,541
—
1,017
31,078
306,705
—
1,709
490,207
150
2,061
22,027
5
9,260
1,240
18,503
--
150
6,713
--
116,818
__
12
69,241
224,176
-_
3753
470,731
220
2,647
17,220
--
9,253
—
37,316
269
—
8,237
—
107,339
—
--
169,394
260,697
__
3.7951*
616,387
--
278
23,496
953
5,150
--
49,706
112
305
7,234
--
89,294
__
11,079
86,757
399,768
__
—
674,132
1,467
2,128
12,300
—
2,296
--
45,246
15
—
13,601
--
597,569
—
5,669
108,580
313,726
—
1,034s
1,103,631
2,180
1,810
12,600
--
1,976
302
10,011
112
2,252
9,229
--
412,166
—
6,554
33,642
189,589
—
9266
683,349
240
1,181
1,620
--
7,873
14,034
5,409
8,170
—
1,199
8,369
--
693,994
550
3,510
50,430
195,336
482
7.9007
1,000,297
—
3,462
--
2,770
8,273
—
2,498
2,313
3,309
4,801
2,680
675,010
__
8,100
10,432
95,955
_-
3,277
822,880
7,187
6,328
1,252
--
—
20,178
10,019
347
3,451
4,556
—
933,411
--
9,770
20,210
158,175
—
9,331
1,174,215
4,768
1,815
--
--
--
9,231
19,090
--
1,452
10,842
--
1,199,601
--
16,912
25,640
130,369
—
14,491
1,434,211
„
12,187
33,452
119
225
17,085
13,551
15,471
--
13,754
6,611
—
1,059,512
193
24,331
15,892
219,493
643
4.7058
1,436,338
Table continued on following page.
Charles
River
Associates
-------
Table 4-12 (Continued)
PLATINUM IMPORTS FOR CONSUMPTION
IN THE UNITED STATES BY COUNTRY,1 1970-1980
(Troy Ounces)
1 Includes unwrought platinum grains, nuggets, sponge, and semimanufactured platinum. Excludes small amounts of platinum contained 1n
sweepings, waste, scrap, and unspecified combinations.
2January-December 1980.
3Includes Denmark.
* Includes Botswana, Brazil, Costa Rica, Finland, Ghana, Malawi, New Zealand, Panama, and Turkey.
5Includes Republic of Korea.
6Includes Costa Rica and Panama.
7Includes Peru and Portugal.
8Includes Argentina, Costa R1ca, Finland, Hong Kong, Namibia, Cyprus, and Israel.
_fi
i SOURCE: U.S. Bureau of Mines. Minerals Yearbook and Mineral Industry Survey, various Issues, Washington, D.C.: U.S. Government Printing
ro Office. " ~~
Charles
River
Associates
-------
Table 4-13
PALADIUM IMPORTS FOR CONSUMPTION
IN THE UNITED STATES BY COUNTRY,1 1970-1980
(Troy Ounces)
i
ro
GO
Country
Austral la
Belgium-Luxembourg
Canada
Colombia
France
Germany, Federal
Republic of
Italy
Japan
Mexico
Netherlands
Norway
South Africa,
Republic of
Sweden
Switzerland
USSR
United Kingdom
Yugoslavia
Other
Total
1970
24
2,275
__
—
4,538
502
3,252
6,969
10,637
_-
28
456,206
287,457
1,929
773,817
1971
8,525
18,279
—
_-
—
__
631
5,840
39,163
888
458
332,909
254,643
2.0123
663,348
1972
974
13,512
--
--
6,840
34,310
— -
520
15,124
111,920
._
2,098
523,112
193,819
--
902,229
1973
5,092
—
--
5,718
5,4%
42,366
6,950
137,615
__
16,450
668,737
265,881
—
1,154,305
1974
665
18,955
—
1,905
4,882
57,628
8
750
7,390
319,854
_-
1,502
763,343
160,172
i;033*
1,338,087
1975
1,124
2,725
11,739
—
1,979
16,604
8,920
2,609
11,555
294,481
4,500
4,447
75,076
117,118
1,225
554,102
1976
2,744
16,773
--
3,363
16,305
260
2,436
9,304
444,119
650
8,431
427,102
187,152
3,922
7505
1,123,311
1977
„
6,592
21,840
—
1,000
6,334
15,730
_-
3,387
8,%2
486,639
--
2,330
514,249
81,416
3,198
1,151,677
1978
32,337
20,821
—
—
—
6,286
4,119
6,454
498,786
__
11,233
503,438
121,134
19,485
1,224,093
1979
12,573
23,800
—
—
—
2,534
18,866
14,657
690,439
~-
19,812
602,307
95,002
24,439
1,504,434
19802
__
51,053
25,736
•* "~
200
32,885
4,549
yi
37,759
5,925
648,987
__.
2,003
339,570
164,231
2,604
9956
1,316,588
Table continued on following page.
Charles
River
Associates
-------
i
r\>
Table 4-13 (continued)
PALADIUM IMPORTS FOR CONSUMPTION
IN THE UNITED STATES BY COUNTRY,1 1970-1980
(Troy Ounces)
'Includes unwrought and semimanufactured palladium. Excludes small amounts of palladium contained In sweepings, waste, scrap, and
unspecified combinations.
2 January-December 1980.
3 Includes Austria and Denmark.
'•Includes Republic of Korea.
5 Includes Peru.
6 Includes Namibia.
SOURCE: U.S. Bureau of Mines, Minerals Yearbook and Mineral Industry Survey, various Issues, Washington, D.C.: U.S. Government Printing
Office.
Charles
River
Associates
-------
Table 4-14
RHODIUM IMPORTS FOR CONSUMPTION IN
THE UNITED STATES BY COUNTRY,1 1970-1980
(Troy Ounces)
Country
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
Table continued on following page.
I9602
-p.
ro
tn
Australia
Belgium-Luxembourg
Canada
Colombia
France
Germany, Federal
Republic of
Italy
Japan
Mexico
Netherlands
Norway
South Africa,
Republic of
Sweden
Switzerland
USSR
United Kingdom
Other
Total
68
3,500
4,160
—
1,465
—
1,320
—
—
--
714
—
--
7,694
22,091
—
41,012
39
--
2,932
—
323
—
--
—
—
--
335
—
--
5,478
25,555
--
34,662
—
428
__
—
27
—
3,213
6
—
--
2,524
—
1
7,139
37,078
3883
50,804
__
._
__
3
—
113
262
—
2,045
_.
—
34,344
56,444
—
93,211
1,210
467
__
300
1,406
-_
6,567
8
96
__
8,622
__.
--
34,646
45,285
—
98,607
— —
897
__
--
3,028
12,831
_.-
300
15,810
__
100
37,977
11,025
6 11*
82,029
2,272
628
1,548
__
— _
-._
26,208
1,200
12,699
19,306
263
64,124
__
2,191
2,946
947
353
— —
342
132
33,690
— -
1,288
19,743
18,308
—
79,940
1,044
510
_._
106
53
2,495
77
53,041
750
23,453
20,893
516
102,938
725
1,186
_._
866
1,559
699
65,157
250
17,310
19,610
1,656
109,018
57
230
643
129
128
914
1,007
81,891
64
50
8,482
15,387
1,295s
110,277
Charles
River
Associates
-------
I
ro
en
Table 4-14 (Continued)
RHODIUM IMPORTS FOR CONSUMPTION
IN THE UNITED STATES BY COUNTRY,1 1970-1980
(Troy Ounces)
Includes unwrought and semimanufactured rhodium. Excludes small amounts of rhodium contained 1n sweepings, waste, scrap, and
unspecified combinations.
2January-December 1980.
3Includes Botswana,, Brazil, Costa R1ca, Finland, Ghana, Malawi, Netherlands-Antilles, Panama, Turkey, and Venezuela.
''Includes Finland.
5Includes Finland and Namibia.
SOURCE: U.S. Bureau of Mines, Minerals Yearbook and Mineral Industry Survey, various Issues, Washington, D.C.: U.S. Government Printing
Office.
Charles
River
Associates
-------
ro
-g
Table 4-15
TOTAL PRIMARY AND SECONDARY PLATINUM-GROUP
METALS REFINED IN THE UNITED STATES, 1970-1979
(Troy Ounces)
1970 1971 1972 1973 1974 1975 1976 1977 1978 1979
Primary Metal :
Nontoll Refined
Toll Refined
Total
19,822
270,335
290,157
21,184
233,850
255,034
15,380
84,219
99,599
19,916
38,566
58,482
13,234
20,107
33,341
16,571
17,174
33,745
7,101
10,232
17,333
5,199
1,083
6,282
8,303
1,354
9,657
8,392
476
8,868
Secondary Metal:
Nontoll Refined 350,176 278,175 255,641 265,901 325,216 270,101 215,355 195,219 257,191 309,022
Toll Refined 1.451.535 1.218.988 1.277.404 1,000.623 1.067.915 1.158.294 859,432 1.003.940 1.021,960 1.090.202
Total 1,801,711 1,497,163 1,533,045 1,266,524 1,393,131 1,428,395 1,074,787 1,199,159 1,279,151 1,399,224
SOURCE: U.S. Bureau of Mines, Minerals Yearbook, various Issues, Washington, D.C.: U.S. Government Printing Office.
Charles
River
Associates
-------
Table 4-16
PRIMARY AND SECONDARY PLATINUM
REFINED IN THE UNITED STATES, 1970-1979
(Troy Ounces)
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
Primary Metal :
Nontoll Refined
Toll Refined
Total
8,036
183,264
191,300
10,198
156,599
166,797
3,708
54,773
58,481
5,560
32,883
38,443
4,103
16,293
20,396
5,292
14,619
19,911
2,748
8,676
11,424
831
466
1,297
1,081
177
1,258
1,980
56
2,036
Secondary Metal:
Nontoll Refined 118,298 103,429
Toll Refined 896.472 625,649
Total 1,014,770 729,078
75,942 94,884
787.697 581,005
863,639 675,889
95,999 103,623
654,156 635,148
750,155 738,771
64,901 50,838 75,585 75,038
494.069 620.848 630.961 585.932
558.970 671,686 706,546 660,970
-^
CO
SOURCE: U.S. Bureau of Mines, Minerals Yearbook, various Issues, Washington, D.C.: U.S. Government Printing Office.
Charles
River
Associates
-------
Table 4-17
PRIMARY AND SECONDARY PALLADIUM
REFINED IN THE UNITED STATES, 1970-1979
(Troy Ounces)
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
Primary Metal :
Nontoll Refined
Toll Refined
Total
10,322
74,953
85,275
10,237
66,467
76,704
10,836
23,752
34,588
13,121
3,972
17,093
8,634
2,784
11,418
10,968
2,002
12,970
4,025
1,063
5,088
4,300
610
4.910
7,222
1.177
8,399
6.412
420
6,832
Secondary Metal:
Nontoll Refined 208,555
Toll Refined 494,758
Total 703,313
162,718
431.248
593,966
150,019
373.396
523.415
149,552 134,747
437.809 311.000
587,361 445,747
134,086 166,371 220,639
327.450 344,022 446.189
461,536 510,393 666,828
SOURCE: U.S. Bureau of Mines, Minerals Yearbook, various Issues, Washington, D.C.: U.S. Government Printing Office.
Charles
River
Associates
-------
I
oo
o
Table 4-18
PRIMARY AND SECONDARY RHODIUM
REFINED IN THE UNITED STATES, 1970-1979
(Troy Ounces)
1970 1971 1972 1973 1974 1975 1976 1977 1978 1979
Primary Metal:
Nontoll Refined
Toll Refined
Total
Secondary Metal:
Nontoll Refined 13,394
Toll Refined 47.861
Total 61,255
64
8,885
8,949
83
8,118
8,201
62
3,354
3,416
88
381
469
38
185
223
28
164
192
35
95
130
6
3
9
8,837
43,173
52,010
11,390
44,065
55,455
11,561
36,865
48,426
11,127
36,1%
47,323
13,683
49,063
62,746
8,058
34,035
42,093
5,011
42,178
47,189
8,266
35,914
44,180
7,964
38,875
46,839
SOURCE: U.S. Bureau of Mines, Minerals Yearbook, various Issues, Washington, D.C.: U.S. Government Printing Office.
Charles
River
Associates
-------
I
CO
r
c
e
n
t
108
88~
68-
48-
28-
8
Figure 4—2
Percentage of Total U.S. Platinum-Group Metal Imports from Various Countries
CPrepared By: Charles River Associates)
II 1^ I I I I I I I
1978 1971 1972 1973 1974 1975 1976 1977 1978 1979
South AfrIca
UnI ted KIngdom
USSR
Other
years
SOURCE" U.S. Bureau of Mines, Minerals Yearbook,
Mineral Industry Surveys, various Issues, Washington, D.C.
Charles
River
Associates
-------
I
00
IN3
r
c
e
n
t
180
88-
60-
40-
20-
0
Figure 4—3
Percentage of Total U.S. Platinum Imports from Various Countries
CPrepared By: Charles River Associates)
\ I I I I I I I I I
1970 1971 1972 1973 1974 1975 1976 1977 1978 1979
South Africa
UnIt«d KIngdom
USSR
Other
years
SOURCE" U.S. Bureau of Mines, Mineral* Yearbook,
Mineral Industry Surveys, various Issues, Washington, D.C.
Charles
River
Associates
-------
-Pi
I
CO
oo
r
c
•
n
t
Figure 4—4
Percentage of Total U.S. Palladium Imports from Various Countries
CPrepared By Charles River Associates)
108
88-
68-
48-
28-
8
1978 1971 1972 1973 1974 1975 1976 1977 1978 1979
South AFrIca
UnIt«d KIngdom
USSR
Othere
SOURCEt U.S. Bureau of Mines, Minerals Yearbook^
Mineral Industry Surveys, various Issues, Washington, D.C.
Charles
River
Associates
-------
I
OJ
r
c
•
n
t
188
88"
68-
48-
28-
8
Figure 4—5
Percentage of Total U.S. Rhodium Imports from Various Countries
CPrepared By: Charles River A««ocfates)
I I I
1978 1971 1972 1973 1974 1975 1976 1977 1978 1979
South Africa
Uni ted K i ngdom
USSR
Other
years
SOURCE* U.S. Bureau of Mines, Minerals Yearbook,
Mineral Industry Surveys, various issues, Washington, D.C.
Charles
River
Associates
-------
Charles
River
Associates
Refining of secondary metal for all PGMs in the United States throughout the
1970s fluctuated within the 1.0 to 1.5 million ounce range, except for 1.8
million ounces in 1970 (Table 4-15). Individually, platinum, palladium, and
rhodium secondary refining followed this fluctuating behavior throughout the
last decade, with no readily discernable upward or downward trend (Tables
4-16 through 4-18).
Table 4-19 identifies trends in U.S. exports of PGMs throughout the 1970s.
The data reveal that exports to Japan, Canada, and the United Kingdom have
increased significantly, while exports to West Germany have trended
downward.
DEMAND
Demand for PGMs fluctuates considerably, due to the overall level of economic
activity, technological changes, and government policy actions (such as auto
emission and lead-free gasoline mandates). This section presents a brief
discussion of PGM industrial uses. In many applications PGMs operate like
capital goods, the purchase of which often can be deferred. This fact
exacerbates fluctuations in demand due to changes in economic activity (a
phenomenon economists refer to as the "acceleration principle"), but also
implies consumption can be forgone during supply disruptions.
CATALYTIC USES
A catalyst is a substance that initiates or speeds up a chemical reaction,
while not being consumed itself. Platinum-group metals are excellent
catalysts that are cost effective for many uses, even compared to other much
less expensive metals such as nickel. A case in point is use of PGMs in
catalytic converters. Currently, no alternative base-metal catalytic
converter appears economical, as discussed at length in a later chapter.
Catalytic uses of various types comprise by far the majority of U.S. PGM
demand. Major catalytic applications are discussed below by consuming
industry.
PETROLEUM INDUSTRY. Platinum metals are used as catalysts in the refining of
petroleum products in three processes: catalytic reforming, hydrocracking,
and isomerization. Reforming is generally the largest use of the three, and
it has recently grown considerably. The platinum catalyst eventually fails
due to high temperatures and contamination; it is then recycled, with about a
3 percent loss rate.
4-35
-------
Table 4-19
U.S. EXPORTS OF PLATINUM-GROUP METALS, BY COUNTRY, 1970-1980
(Troy Ounces)
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
Table continued on following page.
I9601
Argentina
Australia
Belgium-Luxembourg
Brazil
Canada
Colombia
Finland
France
Germany, Federal
Republic of
Hong Kong
Italy
Japan
Mexico
Netherlands
j* Norway
co South Africa,
°> Republic of
Sweden
Switzerland
Taiwan
United Kingdom
USSR
Other
Total
673
2,154
24,818
7,644
11,299
—
—
32,842
155,222
157
16,770
70,811
4,158
12,076
--
—
—
8,415
--
64,212
—
2,515
413,766
1,055
1,932
40,204
1,490
18,506
—
__
35,084
90,616
1,015
10,934
94,265
1,926
15,831
--
1,585
--
4,437
—
80,201
—
5,529
404,610
126
2,990
45,117
5,289
10,069
--
-_
4,386
120,685
1,235
19,221
254,460
2,967
9,715
--
1,106
—
4,964
--
52.105
—
4,559
538,994
„
8,906
69,594
6,057
20.358
—
--
--
103,880
--
--
310,940
5,327
19,669
--
—
—
7,819
--
27,901
—
47.0752
627,526
„
—
49,864
--
30,666
--
-_
9,263
181,553
--
--
247,432
—
24,453
--
25,213
--
95,053
11,267
103,913
__
57.0773
835,754
„
1,780
38,812
6,351
39,607
--
—
10,284
135,754
—
20,387
168,774
--
28,389
--
55,722
--
43,238
—
97,656
—
19.1313
665,885
_ _
3,489
53.342
--
47,854
--
—
10,026
74,716
—
7,007
118,857
2,631
3,097
--
11,190
—
32,785
5,985
132,251
1,603
7, 574"
512,407
__
2,255
54,406
1,176
33,263
--
--
6,279
66,821
--
1,718
110,547
3,431
3,311
--
1,053
—
38,875
1,979
90,303
—
11.1145
426,631
1,391
1,109
36,370
10,147
51,606
11,505
6,140
11,609
73,533
2,906
1,721
225,222
58,099
7,838
3,897
2,307
15,951
26, 963
1,643
146,197
--
6,393
702,547
640
10,008
37,841
6,816
62,396
1,802
2,574
18,753
98,876
1,487
8,774
328,889
55,004
10,610
3,180
6,320
17,739
48,710
116
154,284
—
22,5996
899,598
936
799
32,283
4,634
72,399
—
4,684
li.838
111,175
984
6,393
237,963
6,144
11,786
3,967
3,252
4,249
67,194
—
173,741
—
10.5437
764,964
Charles
River
Associates
-------
I
OJ
Table 4-19 (Continued)
U.S. EXPORTS OF PLATINUM-GROUP METALS, BY COUNTRY, 1970-1980
(Troy Ounces)
January-December 1980.
2Includes People's Republic of China, Israel, and Spain.
3Includes Israel
*• Includes Venezuela.
5Includes Ireland, Republic of Korea, and Peru.
6Inc1udes Greece, Republic of Korea, Singapore, Spain, and Venezuela.
'Includes Greece, Ireland, Republic of Korea, Singapore, and Venezuela.
SOURCE: U.S. Bureau of Mines. Minerals Yearbook and Mineral Industry Survey, various Issues, Washington, D.C.: U.S. Government Printing
Office.
Charles
River
Associates
-------
Charles
River
Associates
Catalytic reforming is the most efficient way known to raise the octane
rating of gasoline without adding lead. The increasing use of lead-free
gasoline in the United States insures that reforming will continue to be a
major use of platinum metals, though sales to the petroleum industry for
reforming may be considerably lower during the 1980s, after initial loadings
of required increased reforming capacity have been completed (the economists'
"accleration principle" in operation). Modern technology in newer plants,
however, can reduce the platinum intensity of petroleum refining.
CHEMICAL INDUSTRY. The chemical industry is a major consumer of platinum,
palladium, and rhodium for catalysis and pollution control. Nitric acid
production is a major component of chemical industry consumption of
platinum-group metals, accounting for about 30 percent of industry use of
PGMs in an average year. Processing losses for PGMs used in nitric acid
production are much higher relative to initial loadings than is the case for
most other chemical uses.
AUTOMOTIVE INDUSTRY. The automotive industry is, in the early 1980s, by far
the largest user of PGMs in the United States, predominantly for use in
catalytic converters. New, highly sophisticated emission-control systems
also make use of platinum in the coating of an exhaust gas oxygen sensor.
The exact amount of platinum used in each sensor is proprietary, but the
amount is much less than is found in converters.
Previously, oxidation catalysts were used to reduce emissions, but with
tighter standards for nitrogen oxide emissions, most U.S. car makers have
switched to the new three-way catalyst. The oxidation catalyst uses
approximately 0.05 troy ounces of platinum and palladium in a roughly 2 to 1
ratio. The three-way catalyst uses platinum, palladium and rhodium.
NONCATALYTIC USES
ELECTRICAL END USES. Platinum, palladium, and rhodium have high electrical
conductivity and hence are used in the electrical equipment industry.
Platinum is used in thermocouples, electrical contacts, and electroplating of
printed circuits, while palladium is used for relays and metal contacts in
telecommunications equipment.
The Bell System uses large amounts of palladium in its switching equipment.
Most contacts are now a silver-palladium alloy rather than all palladium.
Conversion from mechanical to electronic switching by the Bell System is
decreasing its demand for palladium. By the mid-1980s it is anticipated that
the Bell System will be able to meet all its internal needs for palladium
from its own recycling of obsolete equipment.
4-38
-------
Charles
River
Associates
GLASS PRODUCTION. Platinum group metals are used in the manufacture of glass
and glass fiber because of high corrosion resistance, ability to withstand
high temperatures, and compatible expansion coefficients. The glass industry
consumes relatively minor quantities of platinum and palladium, but in 1979
it had the third largest reported sales figures for rhodium of any industry
in the United States.
MEDICAL, DENTAL, AND JEWELRY USES. Very little rhodium is consumed in dental
and medical uses, but substantial amounts of platinum and palladium are used.
In 1979, reported sales of palladium were second highest in the dental and
medical categories. Application of PGMs in dental work include dental
crowns, caps, and bridges. Medical uses for PGMs are quantitatively minor.
They are used in cancer chemotherapy, heart pacemakers, and hypodermic needle
tubing.
On the whole, U.S. consumers have not been as much enamored with platinum
jewelry as with gold and silver jewelry, and U.S. reported sales to this
industrial category are relatively low. Japanese jewelry consumers have
historically favored platinum jewelry, but the relative demand for platinum
in jewelry apparently may be fading there.
CONSUMPTION TRENDS IN THE 1970S
Table 4-20 gives a percentage breakdown of world platinum-group metal
consumption by end use. The figures are estimated by private publications
through 1977, and are probably not as reliable as similar figures for the
United States alone. Two readily discernible worldwide trends from
Table 4-20 are that electrical use has declined relatively and that
automotive use began and climbed rapidly in the mid-1970s. Relative world
shares for petroleum, dental and medical, and glass have remained fairly
constant, while chemical and jewelry shares were erratic from 1972 to 1977.
Tables 4-21, 4-22, and 4-23 and Figures 4-6, 4-7, and 4-8 provide an overview
of U.S. consumption of platinum, palladium, and rhodium. As discussed in the
previous section, the U.S. Bureau of Mines PGM consumption data are not
comprehensive, but are the sum of sales reported to consuming industries.
Total "apparent consumption" figures, based on net import and production
data, are consistently above reported sales, but are not reported on an
industry-specific basis. Reported sales, despite being less than
comprehensive, still can be used reliably to examine changing patterns of
interindustry PGM consumption.
As mentioned in the previous section, sales data reflect not only usage of
PGMs directly or indirectly in production, but also possible changes in shelf
stocks. For example, if the chemical industry embarked on a plan to increase
its PGM shelf stocks in a particular year, reported sales data might look
4-39
-------
Table 4-20
WORLD ESTIMATED CONSUMPTION OF PLATINUM
GROUP METALS BY END USE, 1972-1977
(Percent)
Charles
River
Associates
End Use
1972
1973
1974
1975
1976
1977
Electrical
Chemical
Jewelry
Automotive
Petroleum
Dental and Medical
Glass
Others
Total
33.8
27.5
18.4
-
5.7
6.4
3.8
4.4
100.0
33.5
25.2
19.8
-
5.8
6.7
3.6
5.4
100.0
26.5
23.0
19.4
11.2
6.4
6.7
4.0
2.8
100.0
19.1
22.4
27.4
11.3
5.9
6.5
3.0
4.4
100.0
20.8
21.4
20.8
16.7
4.3
8.4
2.7
4.9
100.0
22.6
23.4
20.5
12.7
5.1
6.3
3.7
5.5
100.0
SOURCE: J. Aron and Roskill estimates from Roskill Information Services,
Ltd., 1979, The Economics of Platinum Group Metals. London.
4-40
-------
Table 4-21
PLATINUM SOLD TO CONSUMING INDUSTRIES IN
THE UNITED STATES, BY END USE,1 1970-1980
(Troy Ounces)
Automotive
Chemical
Dental and Medical
Electrical
Glass
Jewelry and
Decorative
Petroleum
•£» Miscellaneous
i
^ Total
1970
—
148,289
18,302
103,318
46,687
29,203
202,015
18,555
566,369
1971
—
135,112
23,097
51,940
40,703
18,577
141,800
19,859
431,088
'Includes primary and nontoll -refined
2Excludes companies
SOURCE: U.S. Bureau
reporting
of Mines
annually
1972
225,895
30,462
92,381
26,970
20,655
98,847
50,089
545,299
1973
—
238,974
27,887
117,352
72,543
22.433
123,649
55,695
658,533
1974
350,000
215,663
25,513
98,608
74,398
22,968
139,519
17,020
943,689
1975
273,000
148,813
17,097
73,624
33,813
22,900
107,988
21,318
698,553
1976
480,965
83,560
26,858
89,319
41,683
23,371
59,103
46,246
851,105
1977
354,338
84,414
27,083
90,217
59,995
34,650
74,772
64,350
789,819
1978
597 , 538
149,696
44,139
106,422
98,094
25,751
108,365
66,336
1,196,341
1979
803,229
98,600
27,053
115,775
88,594
27,712
170,013
77,949
1,408,925 1
19802
517,143
116,609
26,191
142,442
51,843
38,360
141,197
56,372
,090,157
secondary metals.
•
, Minerals Yearbook and
Mineral
Industry Surveys, various
issues,
Washington
, D.C.: U.
S. Government
Prlntlm
Office.
Charles
River
Associates
-------
Table 4-22
PALLADIUM SOLD TO CONSUMING INDUSTRIES IN
THE UNITED STATES, BY END USE,1 1970-1980
(Troy Ounces)
Automotive
Chemical
Dental and Medical
Electrical
Glass
Jewelry and
Decorative
.p. Petroleum
i
PO Miscellaneous
Total
1970
—
184,618
47,583
429,032
21,147
17,329
15,494
24,140
739,343
1971
--
218,651
61,594
431,505
237
18,752
2,916
26,451
760,106
•includes primary and nontoll -refined
2Excludes companies
SOURCE: U.S. Bureau
reporting
of Mines,
annually
1972
—
292,710
94,274
425,081
2,250
19,375
14,499
27,835
876,024 1
1973
—
259,959
135,060
524,056
1,439
23,052
3,761
65,157
,012,484
1974
150,000
163,205
124,074
390,237
9,549
21,701
14,877
12,420
886,063
1975
97,000
142,975
114,970
132,247
17,633
23,026
1,755
11,942
541,548
1976
194,496
128,229
139,279
152,312
2,989
5,700
7,291
26,766
657,062
1977
125,010
161,234
112,473
223,748
907
15,567
8,507
53,023
700,469
1978
198,809
146,352
206,312
286,574
2,757
12,570
18,909
45,645
917,928
1979
222,156
199,743
243,627
392,372
1,729
11,766
24,588
36,640
1,132,621
19802
176,518
116,515
174,8321
289,797
1,121
12,874
21,391
21,320
814,368
secondary metals.
•
Minerals Yearbook and
Mineral
Industry Surve
ys, various
Issues,
Washington,
D.C.: U.
S. Government
Printing
Office.
Charles
River
Associates
-------
Table 4-23
RHODIUM SOLD TO CONSUMING INDUSTRIES IN
THE UNITED STATES, BY END USE,1 1970-1980
(Troy Ounces)
Automotive
Chemical
Dental and Medical
Electrical
Glass
Jewelry and
Decorative
Petroleum
-P»
4=» Miscellaneous
to
Total
1970
--
26,445
51
9,056
7,138
5,343
59
805
48,897
1971
--
14,910
31
9,084
3,362
5,419
176
1,384
34,366
'Includes primary and nontoll -refined
2Excludes companies
SOURCE: U.S. Bureau
reporting
of Mines
annually
1972
--
15,358
48
7,867
13,923
6,593
149
2,157
46,095
1973
—
23,772
297
13,187
16,689
12,526
3,057
3,987
73,515
1974
—
23,328
373
15,538
7,464
10,460
1,239
3,200
61,602
1975
--
15,440
41
8,252
4,471
4,932
114
3,598
36,848
1976
391
19,225
75
9,062
3,828
5,170
1
3,123
40,875
1977
871
20,245
275
10,758
13,986
5,011
4,070
55,216
1978
2,939
19,397
232
14,329
16,605
9,950
281
5,907
69,640
1979
26,136
11,684
45
16,923
15,376
7,458
1,223
4,625
83,470
19802
37,012
5,174
42
6,646
8,420
9,588
650
1,665
63,197
secondary metals.
•
, Minerals Yearbook and
Mineral
Industry Survej
/s, various
Issues,
Washington,
D.C.: U.S.
Government
Printing
Office.
Charles
River
Associates
-------
t
r
o
o
u
n
o
1508008
1200000-
900000 ~
600000 ~
300000 ~
0
Figure 4—6
Platinum Sold to U.S. Consuming Industries
CPrepared By: Charlie River Associates)
I I I I I
1970 1971 1972 1973 1974 1975 1976 1977 1978 1979
Automot i ve
Chemi caI
Dental & Medical
yean
Electri cal
Petroleum
Other
IIIIIHMII
SOURCE' U.S. Bureau of Mines, Mineral* Yearbook,
Mineral Industry Surveys, various issues, Washington, D.C.
Charles
River
Associates
-------
-pa
I
Figure 4—7
Palladium Sold to U.S. Consuming Industries
CPrepared By: Charles River Associates}
1000000-
t
r
o
V 750808 ~
0
u
0 500000 ~
e
8
250000 -
1
mil
u
1
111
11
[
1
1
1
1
U
1
11
1
1J
1
1
11
1
1
11
i
i
i
i
i
i
i
i
1
i
m
1970 1971 1972 1973 1974 1975 1976 1977 1978 1979
MINI
MM
AutomotIve
ChemIcaI
Dental & Medical
year i
Electr i cal
Petroleum
Other
SOURCE• U.S. Bureau of Mines, Mineral* Yearbook,
Mineral Industry Surveys, various Issues, Wash Inoton, D.C.
Charles
River
Associates
-------
I
-C.
CTl
t
r
o
y
o
u
n
c
Figure 4—8
Rhodium Sold to U.S. Consuming Industries
(Prepared By: Charles River Associates}
67288 ~
58488 -
33688 ~
1 6888 ~
LU
••
III
uuu
mi
1
Uil T1
1 I
" ' T
\
u
y
H
•
•ev
•
•H
1
•
T"
MH
1
1978 1971 1972 1973 1974 1975 1976 1977 1978 1979
MMI
linn
AutomotIve
Chem i caI
Dental & Medical
years
Electrical
Petroleum
Other
SOURCEi U.S. Bureau of Mines, Minerals Yearbook^
Mineral Industry Surveys, various Issues, Washington, D.C.
Charles
River
Associates
-------
Charles
River
Associates
abnormally large relative to changes in the production of the chemical
industry.
Figure 4-6 shows the upward trend in U.S. reported platinum consumption
throughout the 1970s. Figures 4-7 and 4-8 indicate that U.S. reported
palladium and rhodium consumption has been more erratic; stock adjustments
may account for part of this.
From Table 4-21, it can be seen that the industries to whom the largest sales
of platinum were reported in 1979 are automotive, petroleum, electrical, and
chemical. For palladium, reported sales were highest in 1978 to the
electrical, dental and medical, automotive, and chemical industries (Table
4-22). For rhodium, reported sales were highest in 1979 to the automotive,
electrical, glass, and chemical industries (Table 4-23).
PRICES
The market for PGMs has a two-tier pricing structure. In broad terms, there
is one price which primary producers charge their regular contractual
customers, called the producer price, and there is another price that others
must pay on the spot market, called the dealer price.
As Figures 4-9, 4-10, and 4-11 demonstrate, producer and dealer prices for
platinum, palladium, and rhodium have closely tracked one another throughout
the 1970s. This is because the economic forces of supply and demand, which
typically first influence the spot price, are also taken into account during
producer negotiations with the contractual consumers. Occasionally,
short-term forces cause the spot price to deviate substantially from the
producer price.
Figures 4-12, 4-13, and 4-14 plot 1980 monthly producer and dealer prices for
platinum, palladium, and rhodium. Dealer prices are more variable, reacting
constantly to market events that will only slowly influence contractual
arrangements. Especially noticeable is the large excess of dealer prices
over producer prices for platinum in early 1980, which many industry
observers attributed to speculative activity. The producers of platinum-
group metals have kept prices to traditional industrial consumers at more
stable levels so that long-run consumption and profitability for the platinum
producers are not damaged for short-run gains.
4-47
-------
-pi
I
-pi
00
t
r
o
y
o
u
n
c
Figure 4—9
Averooe Annual Platinum Producer & Dealer Price*, 1970-1980
CPrepored By: Char lee River Associates}
1000
800-
600-
400-
200-
0
\ I I I I I I I I I I
1970 1972 1974 1976 1978 1980
1971 1973 1975 1977 1979
Producer PrIce
Dealer Price
years
SOURCE' U.S. Bureau of Mine*, Minerals Yearbook.
Mineral Industry Surveys, various Issues, Washington, D.C.
Charles
River
Associates
-------
t
r
o
y
o
u
n
c
Figure 4—10
Average Annual Palladium Producer & Dealer Prices, 1970-1988
(Prepared By: Charles River Associates)
250
200
150-
100-
50-
0
I I I I I I I I I I I
1970 1972 1974 1976 1978 I960
1971 1973 1975 1977 1979
Producer PrIce
Dealer Price
years
SOURCE* U.S. Bureau of Mines, Mineral* Yearbook,
Mineral Industry Surveys, various Issues, Washlnaton, D.C.
Charles
River
Assonintes
-------
seppossv
' jsAia
S3|JDl|O
"O'Q
8861
SZ.61
9Z61
I I
I I
•oijj ~i»onpojj
U61
ZL&\ 0£61
I I
:Ag
8
-882
-009
L008
0001
0881-8^61 '
|onuuy
o
u
n
o
A
o
j
e
d
$
o
^t-
-------
en
$
p
•
r
t
r
o
y
o
u
n
c
1000
Figure 4-12
Monthly Plallnum Producer and Dealer Price*, 1880
CPrepared Bys Charles River Associates)
800-
600-
400-
200-^
0
I I I I I I I
Jan Feb Mar Apr May Jun Jul
I 1 I I I
Aug Sep Oct Nov Dec
Producer PrIce
Dealer Price
1980
SOURCE' U.S. Bureau of Mines, Mineral* Yearbook,
Mineral Industry Surveys, various Issues, Wash I noton, D.C.
Charles
River
Associates
-------
en
ro
t
r
o
Y
o
u
n
c
Figure 4—13
Monthly Pal I adturn Producer and Dealer prices, 1980
CPrepored By: Chart** River Associates)
580
400-
300-
200-
100-
0
n i i i i i r \ r i i r
Jan Feb Mar Apr May Jun Jul Aua S*p Oct Nov Dec
Producer PrIce
DeaIer Pr i ce
1980
SOURCE' U.S. Bureau of Mines, Minerals Yearbook.,
Mineral Industry Surveys, various Issues, Washington, D.C.
Charles
River
Associates
-------
01
CO
t
r
o
y
o
u
n
c
Figure 4—14
Monthly Rhodium Producer and Dealer Prices, 1980
CPrepared By: Charles River Associates}
1088
900-
800-
700-
600-
508
\ I I I F \ I I I I I I
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Producer PrIce
DeaIer PrIce
1980
SOURCEi U.S. Bureau of Mines, Minerals Yearbook,
Mineral Industry Surveys, various Issues, Washington, D.C.
Charles
River
Associates
-------
Charles
River
Associates
PRIMARY PRODUCERS
This section describes capacity, production, and supply elasticities of
countries mining platinum-group metals. Current conditions are analyzed,
focusing on how each country could respond to increased PGM demand, whether
due to increased consumption for vehicular emissions control or some other
cause.
SOUTH AFRICA
CURRENT PRODUCTION
As Tables 4-8, 4-9, and 4-10 indicate, South Africa produces most of the
world's primary platinum and rhodium, and a significant amount of palladium.
More recent 1979 platinum production figures for the "big three" South
African producers are: Rustenburg Platinum Mines, 1.3 million troy ounces;
Impala Platinum Mines, 900,000 troy ounces; and Western Platinum, 80,000 troy
ounces. Total South African 1979 production of palladium was 906,000 troy
ounces; 1979 rhodium,production was 108,000 troy ounces.
South African PGMs come predominantly from the Merensky Reef in the Bushveld
Complex. About two-thirds of the PGMs in the Merensky Reef by weight is
platinum. The ore grade is approximately 8.1 grams of PGMs per metric ton of
ore. The Merensky Reef is one of the few deposits in the world, and by far
the largest, where PGMs are the principle product. In the Soviet Union and
Canada, PGMs are byproducts from the mining of other metals, mostly nickel.
This fact is important, because the supply of PGMs from byproduct operations
responds more strongly to changes in production of the parent metal than to
changes in PGM prices. Byproduct supply is discussed further below.
Current production technologies in South Africa and elsewhere involve
significant time lags between mining and delivery of refined metal. At
present it requires about seven months after mining to deliver refined
platinum or palladium, and about 16 months for other PGMs. In part, these
lags are due to ores, concentrates, and slimes being shipped to other
countries for refining. In 1978, the Bureau of Mines estimated the cost of
refining platinum and palladium at $7 to $8 per troy ounce, and two to three
times this for other PGMs (Mineral Commodity Profiles, 1978).
New production technologies are being developed which should reduce these
production lags. Texasgulf, Inc. is testing a new plasma smelting process to
obtain PGMs from concentrates containing high amounts of chromite. (See
Engineering and Mining Journal, 1979.) The South African National Institute
Tor Metallurgy has already developed a new PGM extraction process from high
chromite ores which Western Platinum will use to mine the UG2 seam, a high
chromite deposit below the Merensky Reef.
4-54
-------
Charles
River
Associates
The UG2 seam's rhodium concentration is almost three times that of the
Merensky Reef. One source (Buchanan, 1980) puts the rhodium proportion
(percent of PGMs accounted for by rhodium) at 8 percent for UG2 versus
3 percent for Merensky. Consequently, both the platinum/rhodium and
palladium/rhodium ratios are lower for UG2 than for Merensky. The
platinum/rhodium ratio is 5.25 for UG2 and 19.66 for Merensky, while the
palladium/rhodium ratio is 4.375 for UG2 and 8.33 for Merensky.
The high rhodium concentration in UG2 is important because three-way
catalytic converters in 1980 used platinum and rhodium in roughly a 10 to 1
ratio, and there is concern about this ratio being much higher than the
current mining ratios from Merensky of 19 to 1. However, Ford's three-way
catalyst reportedly utilizes rhodium, palladium, and platinum in proportions
"not very far off" from those in which they are currently mined.
As the UG2 deposit comes on stream, platinum/rhodium mining ratios could
approach the 10 to 1 usage ratio. CRA communications with mining companies
in South Africa indicate that Impala and Rustenburg Mines, the two largest
producers, are not planning to mine the UG2 seam in the near future.
However, Western Platinum, due to its smaller holdings in the Merensky Reef,
is planning to recover about 110,000 troy ounces of PGMs annually from UG2.
This quantity, however, is small when compared to total 1979 South African
PGM production of 2.28 million ounces by Rustenburg, Impala, and Western.
Despite the fact that the UG2 deposit lies only 100 to 350 meters under the
Merensky Reef, and the two could be simultaneously mined from a vertical
shaft, Rustenburg and Impala have indicated little interest in developing
UG2. Impala personnel have indicated, however, that should auto emission
standards become tighter and drive up the relative demand for rhodium, they
would consider mining UG2.
It is likely that higher rhodium production would have to come from UG2 due
to rhodium's relative scarcity in Merensky deposits. Working with October
1979 producer prices, Buchanan calculated that platinum accounts for 61.7
percent of PGM dollar revenues from the Rustenburg area of Merensky, compared
to 6.6 percent for rhodium (Buchanan, 1979). The current platinum/rhodium
producer price ratio is higher than in 1979 (0.679 versus 0.475), so revenue
from platinum would now account for an even higher percentage relative to
rhodium.
The economics of increased rhodium production are thus clear. Platinum
prices and not rhodium prices currently control PGM production decisions in
Merensky Reef deposits. To significantly increase the rhodium/platinum
production ratio, the UG2 seam will have to be exploited by more than the
planned 110,000 ounce annual production by Western. To entice development of
UG2 by Rustenburg and Impala, relative rhodium prices probably will have to
4-55
-------
Charles
River
Associates
rise; long-term contracts would maximize the incentive effects of any given
price increase. Because UG2 lies directly below Merensky and could be
simultaneously mined from vertical shafts, additional incentives would not
have to be great to encourage development of UG2, but development could still
require several years.
In 1978 the U.S. Bureau of Mines published data on South African PGM refinery
(as opposed to mining) capacity: 2.1 million troy ounces for Rustenburg; 1.5
million troy ounces for Impala; and 135,000 ounces for Western. Rustenburg's
output is refined by Matthey Rustenburg refiners in South Africa and England.
Impala refines all of its material in South Africa, while Western platinum
ships its concentrate to Falconbridge Nickel's refinery in Norway for removal
of nickel and copper, and then back to South Africa for removal to PGMs.
According to recent news reports, Impala undertook a 10 to 20 percent
increase in refinery capacity from 1978 to 1980, and Western hopes soon to
increase its refinery capacity to 245,000 troy ounces at a cost of
$32.7 million.
T. P. Mohide has published estimated 1978 mining (as opposed to refining)
capacity for South African producers (in troy ounces) (Metal Bulletin
Monthly, 1980) : Rustenburg, 1,516,781 for platinum and 564,383 for
palladium; Impala, 1,093,493 for platinum and 423,288 for palladium; Western,
105,822 for platinum and 56,438 for palladium. While Mohide does not
estimate rhodium capacities, they can be estimated by dividing the platinum
figures by 19.66, the platinum/rhodium production ratio in the Merensky Reef.
Using this method, we arrive at 1978 mining capacity figures for rhodium (in
troy ounces) of 77,151 for Rustenburg, 55,620 for Impala, and 5,383 for
Western.
SUPPLY ELASTICITY
Long-term purchase contracts can play a very important role in inducing South
African producers to increase capacity and production. Increasing mining
capacity is a lengthy and expensive process, and inherently risky from a
business viewpoint since demand may suddenly fall after the expansion. The
price incentives required to induce expansion may not be nearly so large if
long-term contracts allow consumers to share risks with producers.
For PGM producers in South Africa and elsewhere, two facts make mining
expansion particularly risky. First, PGM demand, whether for catalytic
converter, petroleum, chemical, or electronic uses, is largely dependent on
macroeconomic activity in the United States, Japan, and Western Europe. A
second variable in PGM markets is Soviet behavior. While the data discussed
above showed that Soviet PGM sales to the West are declining, South African
producers are still reluctant to expand capacity when large Soviet PGM
supplies could begin flowing to the West again. In addition, the Soviets are
thought to have large stockpiles of platinum and palladium (but no
substantial rhodium stockpile). Mohide estimated that in 1979 the Soviets
4-56
-------
Charles
River
Associates
had 20 metric tons, or 705,479 troy ounces of platinum, and 50 metric tons,
or 1,763,698 troy ounces of palladium stockpiled. These stockpiles loom
large as platinum and palladium sold to U.S. consuming industries in 1979
were 1,408,925 and 1,132,621 troy ounces, respectively.
Currently, most major U.S. PGM users, including General Motors, Ford, and
Chrysler, have long-term contracts with South African producers. General
Motors has a ten-year contract with Impala, while Rustenburg supplies Ford
under long-term contract. The GM contract reportedly includes annual
delivery of 300,000 troy ounces of platinum.
While details of these contracts are proprietary, they generally call for an
annual amount of PGMs to be purchased at a fixed price, usually the producer
price or lower. Often there is an inflation clause in the contract. These
long-term contracts for fixed annual deliveries shift the burden of
fluctuating demand to the auto companies. When auto sales and hence PGM
demand are down, the auto companies incur inventory storage and holding
costs, while the South African producers maintain their revenue flows to
cover production and amortized expansion costs. One implication of this
practice is that auto company stocks of PGMs are presumably inversely related
to sales, allowing for lags involved in routing PGMs through catalytic
converter manufacturers.
According to industry sources, large chemical and petroleum refining firms
often have long-term contracts as well with the South African producers.
Long-term contracts are also common for the electronic industry. This leaves
the jewelry, medical and dental, and miscellaneous other industries to buy
mainly on the spot market, although there also may be some contracting in
these industries. Overall, it appears that approximately 70 percent of
platinum, palladium, and rhodium consumption (not counting reuse of
toll-refined material) is typically sold under long-term contract in the
United States. One implication of this situation for the Environmental
Protection Agency is that it will generally be quite costly to implement new
regulations so quickly that the regulated industry is forced onto the spot
market, without allowing time for long-term contracts and capacity expansion
by South African producers.
The degree to which capacity and production expansion decisions are dictated
by long-term contracts and not short-term PGM prices is illustrated by the
historical behavior of South African producers. Two instances are
particularly illustrative. The first occurred during the 1979 slump in auto
sales. Despite the slump, Rustenburg, with the security of long-term
contracts, undertook to increase output from 1 to 1.2 million troy ounces,
and to aim for 1.4 million ounces for 1981. A company executive stated that
the expansion was undertaken to meet perceived future increased demand from
the auto industry (see American Metal Market. 1979a).
4-57
-------
Charles
River
Associates
A second example illustrates the extent to which production and producer
stocks of PGMs are tied to long-term contracts and thus unavailable on the
spot market. Recently, a major U.S computer manufacturer attempted to
contract for a two-year supply of platinum at producer prices and was unable
to do so, allegedly because all producer inventory was committed in long-term
contracts. One industry observer commented that South African producers
would not increase production for a two-year contract, because this was not
considered a long-term commitment compared to more common ten-year contracts
(see American Metals Market, 1979b). The cost of capacity expansion could
not be amortized over so short a time span.
Under normal circumstances, substantial demands for PGMs are best met from
primary production by the major producers in South Africa, through long-term
contracts of more than two years. Producers are fond of emphasizing this
consideration in press reports or negotiations with potential new customers.
Furthermore, since the above example indicates that producer stocks are
likely to be committed in long-term contracts, a potential purchaser on the
spot market in the United States would have to rely largely on refiner,
importer, and dealer stocks. Earlier analysis of these stocks in the United
States revealed that while physically they could meet a substantial increase
in PGM demand, past behavior suggests that these inventories would be
released only for large price increases.
A further advantage of long-term contracts is that they insulate platinum
producers somewhat from competition with their own material, when it is
recycled. This could be a particularly relevant concern during the 1980s,
should large amounts of PGMs be recycled from vehicular emissions control
devices.
Given the key role of long-term contracts, the question then becomes the
extent and speed with which South African PGM production could be expanded
with such contracts. Discussions with industry sources ,and CRA analyses of
South African production technology, indicate that increases in auto industry
platinum demand on the order of 20 to 40 percent (160,045 to 321,291 troy
ounces over 1979 demand) could be readily met in two to three years without
substantial platinum price increases. Substantially larger production
increases could be implemented over longer periods of time without
substantial increases in the (deflated or "real") price of platinum.
The Merensky Reef is situated such that expansion of capacity and production
is not very difficult. In the past, when platinum demand has increased, the
length and shallowness of the reef have made it possible to increase
production quickly. To extract the ore, surface adits, or openings sloping
down along the reef, are used. In the Rustenburg area, the slope is
generally only 10 degrees, in which case simple incline haulage tracks are
used to remove the ore.
4-58
-------
Charles
River
Associates
Deeper deposits in the reef are opened by sinking vertical shafts from 500 to
3,000 feet down. One difficulty with the vertical shafts and deeper
operations is that the temperature gradient in the Merensky Reef is quite
steep, about one degree Fahrenheit for each 90 feet (compared, for example,
to one degree for each 200 feet in South African gold mines).
One factor that would tend to mitigate price increases in response to higher
PGM demand is the threat of Soviet production and stockpiles entering the
supply stream. Soviet behavior as a PGM supplier is analyzed below.
It was noted above that platinum, palladium, and rhodium accounted for 61.7,
8.2, and 6.6 percent, respectively, of the 1979 revenue generated from
Rustenburg area ores (Buchanan, 1980.) In 1980, palladium's share was up a
bit and rhodium's down due to higher and lower current producer prices, but
the point is the same: increased platinum (not just palladium or rhodium)
demand and prices is usually necessary to encourage increased South African
production from the Merensky Reef.
For South African PGM production ratios to move toward rhodium (and
palladium) and away from platinum, the UG2 seam will have to be mined in
conjunction with the Merensky Reef. On the demand side, a solution is to
produce three-way catalytic converters that utilize platinum, palladium, and
rhodium in a manner more consistent with their mining ratios. GM converter
technology apparently utilizes rhodium more intensively than its current
mining ratio, while Ford has indicated that their three-way catalyst PGM
ratios are not very far from current mining ratios. Given this apparent
flexibility in the ratio in which rhodium and platinum are used, it may not
take large changes in relative prices to induce equilibrium in the platinum,
rhodium, and palladium markets, even if more extensive use of three-way
catalytic converters is required.
The key to increasing South African production of PGMs for emissions control
without undue price disruptions is to announce regulatory changes two or
three years in advance, so the auto companies can negotiate higher output
long-term contracts, and producers can implement expansion plans in an
orderly fashion.
THE SOVIET UNION
Until 1971, the Soviet Union was the world's leading supplier of PGMs, but
conditions have changed throughout the 1970s. Soviet sales to the West of
platinum alone steadily increased from 225,000 troy ounces in 1970 to 631,000
ounces in 1976, then steadily decreased to 300,000 ounces in 1979. PGM
proportions in Soviet production are platinum, 25 percent; palladium, 71
percent; and other PGMs, 4 percent. Soviet PGM production is thus heavily
skewed toward palladium.
4-59
-------
Charles
River
Associates
PGM production in the Soviet Union is largely a byproduct of nickel
operations, and hence responds to decisions concerning nickel production.
Soviet production of PGMs in 1979 can be estimated by applying ore grades and
PGM proportions to estimated Soviet nickel production. Buchanan arrives at
estimates of 757,000 troy ounces of platinum and 2,130,000 troy ounces of
palladium. Extending this methodology, estimated 1979 Soviet production of
rhodium is 90,000 troy ounces.
PGMs are found in three major deposits in the Soviet Union: Norilsk,
Petsama, and the Ural Mountains. Norilsk, located in the north-central part
of the Soviet Union in the Krasnoyarsk Territory, is by far the most
important deposit, accounting for 85 percent of Soviet PGM production.
U.S. imports for consumption of platinum, palladium, and rhodium from the
Soviet Union have displayed a more erratic trend over the 1970s than total
Soviet shipments to the West. From Table 4-15 it can be seen that 1979 U.S.
imports of PGMs from the Soviet Union totaled 693,215 troy ounces, down from
a 1974 high of 1,012,321 ounces. The 1979 figure of 693,215 troy ounces was
20 percent of the 3,479,128 total ounces imported from all countries for
consumption that year in the United States.
Tables 4-12, 4-13, and 4-14 focus respectively on platinum, palladium, and
rhodium imports by the United States. Platinum imports from the Soviet Union
have fallen from the mid-1970s, and in 1979 they accounted for only 1.8
percent of total U.S. platinum imports. Palladium imports from the the
Soviet Union have been erratic, but in 1979 they accounted for a significant
40 percent of total U.S. palladium imports for consumption. Likewise,
rhodium imports from the Soviet Union have been erratic, but accounted for
15.9 percent of 1979 U.S. rhodium imports for consumption.
From these calculations it can be seen that palladium is the only PGM for
which Soviet imports are of much importance to the United States. Late in
1980, the Soviets virtually halted palladium exports to the West for reasons
which are unclear. Despite the cutoff, analysts quoted in press reports were
sanguine about adequate compensatory supplies from South Africa, albeit at a
higher price. The Soviet palladium embargo in 1980 demonstrated that, while
the dealer price of palladium is significantly influenced by Soviet sales,
supplies from South Africa can take up much slack and large increases in the
price of palladium are not required to equilibrate the market. Figure 4-13
shows that the dealer price rose from around $165 an ounce in early June, to
around $220 in mid-September.
SUPPLY ELASTICITY
Having seen that U.S. dependence on imports of platinum, palladium, and
rhodium froi.i the Soviet Union is now quite low, we still must consider the
4-60
-------
Charles
River
Associates
elasticity of Soviet supplies of these three metals should U.S. demand
increase. Soviet actions in PGM markets are difficult to predict. One
theory with some historical merit holds that the Soviets sell metal to the
West mainly to earn foreign exchange funds, so sales are inversely related to
Soviet grain harvests and other determinants of net exports. Soviet export
behavior has followed this pattern less clearly in the 1970s than in earlier
years. In any case, Soviet production of PGMs, and hence potential exports,
are bound by decisions about nickel production. A rough consensus of
industry sources is that Soviet PGM flows to the United States will remain
quite insignificant during the 1980s and will not approach higher historic
levels, unless there is a drastic improvement in East-West relations.
CANADA
Although Canada ranks third in world PGM production, it is dwarfed by South
Africa and the Soviet Union. In 1979, Canada produced 327,000 troy ounces of
PGMs, 44 percent of which was platinum, 46 percent palladium, and 10 percent
other PGMs. By comparison, in 1979 the Soviet Union produced an estimated 3
million troy ounces of PGMs, and South Africa produced almost 3.5 million
troy ounces.
Canada has not played an important direct role in U.S. PGM imports; Tables
4-11 through 4-14 show that throughout the 1970s, Canada supplied only a very
small fraction of U.S. imports of PGMs for consumption.
Canadian production of PGMs, like that of the Soviet Union, comes as a
byproduct of nickel operations, principally by International Nickel (INCO) in
Sudbury, Ontario. Other mining operations are found in Quebec, British
Columbia, and Pickle Lake, Ontario. The PGM concentration in the nickel ore
at Sudbury is quite low, averaging only 0.8 to 0.9 grams per metric ton of
ore.
A deposit that could come on line at much higher PGM prices is Lac des lies,
88 kilometers north of Thunder Bay, Ontario. The deposit contains 3 million
tons of ore with 4.450 grams of PGMs per metric ton, and a platinum/palladium
ratio of 1:8. In close proximity to Lac des lies is the Roly zone,
containing 7 million tons of ore with a PGM concentration of 5.5 grams per
metric ton. However, for these deposits to be profitable, PGM prices would
have to increase by several times current levels. At the Roly zone, PGM
prices would have to increase to 5 to 10 times their current level to make
production profitable.
Because PGMs are produced in Canada as a nickel byproduct, output is
predominantly dependent on events in the nickel market.
4-61
-------
Charles
River
Associates
UNITED STATES
Heretofore, primary production of PGMs in the United States has been quite
insignificant (Tables 4-8 to 4-10). However, PGM deposits have been
discovered in Montana, Minnesota, and Alaska, and it is possible that in 5 to
10 years the majority of U.S. palladium requirements, and a significant
portion of platinum requirements, could be met domestically.
The most promising U.S. deposit, in Still water, Montana, is currently being
explored jointly by the Jonns-Manville and Chevron corporations. Based on
public announcements by Johns-Manville and Chevron executives, the deposit
can be profitable at 1980 PGM prices, but a delay of three to five years to
set up operation is required. Potential annual production has been estimated
at 643,000 ounces of palladium and 225,057 ounces of platinum.
Total PGM reserves at Stillwater are estimated to be around 35 million
ounces. There are three zones in the Stillwater complex: the Basil,
Ultramafic, and Banded. According to a 1978 press release by Johns-Manville,
the Banded zone has a strike length of 18,000 feet and an average grade of 2
grams of PGMs per metric ton of ore, with minor amounts of nickel and copper
across a seam width of 7 feet. There is another section in the Banded zone
with an average grade of 24.7 grams of PGMs per metric ton of ore. In both
sections, the platinum/palladium ratio is 1 to 3.5. In the Ultramafic zone,
PGM occurs from traces up to 8.9 grams platinum and 2 grams palladium per
metric ton.
The Minnamax project, about 60 miles north of Duluth, Minnesota, is being
explored by Amax, Inc. and is believed to contain about 4.4 billion tons of
copper-nickel mineralized rock. Tests on a 120-ton bulk sample by the U.S.
Bureau of Mines found 0.7 to 1.2 grams of platinum , 4.1 to 4.4 grams of
palladium, and small concentrations of other PGMs per metric ton of ore.
Using values from a sample tested by International Nickel Co. (0.037 grams
platinum and 0.1 grams palladium per metric ton of ore), estimated total PGM
resources at Minnamax are 18 million troy ounces. The National Materials
Advisory Board estimated PGM reserves there to be 50 million troy ounces.
In Alaska, the Crillion-La Perouse Complex contains an estimated 100 million
tons of copper-nickel ore with a PGM grade of 0.17 grams per metric ton. The
platinum/palladium ratio is 0.8. Goodnews Bay, Alaska, produced over 500,000
troy ounces of PGMs between 1927 and 1976, but is currently not actively
mined.
In the United States, only the Stillwater and Goodnews Bay deposits could be
mined independently of nickel and copper values. Only Stillwater appears
potentially profitable at 1980 PGM prices. Plans now call for an annual
output at Stillwater of 55,000 troy ounces of platinum, 190,000 of palladium,
22,000 of rhodium, and 2C.OOO of other PGMs. It is economically viable to
4-62
-------
Charles
River
Associates
quadruple these figures at current prices, given enough development time.
Goodnews Bay could produce around 10,000 troy ounces of platinum a year, but
only if prices were to increase tenfold.
If test reports are accurate, the United States for the first time could meet
significant portions of its PGM demand from domestic primary production.
Quadrupling the annual output plans stated above would yield 220,000 troy
ounces of platinum, 760,000 ounces of palladium, and 88,000 ounces of
rhodium. Considering total PGMs sold to consuming industries in the United
States in 1979 (Tables 4-21 through 4-23), the above figures represent
15.6 percent of the platinum demand, 67.1 percent of the palladium demand,
and 105.4 percent of the rhodium demand. Of course, the estimated quadrupled
output should be compared with higher future consumption in most end uses.
Output of the magnitudes discussed above would put the United States squarely
in competition with established South African producers. Whether Still water
is developed to the extent discussed above depends not only on the accuracy
of U.S. production cost estimates, but also on the price and production
decisions in South Africa. South African PGM production costs are believed
to be somewhat lower than those projected at Still water.
COLOMBIA
Production of PGMs in Colombia is a byproduct of gold mining. Currently
25,000 ounces of platinum are produced yearly from placer deposits in
Colombia. Reserves are estimated to be under 1 million troy ounces of
platinum. There is virtually no palladium production in Colombia. Due to
the byproduct nature of PGM productiuon, Colombia would not be a significant
factor in meeting increased PGM demand, short of a price rise to many times
current levels.
OTHER COUNTRIES
Over 98 percent of the world's PGM primary production currently comes from
South Africa, the Soviet Union, and Canada. There is only byproduct PGM
production from nickel-copper ores in Australia, Indonesia, the Philippines,
Zimbabwe, and Finland. At much higher prices, China could produce small
amounts of platinum as a byproduct of copper-nickel production in the Garsu
province.
WORLD RESERVES
Table 4-23 summarizes estimated world PGM reserves by country, with data on
ore grades and proportions for each PGM. The huge reserves in South Africa
indicates that long-run supply response should be very price elastic, as long
as this source of supply remains available.
4-63
-------
Charles
River
Associates
The importance of the South African UG2 seam in terms of world reserves is
clear. Moreover, the table facilitates analysis of the proportions of
various platinum group metals in each deposit. For instance, rhodium is seen
to be in high proportion at UG2 and Stillwater.
It is also important to remember that PGMs are primary products in South
Africa. Unike Canada and the Soviet Union, where PGMs are nickel-copper
byproducts, production in South Africa should respond primarily to events in
PGM markets.
SUPPLY RELIABILITY
Any discussion of the reliability of U.S. platinum and palladium supplies
must, of course, begin with South Africa. Rustenburg and Western Platinum
Mines operate in South Africa, whereas Impala's mines, concentrator, and
smelter are in Bophuthatswana, which became a quasi-independent state in
1977, under the South African Homelands policy. Impala's head office and
refinery are in South Africa proper, however. The fact that much of Impala's
operations are in Bophuthatswana is not necessarily a cause for concern,
since Impala is the country's principal industry and relations with the
homeland government generally have been cordial. A portion of Rustenburg's
ore bodies are also in Bophuthatswana.
Labor conditions in the Republic of South Africa actually may be somewhat
more uncertain than those in Bophuthatswana, at least in the short run. The
most recent labor incident in South Africa affecting mining occurred in March
1979 when the South African Mine Workers Union called on all of its 18,000
members to walk out one week after one-third of its miners struck. The major
issue was removal of job reservations for whites only. The strike ended
after six days, after reasonably organized negotiations among mining
companies, union leadership, and the union rank and file. The South African
Chamber of Mines, which can act as an intermediary in labor disputes, did not
have to intervene in the 1979 incident.
The South African mining industry naturally has been sensitive to concern
about its reliability as a long-run source of supply, and it can present a
persuasive case for confidence. On a recent tour of South Africa, U.S.
Representative James D. Santini (D-Nev.), an advocate of an expanded U.S.
national minerals stockpile policy, found little cause for supply concern in
South Africa (American Metal Market, 1980). Santini serves as chairman of
the House Mines and Mining Subcommittee. He stated after his tour that
"South Africa's Achille's Heel is resolving their racial problems," and he
found that the mining companies themselves were on the cutting edge of social
change.
4-64
-------
Charles
River
Associates
It is certainly outside the scope of the present project to forecast the
likelihood of civil unrest in South Africa comparable to that which has
occurred in outlying areas such as Rhodesia. However, it is worth noting
that in most cases where such unrest has taken place in Africa, mining
operations have not been direct targets for catastrophic sabotage, presumably
because any of the parties vying for power strongly prefer to have the mines'
revenues, should they be successful. However, such historical precedents are
hardly conclusive, and there would be many imponderables even if we had a
great deal more time to give to the issue. We have no immediate reason for
great concern about supplies of PGMs from South Africa.
Canada has experienced rather severe labor problems at the Sudbury nickel
operations. Most recently, in 1979 INCO (Canada's largest producer of PGMs),
reduced customer deliveries to 40 percent of 1978 levels due to an
eight-month strike. Regular INCO customers had to turn to the spot market
and pay premium prices.
The Soviet Union cannot of course be considered a reliable source of supply.
It has cut off supplies of metals to the United States in the past, both for
political reasons (e.g., prior to the Korean War) and for possible economic
gain (e.g., to support much higher export prices for palladium). Even
without such identifiable motivations, Soviet actions on export markets can
seem capricious to Western consumers. The moderate market response to the
cessation of Soviet exports of palladium in 1980, in part because of
compensating supplies from South Africa, is encouraging.
CONSUMPTION ELASTICITY AND SECONDARY RECOVERY
In addition to increased primary production, secondary recovery could help
meet increased demand for PGMs due to tighter emissions standards.
Furthermore, other industries could reduce their consumption if PGM prices
increase sufficently. The central issues then become by how much would
demand by other industries fall and secondary recovery increase, after a
given increase in PGM prices.
CRA has recently designed an engineering-based model of demand elasticity and
secondary recovery for PGMs in each major industrial end use. This analysis
for the U.S. Department of Interior is for the purpose of specifying a
complete model of the platinum and palladium industries. The results of this
work are utilized for an analysis of industry-specific demand elasticities
and secondary recovery below. Analyses of these issues for PGM consumption
by the U.S. automotive industry are treated in some detail in another chapter
of this report.
4-65
-------
Charles
River
Associates
PETROLEUM REFORMING
The demand for PGMs in petroleum processing exhibits low demand elasticity
(and then only after substantial lags) due to lack of substitutes. Our
estimates of the long-run price response in petroleum reforming is tabulated
below:
Relative Price Increase
Long-Run Relative
Consumption Decrease
1
1
2
0.967
5
0.867
10
0.700
As the numbers indicate, a doubling of PGM prices would result in less than a
4 percent decrease in quantity demanded even in the long run. Moderate
additional reductions in PGM demand could be obtained from petroleum
reforming only with a 5- or 10-fold price increase. The recycling recovery
rate for petroleum reforming is quite high, perhaps 97 percent.
PETROLEUM CRACKING
Most platinum purchased by the petroleum industry is used for petroleum
reforming, but moderate amounts used for cracking (a recent development,
about which the industry is still quite secretive) are unusual because there
is no secondary recovery and consumption is more responsive to price
increases than is typical for PGMs. Price responsiveness of demand is
estimated in the table below.
Relative Price Increase
Long-Run Rel
Consumption
lative
Decrease
1
1
0.
2
70
0.
5
44
0.
10
31
NITRIC ACID PRODUCTION
Nitric acid production, which uses platinum in a catalytic process, currently
accounts on average for about 30 percent of U.S. platinum purchases for
chemical processing. (Palladium is used in very small quantities relative to
platinum in nitric acid production; the elasticity figures below refer to
platinum.)
Higher platinum prices can reduce PGM demand in nitric acid production by
encouraging better recycling and usage of the so-called "random pack"
technology, which uses platinum less intensively.
Relative Price Increase 1 2 5 10
Long-Run Relative
Consumption Decrease 1 0.74 0.49 0.37
4-66
-------
Charles
River
Associates
The table above indicates that a doubling of platinum prices will decrease
platinum consumption in nitric acid production by approximately 26 percent in
the long run.
CHEMICAL PROCESSES OTHER THAN NITRIC ACID
In most chemical applications other than nitric acid production there is no
significant loss of platinum or palladium, hence little room to improve PGM
recovery rates. The following very low demand elasticity estimates apply to
both platinum and palladium.
Relative Price Increase
Long-Run Re
Consumption
lative
Decrease
1
1
0.
2
997
0.
5
992
10
0.
989
TELEPHONE SWITCHING EQUIPMENT
Our estimates for price elasticity of demand for palladium in telephone
switching equipment is given below.
Relative Price Increase
Long-Run ReT
Consumption
lative
Decrease
1
1
0.
2
92
0
5
.67
10
0.
25
For a price increase of 100 percent, demand for palladium would decrease by
only 8 percent, but would diminish by 75 percent for a 10-fold price
increase. This is possible because electronic switching, which will be used
for most new switching by the 1990s, could be used immediately in place of
palladium (electromechanical) switches if price incentives were sufficient.
Between 1955 and 1974, the Bell System was the largest purchaser of palladium
in the United States, accounting for around 60 percent of palladium sales
reported by the U.S. Bureau of Mines for the electric category. In 1975 a
silver-palladium alloy was substituted for pure palladium contacts. By 1980
the installation of electronic switching instead of electromechanical relays
was common. By 1980, Bell System purchases of nontoll-refined palladium were
down to around 100,000 ounces per year.
Our estimates are that the amount of recycled palladium from scrapped Bell
equipment throughout the rest of the century will be roughly 100,000 troy
ounces annually. Gross consumption of palladium by the Bell System is
projected to decline from 200,000 ounces in 1980 to 100,000 ounces in 1985,
and 50,000 ounces in 1990. Thus, after 1985, the Bell System will be a net
supplier of palladium.
4-67
-------
Charles
River
Associates
DENTAL AND MEDICAL USES
Gold is the most- generally competitive substitute for platinum and palladium
in dental work. If the price of gold were to remain constant, substitution
away from palladium or platinum in response to increases in their prices
would be rapid and extensive:
Relative Price Increase
Long-Run Re"
Consumption
lative
Decrease
1
1
2
0.15
5
0.01
10
0
(If increased prices for platinum and palladium were caused by a countrywide
disruption in South Africa, it is plausible that gold prices would also
increase, in which case substitution away from platinum and palladium would
be much less extensive.)
ELECTRICAL — OTHER THAN TELEPHONE SWITCHES
Platinum and palladium are used in "electronic inks" which dry into
conductive paths in miniature electronic components. Our estimates for price
elasticity of demand for these two platinum metals are given below:
Relative Price Increase
Long-Run Relative
Consumption Decrease
1
1
2
0.78
5
0.59
10
0.50
Even if PGM prices were to increase 10-fold, PGM consumption in electrical
uses (other than telephone switching) would decrease only 50 percent. It is
not anticipated that recovery of PGMs from scrapped electronic gear will be
profitable unless prices increase greatly.
GLASS
Platinum and rhodium are used extensively in the glass industry, in glass
handling and forming equipment, because of its tendency not to react with
other agents. There are no close substitutes for PGMs in glass production,
and as a consequence price responsiveness is low:
Relative Price Increase 1 2 5 10
Long-Run Relative
Consumption Decrease 1 0.97 0.93 0.90
4-68
-------
Charles
River
Associates
PRICE ELASTICITIES FOR CALCULATING THE CRITICALITY
OF PLATINUM, PALLADIUM, AND RHODIUM
The above estimates of the responsiveness of platinum and palladium
consumption to large price increases were specifically developed for analysis
of severe supply disruptions. Thus, they are designed for the type of
analysis of materials criticality that we carried out in the preceding
chapter.
The only category of platinum and palladium consumption not discussed above
is vehicular emissions control. The concluding chapter of this study
discusses the limited possibilities for reducing consumption of platinum and
palladium, assuming currently promulgated emissions standards are maintained.
For purposes of estimating the criticality of platinum, palladium, and
rhodium, we suppose conservatively that a five-fold increase in the prices of
platinum, palladium, and rhodium would make efficient only a 12 percent
decrease in quantities consumed (beyond decreases that would occur anyway,
due to technological advances, decreases in engine sizes and so on). We
assume the same price responsiveness for rhodium consumption. We do not yet
regard this price responsiveness to be a definitive estimate, but we use it
as a placeholder until further work can be done.
Finally, weighting the price responsiveness of all the categories of platinum
and palladium consumption by their expected relative importance in the
mid-1980s (not discussed here), we obtain the following price elasticities of
U.S. demand that were used in the national criticality measurement in the
second chapter:
• Platinum 0.03
• Palladium 0.045
Without similarly detailed analysis, we specify the price elasticity of U.S.
demand for rhodium to be the same as that for platinum.
SPECULATION AND INCREASED DEMAND FOR
PLATINUM GROUP METALS
In recent years speculating in platinum-group metal spot and futures markets
has become more pronounced. Among the platinum-group metals, only platinum
and palladium are traded in futures contracts on the New York Mercantile
Exchange.
4-69
-------
Charles
River
Associates
Speculating in platinum can be simply defined as buying and selling by those
who otherwise have no use for the metal in anticipation of future price
changes. While speculation more immediately affects the spot price of
platinum, it can also affect the producer price, since underlying forces of
supply and demand are eventually reflected in the provisions of long-term
contracts as well as in spot prices. One industry observer has ventured that
the speculative factor is on average equal to about 25 percent of the
activity in the spot market (see American Metals Market, 1980).
The reasons for speculation in platinum, just as in other metals, vary with
the speculator. Some investors may reason that future demand or supply
events which the current price does not reflect may increase platinum prices,
in which case they can purchase on the spot market and hold the metal, or buy
a futures contract for platinum delivery at a later date at a specified
price. Engaging in a futures contract to buy is referred to as "going long."
On the other hand, if the speculator believes that future market events are
likely to decrease platinum's price relative to the currently quoted futures
trading price, he can "sell short", i.e., engage in a contract to deliver
platinum at a future date at a fixed price. If the speculator has guessed
correctly, he can purchase the platinum to deliver at the future date when
the spot price is lower than the previously contracted price, hence earning a
profit. Only two to three percent of the platinum futures contracts at the
New York Mercantile Exchange are ever consummated with actual physical
delivery of the metal. In most cases, traders cancel their positions on a
daily basis by entering futures contracts on both the buying and selling
sides, incurring either a profit or loss on the difference.
In addition to speculators who engage in the above-described activity, there
are those who desire to hold platinum as security for future severe economic
or political instability, where "hard" currency like gold and silver would be
a more viable medium of exchange. It appears however that currently most
speculators in platinum are there purely for the possibility of short-run
economic gains.
At times, platinum prices follow the speculative trends in gold and silver
prices, allegedly because of concern over the value of paper currency in
inflationary times. Often, however, platinum does not move in tandem with
gold or silver, in part because platinum demand, unlike gold, is
predominantly determined by industrial consumption.
Speculation is often responsible for discrepancies between dealer and
producer prices. As Figures 4-9 through 4-14 indicate, the discrepancies are
usually short-lived but sometimes sharp. The speculative sector in the
platinum market today is credited by some observers with causing the clear
divergence of platinum dealer prices over producer prices during the late
1970s and early 1980s (Figure 4-9). Primary producers of PGMs have kept
prices to their industrial customers significantly below those on the spot
market so as to stabilize demand (and thereby presumably maximize profits in
the long run, if not the short run).
4-70
-------
Charles
River
Associates
The reaction of speculators can be a significant consideration for evaluating
any proposed regulatory policy that would be implemented so quickly as to
force auto manufacturers or their suppliers onto the the spot market to
purchase platinum-grown metals. Even in the absence of speculative activity,
demand and supply for platinum-group metals tend to be insensitive to price
changes, so that relatively large price changes are required to equilibrate
the market when a new demand for PGMs suddenly appears. With speculative
activity, an unanticipated demand can destabilize the market and greatly
exacerbate the resulting price variability.
In the long run, much more platinum and palladium can be supplied from South
Africa and even deposits in the United States without greatly increasing the
metals' prices. If at all possible, regulatory changes requiring sizable new
amounts of PGMs should be planned and promulgated more than two years in
advance, so that long term contracts can be signed with producers and the
required new PGM capacity can be created.
4-71
-------
Charles
River
Associates
Appendix
ANNOTATED BIBLIOGRAPHY AND GUIDE TO SOURCES
OF INFORMATION ON PLATINUM-GROUP METALS
To keep abreast of the PGM market at a general level, or to analyze policy
impacts, there is a nucleus of publications which serve quite well. Below
are listed organizations that publish PGM information, followed by a
discussion of which aspects of the market are adequately covered and which
aspects are poorly covered. Finally, an annotated bibiliography is
provided.
U.S. government publications are readily available and very useful. The U.S.
Bureau of Mines (US BOM) regularly publishes PGM data and market analyses;
these publications are often cited in other PGM reports. The Canada Ministry
of Natural Resources occasionally publishes reports on PGM world markets.
Also, the Mineral Bureau of South Africa publishes regular information.
There are also many private sources of PGM information. Roskill Information
Services in London provides regular newsletters as well as periodic
publications that bring together data from around the world. J. Aron and
Co., New York, is also a source of information on PGM markets.
4-72
-------
Charles
River
Associates
A variety of journals and newsletters publish regular PGM information. Some,
like American Metal Market, provide daily price statistics, while others,
like Engineering and Mining Journal, provide information of a more technical
nature.
Other organizations provide irregular but sometimes very useful publications
on PGMs. The International Precious Metals Institute seminars generate
publications often containing PGM information. Also, the National Material
Advisory Board has published monographs on PGM supply and use patterns.
The increased importance of PGMs has led to publication of increasingly
comprehensive data. Primary production data are generally good, except, of
course, for the Soviet Union. Reasonably reliable data exist for consumption
in industralized countries. In the United States, the Bureau of Mines
estimates "apparent consumption" to compensate for the underreporting of
actual consumption. Trade of PGMs among industrialized countries is fairly
well captured; data on trade with less developed countries are not as good.
The most obvious deficiency is data on PGM stocks. The U.S. Bureau of Mines
publishes data on refiner, dealer, and importer stocks, but these do not
include consumer shelf and in-use stocks, or private investor stocks. Only
unofficial estimates.are available for other countries and other kinds of
private stockpiles.
ANNOTATED BIBLIOGRAPHY
U.S. BUREAU OF MINES
1. Mineral Commodity Profiles, Platinum Group Metals, No. 22, September
1978, 23 pages. An excellent general reference for the structure of PGM
markets, and discussion of major supply and demand issues. U.S. and
world coverage is provided. Price, world production and trade, and
consumption data are included.
2. Mineral Commodity Summaries, annual, Platinum Group Metals chapter. A
brief (usually two page) update of important events in PGM markets.
Data provided include percent breakdowns of U.S. consumption, U.S.
production and recycling, U.S. government stockpiles, and world mine
production and reserves. In addition to these data, a general
discussion of trends and issues is provided. Publishing is prompt.
4-73
-------
Charles
River
Associates
3. Minerals Yearbook, annual, Platinum Group Metals chapter. This is the
most comprehensive yearly US BOM publication on PGMs. Data are provided
on producer and dealer prices, U.S. consumption by end-use and type of
production, foreign trade by country, and world primary production.
Also included are general discussions on recent trends in many aspects
of the PGM market, including technological changes in the use
and production of PGMs. Publication lags are significant, however.
4. Mineral Industry Surveys, quarterly, Platinum Group Metals. A pamphlet
with quarterly U.S. data on PGMs. Included are data on PGMs recovered
by refiners; refiner, importer and dealer stocks; consumption by end
use; imports and exports by country; and producer and dealer prices.
5. Mineral Trade Notes, monthly. This publication compiles PGM trade data
obtained in part from the State Department.
6. Minerals and Materials, monthly edition on Platinum Group metals. Two
or three "pages of charts and tables are provided. Coverage includes
domestic consumption, primary and secondary production, trade, and
prices.
ONTARIO, CANADA MINISTRY OF NATURAL RESOURCES
1. Platinum Group Metals --Ontario and the World by Thomas P. Mohide,
Minerals Resources Branch, March 1979, 162 pages. Description of each
of the PGMs, with emphasis on how they are used. For each PGM, a
complete analysis of where and how primary and secondary production
occurs is included. A good general reference or "textbook."
MINERALS BUREAU OF SOUTH AFRICA
1. The Bureau periodically publishes memoranda and reports dealing with
PGMs. Some relevant publications are Internal Memoranda No. 8,
"Platinum Group Metals in Canada," and No. 25, "Platinum Group Metals in
the People's Republic of China."
ROSKILL INFORMATION SERVICES, LTD. (LONDON)
1. Roskill's Letters from Japan, monthly. Periodic information on PGM
demand by category in Japan, and dealer and producer price is reported.
Also, information is provided on ore content and deposits of leading
world PGM producers.
4-74
-------
Charles
River
Associates
The Economics of Platinum Group Metals, Second Edition, 1979. Useful
compendium of worldwide PGM data"Among the hundreds of tables are
end-use for the world and major countries, voluminous primary and
secondary production data by country, and international trade data. One
difficulty, however, is that the 1979 (second) edition reports data only
through 1978.
J. ARON COMMODITIES CORPORATION
1. J. Aron's Precious Metals Research Department periodically publishes
reports on supply and demand characteristics of PGM markets.
INTERNATIONAL PRECIOUS METALS INSTITUTE
1. The IPMI conducts regular seminars and publishes the papers from these
seminars. Often papers are presented on the economics and metallurgy
of PGMs.
NATIONAL MATERIALS ADVISORY BOARD
1. While the NMAB does not publish regular PGM reports, "Supply and Use
Patterns for the Platinum Group Metals" (1980) discusses the criticality
of PGMs and recommends stockpiling objectives.
JOURNALS
1. American Metals Market (daily) and Metals Week (weekly) provide PGM
price and market information. Several other journals periodically
contain PGM information of a more technical nature: Engineering and
Mining Journal, Mining Journal (London), and World Mining.
4-75
-------
Charles
River
Associates
CHAPTER 4 REFERENCES
1. American Metal Market. 1979a. Statement by Albert Robinson, Chairman,
Rustenberg Platinum Holdings, Ltd. Sept. 28, 1979: 12.
2. . 1979b. December 6.
3. . 1980. Feb. 15: 1.
4. Buchanan, D.L. 1980. "Platinum: Great Importance of Bushveld Complex."
World Mining. (August.)
5. . 1979. "Platinum-Group Metals Production from the
Bushveld Complex and Its Relationship to World Markets." University of
Witwaterstrand, Bureau of Mineral Studies, Johannesburg.
6. Charles River Associates Incorporated. 1976. "Policy Implications of
Producer Country Supply Restrictions: The World Platinum and Palladium
Markets." National Bureau of Standards, U.S. Department of Commerce.
7. "Cobalt, Platinum Scene Stockpile Focus." 1981. American Metal Market
Feb. 5: 1.
8. Engineering and Mining Journal. 1979. September: 38.
9. Mohide, T.P. 1980. Metal Bulletin Monthly. February: 55 (Converted
from metric tons to troy ounces by Charles River Associates).
10. National Materials Advisory Board. 1980. Supply and Use Patterns for
the Platinum Group Metals. Washington, D.C.: 109.
11. U.S. Bureau of Mines. 1978. "Platinum." Mineral Commodity Profiles.
No. 22.
4-76
-------
Charles
River
Associates
RECYCLING OF PLATINUM-GROUP METALS
hRUM CAIALYliC OJNVLHIURS
Since 1975, catalytic converters have been fitted onto the exhaust systems of
American automobiles to reduce the level of hydrocarbon and carbon monoxide
emissions. The converters contain small quantities of platinum-group metals
(PGMs) deposited on a substrate that facilitate the chemical breakdown of
exhaust pollutants into harmless components such as carbon dioxide and water
vapor. The PGM content of each catalytic converter is quite small,
approximately 0.05 troy ounces worth about $20 at 1981 prices. 1981 PGM
prices are relatively high by historical standards, and there is some
question whether market forces will be sufficient to induce recycling of
these metals from spent converters.
Consumption of platinum by the automobile industry beginning in 1974 is shown
in Table 4-21, taken from U.S. Bureau of Mines data. Catalytic converters
purchased by the auto industry in 1980 accounted for approximately 40 to
45 percent of total U.S. consumption of platinum and palladium. By way of
contrast, the chemical and electrical industries each consumed approximately
15 percent, of these precious metals. No major reductions in the use of
platinum-group metals have occurred through 1980. The 1981 model year
employs a "three-way" catalyst nationwide (instituted in 1980 in California)
which contains rhodium in addition to platinum and reduced quantities of
palladium.
5-1
-------
Charles
River
Associates
The recovery and recycling of PGMs from the chemical, petroleum, electrical,
glass and other industries have well-establ ished technologies that can
recover more than 95 percent of the precious metals in scrap residues. On
the other hand, recovery technologies for catalytic converters are still
being developed. We will analyze these issues further after a preliminary
discussion of the functions of catalytic converters in automobiles.
BACKGROUND ON THE FUNCTION OF THE CATALYTIC CONVERTOR
Exhaust gases from the internal combustion engine contain, in the absence
of any controls, four environmentally hazardous components: unburned
hydrocarbons, carbon monoxide, oxides of nitrogen, and lead. The first three
can be controlled using the catalytic converter and the fourth by changing to
unleaded gasolines. There is a clear distinction between catalytic systems
for the oxidation of carbon monoxide and unburned hydrocarbons and
systems for the reduction of the oxides of nitrogen. Problems in developing
exhaust catalyst systems lie not only with the catalyst itself, but also with
support systems and reactors capable of withstanding engine exhaust
conditions as well as thermal shock, vibration, and general misuse.
Emissions are worst at start-up, when the catalyst is cold and below its
effective operating temperature.
THE CHARACTERISTICS OF CATALYST MATERIAL
PGMs are the crucial catalysts being employed in converters. The PGMs are
deposited in a very thin layer onto one of two forms of inert substrate,
either a monolithic honeycomb or pellets. To obtain effective performance as
rapidly as possible after engine start-up, the density of the support
material is kept as low as is practical. Recently, a stainless steel
honeycomb has been developed. The entire catalyst is contained in a
stainless steel casing which is placed into the exhaust system between the
exhaust manifold and the muffler. The casing directs the exhaust flow
through the catalyst bed and protects the catalyst from mechanical damage.
The harmful components in the exhaust gases are converted to carbon dioxide
and water vapor in the convertor.
The PGM loading in each convertor manufactured by GM from 1975 to 1979 was
about 0.05 troy ounces of platinum and palladium, combined in a 5:2 ratio,
amounting to 0.036 troy ounces platinum and 0.014 troy ounces palladium.
Prior to 1980, the total catalyst (including substrate) weighed between 4 and
6.4 pounds. In 1980, the density of the substrate was reduced, lowering the
range of weichts to between 2.8 and 4.4 pounds. In 1981, GM began using the
three-way catalyst with a loading of about 0.05 troy ounces platinum, 0.02
troy ounces palladium, and 0.005 troy ounces rhodium per convertor, which
5-2
-------
Charles
River
Associates
amounts to about $30 worth of PGMs at 1981 prices: ($475 per troy ounce
platinum, $140 per troy ounce palladium, and $700 per troy ounce rhodium).
The total weight of the catalyst system averages 5 pounds in the latest
converters. We present below a breakdown of the costs of a catalytic
converter when it is new and another breakdown of the scrap value of the same
converter after 50,000 miles of use.
COST BREAKDOWN FOR NEW AND USED CONVERTORS
The breakdown of the components of a typical new 1980-1981 three-way pellet
converter are given below:
Part
Converter Assembly
Outer Wrap
Shell
Input/Output pipes
Bed Support
Insulation
Pellets
TOTAL
Material
409 Stainless Steel
409 Stainless Steel
409 Stainless Steel
409 Stainless Steel
Fiberglass
Alumina + PGMs
Manufactured Costs
$ 1.92
5.42
3.00
1.68
2.86
1.85
30.05
Total "aftermarket" selling cost,
including markup by auto parts dealer
46.75
$204.00
SOURCE: Rath & Strong, 1980.
5-3
-------
Charles
River
Associates
The value of materials in a spent three-way converter after 50,000 miles of
operation (not counting recovery costs) is approximately as follows:
Value of
Material
Part Material Convertor After Processing Losses
Converter can and 409 Stainless Steel $ 1.25* $ 1.25
support structures
Pellets PGMs $30.05 $24.04**
$31.30 $25.29
SOURCE: Charles River Associates, 1980.
From the above table, it can be seen that what remains of value from a spent
converter is the can and the PGMs. The can is 409 stainless steel which,
because of contamination, will only be worth about $0.05 per pound. This 409
stainless is not high-quality scrap, especially after contamination by
exhaust gases.
New converters of the major automakers were sold by auto parts dealers
("after-market") for approximately $200 in 1980. (See Chi!tons, 1980.) GM
had the following choices of pellet systems in 1980:
Oxidizing catalysts only $188
Oxidizing and reducing catalyst (Cadillac) $204
For renewing spent converter catalysts, GM had a repair kit containing new
pellets for $38, with a labor charge of $20, bringing the total cost for
replacement to $58.
Ford had a range of honeycomb converter prices ranging from $172 to $320 (on
the Maverick and Granada models). Ford also charged $20 for labor to replace
a spent converter, so the cost of putting in a whole new honeycomb catalyst
totaled between $192 and $340.
*Convertor can weighs approximately 25 pounds, and 409 stainless steel scrap
from this source is valued at approximately $0.05 per pound.
**80 percent PGM recovery assumed (10 percent loss from abrasion during
operation and 10 percent loss during processing). There is not yet an
industry concensus on these loss percentages; these assumptions are a
compromise among industry sources with which CRA talked.
5-4
-------
Charles
River
Associates
Chrysler had honeycomb converters that ranged in price from $176 to $196.
Replacement of a spent converter required about 1.2 hours and cost $24 for
labor, bringing the total for converter replacement to $200 to $220 for an
automobile.
The above prices were typical in 1980. The manufacturing/vendor cost is from
one fifth to one quarter of the $200 selling price, or $40 to $50. The
corresponding actual plant manufacturing cost would be roughly $27 to $34 per
converter. There is, of course, a significant markup over manufacturing
costs by the time the converter reaches the customer. However, a spent
converter, after 50,000 miles of use, will contain scrap metals worth at most
only about $31, including the value of the stainless steel can, which in 1980
was worth about $.05 per pound. If the recovery rate for PGM processing and
recycling is 80 percent, the total scrap value of contained metals drops from
$31 to $25 per converter. (See the table above.)
RECYCLING CATALYTIC CONVERTORS
There are three phases in the lifetime of a catalytic converter when it may
be scrapped, allowing recovery of contained PGMS: 1) after failure to meet
original specifications; 2) after usage makes the performance of the
converter on an operational vehicle unsatisfactory; and 3) when the vehicle
is scrapped. We now consider each of these cases in turn.
PHASE I; REJECTED CATALYSTS
This category consists of calatysts in the form of beads, pellets, monoliths,
honeycomb or biscuits which are supplied by the manufacturers to the
automotive industry, but do not meet specifications for insertion into
convertors. For this classification of scrap, the total amount produced
through 1980 was about 4 million pounds, with roughly an additional 2 million
pounds being generated in 1981.
PHASE II; REPLACEMENT AFTER 50,000 MILES
Catalytic convertors are guaranteed for only 50,000 miles of vehicle
operation. The material from failed convertors scrapped prior to scrappage
of vehicles will be a particularly important source of supply in states with
inspection programs.
The amount of automotive catalyst which will become available will be solely
dependent upon mandatory replacement of spent catalysts after 50,000 miles.
Sebastian Musco of Gemini Industries, a new catalyst recycling company,
5-5
-------
Charles
River
Associates
estimates this market will eventually be 8 million pounds per year,
containing approximately 50,000 troy ounces of platinum and 21,000 troy
ounces of palladium. (See Musco, 1979.) (Replacement catalytic converters
will likely be taken off of late model wrecked cars where the converter is
relatively new and undamaged and resold at prices ranging from $35 to $65
each.)
PHASE III: AUTO CATALYST FROM SALVAGED AUTOS
This phase represents the largest resource for recycled PGMs. Musco at
Gemini Industries has estimated that by 1984, 4 million cars containing 18
million pounds of catalyst will become available. He feels that catalysts
from automobiles in California will begin to surface in significant amounts
by 1981. He projects by 1988 approximately 8.5 million converter equipped
cars will be scrapped, containing 45 million pounds of catalyst, or
approximately 300,000 troy ounces of platinum and 125,000 ounces of
palladium; by 1998, he maintains these figures will double.
Phase I and II converter recovery has been going on for several years, but
Phase III is very much in an embryonic state. Phase III presents a more
complex problem, involving dismantling the converters from salvaged auto
wrecks.
Discussions with several large auto scrap handling companies and precious
metal producers, who are entering the business of PGM recycling from
converters, have indicated that the removal of the converters is quite easy
and fast in most cases. Either the converters are torched off, which takes a
couple of minutes, or a hydraulic cutter is used, which can remove three
converter cans per minute. One scrap yard, when removing the gas tank before
shredding the auto wrecks, uses a fork lift to remove the converter along
with the whole exhaust system in about thirty seconds. There was early
concern that recovering converters is too labor-intensive to be cost
effective. However, we have uncovered no evidence that this is in fact the
case.
One major precious metals producer has estimated that 60 percent of all PGMs
in converters going to scrap yards in 1981 are being recycled in the last two
phases.
AMC and GM converters, containing loose beads or pellets, are simple to open.
The beads are removed easily by a vacuum suction hose and shipped to PGM
processors in drums. The Ford monolithic converter can is more difficult to
shear open, since a torch is not used. The whole converter is usually
shipped to the processor. Probably because of the difficulty with Ford
converters (and the probability of getting some empty cans), Gemini
Industries, a leader in the field, has stopped processing monolithics and now
only handles the beads.
5-6
-------
Charles
River
Associates
The converters being recovered in 1980 came predominantly from 1975 through
1979 model autos, which used the oxidation catalyst only. The typical
loading was a total of 0.05 troy ounces of PGMs in a platinum: palladium
ratio of 5:2, that is, 0.036 ounces platinum and 0.014 ounces palladium.
(The PGM loading varied with model year and engine size, but this was the
average loading level.) At 1980 PGM prices (producer prices) of $475 per
troy ounce for platinum and $140 per troy ounce for palladium, the total
value of PGM in a typical two-way converter was $19.15. Johnson-Matthey, a
large PGM refiner which is getting into converter catalyst recycling, has
reported their refining costs, including PGM losses, to be $8.35 per
converter. This leaves only $10.80 per converter for the purchase and
transportation of used catalysts from the scrap dealers, which, as will be
shown below, is near the breakeven point that makes this activity at all
profitable.
Used converters usually pass through several hands before ending up at the
refinery. The converters are first recovered from salvaged autos at scrap
yards and auto dismantlers, of which there are estimated to be between 15,000
and 20,000 in this country. (See McKinnon, 1978.) Auto dismantlers
typically buy wrecked cars from individuals, auto insurance companies, and
municipal governments. The converters, as well as other major auto parts,
then enter the "core-exchange" market through a core buyer. The core buyer
typically drives a truck around to various scrap yards and picks up barrels
of used converters. He then may sell to a scrap broker, who then sells to
refiners like Gemini, Engelhard, or Johnson-Matthey. If the dismantler is
large enough, he may deal directly with the refiner.
Cohen quoted the following operating cost breakdown for gathering and
processing scrap and catalytic converters (Cohen, 1979):
Range of Cost
Stage of Recycling (dollars)
Auto dismantler 4.00-5.50
Intermediary stages (core buyer 1.50-2.50
and/or core supplier)
Transportation to refinery 0.05-0.12
Refining 0.73-1.22
TOTAL COST $6.28-9.34
We now consider more recent information on this issue.
Information from a 1980 ADRA* survey of scrap dealers in various states has
indicated that dismantling time, the availability of dry storage areas for
300 or more convertors, and shipping costs will be the major costs
determining the profitability of recycling PGMs. If the convertors were
picked up, then $4 per unit would make the activity profitable for scrap
5-7
-------
Charles
River
Associates
dealers. But if the converters had to be shipped, then $7 or $8 would be
required in the case of a Michigan dealer shipping to Gemini Industries in
California. In general, from a survey of Massachusetts auto dismantlers, as
well as the ADRA respondents, prices ranging from $8 to $11 per converter for
bulk quantities delivered to the refiner were found prevailing. Gemini, for
instance, is offering $7.86 for pellet converters and $5.70 for honeycomb
converters at existing platinum prices of $475 per troy ounce. Gemini will
buy only in minimum orders of 1,000 pounds. This is the price for the
catalyst alone. If the can also is bought for its chromium value, another
$1.50 to $2.00 is paid per converter. A Louisiana firm. Southern Scrap
Materials Corporation, now entering the business of refining converter
catalysts, is offering $12.00 to $13.50 per converter.
Transportation costs quoted by Engelhard** are $0.05 per pound of catalyst,
just for shipping from Tennessee to their New Jersey plant, which translates
to $0.30 per pellet converter with 6 pounds of catalyst and $0.20 per
monolithic converter. Assuming 4 pounds for the catalyst system, and using
the refining costs of Johnson-Matthey noted earlier, a revised table of
gathering and processing costs would be as follows for 1980 prices:
Stage of Recycling Pellet Converter Honeycomb Converter
12 12
Auto Dismantler $4.00-6.00 $7.36-13.50 4.00-6.00 $5.70
Intermediary Stages 1.50-2.50 — 1.50-2.50
(core buyer or
supplier)
Transportation to 0.20-0.30 — 0.20-0.30
Refinery
Refining* 8.35 8.35 8.35 8.35
Total Cost $14.05-17.15 $16.21-21.85 14.05-17.15 14.05
1 = Selling to intermediaries from dismantler.
2 = Selling direct to refineriers from dismantler.
*ADRA - Automotive Dismantlers and Recyclers of America, Washington, D.C.
**Personal communication with Englehard Corp., New Jersey.
+See Warwick, 1980.
5-8
-------
Charles
River
Associates
These are average costs. Costs for more specific situations depend on the
origin of the scrapped convertor shipment, the location of the refinery, and
the quality of shipped catalyst material. Moisture, iron, lead, and other
contaminants can reduce the price that refiners will ultimately pay for spent
convertor catalysts.
SPENT CATALYST PGM REFINING
The convertor recyling business is clearly in its infancy, and details on how
the refining companies process the catalyst to recover PGMs is very much
proprietary information right now. However, there are two possible
processing routes that could be taken, which have been discussed publicly in
general terms. The first is by smelting the spent catalyst material,
slagging off impurities, and separating out the PGMs. The second route is by
chemically leaching the PGMs from their alumina substrate and applying
standard recovery methods now being used for primary production of PGMs.
The smelting route appears quite costly in terms of energy consumption, since
platinum melts at 1769°C. Rhodium melts at 1966°C, and the substrate melts
at over 2000°C. It appears that smelting would be the less efficient
processing route.
Chemical leaching of the PGMs leaves behind the substrate. It could be
performed with a number of industrial acids. The spent catalyst would likely
be ground to finer particle size to expose more surface area for more
effective leaching. While no information was obtained on process details,
the outline of a plausible method is as follows: Platinum and palladium are
readily dissolved in aqua regia -- a mixture of hydrochloric and nitric
acids, and HN03, while rhodium will not dissolve and is left behind. This
separates platinum and palladium into a solution where platinum could be
precipitated with ammonium chloride solution, as the impure platinum salt.
Several more steps of redissolving and precipitating platinum would be
required for purification. Subsequently, a pure platinum chloride salt would
be roasted in muffle furnaces at 1000°C to give platinum sponge of 99.99
percent purity.
The palladium remaining in solution after platinum removal might be
precipitated in a manner similar to that for platinum, and ultimately roasted
to form palladium metal sponge.
5-9
-------
Charles
River
Associates
The insoluble rhodium in the catalyst could be removed with molten sodium
trisulfate at 500°C. From the rhodium sulfate solution, rhodium hydroxide
can be precipitated and then dissolved in hydrochloric acid to form rhodium
chloride. Several more redissolving and precipitating stages would be used
to purify the chloride before it goes to a glass lined vessel to be boiled
with formic acid, forming "rhodium black" precipitate. This is then ignited
in a muffle furnace to produce 99.9 percent pure rhodium powder.
This is a hypothetical chemical leaching process based on the current
techniques used in primary PGM production by Johnson-Matthey, INCO, Engelhard
and others. Recovery rates for PGMs from spent catalysts upon processing are
estimated to be approximately 90 percent of the PGMs that physically remain
after use. However, buildup of lead on the catalyst can complicate the
recovery process and reduce the yield. Other losses have been identified by
the PGM refiners, attributed to abrasion of the beads, which removes some of
the PGM surface coating. It apparently is difficult to discern whether loss
is due to abrasion or chemical reactions during operation of the converter.
Some spokesmen for vehicle manufacturers still claim that very little loss of
PGMs occurs, by either volatization or by abrasion. Another type of PGM loss
can occur, called afterburn, which is initiated by the high operating
temperatures inside the converter (over 1000"F). If the engine is run with a
rich fuel mix, unburned gases in the exhaust can ignite in the converter and
even melt down some of the catalyst material, which can make PGM recovery
very difficult or impossible.
The current leader in the technology of recycling PGMs from converters is
generally acknowledged to be Gemini Industries in California. It is a small
company that was retained by General Motors to review the problem of
expensive existing recovery processes and low yields. Gemini claims it has
developed a new, less expensive process to recover the PGMs, and the company
anticipates no difficulties in recovering PGMs from the three-way catalytic
converters going into new automobiles in 1981 (Musco, 1979). Gemini, as of
February 1980, was processing 2 million pounds of catalyst material per year,
and had capacity to process 3.5 million pounds per year. A planned
$2 million capital improvement would expand capacity to 10 million pounds per
year (see Chemical Week, 1980).
Available data on PGM recycling technologies is insufficient to evaluate
independently the summary analyses of their costs provided by refiners such
as Johnson-Matthey. Our best judgment is that industry claims about the cost
of refining converter material are plausible, and improved technologies such
as those claimed by Gemini Industries could indeed determine whether
recycling PGMs in obsolete catalysts would be profitable at PGM prices lower
than those existing today. We do not forecast such substantial decreases in
PGM prices, but they are a possibility, for example, if the Soviet Union
begins exporting much more extensively. At 1981 PGM prices, recycling
appears at least marginally profitable even with "traditional" refining
methods.
5-10
-------
Charles
River
Associates
FUTURE AVAILABILITY OF PGMS FROM SCRAPPED CONVERTORS
During the first three years of substantial catalytic converter use, from
1975 to 1977, approximately 1.1 million troy ounces of PGMs were consumed.
(See Roskill, 1978.) By the end of 1980, the cumulative total had risen to
1,500,000 troy ounces of platinum and 600,000 ounces of palladium (see Musco,
1979). By the early 1980s, 550,000 troy ounces of this PGM content had
become available, according to Sebastian Musco of Gemini Industries.
Engelhard has characterized 1985 as the "kick off" year for PGM recycling,
because at that time most scrapped automobiles coming into auto dismantlers
will have catalytic converters. A significant fraction of new demand for
PGMs by the automotive industry could then be supplied from recycling
catalytic converters.
5-11
-------
Charles
River
Associates
CHAPTER 5 REFERENCES
Chiltons 1980 Labor Guide and Parts Manual. Chi 1 ton Book Co., Radnor, Pa.
Chemical Week. 1980. February 17.
Cohen, Mark A. 1979. "Recycling of Catalytic Converters." Draft. EPA,
Washington. April.
McKinnon, R.F. 1978. "Panel on Trends in the Use of Platinum Group Metals."
Presentation Before the National Academy of Science, Washington, D.C.:
August 30.
Musco, S.P. 1979. "Reclaiming of Precious Metals From Automobile Catalytic
Converters." Third International Precious Metals Institute, Chicago. May:
178.
Roskill. 1978. "Platinum Group Metals." 118.
Wernick, N. 1980. Paper presented at ADRA Annual Convention.
Johnson-Matthey Corp., Honolulu, Hawaii, November.
5-12
-------
Charles
River
Associates
SUBSTITUTES FOR PLATINUM-GROUP CATALYSTS IN VEHICULAR EMISSIONS CONTROL
BACKGROUND
The requirements of the Clean Air Act of 1967, together with the amendments
in 1970 and 1977, provide the framework for the establishment of emissions
standards from automobile engines. These basically set maximum permissible
emission levels for three pollutants, CO, hydrocarbons, and NOX, in terms of
grams per mile, as determined by a standardized test procedure that averages
the representative modes of operation of an automobile.
The emissions standards in effect until 1974 could be met by engine
modifications, but those for the 1975 and later model years required the use
of a catalyst for most cars. For the period of 1975 to 1979, the standards
for NOX (3.1 to 2.0 gm/mile, except for California) could be met by engine
modifications, so the catalyst was an oxidizing catalyst only, designed to
oxidize most of the CO and hydrocarbons in the exhaust to C02 and water.
The standards for 1980 model cars required a further reduction in permissible
CO and hydrocarbon levels. The standards for the 1981 model year reduced the
allowable CO emissions further. The reduction in allowable NOX emissions was
of great importance technically, since it could not be met on all cars by
engine modifications but required the development of a new catalyst system.
The situation with respect to CO emissions is in a state of flux. EPA has
waived the 3.4 g/mile standard for many 1981 and 1982 model cars, allowing
the 7.0 g/mile standard to remain for this period.
6-1
-------
Charles
River
Associates
OXIDIZING CATALYSTS
During the research and development leading to the oxidizing catalyst used
during the period 1975 to 1980, literally thousands of catalyst compositions
were developed and tested by chemical companies, oil companies, catalyst
manufacturers and automobile companies. It was found that no catalyst
compositions could begin to meet the requirements of activity and durability
with the use of gasoline containing lead compounds and the accompanying
halogenated additives (such as ethylene chloride) required to prevent the
buildup of lead deposits in the engine. Hence it was necesesary to require
the use of a lead-free fuel.
Designing catalyst systems to meet mandated emissions standards is a very
intricate matter. The main factors involved are basically the following:
1. Catalyst durability;
2. Emissions before the catalyst reaches operating temperature, that is,
the "cold start" problem; and
3. Intrinsic catalyst activity and response to poisoning from sulfur in the
gasoline.
These factors interact with each other in a complicated manner. An enormous
amount of technical literature exists on various ramifications of the
problem. For example, an excellent, critical and recent review by Kummer of
the Ford Motor Co. (1980) cites 168 references. The key points are the
following:
1. Catalyst Durability
The catalyst is required to meet the required emission levels specified by
the Federal Test Procedure after completing 50,000 miles of driving over a
prescribed route in a specified manner. The catalyst may fail because of
(i) sintering, that reduces active area, (ii) loss of active catalyst by
attrition or spalling, or (iii) deactivation by poisons, especially sulfur.
Catalyst durability is also determined by engine durability. A variety of
engine malfunctions can cause excessive amounts of unburned fuel in the
exhaust that can lead in a short time to excessive catalyst temperatures that
permanently destroy its activity.
2. "Cold Start"
Much of the CO and hydrocarbons that are emitted in the Federal Test
Procedure escape before the catalyst reaches operating temperature, and the
problem is exacerbated by the necessity to choke the engine for start-up,
which unfortunately increases hydrocarbon and CO emissions from the engine.
6-2
-------
Charles
River
Associates
Thus it is vitally important that the catalyst be brought up to operating
temperature as rapidly as possible, a process termed "light-off."
3. Intrinsic Activity
A. Oxidation Catalysis
The best base-metal oxide catalysts, such as copper oxide (CuO) or cobalt
oxide (00304), show activities (per unit surface area) for CO oxidation
comparable to that of noble-metal catalysts, but they are less active for
hydrocarbon oxidation, especially for saturated hydrocarbons. Moreover,
these oxides in an unsupported form sinter readily. When supported on
alumina, which increases the available catalytic area, they tend to react
with the alumina at high temperatures to form less active species, but
alumina is still considered the best available, reasonably economic support.
Non-leaded gasoline contains about 150-600 parts per million of sulfur, which
is converted to S02 during combustion. All base-metal catalysts become
gradually deactivated by S02 in the exhaust, at the temperature range of
400-600QC, as a result of adsorbed sulfate species (Yao, 1975). Those
catalysts containing copper or chromium are least affected and the situation
can be alleviated to some degree by operating at temperatures above about
600oc. The presence of a small amount of noble metal on the surface of a
base metal oxide can also help suppress sulfur poisoning (Gallagher, et al.,
1975). However, supported noble metal catalysts are much less deactivated
than base metal catalysts by S02 at temperatures below about 500°C. Probably
the poisoning of base-metal oxide catalysts could be prevented if the
catalyst were occasionally heated to above 700°C (Fishel, 1974), but
temperatures of 1000°C or so can cause their rapid, irreversible sintering.
In summary, U.S. car manufacturers have gone completely to the use of
platinum or a mixture of platinum and palladium for oxidation catalysts. No
catalysts consisting only of base metals are used. The reasons are as
follows:
• These noble metals are less deactivated by sulfur compounds at
temperatures below 500oC;
• They are more active for hydrocarbon oxidation than base metal oxides;
and
• They are more thermally resistant to sintering.
Development of suitable automobile catalysts requires enormous expense to
demonstrate durability. (It has been estimated that each 50,000 mile
durability test of one car may cost between $50,000 and $100,000.) Therefore
there are very strong incentives to test only those catalysts that clearly
have the potential for the required durability. By about 1973 it had become
6-3
-------
Charles
River
Associates
evident that base metal systems were noncompetitive, and catalyst suppliers
and the automobile companies turned completely to optimizing the performance
of noble metal catalyst systems.
NOX REMOVAL CATALYSTS
The commitment to noble metal catalysts by car manufacturers was further
reinforced by the necessity to develop a new catalyst system to reduce NOx
levels for the 1981 model cars. No known catalyst of requisite activity is
available to decompose NO to N2 and 02. However, by operating a suitable
catalyst within a narrow gas composition range, termed the "window," at
essentially the stoichiometric value* it is possible to markedly lower the
emissions of all three pollutants. This is termed a "three way catalyst
(TWO, because all three pollutants -- CO, hydrocarbons, and NOX -- react
simultaneously. In effect, the exhaust is brought very close to a mixture of
only H20, C02, and N2.
The three-way plus oxidation catalyst system may also be used. It consists
of one bed operated as a three-way converter to reduce NOX, after which
secondary air is added and the mixture passed through an oxidizing converter.
Removal of CO and hydrocarbons is somewhat improved, but NOX conversion is
not as good as with the three-way system, largely because some NH3 is formed
in the first bed and is reoxidized to NOX in the second. However, it can be
operated satisfactorily over a wider range of air-fuel ratios than the TWC.
A three-way catalyst is apparently becoming the preferred system by U.S. car
makers. The fuel-air ratio to the engine is carefully controlled to
essentially the stoichiometric value by use of an oxygen sensor on the engine
exhaust.
The reduction of NOX to nitrogen by CO or H2 is readily catalyzed by base
metal oxide catalysts such as NiO, CuO, or CuCr204, although some of the NOx
may be converted to NH3. NiO is of particular interest in that, in the
presence of H2 or H20, conversion of NO to NH3 is quite low (Shelef and
Gandhi, 1971), and the Ford three-way catalyst used in California in 1979
incorporates NiO as well as Pt and Rh (see the discussion below). However,
these base metal catalysts are severely poisoned by surface sulfides which
are formed under reducing conditions, in contrast to sulfates formed under
oxidizing conditions. Deactivation by sulfur can be lessened by operating at
temperatures above 650°C, but thermal degradation can then be severe.
*The stoichiometric ratio is that ratio of air to fuel at which the oxygen
present is just sufficient to convert the fuel to C02 and H20 if it were
burned completely. If the air-fuel ratio is less than this it is termed
"fuel-rich operation; if the air-fuel ratio is greater than this it is
termed "fuel-lean" operation.
6-4
-------
Charles
River
Associates
For removal of NOX by reduction, the noble metal catalysts are again superior
to the base metal catalysts. Rhodium is more active than platinum or
palladium and produces less NH3, but because of its scarcity and high cost, a
mixture of rhodium with platinum, or with platinum and palladium, is usually
used.
With three-way catalysts, the closed-loop system that controls the air-fuel
ratio to the engine operates essentially by an on-off method that causes this
ratio to oscillate at a frequency that varies moderately above and below
about one Hertz (sec-1). The catalyst thus operates under transient
conditions, the exhaust gas alternately being net oxygen-rich (fuel lean) and
then net oxygen deficient (fuel rich). The performance of noble-metal
catalysts is improved by incorporating an additional component such as cerium
oxide, which is usually described as having an "oxygen storage capacity." In
an oxidized state this component provides oxygen for CO and hydrocarbon
oxidization during the fuel-rich portion of the cycle, when the catalyst is
simultaneously reduced. When the cycle changes to the fuel-lean portion, the
catalyst component is reoxidized from 02 or by NOX; the latter process aids
in NOX removal. The effects are still only partly understood and the water
gas shift reaction, 1^0 + CO -»• H2 = C02, may also play a role in CO removal
(Schlatter and Mitchell, 1980; Hegedus and Gumbleton, 1980).
The monolith three-way catalyst used by the Ford Motor Co. in California in
1979 contained about 14 percent A1203, 1.6 percent Ni02, 0.7 percent Ce02,
0.15 percent platinum and 0.015 percent rhodium. The pelleted catalyst used
by General Motors in California in 1978 contained noble metals at about
0.05 percent of total weight, with a platinum/rhodium ratio of about 2 in one
type of vehicle and about 15 in another (R. Canole, et al., 1978). The noble
metal catalysts can be further improved, at least in principle, by depositing
the noble metal in layers slightly displaced below the outside surface of the
porous support, which gives added protection from poisons such as phosphorus
compounds that come from lubricating oil (Hegedus, et al., 1979; Summers and
Hegedus, 1979; and Hegedus, 1981). The additional expense for development of
the three-way catalysts represents a very large sunk cost that in effect
commits the automobile manufacturers even more deeply to the use of
noble-metal catalysts. (The Japanese car makers have followed essentially
the same path as U.S. car makers, but with slightly different mechanical
methods of controlling the air-fuel ratio with the use of a three-way
catalyst.)
BASE-METAL CATALYST RESEARCH
During recent years, a modest research effort on base metal catalysts has
been maintained at General Motors and by Professor W. Keith Hall (University
of Wisconsin, Milwaukee, Wisconsin), working with G.M. (Hall, 1981). This
work has focused on zeolite catalysts which were not considered in the early
1970s. (A zeolite is a crystalline alumino-silicate containing a very fine
pore structure. A large variety are known and several are currently used
6-5
-------
Charles
River
Associates
commercially as catalysts, primarily in petroleum processing. Some are quite
stable at high temperatures and in the presence of water vapor.) Of a large
variety of compositions considered by Hall and co-workers, a Fe-Y zeolite was
the most attractive (Fu, et al., 1980). It can be reversibly oxidized and
reduced between Fe2+ and Fe3+, and it shows reasonable activity for reaction
of NOx with CO, although not as good activity as that exhibited by a standard
GM oxidizing catalyst. Nonetheless many questions remain, including actual
thermal stability and resistance to poisoning by S02. Hall speculates that
it is possible that Sfy may be excluded by its size from reaction sites
accessible to CO, so that S02 might not be an effective poison in this case,
but this possibility remains to be demonstrated.
A new class of catalysts having strong interactions between a supported metal
and the support have been developed by researchers at Exxon (Tauster, et al.,
1981) (so-called S.M.S.I. catalysts). They consist primarily of group VIII
metals dispersed on transition metal oxides. (Group VIII includes Fe, Co,
and Ni, as well as the six noble metals.) They have not been studied as
automobile catalysts, but they have been shown to have unusual chemisorption
properties which suggests that some members of the class might possibly
adsorb sulfur compounds less readily and hence be more resistant to
poisoning. It is to be emphasized that this possibility is speculative and
no experimental studies have been done. However, it is another direction
that could be explored in the future.
Kummer (1980) emphasizes that one of several requisites for development of a
solely base-metal oxide catalyst is improved stability at high temperatures
and in the presence of water vapor. Some work at Ford Motor Co. was done
with Zr02 as a support, which has a high melting point. New silicalite-type
zeolites developed by Union Carbide (Flanigan, et al., 1978) are stable to
very high temperatures and might be interesting in this regard.
SUMMARY ON POSSIBLE REPLACEMENT OF PLATINUM-GROUP METALS
The above history of the development of automobile catalysts to date shows
the great complexities that have been overcome in order to develop
noble-metal catalyst systems that meet the presently-mandated requirements
for emissions standards and durability. With the present state of the art,
it probably would be possible to design a solely base-metal oxidizing
catalyst system that would at least initially meet (Federal Test Procedure)
standards for emissions of carbon monoxide and hydrocarbons. This system
would have to be somewhat larger than present units utilizing noble metals
and there might be some difficulty with achieving sufficiently fast warm-up.
However, because of sulfur poisoning such a unit could almost certainly not
meet present durability standards and it would not remove NOX, so it could
only meet pre-1981 NOX emissions standards.
6-6
-------
Charles
River
Associates
Technology exists for reducing the sulfur in gasoline to a very low level,
which could substantially prolong the life of a base-metal catalyst. We are
not aware, however, of any analysis of how costly this desulfurizaton would
be. It would probably vary considerably from refiner to refiner and would
depend on the sulfur content of the crude oil. Low sulfur crude oil is even
today a premium material that commands a higher market price than lower
quality, higher sulfur crudes. Unfortunately, crude oils being imported into
the United States increasingly contain higher sulfur contents.
There is no published evidence that a solely base-metal catalyst could match
the performance of the present three-way catalyst system, even initially.
The Ford three-way monolith catalyst incorporating NiO and Ce02 still uses
approximately the same quantity of platinum that is required for a solely
oxidizing catalyst designed to meet 1980 CO and hydrocarbon emission
standards. The very high prices of platinum-group metals of course provide
car makers with a very strong incentive to reduce the cost of their catalyst
units by substituting base metals. Most promising avenues for such
substitutions have been extensively explored.
To reduce the use of platinum significantly further would apparently require
the relaxation of present emissions standards, either with respect to
allowable rates of emission or durability. Exploratory work on certain
zeolite catalysts by General Motors and Professor Hall is interesting, but
much further effort would be required to determine if these catalysts truly
have potential. Considering the fact that all previous base-metal catalysts
have proved inadequate, the like!hi hood of new, solely base-metal catalysts
becoming practical for meeting present standards of durability and emissions
must be considered slight. We have searched the literature and talked to a
number of people knowledgeable in the field of catalysts in general and auto
catalysts in particular, but there appear to be no other leads to solely
base-metal catalysts that would meet present auto emission standards.
6-7
-------
Charles
River
Associates
CHAPTER 6 REFERENCES
1. Canole, R., et al. 1978. S.A.E. Paper 780205.
2. Fishel, N.A., et al. 1974. Environmental Science and Technology, 8,
260.
3. Flanigan, E.M., et al. 1978. Nature, 271.
4. Fu, C.M., M. Deeba and W.K. Hall. 1980. Industrial Engineering &
Chemistry, Product Research and Development, 19, 299.~
5. Gallagher, P.K., etal. 1975. Materials Research Bulletin 10, 623.
6. Hall, W.K. 1981. Personal communication.
7. Hegedus, L.L. 1981. Personal communication.
8. Hegedus, L.L., and J.J. Gumbleton. 1980. Chemtech, 10, 630.
9. Hegedus, L.L.,.etal. 1979. Journal of Catalysis, 56, 321.
10. Kummer, J.T. 1980. Progress in Energy and Combustion Science,
6, 177.
11. Schlatter, J.C. and P.J. Mitchell. 1980. Industrial Engineering &
Chemistry, Product Research and Development!19, 288.
12. Shelef, M. and H.S. Gandhi. 1972. Industrial Engineering & Chemistry,
Product Research and Development, 11, 2.
13. Summers, J.C. and L.L. Hegedus. 1979.
Industrial Engineering & Chemistry, Product Research and Development.
18, 318.
14. Tauster, S.J., etal. 1981. Science, 211, 1121.
15. Yao, Y.Y.F. 1975. Journal of Catalysis, 39, 104.
6-8
-------
Peer Reviewers' Comments on Report Number EPA 460/3-82-012
Larry C. Landman, Project Officer, CTAB
Readers of this Report
Due to the timing of the writing of this final report and the
implementation of EPA's peer review process, it was not possi-
ble to incorporate the reviewers' comments into the final
report entitled "Scarcity, Recycling and Substitution of
Potentially Critical Materials Used for Vehicular Emissions
Control."
The following corrections are applicable to the above cited
report which was prepared under EPA contract number 68-03-2910
and dated February 1982:
Page 2-62
The report now reads as follows:
"...EPA might consider requiring U.S. vehicle manufacturers
to hold specified minimum levels of platinum, palladium,
and rhodium."
EPA has no authority to require manufacturers to do this.
Page 3-5
The report now reads as follows:
"Thus, their [Rath & Strong subproject] report is included
with this study as a separately bound appendix."
At the recommendation of EPA, this subproject report was not
included with this study.
Page 3A-2
The report now reads as follows:
"Grams of Platinum-Group Metals Consumed per Vehicle:"
-------
-3-
Page 4-35
The authors state:
"Currently, no alternative base-metal catalytic converter
appears economical, as discussed at length in a later
chapter."
While base metal catalysts have a number of problems, economics
(i.e., cost) does not appear to be a major difficulty.
------- |