ECONOMIC IMPACT ANALYSIS OF PROPOSED
EFFLUENT LIMITATIONS AND STANDARDS FOR THE
ELECTRICAL AND ELECTRONIC COMPONENTS INDUSTRY
Submitted to:
Environmental Protection Agency
Office of Analysis and Evaluation
Office of Water Regulations and Standards
401 M Street, S.W.
Washington, D.C. 20460
Submitted by:
JRB Associate?
Quantitative Analysis and Evaluation Division
8400 Westpark Drive
McLean, Virginia 22102
EPA Contract No. 68-01-6348
JRB Project No. 2-834-03-760-14
July 23, 1982
-------
PREFACE
This document is a contractor's study prepared for the Office of Water
Regulations and Standards of the Environmental Protection Agency (EPA). The
purpose of the study is to analyze the economic impact which could result from
the application of effluent standards and limitations issued under Sections 301,
304, 306, and 307 of the Clean Water Act to the electrical and electronic
components industry.
The study supplements the technical study (EPA Development Document) sup-
porting the issuance of these regulations. The Development Document surveys
existing and potential waste treatment control methods and technology within
particular industrial source categories and supports certain standards and
limitations based upon an analysis of the feasibility of these standards in
accordance with the requirements of the Clean Water Act. Presented in the
Development Document are the investment and operating costs associated with
various control and treatment technologies. The attached document supplements
this analysis by estimating the broader economic effects which might result
from the application of various control methods and technologies. This study
investigates the effects in terms of product price increases, effects upon
employment and the continued viability of affected plants, effects upon foreign
trade, and other competitive effects.
The study has been prepared with the supervision and review of the Office
of Water Regulations and Standards of EPA. This report was submitted in
accordance with Contract No. 68-01-63A8, Work Assignment 14, by JRB Associates
and was completed in June 1982.
This report is being released and circulated at approximately the same
time as publication of a notice of proposed rulemaking in the Federal Register.
It will be considered along with the information contained in the Development
Document and any comments received by EPA on either document before or during
final rulemaking proceedings.
-------
TABLE OF CONTENTS
SECTION TITLE PAGE
PREFACE
SUMMARY S-l
PART I - ELECTRONIC CRYSTALS SUBCATEGORY
Chapter
INTRODUCTION 1-1
1.1 PURPOSE 1-1
1.2 SCOPE 1-1
1.3 TECHNICAL AND ECONOMIC SUBCATEGORIZATION 1-2
1.4 ORGANIZATION OF PART I OF THIS REPORT 1-3
STUDY METHODOLOGY 1-4
2.1 OVERVIEW 1-4
2.2 STEP 1: DESCRIPTION OF INDUSTRY CHARACTERISTICS 1-6
2.3 STEP 2: SUPPLY-DEMAND ANALYSIS 1-6
2.4 STEP 3: COST OF COMPLIANCE ESTIMATES 1-9
2.5 STEP 4: PLANT LEVEL SCREENING ANALYSIS . 1-9
2.6 STEP 5: PLANT LEVEL PROFITABILITY ANALYSIS 1-10
2.7 STEP 6: CAPITAL REQUIREMENTS ANALYSIS 1-11
2.8 STEP 7: PLANT CLOSURE ANALYSIS 1-13
2.9 STEP 8: OTHER IMPACTS 1-14
2.10 STEP 9: SOCIAL COST ANALYSIS 1-15
2.11 STEP 10: SMALL BUSINESS ANALYSIS 1-16
INDUSTRY DESCRIPTION 1-19
3.1 MANUFACTURING PROCESSES 1-19
3.2 FIRM AND PLANT CHARACTERISTICS 1-19
3.3 PRODUCT CHARACTERISTICS AND USES 1-20
3.4 TRENDS 1-24
3.5 FOREIGN TRADE 1-29
BASELINE PROJECTIONS OF INDUSTRY CONDITIONS 1-30
4.1 U.S. ELECTRONIC CRYSTAL SALES PROJECTIONS 1-30
4.2 CAPITAL EXPENDITURES PROJECTIONS 1-32
4.3 EMPLOYMENT PROJECTIONS 1-32
4.4 SUMMARY 1-33
COST OF COMPLIANCE 1-34
5.1 OVERVIEW 1-34
5.2 POLLUTANT PARAMETERS - 1-34
5.3 RECOMMENDED TREATMENT TECHNOLOGIES 1-35
5.4 TREATMENT COST ESTIMATES 1-36
5.4.1 Existing Sources 1-37
5.4.2 New Sources 1-37
-------
TABLE OF CONTENTS (Continued)
SECTION
PAGE
PART I - ELECTRONIC, CRYSTALS SUBCATEGORY (Continued)
5.5 SOCIAL COST ESTIMATES
5.5.1 Conceptual Framework
5.5.2 Social Cost Analysis
6 ECONOMIC IMPACTS
6.1 PRICE AND QUANTITY CHANGES
6.2 RESULTS OF SCREENING ANALYSIS
6.3 PROFIT IMPACT ANALYSIS
6.4 CAPITAL REQUIREMENTS ANALYSIS
6.5 POTENTIAL PLANT CLOSURES
6.6 EMPLOYMENT EFFECTS
6.7 SUBSTITUTION EFFECTS
6.8 FOREIGN TRADE IMPACTS
6.9 NEW SOURCE IMPACTS
7 SMALL BUSINESS ANALYSIS
8 LIMITATIONS OF THE ANALYSIS
8.1 DATA LIMITATIONS
8.2 METHODOLOGY LIMITATIONS
8.2.1 Price Increases Assumption
8.2.2 Profit Impact Threshold Assumptions
8.2.3 Capital Availability Threshold Assumptions
8.2.4 Plant Closure Assessment
8.3 SUMMARY OF LIMITATIONS
PART II - SEMICONDUCTOR SUBCATEGORY
1 INTRODUCTION
1.1 PURPOSE
1.2 SCOPE
1.3 ORGANIZATION OF PART II OF THIS REPORT
2 STUDY METHODOLOGY
2.1 OVERVIEW
'. 2.2 STEP 1: DESCRIPTION OF INDUSTRY CHARACTERISTICS
2.3 STEP 2: SUPPLY-DEMAND ANALYSIS
2.4 STEP 3: COST OF COMPLIANCE ESTIMATES
• 2.5 STEP 4: PLANT LEVEL PROFITABILITY ANALYSIS
2.6 STEP 5: CAPITAL REQUIREMENTS ANALYSIS
2.7 STEP 6: PLANT CLOSURE ANALYSIS
2.8 STEP 7: OTHER IMPACTS
2.9 STEP 8: SOCIAL COST ANALYSIS
2.10 STEP 9: SMALL BUSINESS ANALYSIS
1-37
1-42
1-42
1-45
1-45
1-46
1-46
1-50
1-54
1-54
1-54
1-54
1-55
1-56
1-59
1-59
1-60
1-60
1-61
1-61
1-62
1-62
II-l
II-l
II-2
II-3
11-4
II-4
II-6
II-6
11-10
11-10
11-12
11-13
11-14
11-15
11-17
11
-------
TABLE OF CONTENTS (Continued)
SECTION TITLE PAGE
PART II - SEMICONDUCTOR SUBCATEGORY (Continued)
3 INDUSTRY DESCRIPTION 11-18
3.1 MARKET CHARACTERISTICS 11-18
3.1.1 Major Product Groups and Trends 11-18
3.1.2 End-Use Markets 11-20
3.2 INDUSTRY STRUCTURE 11-22
3.2.1 Plant Characteristics 11-22
3.2.2 Industry Concentration 11-25
3.3 FOREIGN TRADE 11-25
3.4 PRICING BEHAVIOR AND TRENDS 11-27
4 BASECASE PROJECTIONS 11-32
4.1 U.S. SEMICONDUCTOR SALES PROJECTIONS 11-32
4.2 CAPITAL EXPENDITURES PROJECTIONS 11-34
4.3 EMPLOYMENT PROJECTIONS 11-34
5 COST OF COMPLIANCE 11-38
5.1 OVERVIEW 11-38
5.2 POLLUTANT PARAMETERS . 11-39
5.2.1 Pollution Parameters Analyzed 11-39
5.2.2 Pollutants to be Regulated 11-39
5.3 RECOMENDED TREATMENT TECHNOLOGIES 11-40
5.4 TREATMENT COST ESTIMATES 11-41
5.4.1 Existing Sources 11-43
5.4.2 New Sources 11-47
6 ECONOMIC IMPACTS 11-48
6.1 PRICE AND QUANTITY IMPACTS 11-48
6.2 PLANT LEVEL PROFIT IMPACT ANALYSIS 11-49
6.3 CAPITAL REQUIREMENTS ANALYSIS 11-52
6.4 POTENTIAL PLANT CLOSURES 11-52
6.5 EMPLOYMENT IMPACTS 11-56
6.6 FOREIGN TRADE IMPACTS 11-56
6.7 NEW SOURCE IMPACTS 11-56
6.8 LONG-TERM IMPLICATIONS AND OTHER IMPACTS 11-57
6.9 SMALL BUSINESS ANALYSIS 11-57
6.9.1 Definitions of Small Business 11-58
I 6.9.2 Impacts on Small Entities 11-58
6.10 SOCIAL COSTS 11-61
7 • LIMITATIONS OF THE ANALYSIS 11-63
7.1 DATA LIMITATIONS . 11-63
7.2 METHODOLOGY 11-64
7.2.1 Price Increases 11-64
7.2.2 Sensitivity Analysis 11-65
7.3 PLANT CLOSURES 11-68
7.4 SAMPLING 11-68
7.5 SUMMARY OF LIMITATIONS 11-68
iii
-------
.1ST OF TABLES
NUMBER _TITLE PAGE
S-l INDUSTRY COMPLIANCE COSTS OF ALTERNATIVE TREATMENT OPTIONS S-ll
(In Thousands of 1979 Dollars)
S-2 INDUSTRY COMPLIANCE COSTS OF ALTERNATIVE TREATMENT OPTIONS S-12
(In Thousands of 1982 Dollars)
S-3 SOCIAL COSTS OF ALTERNATIVE REGULATORY OPTIONS S-13
(Millions of 1982 Dollars)
S-4 SUMMARY OF ESTIMATED ECONOMIC IMPACTS FOR THE ELECTRONIC S-15
CRYSTALS SUBCATEGORY
S-4 SUMMARY OF ESTIMATED ECONOMIC IMPACTS FOR THE SEMICONDUCTORS S-16
SUBCATEGORY
PART I - ELECTRONIC CRYSTALS SUBCATEGORY
3-1 NUMBER OF ELECTRONIC CRYSTAL PLANTS IN THE UNITED STATES 1-21
3-2 GEOGRAPHICAL DISTRIBUTION OF ELECTRONIC CRYSTAL PLANTS 1-22
3-3 DISTRIBUTION OF PLANTS BY PLANT SIZE 1-23
3-4 MAJOR APPLICATIONS FOR ELECTRONIC CRYSTALS , 1-25
3-5 U.S. SHIPMENTS OF PIEZOELECTRIC CRYSTAL DEVICES, 1973-1978 1-27
3-6 U.S. SHIPMENTS OF SEMICONDUCTOR DEVICES 1969-1979 1-28
4-1 PROJECTIONS OF U.S. ELECTRONIC CRYSTAL SHIPMENTS 1-31
4-2 PROJECTIONS OF U.S. ELECTRONIC CRYSTAL INDUSTRY NEW CAPITAL 1-32
EXPENDITURES
4-3 PROJECTIONS OF EMPLOYMENT IN THE U.S. ELECTRONIC CRYSTAL 1-33
INDUSTRY
-5-1 POLLUTANTS COMPRISING TOTAL TOXIC ORGANICS 1-35
5-2 PROFILE OF NON-ARSENIDE METAL CRYSTAL MODEL PLANTS 1-38
5-3 COMPLIANCE COSTS OF INDIRECT DISCHARGER NON-ARSENIDE METAL 1-39
CRYSTAL MODEL PLANTS
5-4 TOTAL COMPLIANCE COSTS FOR 53 INDIRECT DISCHARGER NON-ARSENIDE 1-40
CRYSTAL PLANTS
IV
-------
LIST OF TABLES (Continued)
NUMBER TITLE PAGE
PART I
5-5
5-6
6-1
6-2
6-3
- ELECTRONIC CRYSTALS SUBCATEGORY (Continued)
COMPLIANCE COSTS OF EIGHT ARSENIDE CRYSTAL PLANTS
PRESENT VALUE OF SOCIAL COSTS OF PSES REGULATIONS FOR THE
ELECTRONIC CRYSTALS INDUSTRY
SCREENING ANALYSIS
SUMMARY OF PROFIT IMPACT ASSESSMENT FOR NON-ARSENIDE CRYSTAL
MODEL PLANTS
PROFIT IMPACT ASSESSMENT OF POTENTIAL "HIGH IMPACT" ARSENIDE
CRYSTAL PLANTS
1-41
1-44
1-47
1-48
1-49
6-4 COMPLIANCE CAPITAL REQUIREMENTS ANALYSIS - NON-TOXIC METAL 1-51
CRYSTAL MODEL PLANTS
6-5 COMPARISON OF CASH FLOW AND TOTAL CAPITAL REQUIREMENTS 1-52
6-6 COMPLIANCE CAPITAL REQUIREMENTS ANALYSIS - ARSENIDE 'CRYSTAL 1-53
PLANTS
7-1 DISTRIBUTION OF NUMBER OF PLANTS, PLANT REVENUES, COMPLIANCE 1-57
COSTS AND PLANT CLOSURES BY PLANT SIZE - INDIRECT DISCHARGERS
PART II - SEMICQNDUCTOR SUBCATEGORY
3-1 VALUE OF SHIPMENTS OF THE U.S. SEMICONDUCTOR INDUSTRY 1974 11-19
TO 1979
3-2 DISTRIBUTION OF SEMICONDUCTOR PRODUCTS BY MAJOR END-USERS 11-21
(1973-1975)
3-3 DISTRIBUTION OF SEMICONDUCTOR PLANTS AND VALUE OF SHIPMENTS 11-23
BY SIZE OF PLANT
3-4 GEOGRAPHICAL DISTRIBUTION OF DOMESTIC SEMICONDUCTOR PLANTS IN 11-24
1977
3-5 CONCENTRATION RATIOS OF U.S. DOMESTIC SEMICONDUCTOR SHIPMENTS 11-26
' 1957, 1965, AND 1972
3-6 U.S. FOREIGN TRADE OF SEMICONDUCTORS, 1967-1977 11-28
3-7 PRICE INDEXES FOR SEMICONDUCTOR DEVICES (1975-1980) 11-29
-------
LIST OF TABLES (Continued)
NUMBER TITLE PAGE
PART II - SEMICONDUCTOR SUBCATEGORY (Continued)
4-1 U.S. SEMICONDUCTOR SHIPMENT FORECASTS 11-33
4-2 NEW CAPITAL EXPENDITURES FOR THE SEMICONDUCTOR INDUSTRY, 11-35
HISTORICAL AND FORECAST VALUES
4-3 SEMICONDUCTOR INDUSTRY EMPLOYMENT FORECASTS 11-36
5-1 SEMICONDUCTOR PLANT OPERATING STATISTICS 11-44
5-2 MODEL PLANT COMPLIANCE COSTS 11-45
5-3 TOTAL INDUSTRY COMPLIANCE COSTS 11-46
6-1 SALIENT STATISTICS FOR THE SEMICONDUCTOR SAMPLE PLANTS 11-50
6-2 SEMICONDUCTORS - ANNUAL COMPLIANCE COSTS AS A PERCENTAGE OF 11-51
REVENUES
6-3 SEMICONDUCTORS - POLLUTION CONTROL INVESTMENT COSTS AS A 11-53
PERCENT OF ANNUAL CAPITAL EXPENDITURES
6-4 POTENTIAL PLANT CLOSURES AND EMPLOYMENT IMPACTS FOR 11-55
SEMICONDUCTOR PLANTS
6-5 SEMICONDUCTOR - AVERAGE PLANT AND TOTAL INDUSTRY FLOW RATES 11-59
6-6 SUMMARY PROFIT AND CAPITAL IMPACT ANALYSIS BY EMPLOYEE SIZE 11-60
CLASSIFICATION
6-7 SEMICONDUCTORS - SOCIAL COSTS FOR POLLUTION CONTROL 11-62
7-1 SEMICONDUCTORS SENSITIVITY ANALYSIS - ANNUAL COMPLIANCE COSTS 11-66
AS A PERCENTAGE OF REVENUES
7-2 SEMICONDUCTORS SENSITIVITY ANALYSIS - POLLUTION CONTROL 11-67
INVESTMENT COSTS AS A PERCENTAGE OF NEW CAPITAL EXPENDITURES
VI
-------
LIST OF FIGURES
NUMBER TITLE PAGE
PART I - ELECTRONIC CRYSTALS SUBCATEGORY
2-1 ECONOMIC ANALYSIS STUDY OVERVIEW 1-5
PART II - SEMICONDUCTOR SUBCATEGORY
2-1 ECONOMIC ANALYSIS STUDY OVERVIEW II-5
3-1 INTEGRATED CIRCUIT LEARNING DRIVE 11-31
VII
-------
SUMMARY
-------
SUMMARY
1. INTRODUCTION
1.1 Purpose
This report identifies and analyzes the economic impacts which are likely
to result from the promulgation of EPA's effluent regulations on the Electronic
Crystals and Semiconductors subcategories of the Electrical and Electronic
Components point source category. These regulations include effluent limita-
tions and standards based on Best Practicable Control Technology Currently
Available (BPT), Best Available Technology Economically Achievable (BAT),
Best Conventional Pollutant Control Technology (BCT), New Source Performance
Standards (NSPS), and Pretreatment Standards for New and Existing Sources
(PSNS and PSES) which are being proposed under authority of Section 301, 304,
306, 307, and 501 of the Federal Water Pollution Control Act, as Amended (the
Clean Water Act of 1977, Public Law 92-500). The primary economic impact
variables assessed in this study include the social costs of the proposed
regulations and potential for these regulations to cause plant closures, price
changes, unemployment, changes in industry profitability, structure and com-
petition, shifts in the balance of foreign trade, and impacts on small busi-
nesses.
1.2 jndustry Coverage
This study is concerned with two subcategories of the Electrical and
Electronic Products point source category. These industry segments are:
•• Electronic crystals
• Semiconductors.
S-l
-------
The Electronic Crystals subcategory as defined for this study includes
manufacture of piezoelectric, semiconductor, and liquid crystals from polycrys-
talline raw materials. This industry segment does not include establishments
that manufacture polycrystalline materials or finished electronic components.
The major products manufactured are:
• Piezoelectric crystals which include quartz, ceramic,
yittrium iron garnet, and lithium niobate crystals
• Semiconducting crystals which include silicon, gallium,
arsenide, gallium phosphide, indium arsenide, indium,
antimonide, bismuth telluride, and sapphire
• Liquid crystals.
For the purpose of setting effluent standards, the Electronic Crystals
subcategory was organized into the two following product groups:
• Plants which fabricate gallium arsenide and/or indium
arsenide crystals
• Plants which fabricate all other electronic crystals.
The Semiconductor subcategory includes establishments that manufacture
semiconductors and related devices (SIC 3674) and electronic components not
elsewhere classified (SIC 3679). The growing and fabrication of crystals and
the assembly of finished electronic products are not included in this sub-
category. The major product areas of the SIC 3674 industry segment include:
• Hybrid Integrated Circuits - thick filament, thin
film, and multichip devices
• Bipolar Integrated Circuits
• Metal Oxide Silicon Devices
• Transistors
• Diodes and Rectifiers
S-2
-------
• Selenium Rectifiers
• Light Sensitive and Light Emitting Devices (solar
cells and light emitting diodes)
• Thyristors.
The major product areas of the SIC 3679 industry segment include:
• Magnetic Bubble Memories
• Liquid Crystal Displays.
2. METHODOLOGY
The approach used to asses the economic impacts likely to occur as a
result of the costs of each regulatory option is to (1) develop an opera-
tional description of the price and output behavior of the industry and (2)
assess the likely plant-specific responses to the incurrence of the compliance
costs enumerated in the body of this report. Thus, industry conditions before
and after compliance with the proposed regulations are compared. Supplemental
analyses are used to assess linkages of the Semiconductor and Electronic Crys-
tals industries' conditions to other effects such as employment, community,
and balance of trade impacts. These analyses were performed for each of four
regulatory options considered by EPA. The methodology of the study includes
ten tasks that can be grouped into eight major steps. Although each step is
described independently, there is considerable interdependence among them.
Because each of the two subcategories (electronic crystals and semiconductors)
were analyzed separately, specific analytical techniques used in some of the
steps differ somewhat between the two subcategories. Specifically, the study
proceeded using the following eight steps:
iStejp 1: Description Industry Characteristics
The first step in the analysis is to develop a description ol basic
industry characteristics such as the determinants of demand, market structure,
S-3
-------
the degree of infra-industry competition, and financial performance. The
resulting observations indicated the type of analytical methodology needed
for the industry. The sources for this information include government reports,
trade association data, discussions with various trade associations and indus-
try personnel, and an EPA survey of firms in the industry.
Step 2: Industry Supply_^Demand Analysis
The second step in the analysis is a determination of the likely changes
in market prices and industry production levels resulting from each regulatory
option. The estimates of post-compliance price and output levels are used in
the plant-level analysis (steps 4 and 5) to determine post-compliance revenue
and profit levels for specific plants in each product group.
The supply-demand analysis for this study assumes that in the short to
intermediate time periods electronic crystal and semiconductor manufacturers
would attempt to absorb all of the compliance costs and, therefore, there will
be no price increase because of the regulations. This assumption derives from
an evaluation of the structure, operations, and trends of the two industries.
The post-compliance market price levels (i.e., zero price increases
resulting from the regulation) are used, in a later step, to assess the finan-
cial conditions of individual electronic crystals and semiconductor manufac-
turing facilities.
Step 3; Compliance Cost Estimates
Investment and annual compliance costs for the recommended treatment
op_tions were estimated by EPA's Effluent Guidelines Division for treatment
systems of various selected sizes. Based on these cost estimates, compliance
cost curves were developed and then used to estimate plant specific compliance
costs.
S-4
-------
Step A: Plant Level Financial Analysis
Two basic financial characteristics are examined: profitability and
capital requirements. Both characteristics are examined through standard finan-
cial analysis techniques. That is, profitability is measured by both return
on assets (crystals) and return on sales (semiconductors). Capital requirements
of the proposed regulations are evaluated in terms of the amount of the initial
capital investment in relation to revenues (crystals) and in relation to average
industry levels of annual new plant and equipment expenditures (semiconductors).
For the Semiconductors subcategory, the use of these techniques was
hampered by a lack of plant-specific data on a number of input variables,
such as value of shipments, plant asset values, and profit margins. However,
using a number of industry-wide parameters from various published sources
in combination with the limited firm-specific data available, estimates were
developed for each of the key parameters.
Step 5: Estimation of Plant Closures
This step involves the assessment of the degree of impacts on individual
plants. The decision to close a plant, like most major investment decisions,
is ultimately judgemental because it involves a wide variety of considerations,
many of which cannot be quantified or even identified. These assessments were
made by evaluating the above financial variables in conjunction with non-
financial and non-quantifiable factors, such as substitutability of products,
plant and firm integration, the existence of specialty markets, and expected
market growth rates.
Step 6: A,ssessment °l Other Impacts
Once the assessment of plant closures and price and quantity changes are
made, other variables which flow from these are analyzed including employment,
industry structure, and imports and exports. These impacts are assessed
through the use of industry-wide and firmwide ratios calculated from the
available data sources (e.g., value o,f shipments per employee).
S-5
-------
Step 7: Small Business Analysis
This analysis identifies the economic impacts which are likely to result
from the promulgation of the regulations on small businesses in the Electronic
Crystals and Semiconductors manufacturing industries. The primary economic
variables covered are those analyzed in the general economic impact analysis
such as plant financial performance, plant closures, and unemployment and com-
munity impacts. Most of the information and analytical techniques in the
small business analysis are drawn from the general economic impact analysis.
The specific conditions of small firms are evaluated against the background
of the general conditions in each product group's market. In addition to
the general economic impacts discussed in previous steps, two analytical
problems are central to this portion of the analysis: the definition of
small businesses in the Electronic Components industry and the considerations
of alternative regulatory options that might mitigate potential impacts on
small businesses.
Size definitions were sought which would account for firm size in com-
parison to total industry size and in comparison to unit compliance costs
(unit compliance costs increase significantly in reverse proportion to plant
size). Since available data on compliance cost and production is on a plant
basis, the individual production facility, rather than firm is used as the
basis for the analysis. For electronic crystals manufacturing plants four
size definitions based on plant revenues were examined to provide EPA with
alternative definitions of small plants. These size categories are: plants
with revenues under $1 million, $1-3 million, $3-5 million, and greater then
$5 million. For semiconductor manufacturing plants the number of employees
was the primary variable used to distinguish size. Plants with fewer than
100 employes are considered small and those with 100 to 250 employees are
considered medium-small.
The consideration of alternative regulatory options is considered in
terms of the most extreme special consideration — the exemption of small
entities from the regulations. The effects of these exemptions on total
S-6
-------
compliance costs, the economic impacts and the effectiveness of the regula-
tions are assessed.
Step 8; Estimation of Social Costs
Social costs measures the value of goods and services lost by society
V
because of the regulatory action. These costs generally include the use of
resources needed to comply with a regulation, the use of resources to implem-
ent or enforce a regulation, plus the value of the output that is forgone
because of the regulation. A reasonable estimate of these costs is provided
by the present value of direct compliance costs, which generally account for
most of the social costs.
3. INDUSTRY CHARACTERISTICS
3.1 Electronic Crystals
In 1978, U.S. production of piezoelectric crystals was estimated to be
around $110 million, while that for semiconducting and liquid crystals was
$450 million. Shipments of peizoelectric crystals have been growing at an
annual rate of 5 percent since 1973, while shipments of semiconducting and
liquid crystals have been growing about 17 percent a year between 1969 and
1979. U.S. production of piezoelectric crystals is projected to continue to
grow about 3 percent a year to reach $130 million and $160 million in 1983
and 1990, respectively. Meanwhile, shipments of semiconducting crystals are
projected to increase rapidly through 1983 (about 25 percent a year) to
reach $1.1 billion and then average about 8 percent a year for the rest of
the decade to reach $1.9 billion in 1990.
EPA identified 70 plants that manufacture electronic crystals in the
United States, employing about 10,000 employees. The majority of the plants
are concentrated in four states: California, Ohio, Pennsylvania, and Texas.
More than half of the plants are small (i.e., have less than $3 million in
S-7
-------
product shipments or less than 200 employees). Most electronic crystal pro-
ducers are diversified, producing other electronic devices in addition to
crystals.
The electronic crystals industry exhibits some characteristics of noncom-
'petitive markets. There is a significant amount of industry concentration, low
demand elasticities, and a high degree of capital intensity. However, the
existence of the threat of foreign competition appears to be a deterent to
noncompetitive pricing behavior. For these reasons, most firms in the indus-
try appear to have both the capability and the incentive for absorbing small
and moderate amounts of production cost increases in the short and intermediate
terra.
3.2 Semiconductors
The value of shipments for all semiconductor devices in the U.S. domestic
industry was $6.9 billion in 1979. This represented an increase of 20 percent
over the 1978 level of shipments. The rapid growth in this industry segment
has been a result of the increasing demand for integrated circuit products,
which in 1979 represented approximately 70 percent of semiconductor industry
sales.
Approximately a third of these plants were located in California. Most of
the largest semiconductor companies have several domestic plants (usually in
the same region of the country) as well as several foreign plants. The semi-
conductor industry is comprised of a large number of very small and medium-
size plants, which are concentrated in a few states. The 1977 Census of Manu-
factures reports 545 establishments whose primary activity is the manufacture
of_ semiconductors. A significant portion of these plants, especially the
smaller ones, may be R&D and other facilities that are not necessarily involved
in wafer manufacturing and, consequently, are not covered by the proposed
regulation. It is estimated that 257 semiconductor plants will be affected
by the regulation. Of these, 77 plants are direct dischargers and 180 are
indirect dischargers.
S-8
-------
The principal markets for semiconductor devices include the four major
market categories:
• Computer
• Industrial
• Consumer
• Government, including military and space.
These markets, with few exceptions, are outperforming the overall economy.
They are expected to continue to grow in the future.
The specific nature of the pricing mechanism in the semiconductor indus-
try is not clear, since the avaiable evidence show some characteristics that
are indicative of noncompetitive markets and some that are indicative of
competitive markets. Although the market pricing mechanism is not precisely
determined, it has been shown that the industry has both the incentive and
the capability to not raise prices in response to small and moderate levels
of mandated pollution control costs.
4. COMPLIANCE COSTS
Based on the analysis of the potential pollutant parameters and treatment-
in-place in the Electronic Crystals industry, EPA identified six treatment
technologies that are most applicable for the reduction of the selected
pollutants. These treatment technologies are:
• Option 1: Segregation and collection of spent solvents
containing toxic organics for reuse, resale,
or contract hauling (referred to as solvent
management), plus end-of-pipe treatment for
pH control.
• Option 2: Option 1 plus end-of-pipe precipitation/
clarification for control of fluoride, arsenic,
and suspended solids.
S-9
-------
• Option 3: Option 1 plus precipitation/clarification of
concentrated fluoride wastes for control of
fluoride.
0 Option 4: Option 2 plus 75% recycle of treated effluent
to further reduce pollutant discharges.
• Option 5: Option 2 plus filtration to further reduce
pollutant discharges.
• Option 6: Option 5 plus activated carbon to further
reduce toxic organics.
The economic impact analysis does not consider treatment options 4 and
6 because they are not technically feasible nationwide. For this reason,
the economic impact analysis concentrated on treatment options 1, 2, 3, and
5 only. Table S-l presents the estimated investment and annual compliance
costs for these treatment options for the Electronic Crystals and Semicon-
ductors subcategories.
Table S-2 summarizes the compliance costs of the s-slected BPT, BAT,
NSPS, PSES, and PSNS limitations. All except the selected PSES regulations
for the arsenide crystals and BAT regulations for the Semiconductors subcate-
gory will have zero compliance costs. The total investment and annual com-
pliance costs for the proposed regulations are $5.1 million and $3.4 million
(1982 dollars), respectively. The social costs of these regulations are
$58.5 million (1982 dollars) as shown in Table S-3.
5. ECONOMIC IMPACTS
As described previously, the primary economic impact variables assessed
are industry financial performance, plant closures, unemployment community
ef-fects, changes in imports and exports, and industry structure. The findings
regarding these variables are reported first for the Electronic Crystals sub-
category and then for the Semiconductors subcategory.
S-10
-------
CO
z
o
M
H
OH
O
H
z
w
5C
g^
<
W
OS
H
W
^
»— 1 /— -
H tn
< u
Z OJ
Pi l-H
W i-l
H 0
,-4 0
<
ON
fa r-
O ON
l-H
CO
H U-i
00 O
O
CJ> (0
•o
W C
O to
Z W
<: 3
M O
J .C
OH H
s
O C
U -H
>> ^
0-
H
CO
S
Q
Z
Z to
O 3
M C
EH C
OH <
O
H
W
S •
H >
< C
W M
PC
H
CO
i— i
Z co
C 3
n C
EH C
OH <
O
H
Z
H >
< C
W M
PC
H
CN
^•^
Z to
O 3
1-4 C
EH C
OH <
0
H
Z
§ *
H >
< C
W M
OS
EH
i-H
i— 1
z to
O 3
M C
H C
OH <
O
r*. o r»-
ON ON.
in m
in O m
•* «*
CO CO
CM
I
CO
w
a >
H C
< M
U
oo o oo
-* -»
oo oo
r- o
oo o oo
m
co
cs
r-» O
Vt -H
in oo
en cvj
m CN
cs
oo
t~- en
in
CO
oo
ON
Z Z Z
z z z
IT)
O CO
oo o co
CN CN
m o m
r^ r~-
O O
r— o r-~
in in
00
r-
cs
o
^^
cs
*
in
CN
t A
CM
vo r~
co r^«
cs in
A •>
CN CO
ON 00
O\ vO
cr oo
A *
00 O
«J CN
SO O>
in r*-
\D CN
O O O
O O O
O O O
o o o
O O
O O
H
Z
o
w
CO
>H
PC
H
to
E>
a
z
M
to
^
E-i
to
>-
PS
a
0
M
Z
o
as
H
o
M
J
U
m Arsenide
3
•H
•o
c
M
•*^^^
s
•I-l
1— 1
u "— '
to
O
schargers
hargers^.'
• rJ
O
4J
0
41
Vi
• i-l
T3
C
M
U
CO
• H
G
4J
U
4)
lj
•H
Q
rH
CO
4-1
O
4J
J5
3
to
CA
i— 1
CO
4->
CO
>>
u
o
4>
•o
• -I
C
4)
to
u
<;
i
c
o
z
schargers
hargers
•H U
O to
• H
i-> &
O
4) 4-1
I-l U
• ^ 41
-O M
C -H
M C
r-l
CO
U
O
4J
J3
3
to
CN
A
\D
O
O
O
CO
i—I
00
A
m
oo
A
00
CO
U
o
60
41
4-J
cfl
o
.0
3
CO
to
rH
to
4_>
»
>,
U
u
I—I
to
4J
o
EH
CO
Ui
41
60
In
Cfl
jr
o
to
CO
PS
O
H
U
t— '
o.
z
o
u
1— 1
s
w
CO
• f-i
Q
u
O
41
W
• i-l
TT
C
M
CO
Vj
41
60
U
to
JS
o
CO
• I-l
Q
4_>
U
41
U
•H
O
CO
|H
o
u
-,
1_!
o
60
4)
4-1
CO
U
XI
3
to
3
O
lH
60
4_l
U
3
•o
O
u
CX
CO
• H
J=
|H
O
ll .
<
OH
U
>,
Xl
•o
41
•o
C
41
£2
O
o
4)
U
60
C
• H
41
£>
4J
O
c
c
o
« f— (
4J
a.
o
c
41
g
CO
41
1-
H
•
41
r— 1
^3
«
0
• r-l
I— 1
a.
a.
to
4-1
O
z
If
c
o
• rt
4~J
CO
l-H
3
60
41
lH
4)
jT
4J
»C
4J
• i-l
*
4)
U
C
CO
• H
a.
e
o
u
c
• r-t
^
T3
CO
41
1-
r~H
to
41
lj
CO
CO
IH
41
6£
U
CO
r-
*o
to
• H
T3
4-1
U
41
L.
M
•i-4
•o
l-H
CO
^J
CO
>n
I-l
O
CJ
. -J
C
o
u
J-J
0
41
r— 1
U
S-ll
-------
TABLE S-2. INDUSTRY COMPLIANCE COSTS OF SELECTED TREATMENT OPTIONS
(In Thousands of 1982 Dollars!/)
RECOMMENDED REGULATIONS
ELECTRONIC CRYSTALS SUBCATEGORY
Gallium/Indium Arsenide Crystals
BPT
BCT
BAT
NSPS
PSES
PSNS
Non-Arsenide Crystals
BPT
BCT
BAT
NSPS
PSES
PSNS
SELECTED
TREATMENT
TECHNOLOGIES
Option 2
Option 2
Option 2
Option 2
Option 2
Option 2
Option 2
Option 2
Option 2
Option 2
Option 1
Option 1
COMPLIANCE COSTS
INVESTMENT ANNUAL
0
0
0
0
893
0
0
0
0
0
0
0
0
0
0
645
0
0
0
0
0
0
SEMICONDUCTORS SUBCATEGORY
BPT
BCT
BAT
NSPS
PSES
PSNS
Option
Option
Option
Option
Option
Option
0
0
4,242
0
0
0
0
2,719
0
0
TOTAL
5,135
3,364
Adjusted to 1982 dollars using Engineering News Reciord Construction Cost
indexes.
NSPS/PSNS costs are defined as incremental costs from BAT/PSES. Since
the recommended NSPS/PSNS limitations are the same as those proposed for
BAT/PSES, NSPS/PSNS costs are zero.
S-12
-------
TABLE S-3. SOCIAL COSTS OF ALTERNATIVE REGULATORY OPTIONS
(Millions of 1982 Dollars.!)
TREATMENT ALTERNATIVES
Electronic Crystals
Gallium Arsenide/Indium
Arsenide Crystals
Treatment Option 1
Treatment Option 2
Treatment Option 5
Other Crystals
Treatment Option 1
Treatment Option 2
Treatment Option 3
Treatment Option 5
Semiconductors
Treatment Option 1
Treatment Option 2
Treatment Option 3
Treatment Option 5
DIRECT
DISCHARGERS
INDIRECT
DISCHARGERS
0
0
0
$ 0
0
0
0
$ 0
154.2
15.4
166.0
$ 0
4.0
4.3
$ 0
22.1
14.1
23.9
$ 0
361.7
54.5
390.2
jl' Adjusted to 1982 dollars using Engineering News Record Construction Cost
indexes.
S-13
-------
ic Crystals
The estimated economic impacts for the Electronic Crystals subcategory
are summarized in Table S-4. Only the proposed PSES regulations for arsenide
crystal plants require treatment costs. Just one of the five existing
arsenide crystal indirect dischargers has annual compliance costs greater
than one percent of its revenues. The profit impact and capital requirements
of the proposed regulations are significant for this plant. However, this
plant is owned by a large integrated electronic company and supplies crystals
to be used in the production of LEDs at other production facilities of the
firm. Since the plant remains profitable, it is expected that the firm will
keep this plant operating to maintain control of the source of supply for
its raw material. Consequently, no plant closure is expected.
The primary determinants of balance of trade impacts are relative prices
between domestic and foreign firms. Since no price changes are estimated to
result from the regulations, no balance of trade shifts are expected.
The recommended effluent guidelines and associated technologies for new
sources are identical to those of existing sources. Consequently, the economic
impacts for new sources will mirror those of existing sources and the proposed
regulations are not expected to foster competitive advantages or disadvantages
between new and existing sources.
Compliance costs for the non-selected options are higher, but not enough
to cause any plant closures.
5 .2 Semiconductors
The estimated economic impacts for the Semiconductors subcategory are
summarized in Table S-5. Fifty-two direct dischargers and 180 indirect
dischargers will incur compliance costs as a result of these regulations.
Another 25 direct dischargers were found by EPA to be in compliance with the
S-14
-------
TABLE S-4. SUMMARY OF ESTIMATED ECONOMIC IMPACTS
FOR THE ELECTRONIC CRYSTALS SUBCATEGORY
ECONOMIC IMPACT VARIABLES
Gallium/Indium Arsenide Crystals
Annual Compliance Costs/Revenues I
Indirect Dischargers
Direct Dischargers
Change in Price (%)
Change in Quantity (%)
Change in Profitability
Indirect Dischargers
Direct Dischargers
Capital Requirements
Indirect Dischargers
Direct Dischargers
Plant Closures due to Regulations
Employment at Closed Plants
Balance of Trade Changes
Industry Structure Changes
Non-Arsenide Crystals
Annual Compliance Costs/Revenues (%)
Indirect Dischargers
Direct Dischargers
Change in Price (%)
Change in Quantity (%)
Change in Profitability
Indirect Dischargers
Direct Dischargers
Capital Requirements
Indirect Dischargers
Direct Dischargers
Plant Closures due to Regulations
Employment at Closed Plants
Balance of Trade Changes
Industry Structure Changes
OPTION 1
0
0
0
0
0
0
0
0
0
0
0
None
None
0
0
0
0
0
0
0
0
0
0
0
None
None
OPTION 2 OPTION 3
0-3.4
0
0
0
Moderate
0
Moderate
0
0
0
None
None
0.6-5.6
0
0
0
Signif .
0
Moderate
0
0
0
None
None
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
0-2.8
0
0
0
Moderate
0
Moderate
0
0
0
None
None
SELECTED
OPTION 5 OPTION
0-3.7
0
0
0
Moderate
0
Moderate
0
0
0
None
None
0.6-5.6
0
0
0
Signif.
0
Moderate
0
0
0
None
None
0-3.4
0
0
0
Moderate
0
Moderate
0
0
0
None
None
0
0
0
0
0
0
0
0
0
0
None
None
S-15
-------
TABLE S-5. SUMMARY OF ESTIMATED ECONOMIC IMPACTS
FOR THE SEMICONDUCTORS SUBCATEGORY
ECONOMIC IMPACT VARIABLES
Annual Compliance Costs/Revenues (
Indirect dischargers
Direct Dischargers
Change in Price (%)
Change in Quantity (%)
Change in Profitability
Indirect Dischargers
Direct Dischargers
Capital Requirements
Indirect Dischargers
Direct Dischargers
Plant Closures due to Regulations
Indirect Dischargers
Direct Dischargers
Employment at Closed Plants
Indirect Dischargers
Direct Dischargers
Balance of Trade Changes
Industry Structure Changes
OPTION
;)
0
0
0
0
0
0
0
0
0
0
0
0
None
None
SELECTED
1 OPTION 2 OPTION 3 OPTION 5 OPTION
0.1-11.0
0.1-11.0
0
0
Signif .
Signif .
Signif.
Signif.
34
14
1,666
686
None
Signif.
0.17
0.17
0
0
Low
Low
Low
Low
0
0
0
0
None
None
0.1-12.0
0.1-12.0
0
0
Signif.
Signif.
Signif.
Signif.
34
14
1,666
686
None
Signif.
0
0.17
0
0
0
Low
0
Low
0
0
0
0
None
None
S-16
-------
selected BAT option (i.e., option 3). Annual compliance costs as a percentage
of revenues are not significant for the selected option. These costs will
result in reduction in the profitabilities of these plants; however, these
profit declines are not great enough to cause any plant closures.
The primary determinant of balance of trade impacts are relative prices
between domestic and foreign firms. Since no price changes are estimated to
result from the regulations, no balance of trade shifts are expected.
The recommended effluent guidelines and associated technologies for new
sources are identical to those of existing sources. Consequently, the incre-
mental costs or new source performance standards (NSPS) and pretreatment
standards for new sources (PSNS) are zero. For this reason, the economic
impacts of proposed regulations for new sources will mirror those of existing
sources and are not expected to foster competitive advantages or disadvantages
between new and existing sources. Moreover, since the impacts on industry
profitability are small, no significant change in industry growth rates is
expected to result from the regulations.
The estimated pollution control costs associated with options 2 and 5
(these options are not being proposed) indicate that the annual profit margin
reductions combined with the capital requirements will cause 14 direct dis-
chargers and 34 indirect dischargers to close. These 48 plants are 19 percent
of the semiconductor manufacturing plants identified. Employment at these 48
plants is estimated at 2,352.
6. SMALL BUSINESS IMPACTS
The Regulatory Flexibility Act (Public Law 96-354) requires EPA to deter-
mine if a significant impact on a substantial number of small entities occurs
as a result of the proposed regulations. This analysis may be done in conjunc-
tion with or as a part of any other analysis conducted by the Agency. The
economic impact analysis contains this analysis which indicates that for the
S-17
-------
selected options, some reduction in the profitability of small plants occurs
in both subcategories. However, these declines in financial performance
will not be enough to cause any plant closures or job losses for any segment
of the regulated population, large or small.
For the Semiconductor industry the pollution control treatment options
2 and 5 would cause 48 small semiconductor plants, or 19 percent of the
semiconductor plants identified, to close, and approximately 2,352 people
may become unemployed. Because these options were not selected and because
the selected options do not have a significant impact on small plants, a
formal regulatory flexibility analysis is not required.
7. LIMITATIONS TO THE ANALYSIS
7.1 ELECTRONIC CRYSTALS
The major limitations relate to the data used in the analysis. Limited
economic data was available on a sample of plants in the industry. However,
this data was insufficient to develop a formal discounted cash flow analysis
to assess plant closure potential. Instead, comparative analysis of financial
ratios such as return on investment was used to assess plant closures. Analy-
ses based on this approach do not consider the timing of cash flows and are
based on accounting income rather than cash flows. Nevertheless, return on
investment analysis is widely used in the financial analysis community and
can be expected to provide consistent results for industry-wide assessments.
A number of assumptions were necessary to implement the analysis and to
overcome the data limitations. These are:
• There will be no price increase resulting from the
regulations
• Financial performance in the survey year (1978) is
fairly typical of what might be expected in the year
the regulations become effective (1984)
S-18
-------
• Baseline data was based on sample sizes which varied
in industry representation from one variable to
another
• The yield on U.S. treasury bonds is a threshold value
for plant closures.
To assess the potential effects of violations of these assumptions or
of errors introduced because of the aforementioned methodology limitations
a sensitivity analysis of plus and minus 20 percent is performed on a number
of key variables, such as compliance costs and baseline profit rates. It
is concluded that there would be no plant closures under these conditions.
Despite the above limitations, the economic impact analysis conclusions
appear to be valid within a rather liberal range of potential error.
7.2 SEMICONDUCTORS
As noted above for the electronic crystals industry, the major limita-
tions relate t.o the data used in the study. A survey to collect plant-
specific financial and economic data was not conducted, and these types of
data are not in the public domain for this industry. Instead, most of the
plant-specific data used are estimates derived from industry average operating
and financial ratios. The plant-specific data must, therefore, be considered
order-of-magnitude estimates, and not actual data for specific semiconductor
plants. As a result of the lack of adequate data the methodology to assess
plant closure is limited to comparisons of simple operating ratios. A more
appropriate closure analysis would be based on more formal investment analy-
sis techniques, which rely on precise data on the long-term capital structure,
cash flow, and the profitability of plants in the semiconductor industry.
A sensitivity analysis was performed to examine the sensitivity of
the conclusions to some of the parameters estimated for the plant closure
analysis. The primary variables considered were compliance cost estimates,
baseline profit rates (as measured by before tax return on sales) and plant
S-19
-------
closure threshold criteria. A 20 percent variation in any of these variables
would not significantly change the results for the selected option. However,
-the number of plant closures due to the regulation under options 2 and 5
would increase by 50 percent, to 72 plants.
For these reasons, it is concluded that although the above assumptions
and limitations may bias the analysis somewhat, the potential changes to the
conclusions resulting from elimination of these biases are small.
S-20
-------
PART I
ELECTRONIC CRYSTALS SUBCATEGORY
-------
1. INTRODUCTION
1.1 PURPOSE
The purpose of this report is to identify and analyze the economic
impacts that are likely to result from the promulgation of regulations for
Best Practicable Control Technology (BPT), Best Conventional Pollutant Control
Technology (BCT), Best Available Technology Economically Achievable (BAT), New
Source Performance Standards (NSPS), and Pretreatment Standards for New and
Existing Sources (PSNS and PSES) on the Electronic Crystals subcategory of the
Electrical and Electronic Components Point Source Category. The study focuses
primarily on the impacts of the proposed EPA regulations on industry profit-
ability, ability to raise capital to finance the-proposed treatment equipment,
employment, plant closures, imports and exports substitution effects, and
social costs. In addition, impacts on small business entities are also
analyzed.
1.2 SCOPE
The Electronic Crystals subcategory as defined in this study includes
establishments that grow and fabricate (i.e., grind, slice, lap, and polish)
piezoelectric and semiconducting single crystals from polycrystalline raw
materials, and establishments that fabricate liquid crystals. For the
purposes of this study, the manufacturing of polycrystalline materials and
finished electronic components are not included. The former is considered to
be part of the Inorganic Chemical point source category and the latter as part
of other subcategories of the Electrical and Electronic Components point
source category such as Semiconductors and Ferrite Electronic Products.
Piezoelectric crystals are crystals that have the ability to generate a
voltage when a mechanical force is applied, or produce a mechanical force when
Polycrystalline is a body of crystalline material with many small, distinct
crystalline regions of different orientations; as opposed to single crystal
which has a single, periodic, ordered arrangement of atoms throughout.
1-1
-------
a voltage is applied. The major types of piezoelectric crystals are quartz,
ceramics, yttrium-iron-garnet (YIG), and some other less common types.
2 /
Semiconducting crystals are crystals whose electrical resistivity is
intermediate between that of metals (conductors) and insulators (noncon-
ductors) . The major types of semiconducting crystals are silicon, gallium
arsenide, gallium phosphide, indium arsenide, indium antimonide, bismuth
telluride, sapphire, and gallium gadolinium garnet (GGG).
Liquid crystals are organic compounds that exhibit properties of fluidity
and molecular order simultaneously over a small temperature range. An
electric field can disrupt the orderly arrangement of liquid crystal molecules
and darken the liquid enough to form visible characters in a display assembly.
1.3 TECHNICAL AND ECONOMIC SUBCATEGORIZATION
For the purpose of setting effluent standards, EPA proposed the following
two product groups for the Electronic Crystals subcategory:
• Plants that fabricate gallium arsenide and/or indium
arsenide crystals
• Plants that fabricate all other electronic crystals
("non-arsenide" plants).
EPA identified only eight plants in operation in the U.S. which produce
gallium arsenide and indium arsenide crystals. Analyses were performed on
each of these plants to assess the potential impacts of the proposed
regulations. EPA estimated that 62 plants produce non-arsenide crystals. The
assessment of the economic impacts of the proposed regulations for these
plants was based on model plants which were developed to represent the
62-plant population. The representative model plants were developed from EPA
survey data and were based on plant sizes and wastewater flow rates.
21
Resistivity is a property of an electrically conducting material- which
expresses the relative ease (low resistivity) or difficulty (high
resistivity) of passage of an electric current.
1-2
-------
1.4 ORGANIZATION OF PART I OF THIS REPORT
The remainder of Part I of this report is organized into seven sections.
Section 2 outlines the study methodologies and Section 3 presents a
description of the industry. Section 4 projects the baseline conditions of
the industry assuming no further water pollution control requirements.
Section 5 details the costs of the alternative treatment technologies being
considered. Section 6 presents the economic impacts of pollution control
costs. Section 7 addresses the impacts of the proposed regulations on small
businesses. Finally, Section 8 discusses the assumptions and limitations of
this study.
1-3
-------
2. STUDY METHODOLOGY
2.1 OVERVIEW
Figure 2-1 shows an overview of the analytical approach used to assess
the economic impacts likely to occur as a result of the costs of each proposed
regulatory option. The approach used in this study is to (1) develop an
operational description of the price and output behavior of the industry, and
(2) assess the likely plant-specific responses to the incurrence of the
compliance costs enumerated in Chapter 5.
The operational description of the price and output behavior is used, in
conjunction with compliance cost estimates supplied by EPA, to determine new
post-compliance industry price and production levels for each regulatory
option and for each of the two electronic crystal product groups. Individual
plant data is then analyzed under conditions of the post-compliance industry
price levels, for each regulatory opcion, to isolate those plants whose
production costs would appear to change significantly more than the estimated
change in their revenues. These identified plants are subjected to a
financial analysis that uses capital budgeting techniques to determine likely
plant closures. The industry description is then revised, for each regulatory
option, to incorporate the reduced supply into the analysis. Finally, other
effects which flow from the basic price, production, and industry structure
changes are determined. These include employment, community, and foreign
trade impacts. Specifically, the study proceeded in the following ten steps:
1. Description of industry characteristics
2. Industry supply and demand analysis
3. Analysis of cost of compliance estimates
4. Plant level screening analysis
5. Plant level profitability analysis
6. Plant level capital requirements analysis
7. Assessment of plant closure potential
8. Assessment of other impacts
9. Social costs analysis
10. Small business analysis.
1-4
-------
o <
CO
. UJ
1
V)
i_> o o z
— — < LJ
u. = 0. r
— EO
(_ u_ — LU
Z O CO
UJ
a
Z CO
O uJ I—
— Q 'O
LJ < _LJ
c x-
Z ^
> o
a: — co
H- i- _i
CO O UJ
H ^2 UJ
Z O -J
— ce
a.
z co
O t-
— CO
1- O
3 !->
O O
Q- C£
f. Z
Q. O
- 5
> ^ y) uj«
a: 5? —• cc —
H- ^ 00 O CO
t/> g > 2: >-
i S< "<
Z g Z LJ Z
- g < 0<
IT o:
*• Q-
W
M
>
w
o
Q
H
CO
M
CO
J
z
1—I
*-i
o
c
u
I
cs
w
Oi
o
f—I
fe
I— <
co t—
= z
O LU
z s;
— o
LJ
CO
z £
1-5
-------
Although each of these steps is described separately in this section, it is
important to realize that there are significant interactions between them, as
shown in Figure 2-1.
2.2 STEP 1: DESCRIPTION OF INDUSTRY CHARACTERISTICS
The first step in the analysis is to develop a description of the basic
industry characteristics that would enable estimation of key parameters which
describe the initial impacts of the regulation. These characteristics, which
include the determinants of demand, market structure, the degree of intra-
industry competition, and financial performance, are described in Chapter 3 of
this report. The resulting observations indicate the type of analysis needed
for the industry.
The sources for this information include goverameul reports, trade
association data, discussions with various trade association representatives
and individuals associated with the industry, and an EPA industry survey.
2.3 STEP 2: SUP PLY--DEMAND ANALYSIS
The purpose of the supply-demand analysis, step 2 of the study approach,
is to determine the likely changes in market prices and industry production
levels resulting from each regulatory option. The estimates of post-
compliance price and output levels are used in the plant-level analysis to
determine post-compliance revenue and profit levels for specific plants in
each product group. If prices are successfully raised without significantly
reducing product demand and companies are able to maintain their current
financial status, the potential for plant closings will be minimal. If prices
cannot be raised to fully recover compliance costs because of the potential
for a significant decline in product demand or because of significant intra-
industry competition, the firms may attempt to maintain their financial status
by closing higher cost/less efficient plants. The supply-demand analysis was
divided into four basic components: determination of industry structure,
projection of possible changes in industry structure to 1984 (the expected
effective date for the proposed regulation), determination of plant- and firm-
specific operational parameters (e.g., production costs, profit rates, etc.),
and development of price-quantity algorithms.
1-6
-------
Short-run pricing behavior depends upon the market structure of the
industry, which can range from competitive to oligopoly, and to monopoly
situations. Many economic impact studies begin by assuming perfect
competition. However, the product groups covered in this study exhibit some
characteristics that are indicative of non-competitive pricing mechanisms.
The perfectly competitive market structure is one in which there are many
buy rs and sellers and the actions of any one of these do not significantly
affect the market. Firms in a competitive market generally earn a "normal"
rate of return on their assets. If it is assumed that (1) the market for a
competitive good is currently at equilibrium, or will be when the regulations
become effective, and (2) firms will attempt to maintain their current
financial status by passing through industry-wide cost increases in the form
of higher prices, the post-compliance equilibrium price and quantity level can
be derived from the interaction of the elasticities of supply and demand.
That is, the amount of the cost increase that will be passed through into
higher prices is
(1) E
s
where E is the elasticity of supply and E is the elasticity of demand.
The high concentration ratios for some of the product groups and the
existence of speciality markets could cause non-competitive pricing behavior.
For example, if foreign competition were not a pro-olem, full cost pricing
might be assumed to characterize the behavior of oligopolistic firms. The
price could be assumed to cover average total cost (the sum of variable cost,
average fixed cost, and return on investment) of the "price leaders" in the
industry. Thus, one or a few firms may have the ability, because of their
-In a competitive market price = marginal cost, therefore
E = . = S . , where me = marginal cost, P = market price, and
dP Qs dmc Qs Qs = quantity supplied.
See Levenson, Albert M., and Solon, 3.S., Outline of Price Theory, Holt,
Rinehart and Winston, Inc., pp 56-59, 1964.
1-7
-------
market power, to impose their most desired price and output strategy on the
other firms in the market.
Oligopolistic pricing schemes are applicable for those product categories
which exhibit characteristics of oligopoly markets, such as the following:
• Few firms in the product group
• High industry concentration
» Low degree of foreign competition (less than 10 percent of U.S. sales)
• Abnormally high profitability
« Low demand elasticities
• Highly capital intensive
« Large degree of integration of production, marketing, and distribution
* Large degree of specialized knowledge.
Industries which exhibit the first three of these characteristics are
those in which the pricing and output actions of one firm will directly affect
those of other firms in the industry. While these conditions do not guarantee
oligopolistic behavior, they are necessary conditions for an olligopoly and
good indicators that one exists. Abnormally high profits in an industry
would, in time, normally attract new entrants to the industry, thereby
increasing price competition. However, very high profits over long periods of
time which are not explained by such factors as excess risk, unusual amounts
of technological innovation, or firm size may be an indicator that an
imperfect market structure exists. Such conditions may occur when entry into
an industry is difficult. The last three of the above points are indicators
of difficulty of entry into the market.
Although the domestic electronic crystals industry exhibits some of these
characteristics of non-competitive markets, the existence of the threat of
foreign competition changes their pricing behavior. That is, if the differen-
tial between foreign and domestic prices is in equilibrium, an increase in the
domestic price would result in lost market share. Since the domestic producer
1-8
-------
profits are above "normal" profit rates, they have the capability of absorbing
the cost increases without increasing product prices and would probably do so.
Having described the current industry structure, it was necessary to
determine if the key parameters would change significantly by 1984.
Projections of industry conditions begin with a demand forecast. The demand
in 1984 is estimated via trend analysis and market research analysis. An
examination was also made of the factors which might affect the real cost of
manufacturing electronic crystals. No reason was found to expect the real
price of these products to increase between now and 1984. It was concluded
from the projections of industry conditions that only minor changes in market
structure would occur in the base case. For this reason it was concluded that
the market structure previously described can be used to determine price
changes due to the regulation.
The post-compliance market price. Ir.Vels (i.e., zero price increases due
to the regulation) are used, in a later step, to assess the financial
condition of individual electronic crystals manufacturing facilities.
2.4 STEP 3: COST OF COMPLIANCE ESTIMATES
Investment and annual compliance costs for the recommended treatment
options 1, 2, 3, and 5 were estimated by EPA's Effluent Guidelines Division
for treatment systems of various selected sizes. Based on these cost
estimates, compliance cost curves were developed and then used to estimate
plant-specific compliance costs. A description of the control and treatment
technologies and the rationale behind these compliance cost estimates appear
in chapter 5.
2.5 STEP 4: PLANT-LEVEL SCREENING ANALYSIS
The screening analysis uses a basic criterion to separate those plants
with obviously small impacts from those with potentially significant impacts.
The criterion uses the ratio of total annual compliance cost to revenue for
each plant; if the ratio is less than one percent, that plant is considered
"low impact." For such plants, the impacts on plant profitabilities are very
1-9
-------
small. Although some of these plants will probably experience some drop in
their profitabilities, they would not be considered candidates for closure.
Estimates of profitability changes for these plants were recorded, but they
are not subject to a detailed financial analysis. Plants with compliance
costs to revenue ratios greater than 1 percent were subjected to a more
detailed financial analysis to determine if they are likely plant closures.
2.6 STEP 5: PLANT-LEVEL PROFITABILITY ANALYSIS
The basic measure of financial performance used to assess the impact of
the proposed regulations on the profitabilities of individual plants is return
on investment (ROI). The use of this technique involved a comparison of the
measure with a critical return.
The return on investment is defined as the ratio of annual profits before
taxes to the gross book value of fixed assets of a plant. The principal
virtue of this technique is its simplicity and its common use in comparative
analyses of the profitability of financial entities. Its principal short-
comings are that it is based on accounting income rather than cash flows and
that it fails to account for the timing of cash flows, thereby ignoring the
time value of money.
The profit impact assessment is determined by calculating the after-
compliance ROI for each plant and comparing them to the average yield of U.S.
Treasury bonds. As discussed in Section 2.3, it is expected that the
electronic crystals producers will absorb the costs of compliance and prices
will not be raised. As the result, the after-compliance KOI can be estimated
as follows:
(2)
A. + CGI .
i i
PROFIT^ = Ru x PMU (3)
where ROI,.. = After compliance ROI of plant i
PROFIT . = Pre-compliance profit of plant i
ACC, = Annual compliance cost for plant i
A. = Pre-compliance capital investment for plant i
1-10
-------
CCI. = Compliance capital investment for plant i
Rl. = Pre—compliance revenue of plant i
= Pre-compliance profit margin of plant i
PM
li
Plants with after-compliance ROI below 11.5 percent, the average yield of
U.S. Treasury bonds for 1980, are considered potential plant closures. The
underlying assumption is that plants cannot continue to operate as viable
concerns if they are unable to generate a return on investment that is at
least equal to the opportunity cost of other lower risk investment alter-
natives, which in this case is defined as the average yield of the U.S.
Treasury bonds.
Most of the data used in this analysis were estimated from a combination
of publicly available information and the EPA 308 Survey. The following are
the most important variables:
• Baseline financial profiles (i.e., revenues, assets value, profit
margin) of individual plants/model plants are obtained from the EPA
308 Survey
• Pollution control costs are estimated as described in Step 3.
2.7 STEP 6: CAPITAL REQUIREMENTS ANALYSIS
This section analyzes the ability of firms to make the initial capital
investment needed to construct and install the required treatment systems.
Some plants which are not initially identified as potential closures in the
profitability analysis may encounter problems raising the amount of capital
required to install the necessary treatment equipment. The limit on a given
firm's ability to raise capital to finance investment expenditures is quite
variable, depending upon factors such as the firm's capital structure, profit-
ability, future business prospects, the industry's business climate, the
characteristics of the financial markets and the aggregate economy, and the
firm management's relationships with the financial community. The precise
limit, considering all these factors, is ultimately judgmental. Even with
firm-specific data, a limit on a firm's ability (or willingness) to raise
1-11
-------
funds for capital investment would be difficult to estimate. Because firm-
specific data for this study is scarce, the analysis of capital availability
was based on the following ratio which provide? a good indication of the
relative magnitude of the capital requirements for pollution control:
CCI
where
K = plant revenues
CCI = capital compliance investment
The CCI/R ratio was calculated for each sample plant.
Plants with aft-°r-compliance ratios greater than the difference between
the highest and lowest ratios of annual capital expenditures to value of
shipments (CE/VS) between 1970 and 1977 for SIC 3679 (Electronic Components,
n.e.c.) as reported by the Census of Manufactures are considered to be
potential plant closures. The difference between the highest and lowest CE/VS
ratios is considered to be an indicator of the maximum amount of capital
available for pollution control investment. That is, the highest CE/VS ratio
of 5.1 percent represents the maximum capital expenditure rate that has
historically been made available for capital investment. It is assumed that
CE/VS ratios below this amount can be successfully attained again by the firms
in the industry, since it has been achieved before. The lowest CE/VS ratio of
2.8 percent represents the minimum amount necessary to maintain a targeted
rate of industry growth. If the ratio CCI/R for a given plant is greater than
the difference between the industry's highest and lowest CE/VS ratios,
2.3 percent, then the plant could experience some difficulties in making the
capital investment for pollution control. Although this analytical technique
does not precisely indicate whether or not a given firm will be able to raise
the" necessary funds, it does provide an indication of the relative magnitude
of the compliance investment.
1-12
-------
Plants with CCI/R ratios greater than the difference between the highest
and lowest CE/VS ratios were then subjected to the following cash flow test.
Cash flow is defined as net after tax profits plus depreciation charges.
Depreciation is assumed to be 10 percent of fixed assets and the tax rate is
assumed to be 40 percent. If the CCI/R ratio is less than the cash flow, the
investment may be financed out of a single year's cash flow without additional
long-term debt burden.
Although the above analyses provide a good indication of the relative
burden created by the compliance requirement, they do not precisely indicate
whether or not firms can afford to make the investments. If, for example, the
same investment requirements were placed on a firm which is already highly
leveraged (as indicated by a high debt/equity ratio) and a firm which is not
leveraged (as indicated by a debt/equity ratio of zero), the highly leveraged
firm is likely to experience the most significant impact. In addition, the
capital requirements must be evaluated together with other factors, such as
profitability. For example, a plant that is extremely profitable would
consider the risk of more leverage or increased cost of capital resulting from
investment more worthwhile than would a less profitable plant.
2.8 STEP 7: PLANT CLOSURE ANALYSIS
The plant-level analysis examined the individual production units in each
product group to determine the potential for plant closures and profitability
changes. The decision to close a plant, like most major investment decisions,
is ultimately judgmental. This is because the decision involves a wide
variety of considerations, many of which cannot be quantified or even identi-
fied. Some of the most important factors are;
0 Profitability before and after compliance
• Ability to raise capital
'_ 0 Market and technological integration
0 Market growth rate
0 • Other pending Federal, state, and local regulations
0 Ease of entry into market
1-13
-------
• Market share
« Foreign competition
* Substitutability of the product
» Existence of specialty markets.
Many of these factors are highly uncertain, even for the owners of the
plants. However, this analysis was structured to make quantitative estimates
of the first two factors, as described above, and to qualitatively consider
the importance of the others. In this analysis, the first two f.actors are
given the greatest amount of weight, while the importance of the other factors
varies from plant to plant.
2.9 STEP 8: OTHER IMPACTS
This analysis addresses economic impacts '-'hich flow from the basic price,
production, and plant level profitability changes. These impacts include
impacts on employment, communities, industry structure, and balance of trade.
The estimate of employment effects flows directly from the outputs of the
industry level analysis and the plant closure analysis. Employment estimates
for the production facilities proiected to close are available from the EPA
308 Survey. Meanwhile, employment in the remaining plants is not expected to
decrease because there will be no price increase as the result of the
regulations.—-
Community impacts result primarily from employment impacts. The critical
variable is the ratio of Electronic Crystals industry unemployment to total
employment in the community. Data on community employment are available
through the Bureau of the Census and the Bureau of Labor Statistics. Some-
times county-wide data will have to be relied upon in the absence of
community-specific data, although this was not necessary in this study.
21
— Some .increase in employment might occur at the remaining facilities as
substitutes for the eliminated capacity from the closed plants. .
Quantitative evaluation of this effect was not developed in this study.
1-14
-------
The assessment of industry structure changes is based on examination of
the following before and after compliance with the regulation:
• Numbers of firms and plants
* Industry concentration ratios
• Variance of average total production cost per unit among plants
• Effects of plant closures on specialty markets.
Decreases in the first two factors would indicate an increase in industry
concentration and may change the pricing behavior of the industry. Such
potential changes were qualitatively evaluated. An increase in variability of
average costs would indicate that some firms have become more competitive than
others, as a result of the regulation. The long-term implications of such
developments were examined.
Imports and exports are important determinants of pricing behavior in the
Electronic Crystals industry. The role of these variables is qualitatively
evaluated in Chapter 3 of this report. Basically, the threat of imports
appears to deter domestic producers from excessive non-competitive pricing
behavior.
2.10 STEP 9: SOCIAL COST ANALYSIS
This analysis assesses the total social costs that can be associated with
the EPA effluent regulations. The social costs measure the value of goods and
services lost by society due to a given regulatory action. These costs
generally include the use of resources needed to comply with a regulation, the
use of resources to implement and enforce a regulation, plus the value of the
output that is forgone because of a regulation.
The partial equilibrium analytical framework is conceptually the most
practical means for estimating total social costs. This framework, in its
most sophisticated form, is based on an analysis of supply and demand rela-
tionships in the directly-affected markets. When an industry is regulated,
compliance requirements result in increased unit costs of production. This
1-15
-------
increase, in turn, leads to an upward shift in the industry's supply curve.
The supply curve shift normally results in higher prices and a lower produc-
tion level. Compliance costs, production losses, and net welfare losses
incurred by producers and consumers due to decreased output are measurable
within this framework. There are other costs that are not measurable within
this framework. Costs of implementing and enforcing a regulation must be
added. Also, other social costs do not appear in this static analysis such as
productivity effects, innovation impacts, and costs of reallocating resources
that become unemployed. Unfortunately, the data do not exist to carry out
such analysis at this time, and a compromise which captures the major costs to
society was performed.
For this analysis only the real resource costs are considered. This
provides a reasonable estimate of social costs, since EPA studies show than 95
percent or more of the social costs are directly related to compliance expen-
ditures by the regulated entities. Consequently, the present value of social
costs (PVSC) of regulations can be approximated by the following equation:
pvsc = i/d + .Dn + (OM/.D/(i + .Dn
where: PVSC = present value of social costs
I = investment cost
OM = annual operating and maintenance cost
n = number of years between now and year of investment
The above equation assumes that:
» The regulations will be in effect in perpetuity
• Operating and maintenance costs will be incurred in the first year of
investment
e The real discount rate is 10 percent.
2.11 STEP 10: SMALL BUSINESS ANALYSIS
- The Regulatory Flexibility Act (RFA) of 1980, (P.L. 96-354) which amends
the Administrative Procedures Act requires Federal regulatory agencies to
consider "small entities" throughout the regulatory process. The RFA requires
an initial screening analysis to be performed to determine if a substantial
1-16
-------
number of small entities will be significantly impacted. If so, regulatory
alternatives that eliminate or mitigate the impacts must be considered. This
step in the study addresses these objectives by identifying the economic
impacts which are likely to result from the promulgation of BPT, BCT, BATEA,
NSPS, PSES, and PSNS regulations on small businesses in the electronic
crystals manufacturing industry. The primary economic variables covered are
those analyzed in the general economic impact analysis such as plant financial
performance, plant closures, and unemployment and community impacts. Most of
the information and analytical techniques in the small business analysis are
drawn from the general economic impact analysis which is described above and
in the remainder of this report. The specific conditions of small firms are
evaluated against the background of general condition in the electronic
crystal markets.
A specific problem in the methodology was developing an acceptable
definition of small entities. 'The Small Business Administration (SBA) defines
small entities in SIC 3674 (Semiconductors and Related Devices and SIC 3679
(Electronic Components, n.e.c.) as firms of fewer than 500 employees. The SBA
definition was found to be inappropriate as a basis for defining small
entities in the electronic crystal manufacturing industry for purposes of
developing water pollution regulations. Instead, a definition was sought
which would account for firm size in comparison to total industry size and in
comparison to unit compliance costs (unit compliance costs increase signif-
icantly in reverse proportion to plant size). Moreover, since the available
data on compliance cost and production was on an individual plant basis, the
individual production facility rather than the firm was used as the basis for
the analysis.
In the absence of an appropriate definition for a small business in the
electronic crystals manufacturing industry, four size definitions for elec-
tronic crystal manufacturing plants based on plant revenues were selected for
examination. Plant revenues were selected to distinguish plant size because
plant level employment data were considered less reliable and plant production
volume 'data did not allow consistent comparisons across the product groups.
1-17
-------
The four size categories are plants with revenues less than $1 million,
$1-3 million, $3-5 million, and greater than $5 million. The use of several
different size definitions provides EPA with alternatives in defining small
electronic crystal plants.
The impacts on small plants under each definition were assessed by
examining the distribution by plant size of the number of electronic crystal
plants, plant revenues, wastewater volumes, compliance costs and potential
closures from the regulations.
1-18
-------
3. INDUSTRY DESCRIPTION
This section describes the manufacturing processes and the plant and firm
characteristics of the electronic crystals industry. In addition, the
characteristics and end uses of crystals and the extent of foreign competition
are also examined.
3.1 MANUFACTURING PROCESSES
As indicated earlier, the scope of this study includes the plants that
grow and fabricate (i.e., grind, slice, L?p, and polish) piezoelectric,
semiconducting and liquid crystals. The manufacturing processes for these
electronic crystals are generally very similar. First, single crystals are
grown from polycrystallin.2 raw materials. The most common technique for
growing single crystals is the Czochralski method in which the raw material is
melted; a single crystal bar (seed) is then dipped into the melt and raised
slowly with rotation to allow the molten material to freeze onto the seed to
form a single crystal rod. Other methods based on some variations of the
Czochralski method are also used for growing electronic crystals and are
described in the Draft Development Document.
After the single crystal rod has been grown, it must be fabricated into a
suitable form, which for most applications is a wafer. Generally, the crystal
rod is ground to a preferred shape. It is then sliced into thin pieces using
diamond blade saws or slurry saws. Once sliced, the crystal wafers are then
lapped to obtain a desired thickness and a smoother finish. Finally, the
slices are polished to a mirror finish, and are then ready for further
processing to produce various electronic components.
3.2 FIRM AND PLANT CHARACTERISTICS
According to EPA estimates, there are approximately 70 producers of elec-
2/
tro'hic crystals in the United States , with total employment of the industry
EPA, Draft Development Document for Effluent Limitation Guidelines
for t?ie Electrical and Electronic Components Point Sourc_e_
Category, April 1982.
2/Ibid.
1-19
-------
estimated at around 10,000 employees. Table 3-1 lists the number of plants by
type of crystals. As indicated in the table, with the exception of quartz
crystal the production of electronic crystals is highly concentrated with a
small number of firms accounting for most of the production of each type of
crystal.
The geographical distribution of the plants is shown in Table 3-2. Four
states account for 57 percent of the number of plants identified (California,
Ohio, Pennsylvania, and Texas, with 26, 12, 12, and 7 percent of the number of
plants, respectively), with the remainder widely dispersed throughout the
country.
Based on the survey data obtained on 52 plants, about 60 percent of the
plants are small (i.e., plants with less than ?3 million in product shipments
or with less than 100 employees). Table 3-3 presents a distribution of the
electronic crystals plants by plant size. Most of the plants appear to be
highly specialized in the fabrication of electronic crystals. Only 12 plants
report integration to the manufacture of electronic devices at the same pro-
duction facilities. The companies that own the crystal plants, on the other
hand, are generally more diversified and often produce other electronic
devices in addition to crystals, usually at other production facilities.
3.3 PRODUCT CHARACTERISTICS AND USES
As indicated in Section 1, the Electronic Crystals subcategory as defined
in this study includes three types of crystals:
® Piezoelectric crystals which include quartz, ceramic,
yittriura-iron-garnet, and lithium niobite crystals
* Semiconducting crystals, which include silicon, gallium
arsenide, gallium phosphide, indium arsenide, indium anti-
monide, bismuth telluride, and sapphire
a Liquid crystals.
1-20
-------
TABLE 3-1. NUMBER OF ELECTRONIC CRYSTAL PLANTS IN THE UNITED STATES
ESTIMATED
TOTAL NUMBER
OF PLANTS
TYPES OF CRYSTALS
Piezoelectric Crystals
Quartz 40
Ceramics 8
Yittrium Iron 3
Garnet (YIG)
YAG 2
Lithium Niobate 3
Semiconducting Cry_st.a1 s
Silicon 8
Gallium Arsenide/Photphide 6
Indium Arsenide/ ' 1
Indium Antimonide/
Bismuth Telluride
Gallium Gadolinium 3
Garnet (GGG)
Sapphire 1
Liquid Crystal 2
TOTAL 70
a/
a/b/
NUMBER OF PLANTS
WITH RETURNED
QUESTIONNAIRES
27
8
3
2
3
3
1
2
57
a/
b/
Columns do not add up because some plants make more than one type of
product.
Projected by JRB assuming that non-respondent plants are single product
plants.
Source - EPA, Draft Development Document, April 1982
1-21
-------
TABLE 3-2. GEOGRAPHICAL DISTRIBUTION OF ELECTRONIC CRYSTAL PLANTS
STATE NUMBER OF PLANTS
Source: EPA 308 Survey
Arizona 1
California 15
Colorado 1
Connecticut 2
Florida 1
Illinois 1
Kansas 1
Massachusetts 1
Missouri 1
Montana 1
New Jersey 3
New York 2
North Carolina 1
Ohio 7
Oklahoma 2
Oregon 1
Pennsylvania 7
South Dakota 1
Texas 4
Utah ' 2
Not Reported _2_
TOTAL 57
1-22
-------
TABLE 3-3. DISTRIBUTION OF PLANTS BY PLANT SIZE
Plant Size3
Gallium Arsenide/Indium
Arsenide Crystals
Extra Small
Small
Medium
Large
Number of
Sample, ,
Plants
2
3
-
3
Percent of
Total
25
38
-
38
Estimated Number of
Plants for
Total Industry
2
3
-
3
100
Other Electronic Crystals^
Extra Small
Small
Medium
Large
10
16
10
44
23
36
23
18
100
14
23
14
62
Extra Small Plant = Plants with less than $1 million in value of shipments
or with less than 20 employees.
Small Plant
Medium Plant
Large Plant
b/
- Plant with $1 million to $3 million in value of
shipments cr with 20 to 100 employees .
= Plant with $3 million to $5 million in value of
shipments or with 100 to 250 employees.
= Plant with over $5 million in value of shipments or with
more than 250 employees.
Data on plant size are available only for 52 of 57 plants with survey
responses .
Source: EPA Survev
1-23
-------
Quartz and silicon crystals are by far the most widely used crystals. The
principal end-uses of various electronic crystals are summarized in Table 3-4,
The potential for substitution of electronic crystals in their current
applications is limited. For instance, silicon's exceptional physical and
chemical properties make this crystal the best and most important material for
semiconductor applications. Similarly, the special characteristics of the
other crystals make them hard to be replaced in most of their applications.
Potential substitution among various crystals is also limited. Except
for gallium arsenide, gallium phosphide, and liquid crystals, which are all
used in the manufacturing of display devices, each electronic crystal has
quite distinctive applications for which it cannot be easily replaced (see
Table 3-4), Gallium arsenide and gallium phosphide can substitute for each
other in the fabrication of light emitting diodes (LEDs)—gallium arsenide for
red displays and gallium phosphide for green displays—however, each manufac-
turer can easily produce either or both materials. Therefore, this substi-
tution does not present any problem to the industry. Meanwhile, LEDs and
liquid crystal displays (LCDs) compete against each other; however, the
potential for substitution is limited due to the special characteristics of
each product. Because it requires minimum use of power, LCD has replaced LED
in many battery-operated devices such as electronic watches and calculators;
however, LCD is not as visible as LED, and thus its substitution for LED is
limited.
3.4 TRENDS
Data are not available on U.S. total production of fabricated (i.e.,
sliced, polished, lapped) electronic crystals. However, the U.S. Department
of Commerce reported that shipments of quartz and other crystal filters and
frequency control devices (which account for most of the demand for piezo-
electric crystals) totaled $148.5 million in 1978. It was estimated that the
value of the crystal accounted for about 75 percent of the value of the
finished devices, thus, U.S. shipments of piezoelectric crystals in 1978
were valued around $110 million. According to the U.S. Department of Commerce,
Interview with Jack Clifford of U.S. Department of Commerce.
1-24
-------
TABLE 3-4. MAJOR APPLICATIONS FOR ELECTRONIC CRYSTALS
TYPE OF CRYSTAL^
Piezoej.ectrie Crystals
® Quartz
» Ceramic
• Yittrium-Iron-Garnet
(YIG)
Semiconducting^jCrjystals
» Silicon
Gallium A.rsenide,
Gallium Phosphide
Indium Arsenide, Indiun
Antimonide
Bismuth Telluride
Sapphire
a Gal liura-Gadolinium-
Garnet (GGG)
MAJOR APPLICATIONS
Timing devices in watches; frequency
control, modulation, and demodulation in
oscillators ,ind filters
Used in transducers, oscillators, ultrasonic
cleaners, gas igniters, audible alarms,
keyboard switches, medical electronics
High-frequency control in microwave circuits
for electronic devices such as sonar
equipment.
Wafers for semiconductor devices such as
integrated circuits, diodes, rectifiers,
transistors and other circuit elements
Light emitting diodes (LEDs) used as display
devices in calculators, digital watches and
other electronic equipment; high-performance
transistors such as field effect transistors
(FET)
f power measuring devices
Devices to cool small components of
electrical circuits
Substrate for silicon-on-sapphire (SOS)
semiconductor devices; infrared detector
cell windows; ultraviolet windows and
optics; high power laser optics;
ultracentrifuge cell windows
Substrate for magnetic garnet films used in
the manufacture of magnetic bubble memories
Liquid Crystals
...-,, — ,_-—„-..-,. . . ___._.„ ._„ .1 I,
Source: EPA, Draft Development Document
Liquid crystal displays (LCDs) for
calculators, digital watches, etc.
1-25
-------
U.S. ship-dents o£ piezoelectric crystals had been growing slowly. The reason
for this modest growth was the inability of U.S. companies to compete with the
Japanese, in several f =15 ir-growi ng nsrVets such as those for electronic watch
and CB radio crystals (for which the average annual growth rates between 1973
4/
and 1978 were 26 and 24 percent respectively ). As a result, U.S. fines are
generally more specialized in the market for high performance crystals where
foreign competition is not as strong. Although the number of devices that use
these high-quality crystals has been increasing significantly, the demand for
the crystals has been growing slowly because technological advances have
reduced the number of crystals required in each device. For instance, the
introduction of phase lock loop system has reduced the number of crystals used
in communications equipment. U.S. Department of Commerce data show that
U.S. shipments of crystal filters and frequency control devices fluctuated
widely between 1973 and 1978, and averaged about 5 percent annual growth rate
during that period (see Table 3-5). However, the validity of the data
reported is very questionable, especially for earlier years.
It was estimated that the cost of a semiconductor wafer represented about
5 percent of the value of a semiconductor discrete device, and about
7 to 8 percent of the value of an integrated circuit. Based on the value of
shipments of semiconductor devices reported in Table 3-6, the U.S. shipments
of semiconducting and liquid crystals in 1979 were valued around $450 million.
Between 1969 ?nd 1979, the demand for semiconducting and liquid crystals had
been growing rapidly as indicated by' the 17 percent average annual growth rate
of U.S. shipments of semiconductor devices (see Table 3-6).
4/
U.S. Department of Commerce, U.S. Industrial Outlook 1980, p. 400.
Interview with Jack Clifford of U.S. Department of Commerce.
6/
Ibid.
U.S. Department of Commerce, A Report on the U.S. Semiconductor_Indust_ry_,
September, 1979. "~~'
1-26
-------
TABLE 3-5. U.S. SHIPMENTS OF PIEZOELECTRIC CRYSTAL
DEVICES, 1973 - 1978 ($ MILLIONS)
Year
1973
1974
1975
1976
1977
1978
Frequency
Control Devices
98.
90.
94.
143.
111.
107.
2
6
3
3
1
1
Filters
22.
29.
114.
76.
40.
41.
1
9
3
9
3
5
Es t irnat ed
Value of
Total Crystal
120
120
208
220
151
148
Q —
C _
.6
.2
.4
.6 110. 03/
AVERAGE ANNUAL
COMPOUNDED
GROWTH RATE
b/
4.3%
10. 1%
5 3%
a/ About 75 percent of value of finished devices (Interview with Jack
Clifford of U.S. Department of Commerce).
b/
Calculated using semi-log regression technique.
Source: U.S. Department of Commerce, Current Industrial Reports - Selected
Electronic and Associated Products, Including Telephone and Telegraph
Apparatus, selected issues.
1-27
-------
TABLE 3-6. U.S. SHIPMENTS OF SEMICONDUCTOR DEVICES
1969 - 1979 ($ MILLIONS)
Year
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
Discrete
Devices
1013
987
870
1093
1469
1519
1320
1594
1835
1880
1957
Integrated
Circuits
480
594
696
1268
1709
2056
1718
2598
2697
3950
4671
Total
1493
1581
1566
2361
3179
3575
3038
4292
4532
5830
6628
AVERAGE ANNUAL
COMPOUNDED
GROWTH RATE
a/
25%
17%
a/
Calculated using semi-log regression technique,
Sources: U.S. Department of Commerce
Semiconductor Industry Association
1-28
-------
3.5 FOREIGN TRADE
Data on U.S. imports and exports of electronic crystals are not available
directly. However, the available evidence indicates that the U.S. manufac-
turers of semiconducting crystals seem to be able to successfully compete with
their foreign producers. According to a recent Department of Commerce
8/
Report, many of the U.S. .semiconductor device manufacturers have transferred
their assembly operations overseas to take advantage of lower wage rates.
However, these same companies still maintain the production of semiconductor
wafers in the country. The primary reason for this is the capital intensive
nature of the growing and fabricating process for wafers. However, this does
not mean that there is no potential for competition from abroad. In fact, a
number of firms in Japan and Europe have the necessary technology and capacity
and, consequently, may serve to limit the ability of the domestic producers of
9 /
semiconductor wafers to pass through the costs of environmental regulations.
For the manufacture of piezoelectric crystals, foreign competition is not
expected to be too significant. As indicated earlier, the piezoelectric
crystals currently produced by U.S. manufacturers are generally high-perfor-
mance crystals and competition from foreign producers in this market does not
seem to be very important. In the past, U.S. firms had tried to penetrate the
booming market for quartz crystals for electronic watches and CB radios.
However, aggressive competition from the Japanese producers, who now dominate
the world market for these lower quality crystals, has driven most of these
U.S. companies out of the market and forced them to specialize in the low-
volume, high-quality crystal market instead.
87
U.S. Department of Commerce, A Report on the U.S. Semiconductor Industry,
September 1979.
9/ . . . -
Interview with Jack Clifford of U.S. Department of Commerce.
10
1-29
-------
4. BASELINE PROJECTIONS OF INDUSTRY CONDITIONS
This section provides projections of conditions in the U.S. electronic
crystals industry to 1990 in the absence of further water pollution control
requirements resulting from the Clean Water Act. The baseline projections in
this report provide a general point of reference for the analysis and are not
intended to be a comprehensive, authoritative forecast of future industry
conditions. These projections provide a plausible picture of future develop-
ments, and thus can be used as a benchmark for comparison. The primary vari-
ables of interest are the industry's projected sales, capital requirements,
and employment.
4.1 U.S. ELECTRONIC CRYSTAL SALES PROJECTIONS
As indicated in Table 3-5, shipments of piezoelectric crystals grew at a
modest 5 percent average annual rate between 1973 and 1978. The U.S. Bureau
of Mines projected that U.S. demand for lasca, a feedstock for growing
cultured (i.e., artificial) quartz, will increase at an annual rate of
approximately 3 percent through 1990. Since quartz crystal is the most common
piezoelectric crystal and cultured quartz represents 95 percent of all
electronic grade quartz, it is projected that U.S. annual production of piezo-
electric crystals will grow about 3 percent a year and will total $130 million
in 1983 and $160 million in 1990.— These projections are shown in Table 4-1.
U.S. production of semiconducting and liquid crystals is expected to grow
more rapidly in the coming years. As summarized in Table 4-1, shipments of
semiconductors by U.S. manufacturers were projected to be $15.7 billion for
1983 and $27.6 billion for 1990, up from $6.6 billion in 1979. Assuming that
the crystals used in the manufacturing of semiconductor components would
continue to be supplied mainly by domestic producers, U.S. production of
semiconducting and liquid crystals is projected to grow from an estimated
$45D million in 1979 to $1.1 billion in 1983 and $1.9 billion in 1990.
— U.S. -Bureau of Mines, Mineral Commodity Summaries 1981, pp. 120-121.
1-30
-------
TABLE 4-1. PROJECTIONS OF U.S. ELECTRONIC CRYSTAL SHIPMENTS ($ MILLIONS)
Piezoelectric Semiconductor
Year Crystals and Liquid Crystals Semiconductors
1978 (Estimated)
1979 (Estimated)
1983 (Projected)
1990 (Projected)
(NA) = Not available
a/ JRB estimates (see Text)
b/ Semiconductor Industry Association's estimates
c/ Assume 3 percent annual growth rate
d/ Assume same growth rate as semiconductor components.
110a/
(NA)
130C/
160C/
390a/
450a/
l,100d/
l,900d/
5,830
6,628
15,690b/
27,600b/
1-31
-------
4.2 CAPITAL EXPENDITURES PROJECTIONS
Plants that produce semiconducting and piezoelectric crystals are
classified under SIC 3674 (Semiconductors and Related Services) and SIC 3679
(Electronic Components, Not Elsewhere Classified), respectively. Between 1972
and 1977, annual capital expenditures of the plants under SIC 3674 ranged
between 6 percent and 11 percent of the industry's value of shipments. A
large share of the capital expenditures was for plant and equipment used in
the manufacture of semiconductor devices which have experienced rapid
obsolescence due to rapid changes in production technology. Since the manu-
facturing processes of crystals are not expected to change as rapidly, the
annual capital expenditure rates are estimated to be lower and similar to
those of the more stable segment of the industry, SIC 3679, which ranged from
3 percent to 5 percent between 1972 and 1977. For this reason, the annual
capital expenditures rates were projected to average about 4 percent of the
industry's value of shipments in the 1980s.
Table 4-2 below presents the projections of the capital expenditures
required to maintain the projected levels of sales for the 1980s. It shows
that new capital expenditure requirements in 1990 for the electronic crystal
industry would be around $80 million.
TABLE 4-2. PROJECTIONS OF U.S. ELECTRONIC CRYSTAL INDUSTRY
NEW CAPITAL EXPENDITURES ($MILLIONS)
Industry Value of Industry New Capital
Year Shipments Expenditures
1983 1,230 50
1990 2,060 80
Source; JRB Associates estimates.
4.3 EMPLOYMENT PROJECTIONS
Based on EPA survey data, total employment in the industry was estimated
at 10,000 in 1978. Since the manufacturing processes of electronic crystals
are already very capital-intensive, it appears.likely that the value of ship-
ment per employee ratio would remain fairly stable in the 1980s.
1-32
-------
Table 4-3 presents the projections of employment for the electronic
crystal industry. The employment projections are based upon the ratio of
value of shipments per employee and show that total employment in 1990 would
be about 41,000.
TABLE 4-3. PROJECTIONS OF EMPLOYMENT IN THE
U.S. ELECTRONIC CRYSTAL INDUSTRY
Industry Value of
Shipments Industry Employment
($Millions)
1978 500 10,000
1983 1,230 24,600
1990 ' 2,060 41,000
Source: JRB Associates estimates.
4.4 SUMMARY
Projections of conditions in the U.S. electronic crystal industry have
been estimated in this section as a means of looking at future expectations.
Our estimates show that by 1990, U.S. electronic crystal industry value of
shipments will be $2.06 billion (an increase of $1.56 billion from 1978),
industry new capital expenditures will be $80 million, and industry employment
will be 41,000 (an increase of 31,000 from 1978).
1-33
-------
5. COST OF COMPLIANCE
5.1 OVERVIEW
The recommended water treatment control systems, costs, and effluent
limitations for the Electronic Crystals subcategory are enumerated in the
Draft Development Document for Effluent Limitation Guidelines for the
Electrical and Electronic Components Point Source Category. This document
identifies various characteristics of the industry, including manufacturing
processes, products manufactured, raw waste characteristics, supply, volume,
and discharge destination of water used in the production processes, sources
of waste and wastewaters, and the constituents of wastewaters. Using that
data, pollutant parameters requiring effluent limitations or standards of
performance were selected by EPA.
The EPA Draft Development Document also identifies and assesses the range
of control and treatment technologies which apply to the Electronic Crystals
subcategory. This assessment involved an evaluation of both in-plant and
end-of-pipe technologies that could be designed for this subcategory. This
information was then evaluated for existing direct dischargers to determine
the effluent limitations achievable based on the best practical control
technology currently available (BPT), and the best available technology
economically achievable (BAT). Similar evaluations were performed for new
direct dischargers to develop new source performance standards (NSPS).
Finally, pretreatment standards for existing sources (PSES) and pretreatment
standards for new sources (PSNS) were developed for dischargers to publicly
owned treatment works (POTWs). The technologies identified were analyzed to
calculate cost and performance of each. Cost data were expressed in terms of
investment, operating and maintenance costs, depreciation, and interest
expense.
5.2- POLLUTANT PARAMETERS
The selection of pollution parameters to be considered for regulation was
based primarily on laboratory analyses of wastewaters sampled from several
electronic crystals plants and responses to a mail survey of electronic
crystals manufacturers. The specific approach to selecting the pollutant
1-34
-------
parameters was presented in Sections 5 and 6 of the Draft Development
Document. The specific pollutants selected for regulations are pH, total
suspended solids (TSS), fluoride, total toxic organics (TTO), and arsenic.
Arsenic is to be regulated only at facilities that produce gallium arsenide or
indium arsenide crystals. The pollutants comprising TTO are listed in
Table 5-1.
TABLE 5-1
POLLUTANTS COMPRISING TOTAL TOXIC ORGANICS
1,2,4-trichlorobenzene methylene chloride
1,1,1-trichloroethane naphthalene
2,4,6-trichlorophenol 2-nitrophenol
chloroform 4-nitrophenol
2-chlorophenol pentachlorophenol
1,2-dichlorobenzene phenol
1,3-dichlorobenzene bis(2-ethylhexly)phthalate
1,4-dichlorobenzene di-n-butyl phthalate
1,1-dichloroethylene tetrachloroethylene
2,4-dichlorophenol toluene
ethylbenzene trichloroethylene
5.3 RECOMMENDED TREATMENT TECHNOLOGIES
Based on the analysis of the potential pollutant parameters and treatment-
in-place in the electronic crystals industry, EPA identified six treatment
technologies that are most applicable for the reduction of the selected pol-
lutants. These treatment technologies are described in detail in Section 7 of
the Draft Development Document and are listed below:
• Option 1: Segregation and collection of spent solvents containing
toxic organics for reuse, resale or contract hauling
(referred to as solvent management) plus end-of-pipe
treatment for pH control
• Option 2: Option 1 -plus end-of-pipe precipitation/clarification for
_ control of fluoride, arsenic, and suspended solids
• Option 3: Option 1 plus precipitation/clarification of concentrated
fluoride wastes for control of fluoride
• Option 4: Option 2 plus 75% recycle of treated effluent to further
reduce pollutant discharges
1-35
-------
• Option 5: Option 2 plus filtration to further reduce pollutant
discharges
• Option 6: Option 5 plus activated carbon to further reduce toxic
organics.
The Agency's evaluation of treatment options 4 and 6 concluded that these
technologies are not feasible on a nationwide scale. For this reason, the
economic impact analysis concentrated on treatment options 1, 2, 3, and 5
only.
5.4 TREATMENT COST ESTIMATES
Costs of compliance were estimated by EPA for treatment options 2, 3,
and 5. Treatment costs of option 1 are assumed to be negligible and are not
estimated by EPA, because information available indicates that most electronic
crystal plants are practicing good solvent management techniques to control
toxic organics. Moreover, many of these plants are also controlling the pH of
discharges by end-of-pipe neutralization.
In developing the compliance cost estimates for treatment options 2, 3,
and 5 the following major assumptions were made by EPA:
• All costs are expressed in end-of-year 1979 dollars.
• The treatment facilities were assumed to operate 8 hours per
day, 260 days per year for plants with discharge less than
60,000 gpd; 24 hours per day, 260 days per year for plants
with 60,000 gpd to 200,000 gpd; and 24 hours per day, 350
days per year for plants with more than 200,000 gpd.
• Labor costs are based on an hourly rate of $20, including
fringe benefits and plant overhead.
• The cost of land is valued at $12,000 per acre.
• Energy costs are based on $306 per horsepower.
• Sludge disposal costs are included. Available information
indicates that the sludge to be disposed is not defined as
^hazardous wastes by RCRA.
• Capital costs are amortized at 5 years and 13 percent
interest.
1-36
-------
5.4.1 Existing Sources
Based on the above assumptions, capital investment costs and annual
compliance costs at several selected flow rates were estimated for treatment
options 2, 3, and 5. Figures 9-1 to 9-3 in the Draft Development Document
present the cost curves for these three options. These cost curves were used
to estimate compliance costs for the seven model non-arsenide crystal plants
and the eight actual arsenide crystal plants. The key operating character-
istics of the seven model plants are summarized in Table 5-2. The seven model
plants (non-arsenide crystal) were based on the plant sizes defined in Table
3-3, wastewater flow rates, and other data developed from the EPA 308 survey.
Information obtained by EPA indicates that all the identified non-
arsenide direct dischargers are already in compliance with the proposed
effluent standards. Tables 5-3 and 5-4 present the estimated investment and
annual cost of compliance for the 53 non-arsenide indirect dischargers. Table
5-5 summarizes the compliance costs for the eight arsenide crystal plants.
5.4.2 jfew Sources
The basis for the new source standards (NSPS and PSNS) is the best
available demonstrated technology (BDT), including in-plant controls and end-
of-pipe treatment technologies that provide the maximum pollution reductions
feasible. For the Electronic Crystal subcategory, the proposed NSPS and PSNS
discharge limitations are the same as those proposed for BAT and PSES. Since
new source costs are defined as incremental costs from BAT and PSES, costs of
NSPS and PSNS are zero.
5.5 SOCIAL COST ESTIMATES
This section assesses the total social costs that can be associated with
the EPA effluent regulations. The social costs measure the value of goods and
services lost by society due to a given regulatory action. These costs
generally include the use of resources needed to comply with a regulation, the
use of resources to implement and enforce a regulation, plus the value of the
output that is forgone because of a regulation.
1-37
-------
O CD
JJ
1) CU
3 co
i—I CO
05 <
jj
Jd C
O M
O —I
co a,
o
e/>
o
cs
CS
O
o
CM
O
o
o
o
o
o
CS
o
o
JJ C
•H "4
UJ. 00|9>S
O -
in
CM
O
CN
m
tN
m m o
-H —t CS
O C
a)
a)
3
> W
8 8
O
o
o
ft
es
O
o
r-
o o
o o
r- O
o
o
o
w
en
z
i
O
in
vO
o
m
m
o
o
m
o
o
in
CM
cs
a
M
1
OH
CS
I
in
a
PQ
u-i m
o 0)
v
i- >>
4) O
JD f-l
S i
z w
o
es
o
oo
O
co
o m o
oo vO m
i—i en
C
(0
CO
at
•B
en
X
Qj
4J
(0
O
JJ
C
CO
1-4 CO
p-l JJ
CO
<-( OS
CO
t- £
JJ 00
es
en
X
I
§
• H
•o
cu
sc
£
jj
•<-(
»
jj
C
HI CO 0)
JJ I—I JJ
CO
J
CO O
cs
en
00
4J
CO
-------
•H *H —c -« r-. eo
cs
•*
Bu
V
B.
i
c
•9
V
e
'a
<3
4)
U
1
a.
§
e
o
4J
a
o
^ ^^. ^^ in
999
O O O —
*O
CM
e
o
°
o o o o
oo co oo co
•V.
**
§
*
o
e e o o
O CM
O •"
01
H
.S
o
4J
•o
I
T3
C
3 I
e c
01
> 0) O
— o o
Ifc OS —•
£ S
m
8
o
8
«M
01 <
u
s s
-* C4
- , vi ca -< ts en
p» Hi x x co co ca
1-39
-------
a
o
a.
O
3
x CS
8
00
vO
s
ON
o
o
V O
to r*
m 00
in
-ar
8
I
^
in
1
v a
*j x
a B
e -< J 03 (0 —I
p» HI X M «
O
to
1-40
-------
0
b
eg
.N
"-
ti
•O
c
CO
BD
3
O
p4
Vn'
CO
g
^
J
a.
H
w
u
u
o
u
to
OS
<
X
u
S3
S
8
u
y
<
2
1
u
J
<
H
in
e
o
••4
CO 4J
• O
s
it
O CM
e
5 e
•H O
f4J
*
•^
I"-1
e
o
a
in
e
o
••4
e o.
V 0
4->
CD
9)
O
eg w
x&
o.
CO
u
See?
e — *
a
•H C
— 0
&•••<
o a.
O o
•
IM U
11
3 5.
^4 ••«
> <»
(V
4J
S
|
(h
"o S
b X
4, o
-O *-i
8 o-
££
41
00
b
.c 3
•*4 4J
O CO
O
1-1
c
04
^ ^ ^
XI U • • O
O O O O 00 O CO
oo ON —
~< en
— ». -•. o m !»••»»
ooooomooo
oo r- ao
-H CM
oooooooo
^^,
^^ ^^ ^ i*^ ^ ^3
O O O O O 0> r»
f»> vO
OOOONO^T— • OOO
CM CM m in oo
CM -* r-
oooooooo
in o O o O o
^N CM •& CM CM O O
e . ....
pu « i>» m in o
O ^f C^ QO
*-> -4 CM
— * f^ *^ «Cf O ^*
oo in
.
MM MMMI-II-IO
.^CM m^in\o^oo
HHHHHHHH
CM CM
fT |«J o
in in
0 0
•* •* O
^ -»
in m
OOO
«—S «N
oo oo o
OO 00
ON*
« •
CM CM O
in in
i-» r«-
O O O
u
u
V w
b u
.rf «
« e "4
4J M O
5
a
3
O
b
M
4J
U
3
•§
b
a
•
••4
X
4J
£
h
HH
X
XI
"S
1
s
i
U
£
eo
e
• f4
A
4J
§
O
•V*
O.
o
f*
01
w
a
V
£
.
V
41 —
— Xt
Xl eg
CO U
•t"4 r4
iT
•o
41
U
3
01
b
£
o
e
u
• •4
e
00
o
u
••4
X
4J
x<-^
I'M n^1
57
in
« — i
•0 XI
eg H
? -
« u
u c
" I
Ji
IM e
B
u E.
•rt o
c •->
V 41
b 41
CO 0
*Q w
4 CO
b
41 O
"O
•H 41
b 01
3
«
qj
a
•*
a
01
«
01
••4
U
o
a
a
as
01
b
3
S
1-41
-------
5.5.1 Conceptual Framework
The partial equilibrium analytical framework is conceptually the most
practical means for estimating total social costs. This framework, in its
most sophisticated form, is based on an analysis of supply and demand rela-
tionships in the directly-affected markets. When an industry is regulated,
compliance requirements result in increased unit costs of production. This
increase, in turn, leads to an upward shift in the industry's supply curve.
The supply curve shift normally results in higher prices and a lower produc-
tion level. Compliance costs, production losses, and net welfare losses
incurred by producers and consumers due to decreased output are measurable
within this framework. There are other costs that are not measurable within
this framework. Costs of implementing and enforcing a regulation must be
added. Also, other social costs do not appear in this static analysis, such
as productivity effects, innovation impacts, and costs of reallocating
resources that become unemployed. Unfortunately, the data does not exist to
carry out such analysis at this time, and a compromise which captures the
major costs to society was performed.
5.5.2 Social Cost Analysis
For this analysis only the real resource costs are considered. This
provides a reasonable estimate of social costs, since EPA studies show that
95 percent or more of the social costs are directly related to compliance
expenditures by the regulated entities. Consequently, the present value of
social costs (PVSC) of regulations can be approximated by the following
equation:
PVSC = I/(l + .l)n + (OM/.
where: PVSC * present value of social costs
I * investment cost
OM » annual operating and maintenance cost
_ n » number of years between now and year of investment
The above equation assumes that:
• The regulations will be in effect in perpetuity
1-42
-------
• Operating and maintenance costs will be incurred in the first year of
investment
• The real discount rate (imposed by OMB) is 10 percent.
Assuming that the industry will have to be in compliance by 1984 and
compliance expenditures will begin in 1984, n would equal 2. Social costs of
the proposed PSES regulations for the Electronic Crystals subcategory are
calculated and presented in Table 4-6. The results indicate that the social
costs of the alternative PSES vary from zero cost at treatment option 1 to
$3.6 million at option 5 for the arsenide crystal plants. Social costs for
the non-arsenide crystal plants range between zero cost and $20.2 million for
treatment options 1 and 5 respectively.
Since the existing direct dischargers are already in compliance, social
costs of the alternative BAT treatment options are zero. Similarly,
compliance costs of NSPS and PSNS treatment options are zero, thus social
costs are also zero.
1-43
-------
TABLE 5-6. PRESENT VALUE OF SOCIAL COSTS OF PSES REGULATIONS
FOR THE ELECTRONIC CRYSTALS INDUSTRY (Thousand of 1979 Dollars)
Treatment Alternatives Social Costs
Gallium Arsenide/Indium
Arsenide Crystals
Treatment Option 1 0
Treatment Option 2 3,350
Treatment Option 5 3,643
Other Crystals
Treatment Option 1 0
Treatment Option 2 18,626
Treatment Option 3 11,858
Treatment Option 5 20,155
1-44
-------
6. ECONOMIC IMPACT
This section provides an assessment of the economic impacts which are
likely to occur as a result of the costs of the proposed effluent treatment
technologies described in Chapter 5. It is based upon an examination of the
estimated compliance costs and other economic, technical, and financial
characteristics of both model and actual plants of various sizes and the
analytical methodology described in Chapter 2. The primary economic impact
areas discussed include changes in industry profitability, plant closures,
substitution effects, changes in employment, shifts in imports and exports,
etc.
6.1 PRICE AND QUANTITY CHANGES
As described in Chapters 2 and 3, the Electronic Crystals industry
exhibits characteristics associated with non-competitive markets. For
example, the industry has unusually high profit rates and most of the product
groups have high concentration ratios and low demand elasticities. These
three characteristics are often indicative of non-competitive markets. All
other things being equal firms in this type of market would be able to recover
part or all of the compliance costs in the form of higher prices.
However, as previously mentioned, the existence of the threat of foreign
competition serves to mitigate the ability of firms in the industry to raise
prices. Such price action would cause the domestic industry to lose market
share. Consequently, it is assumed that there is an equilibrium trade-off
between domestic and foreign market share which is determined largely by the
price differentials; and that domestic producers would like to maintain the
baseline equilibrium by maintaining the price spread between domestic and
foreign producers. Since the domestic producers have high profit rates they
can afford to maintain the spread by absorbing the increased costs. For this
reasxm, it is assumed that for small and moderate levels of compliance cost
domestic electronic crystals producers will absorb the costs and not adjust
prices. " Consequently, the price changes due to the regulations would be zero.
It follows, also, that quantities demanded will not change because of the
regulations. ,
1-45
-------
6.2 RESULTS OF SCREENING ANALYSIS
As described in Chapter 2 a ratio of annual compliance cost to plant
revenues of one percent was used as a threshold value below which plants are
considered to have low impacts from the regulations. The calculated ratios
are shown in Table 6-1. Most non-arsenide model plants have compliance costs
above this value for option 5. Therefore, the profitability analysis was
conducted for all non-arsenide model plants. For the arsenide metal crystal
plants only one plant (model plant number T5) has compliance costs greater
than one percent.
The one arsenide metal crystal plant and all of the non-arsenide plants
were subjected to a more detailed analysis of plant profitability.
6.3 PROFIT IMPACT ANALYSIS
As indicated in Chapter 2 above, the assessment of the impact of com-
pliance on plant profitability is based on the plants' after-compliance return
on investment (ROI) ratios.
Table 6-2 summarizes the estimated profit impact for the seven non-
arsenide model plants. As indicated in the table, the profit impacts do not
appear to be very significant for all four options. All non-arsenide model
plants remain profitable, their after-compliance ROIs ranging between 20.4
percent and 40.1 percent at treatment option 5 as compared to 11.5 percent
average yield of U.S. Tre<
potential plant closures.
average yield of U.S. Treasury bonds for 1980 , and therefore are not
Table 6-3 presents the estimated profit impact on the potential "high
impact" arsenide metal crystal plant. This table indicates that profit
impacts of treatment options 2 and 5 are high for plant T5 (i.e., after-
compliance ROIs are below the 11.5 percent threshold level).
Standard and Poors Corp., The Outlook
1-46
-------
TABLE 6-1. SCREENING ANALYSIS
ANNUAL COMPLIANCE COSTS TO VALUE OF SHIPMENTS (%)
PLANT
Arsenide Crystals
Tl
T2
T3
T4
T5
T6
T7
T8
Non-Arsenide Crystals
XS1
XS2
SI
S2
S3
M
L
Option 1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Option 2
0
0
0
0
3.4
0.8
0.6
0
5.6
3.8
0.6
0.8
4.0
1.9
1.2
Option 3
NA
NA
NA
NA
NA
NA
NA
NA
0
0
0
1.7
2.8
1.2
0.7
Option 5
0
0
0
0
3.7
0.8
0.7
0
5.6
3.8
0.6
0.8
4.5
2.1
1.3
NA = Not applicable. Treatment option not being recommended by EPA for this
product group.
Sour.ce: JRB Associates estimates.
1-47
-------
CO
1
J
jj
w
§
S
*J
<;
H
^4
C6
CJ
W
O
HH
z
w
CO
es
i
z
o
z
OS
o
Pu
H
Z
W
£
co
co
W
CO
CO
H
CJ
p-
z
M
H
M
to
O
W
cm
to
O
5>-
os
5
5
5
eo
,
es
1
vO
a
CO
in
c
o
•H
U
a.
o
CO
C
o
• H
JJ
a
o
es
e
o
•H
4J
a
0
1-1
c
0
• r>4
4J
a
o
M
O
G6
09
-l
0)
t5
O
vO
»
*^
CM
r^
•
/~v in
K CO
**^
vO
•
-x CO
0
•
m
/-N es
X
^v O
^3 c^f
o -<
o
v>
*~s
O
o
x-> es
0
0
*-/
c-4
CO
X
<^
•
o
CN
f*^
•
r^-
C*4
^r
•
o
• t • V
O CO CO CO —(
-* co es es co
r*" «^ "^ \^ O^
*»H vO sO 00 <
jj
'H
o
0
CO
CO
«
05
.«
0)
CJ
^J
3
Q
c/5
1-48
-------
co
_]
PH
hJ
O
C
CO
o.
g
o
0
1
»-v
o
o
o
-^
co
^4 4iJ
o e
• S
3 &
i— i -i-i
« j:
> en
*^
o
o
o
^x
Q
4J
e
CO
^H
Bri
,— 1
•
to
0
ol
o
•
CM
o
^^
o
o
CO
A
•*
,^
co
CO
««
CM
in
H
CO
0)
4J
a
e
• H
4J
CO
CO
01
4-1
CO
O
O
CO
CO
CO
06
n
0)
o
3
O
CO
1-49
-------
6.4 CAPITAL REQUIREMENTS ANALYSIS
As described in Chapter 2, the assessment of the firms' ability to raise
the required capital is based upon ratios of compliance capital investment to
revenues (CCI/R) for specific plants compared with the difference between the
industry's highest and lowest ratios of average annual capital expenditures to
average value of shipments between 1970 and 1977. The highest and lowest
ratios for SIC 3679 (Electronic Components, n.e.c.) were 5.1 percent and 2.8
percent over the 1970-1977 time period respectively;2 therefore, 2.3 percent
was selected as the threshold level for the capital availability analysis.
Plants whose CCI/R ratios are lower than 2.3 percent are considered to have
low levels of impacts. Those plants with higher CCI/R ratios are subjected to
the cash flow test described in Chapter 2.
Table 6-4 presents the compliance capital investment to plant revenue
ratios for the non-arsenide crystal plants. As indicated in the table, com-
pliance capital requirements for the proposed treatment options appear to be
high for model plants XS1, XS2, S2, S3, and M. The cash flow test for these
five plants is shown in Table 6-5. The plant annual cash flow (i.e.,
depreciation plus net profit after taxes) seem to be sufficient to meet total
capital requirements (compliance capital investment plus minimum annual
capital expenditures for plant and equipment) of all plants at each treatment
option.
Table 6-6 summarizes the capital requirements analysis for the eight
arsenide crystal plants, where plant T5 is predicted to have some difficulties
financing treatment options 2 or 5. However, plant T5 is owned by a large
integrated electronics company and the required compliance capital investments
for both treatment options represent less than 0.1 percent of this firm's
total annual capital expenditures in 1978. ' Since the capital requirements
is so small relative to total capital expenditures for this company, it is
assumed that the company can readily finance the investment.
2 /
U.S. Department of Commerce, 1977 Census of Manufactures.
Standard and Poor's Corp., Stock Market Encyclopedia
1-50
-------
CO
H
£
w
Q
Q
3E
j
<4
&
g
j
H
g
O
H
H
jg;
o
Z
1
CO
H
CO
>"
nJ
<
**•
CO
z
£
U
Erf
M
&
8
j
e
M
i— 1
AJ
(f^
a
CO
^
^
>._^
CO
AJ
C
• H
.C
CO
U-l
O
0)
3
CO
>
m
c
o
i-4
a
o|
CO
c
o
I-l
AJ
a
ol
CN
C
o
AJ
a
0
^4
e
o
a
0
s*+^
**
I-l
0
^N
8s?
\^f
G
•^
00
CO
*•
AJ
• H
|ij
O
(y
3
1-4
CO
>
^
0
o
o
v>
"-'
09
^j
C
•
t^^ \O ^D ^^ ^5 f^ *""*
* *
V.O ^0 itO ^^
• • « •
•O O ^D f^ f^ *^ ^3
* * * *
E^J ^"^ 5^ fi»4 C^ QQ f^
o\ \o O "— i in CN i— i
O O O O O O O
^^ CO ^^ ^^ >^f ^O ^^
tn r^ CD vo \o eo -*
t*i CM m in co CN fi
o o O O o o O
in o o m m m o
CN CM CO — 1 «— 1 •— 1 CN
o o o o o o o
-------
TABLE 6-5. COMPARISON OF CASH FLOW
AND TOTAL CAPITAL REQUIREMENTS
(as a Percent of Plant Revenues)
Internally Total Capital Requirements
Generated
Funds3
Model Plants (percent) Option 2 Option 3 Option 5
XS1 18.5 11.8 0 11.8
XS2 15.7 8.8 0 8.8
S2 11.1 3.9 6.4 3.9
S3 li.l 8.7 6.4 9.5
M 11.6 5.6 4.3 5.9
a
Equivalent to net profit margin, (assuming an average corporate income tax
rate of 40 percent) plus depreciation (10 percent of depreciable fixed
assets, assumed to be 50 percent of total assets).
Compliance capital investment plus minimum annual capital expenditures for
plant and equipment to maintain a targeted production level (assumed to be
2.8 percent of plant revenues).
1-52
-------
H
3
^
Oi
^
H
co
i
w
o
M
z
w
CO
3
i
CO
h-1
CO
2
^c
%
to
H
SB
w
s
S
M
£3
cy
pj
w
,
<:
Fj^
M
DH
3
S
5£
•
1
>o —
—
W
>J
tfl
^C
H
CO
TJ
C
a
JS
CO
U-l
O
0)
3
^H
CO
o
H
4J
C
i
4J
CO
4)
c
M
CO
4J
a
a
0
0)
u
c
CO
1— 1
a
e
o
o
to
4J
C
a.
CO
IM
o
01
3
i— i
CO
^>
O
M
JJ
C
CO
DL,
in
C
o
.1-1
4J
a
o
oo
'
HHHHHHHH
en
v
g
M
• i-l
U-l
O
^J
C
0)
u
01
a.
^H
,
O
e
CO
J=
4J
CO
CO
01
1— 1 •
00
CO f».
e -H
CO 0)
oi co c
Ol 01 *i-<
O a- en
^-l O) 4)
a. M u co
§3 0)
^j ±J 4J
C -H CO
U-l 0) T3 g
o 6 c -5
4J 0) 4J
M 00 &- 03
0> 01 X 01
•s > *
Be co
3 'H i— * 0)
e co jj
^-i JJ CO
e « • i-< -H
O 4J Q. U
•i-l CO O
T3 fr O CO
01 CO CO
CD U i-l •<
CO CO
43 0) 3 «
U C OS
oi c c "n
4J CO CO
CO .H
€ !— 1 ,— I
•i-l O. to OI
4J g *J O
co O 0 >-i
W O JJ 3
*•>• -^ o
to 40 co
1-53
-------
6.5 POTENTIAL PLANT CLOSURES
Based on the profit impact assessment and capital requirements analysis
discussed in Sections 6.2 and 6.3, only one arsenide crystal plant, model
plant T5, might be considered to be a potential closure at treatment options 2
and 5. (Since option 1 compliance costs were insignificant, no plant closures
are expected from option 1). However, plant T5 is owned by a large integrated
electronic company and supplies gallium arsenide and gallium phosphide
crystals to be used in the production of LEDs at other production facilities
of the firm. Since the operations at plant T5 continue to contribute to the
firm's total profitability, it is expected that the firm will keep this plant
operating to maintain control on the source of supply for its raw material.
Furthermore, the plant's low reported profitability may be due to the firm's
intercompany transfer pricing policy and actual profit at market price may be
very satisfactory.
6.6 EMPLOYMENT EFFECTS
The preceding section projects no closure potential for the proposed
treatment technologies. Thus, there will be no employment effects from the
proposed regulations.
6.7 SUBSTITUTION EFFECTS
As described in Section 2.2, because of their special properties the
potential for substitution of electronic crystals in their current applica-
tions is limited in the base case. In addition, the preceding impact analysis
assumes that there will be no price changes resulting from the proposed
regulations. For these reasons, no substitution effects will result from the
proposed regulations.
6.8 FOREIGN TRADE IMPACTS
_ As described in chapters 2 and 3 the United States is currently very
competitive with foreign producers on the world market for electronic
crystals. However, there exists considerable technical expertise among
foreign producers, which is perceived as a potential source of competition.
If there was a significant price effect from the regulations, the U.S.
1-54
-------
producers could lose market share. However, as indicated in the profit impact
analysis described in Section 6.4, U.S. producers appear to be able and likely
to absorb the additional pollution control costs. Consequently, no foreign
trade impacts are expected to result from the proposed regulations.
6.9 NEW SOURCE IMPACTS
As indicated in Section 5.4.2, costs of NSPS and PSNS are zero.
Consequently, there will be no economic impact from NSPS and PSNS require-
ments. Also, because there is no difference in compliance costs between
existing and new sources, the proposed regulations would not foster any
competitive advantage or disadvantage between new and existing sources as long
as treatment requirements for each are the same.
1-55
-------
7. SMALL BUSINESS ANALYSIS
The Regulatory Flexibility Act (RFA) of 1980 (P.L. 96-354), which amends
the Administrative Procedures Act, requires Federal regulatory agencies to
consider "small entities" throughout the regulatory process. The RFA requires
an initial screening analysis to be performed to determine if a substantial
number of small entities will be significantly impacted. If so, regulatory
alternatives that eliminate or mitigate the impacts must be considered. This
analysis addresses these objectives by identifying and evaluating the economic
impacts of the aforementioned regulations on small electronic crystal
manufacturers. As described in Chapter 2, the small business analysis is
developed as an integral part of the general economic impact analysis and is
based on the examination of the distribution by plant size of the number of
electronic crystal plants, plant revenues, wastewater volumes, compliance
costs and potential closures from the regulations.
In the absence of an appropriate specific definition for a small
business , four plant size definitions based on plant revenues were used to
provide the EPA possible alternative definitions of small electronic crystal
plants. These are:
• Plants with less than $1 million in value of shipments
• Plants with $1 million to $3 million in value of shipments
• Plants with $3 million to $5 million in value of shipments
• Plants with over $5 million in value of shipments.
The number of electronic crystals plants falling into each size category
is shown in Table 7-1. The table also shows the plant revenues of the
different sized plants, along with the percentages of the industry total for
each. As the table shows, the industry is characterized by a. few large plants
while account for most of the production and many smaller plants producing a
smaller portion of industry output.
I/
See Chapter 2 for the discussion of the plant size definition
1-56
-------
•
V
fa
£
8
a
|
'^
e
i
S
W
e
•*
£
b
9
O
e
ft.
1
i
|
!
j
i
e
m
o>
*
e
D.
o
hi
J
"o
1
9
W
.
T
r*.
V
3
!
3
e
o
J
?]I
!u
H e
• Q
* o
s*
9
*
** 3
§
O 3 5
el*
8
1 18
• o
S*
M
b
O
I5
*-g
IMS
«
1i
V <•*
1
*
w
9
O
1-
Ir
5 Q
• o
w *»•
I
-il
5 S =
• w
lh
e u
"•
5
M
g
•
ft*
ooooo ooeoo
*«* • ^^ O «M « *^. o ff* « r* o
"S ~ S2 R2 2~ 3= 22 32 S2
^ A • •«•«
•> * ^ n« — o n e o> • O
OO^OO^VnQ oin««nin'a-!»g
e e o o o
^. ^% ^^ ^^
V • ~ - °.
«d OO«»M«|A«M0O
• wmc««•«< me
«^«<« >^t0«^O O-^W—^M-t^^O
<^ ^«^M^*« ^ «^x^O«% o?> a>3 3 « S o 3
.
ooeoo eoooo
00000 ooooo
ooooo ooooo
«— o*«e f»—>-r>o«o«mo
oe *» o •> jj «j o « — >e jj » jj o jj go
«« S3 — ^, „ „ „
oo e-« eo o* So oo m» owi $•* '"S
•^ . 5 •** >* t^-** «« £**O*~5r«O'0«4*OO0<"1O
• ^ m ^ o^o ••Nn«Mfn«*««iAO
s - c
1 .. i ... 8
C • — -< • ""
! I = . 21 =
< « • j » J • «
JS58JS «!S8j =
^ i C — — "J -8 J C — — 'i
2 S • 3 3 J, 'S " 5 3 i ^
"•vnmM b««f*1if\b
2eS-A ? < * e ! — A »«
S Z g £ £ I
I
w
o
9
I
•
*
& '
<
b
1
S
I
f
J
1 £
I 1
u •
* >
V C
1-57
-------
The information available for this study precluded a forecast to the
1985-1990 period of change in the distribution of industry output by plant
size. However, no reason was found to expect the proportion of industry
output produced by these small plants to change. For these reasons, the
assessment of impacts on small business were developed from the data on
existing plants.
The assessment of economic impacts on small plants begins with an
evaluation of compliance costs incurred at the plant level. Table 7-1 also
shows the estimated investment and total annual compliance cost for electronic
crystal manufacturing by plant size. The table also shows the proportion of
electronic crystal output and industry compliance costs attributed to both
small and large plants. For example, at treatment option 2 the nine non-
arsenide crystal plants with less than $1 million in revenues discharge only
0.2 percent of the total wastewater discharged while they account for 3.3
percent of total annual compliance costs. In contrast to these small plants,
the ten plants with sales greater than $5 million discharge 64.1 percent of
the wastewater and incur only 42.9 percent of total annual compliance costs.
For all regulatory options, compliance cost as a percent of revenues is
significantly larger for the smaller than for the larger plants.
Because the compliance costs of small plants are greater than that of
larger plants and because they compete for the same markets as larger plants,
their profits as measured by return on assets would decrease more than those
of larger plants. Likewise, their capital requirements for compliance
relative to their revenues will be greater than that of larger plants.
However, as shown in Chapter 6, the magnitude of these impacts are not enough
to cause any closures among these smaller plants. Consequently, there will be
no employment, community impacts, or industry structure changes resulting from
the regulations.
As described in Chapters 2 and 6, the regulations will not cause changes
to the prices of electronic crystals. As a result, manufacturers will absorb
the full amount of the compliance costs and, therefore, will experience lower
profits from electronic crystal manufacturing.
1-58
-------
8. LIMITATIONS OF THE ANALYSIS
This section discusses the major limitations of the economic impact anal-
ysis. It focuses on the limitations of the data, methodology, assumptions,
and estimations made in this report. Information pertaining to the estimation
of the compliance costs for specific plants and its limitations are outlined
in Section 9 of the EPA Development Document.
The economic impacts assessed are the result of the proposed water
effluent requirement only. The assessment does not include the economic
impacts from such items as air pollution control, OSHA requirements, solid
waste requirements, and cost resulting from other regulations.
8.1 DATA LIMITATIONS
The accuracy of the conclusions of this report depends largely on the
accuracy of the data used in the analyses, especially that of the estimated
compliance costs, and plant financial and economic characteristics.
A critical data input to this study is the compliance cost estimates.
The compliance costs used were developed from compliance cost curves provided
by EPA. The assumptions relating to the estimation of compliance costs are
outlined in Section 9 of the EPA Development Document and summarized in
Section 5.4 of this report.
In addition to the compliance cost assumptions, plant financial profiles
used in the analysis are subject to the following major assumptions and
limitations:
• Baseline financial data of the eight existing arsenide crystal plants
are obtained from EPA 308 survey and follow-up telephone interviews
with plant representatives
• Plant characteristics such as size and flow rate of the 53 non-
arsenide crystal plants are inferred from survey data collected by EPA
• on 38 existing plants
• Baseline financial data of the 53 non-arsenide crystal plants are
inferred from survey data on 26 plants
1-59
-------
• The impact analysis for non-arsenide crystal plants was based on seven
model plants which were developed from the above survey data
• Plant financial data are based on 1978 economic conditions and will
remain essentially the same over the period of the study.
Because individual plant financial characteristics vary from the model
plant profiles, the model plants developed for this study should be considered
order of magnitude estimates of the typical operating characteristics of the
non-arsenide crystals manufacturing industry. However, it is believed that
the sample of plants was sufficient to reflect considerable intra-industry
variations in conditions. Therefore, although the confidence interval of the
conclusions are indeterminate, there is no reason to suspect that the
conclusions are biased toward overestimating or underestimating the degree of
potential impact.
8.2 METHODOLOGY LIMITATIONS
In addition to the data limitations described above, this report is also
subject to the following limitations of the methodology used:
• It is assumed that firms will not attempt to pass through the
compliance costs to their customers by raising their product prices
• The required rate of return on investment (i.e., profit impact
threshold value) is estimated to be 11.5 percent, the 1980 average
yield of U.S. Treasury bonds
• The threshold values of a firm's capital availability are assumed to
be the plant's estimated annual cash flow and/or the industry
historical annual rates of capital expenditures as reported by the
Census of Manufactures
• Plant closure criteria do not include the financial structure of the
plants, and the time value of money, or special plant-specific
conditions.
•8.2.1 Price Increases Assumption
The analysis assumes that the electronic crystals industry will not
increase the price of their products and pass through any portion of the
compliance costs to their customers. While this is a reasonable assumption
for some types of crystals it may not be true for the entire industry. It is
1-60
-------
possible that the price pass-through percentage will vary among the various
types of crystals. Adequate information on these segments is not available.
Therefore, no further analysis is carried out to measure the extent of a price
increase which may occur as the production costs of the plants increase with
the installation of pollution control.
The estimation error which might result from the price change assumptions
depends primarily on the cross elasticity of demand with respect to the ratio
of domestic to foreign prices. That is, if increased domestic prices shifts
demand production overseas, there would be a drop in aggregate quantities
produced by domestic producers and, consequently, unemployment, and possibly
other impacts. Thus, the remaining domestic crystal producers would have
healthy profits, but account for a smaller portion of world market share. On
the other hand, if such shifts were smaller, then domestic producers could
raise prices to cover compliance expenditures without losing market share.
If the latter situation prevailed in the real world, the economic impacts
estimated in this report are overstated. If the former situation prevailed,
the impacts would take a different form, but would probably not exceed the
levels estimated in Chapter 6 of this report.
8.2.2 Profit Impact Threshold Assumptions
The profit impact assessment is determined by company the plant's after
compliance return on investment (ROI) to the average yield of U.S. Treasury
bonds which is 11.5 percent in 1980. Plants with after-compliance ROI below
11.5 percent are assumed to be potential plant closures.
Although investment decisions based on discounted cash flow techniques
generally provide more consistent and theoretically correct investment
decisions than those based on ROI analysis, ROI analysis is widely used by the
investment community and does provide acceptable results in comparative
analyses of industry-wide conditions, such as that presented in this report.
It is reasonable to assume that plants cannot continue to operate as viable
concerns if they are unable'to generate a ROI that is at least equal to the
opportunity cost of other lower risk investment alternatives (e.g., U.S.
Treasury bonds)
1-61
-------
8.2.3 Capital Availability Threshold Assumptions
A firm's ability to make the required pollution control investment is
evaluated by comparing the required compliance capital investment to the
industry historical annual rates of capital expenditures and the plant's
estimated internally generated annual cash flow. Although the above analyses
do not precisely indicate whether or not firms can afford to make the
investments, they do provide a good indication of the relative burden created
by the compliance requirement.
8o2.4 Plant Closure Assessment
The criteria for plant closure are not all inclusive. They do not
address the financial structure of the plants, the time value of money, or
other factors which affect closure decisions. A more detailed closure
analysis would require more precise data on the capital structure, and cash
flow of plants in the electronic crystals industry. However, the costs of the
recommended treatment technology are small, and a more detailed analysis will
not necessarily yield different results.
8.3 SUMMARY OF LIMITATIONS
Although a number of assumptions and data limitations may significantly
bias the economic impact conclusions, the potential changes to the conclusions
resulting from the elimination of these biases are probably small. Sensitiv-
ity analyses of plus and minus 20 percent were performed for each of the above
assumptions. These analyses concluded that if each of the aforementioned
variables changed by plus or minus 20 percent, there would be no plant
closures as a result of the regulations and the other results of the study
would not be significantly impacted. Consequently, the conclusions of this
study appear to represent a reasonable industry-wide assessment of potential
impacts.
1-62
-------
PART II
SEMICONDUCTOR SUBCATEGORY
-------
1. INTRODUCTION
1.1 PURPOSE
The purpose of this report is to analyze the economic impacts that are
likely to result from the promulgation of EPA's effluent regulations on the
semiconductor industry subcategory of the Electrical and Electronic Products
point source category. The effluent standards and limitations which are
issued under Section 30l(b) of PL 92-500, the Clean Water Act of 1977, require
the Administrator (EPA) to establish the following:
1. Effluent limitations based on the Best Practicable
Technology Currently Available (BPT), Best Conventional
Pollutant Control Technology (BCT), or Best Available
Technology Economically Achievable (BATEA or BAT) to
be met by industrial dischargers
2. Pretreatment standards for existing and new dis-
chargers (PSES and PSNS) to publicly owned treat-
ment works (POTWs)
3. New Source Performance Standards (NSPS) to be met
by new source industrial dischargers.
In order to assist in the development of these limitations this study explic-
itly assesses the effects of various pollution abatement regulatory options
upon the costs of production, capacity expansion and replacement, profita-
bility, and the potential for plant closures in the semiconductor industry.
In addition, the likely effects on employment, communities, foreign trade,
small business, and the welfare of society are estimated.
II-l
-------
1.2 SCOPE
The semiconductor subcategory as defined for this study includes the
manufacture of electronic components made from elements such as silicon and
germanium that contain small amounts of impurities which make them neither
good electrical conductors nor good insulators. Among other functions,
semiconductors can amplify, switch, and rectify electronic current. They
are increasingly used in place of vacuum tubes or electron receiving tubes,
mechanical relays, switches, and ferromagnetic memory coils (in computers).
Semiconductors are used in many electrical and electronic products, such as
radio and television receivers, and in transmitters, amplifiers, calculators,
computers, auto ignitions, telephone switching systems, cameras, and digital
watches. The growing and fabrication of crystals and the assembly of finished
electronic products are not included in this subcategory.
The industry can be segmented into four separate product classes. These
product classes and their associated standard industrial classification (SIC)
codes are shown below.
SIC INDUSTRY PRODUCT CLASS
3674 Semiconductor and Related Devices
36741 Integrated Microcircuits
36742 Transistors
36745 Diodes and Rectifiers
36749 Semiconductor Devices, n.e.c., and parts
The semiconductor industry includes both discrete or individual function
devices such as diodes, rectifiers, and transistors; and integrated circuits,
including both monolithic (which may contain hundreds or thousands of tran-
sistors and diodes on a single chip 1/4 inch square) and hybrid integrated
circuits (on which discrete semiconductors have been attached). .Also included
in this industry are such specialized devices as light-emitting diodes (used
II-2
-------
as the display device in many calculators and digital watches) and light
sensitive devices (such as photo diodes). Moreover, there are thousands of
different kinds of diodes, transistors, and integrated circuits within each
of these groups.
"1.3 ORGANIZATION OF PART II OF THIS REPORT
The remainder of Part II of this report is organized in the following
manner. Chapter 2 provides the methodology used to analyze the economic
impacts of the regulations. Chapter 3 describes the semiconductor industry
characteristics and performance in the domestic and international markets.
Chapter 4 provides projections of the growth and performance of the semicon-
ductor industry over the regulatory impact period assuming there are no
additional water pollution control requirements. In Chapter 5 the cost of
pollution control for the Treatment options are presented. Chapter 6 pre-
sents the results of the economic impact assessment. Finally, Chapter 7
discusses the limitations of the economic impact analysis.
II-3
-------
2. STUDY METHODOLOGY
2.1 OVERVIEW
Figure 2-1 shows an overview of the analytical approach used to assess
the economic impacts likely to occur as a result of the costs of each proposed
regulatory option. For the Semiconductor subcategory, four regulatory options
are considered in the economic impact study. The basic approach used in this
study is to (1) develop an operational description of the price and output
behavior of the industry, and (2) assess the likely plant-specific responses
to the incurrence of the compliance costs enumerated in Chapter 5.
The operational description of the price and output behavior is used, in
conjunction with compliance cost estimates supplied by EPA, to determine new
post-compliance industry price and production levels for each regulatory
option and for each of the two semiconductor product groups. Individual plant
data is then analyzed under conditions of the post-compliance industry price
levels, for each regulatory option, to isolate those plants whose production
costs are projected to change significantly more than the estimated change in
their revenues. These identified plants are subjected to a more detailed
financial analysis to determine likely plant closures. The industry
description is then revised, for each regulatory option, to incorporate the
reduced supply into the analysis. Finally, other effects which flow from the
basic price, production, and industry structure changes are determined. These
include employment, community, and foreign trade impacts. Specifically, the
study proceeds in the following steps:
1. Description of industry characteristics
2. Industry supply and demand analysis
3. Analysis of cost of compliance estimates
4. Plant level profitability analysis
I 5. Plant level capital requirements analysis
6. Assessment of plant closure potential
7. Assessment of other impacts
8: Social costs analysis
9. Small business analysis.
II-4
-------
z;: u
g
z -
—• O O
uj< LLI
O 1— U.
U. UJ
a
H
cn
CO
o
ZWl
o>-
— in
I-O
30
O O
a. ce
>-
< z
a.9
UJU
s
§
U
S£
«_j
S<
UJ
at
at —
u vt
<
UJZ
0<
e
a.
en i— in
3 Z =
a uj c ,
5§ S1
UJ
in
- 55
-------
Although each of these steps is described separately in this section, it is
important to realize that there are significant interactions between them, as
shown in Figure 2-1.
2.2 STEP 1: DESCRIPTION OF INDUSTRY CHARACTERISTICS
The first step in the analysis is to develop a description of the basic
industry characteristics that will enable estimation of key parameters which
describe the initial impacts of the regulation. These characteristics, which
include the determinants of demand, market structure, the degree of intra-
industry competition, and financial performance, are described in Chapter 3 of
this report. The resulting observations indicate the type of analytical
methodology needed for the industry.
The sources for this information include government reports, trade
association data, discussions with various trade association representatives
and individuals associated with the industry, and an EPA industry survey.
2.3 STEP 2: SUPPLY-DEMAND ANALYSIS
The purpose of the supply-demand analysis, step 2 of the study approach,
is to determine the likely changes in market prices and industry production
levels resulting from each regulatory option. The estimates of post-
compliance price and output levels are used in the plant-level analysis to
determine post-compliance revenue and profit levels for specific plants in
each product group. If prices are successfully raised without significantly
reducing product demand and companies are able to maintain their current
financial status, the potential for plant closings will be minimal. If prices
cannot be raised to fully recover compliance costs because of the potential
for a significant decline in product demand or because of significant intra-
industry competition, the firms may attempt to maintain their financial status
by closing higher cost/less efficient plants. The supply-demand analysis was
divided into four basic components: determination of industry structure,
projection of possible changes in industry structure to 1984 (the expected
effective date for the proposed regulation), determination of plant- and
II-6
-------
firm-specific operational parameters (e.g., production costs, profit rates,
etc.), and development of price-quantity algorithms.
Short-run pricing behavior depends upon the market structure of the
industry, which can range from competitive, to oligopoly, and to monopoly
situations. Many economic impact studies begin by assuming perfect
competition. However, the product groups covered in this study exhibit some
characteristics that are indicative of non-competitive pricing mechanisms.
The perfectly competitive market structure is one in which there are many
buyers and sellers and the actions of any one of these do not significantly
affect the market. Firms in a competitive market generally earn a "normal"
rate of return on their assets. If it is assumed that (1) the market for a
competitive good is currently at equilibrium, or will be when the regulations
become effective, and (2) firms will attempt to maintain their current
financial status by passing through industry-wide cost increases in the form
of higher prices, the post-compliance equilibrium price and quantity level can
be derived from the interaction of the elasticities of supply and demand.
Because of the existence of speciality markets, non-competitive pricing
behavior might be expected. In this situation, a number of pricing schemes
may be used to characterize the behavior of oligopolistic firms, depending on
specific industry characteristics. The oligopolistic pricing scheme is
applicable for those industry groups which exhibit characteristics of
oligopoly markets, such as the following:
• Few firms in the product group
• High industry concentration
• Low degree of foreign competition (less than 10 percent of U.S. sales)
• Abnormally high profitability
T • Low demand elasticities
• Highly capital intensive
• .Large degree of integration of production, marketing, and distribution
• Large degree of specialized knowledge.
II-7
-------
Industries which exhibit the first three of these characteristics are
those in which the pricing and output actions of one firm will directly affect
those of other firms in the industry. While these conditions do not guarantee
oligopolistic behavior, they are necessary conditions for an oligopoly and
good indicators that one exists. Abnormally high profits in an industry
would, in time, normally attract new entrants to the industry, thereby
increasing price competition. However, very high profits over long periods of
time which are not explained by such factors as excess risk, unusual amounts
of technological innovation, or firm size may be an indicator that an
imperfect market structure exists. Such conditions may occur when entry into
an industry is difficult. The last three of the above points are indicators
of difficulty of entry into the market.
Although the domestic semiconductor industry exhibits some of these
characteristics of non-competitive markets, the specific nature of the pricing
mechanism is not clear from the available evidence. That is, the existence of
specialty markets, low-demand elasticities, and high levels of specialized
knowledge characteristic of this industry are indicative of non-competitive
markets. On the other hand, there are a number of characteristics operating
which are indicative of competitive markets, such as the existence of many
firms in the industry, a significant amount of competition from foreign
producers, and the fact that barriers to entry into the industry are not
particularly severe. Little can be concluded from the four-firm and eight-
firm concentration ratios, because they are in a middle (moderate) range. For
these reasons and because of the dynamic nature of this new industry, the
pricing mechanisms that dominate the industry cannot be specified with
certainty. Instead, it is assumed that semiconductor plants will attempt to
absorb all of the costs of compliance and that in the short to intermediate
period there will be no change in semiconductor prices resulting from the
regulations. This assumption derives from two observations:
• The international competitiveness in the industry has been increasing
in recent years and domestic manufacturers are quite sensitive to this
'fact.
II-8
-------
• If, at any point in time, it were possible to determine profit rates
attributable to different semiconductor product lines within a given
plant, there would probably be a wide variation in the profit rates.
The U.S. semiconductor manufacturers operate in a very competitive world
market. There is significant trade and competition among the western
developed nations for market share. Consequently, U.S. producers cannot
unilaterally attempt to pass additional regulatory costs through to their
customers without losing significant portions of the domestic and inter-
national markets. If they do atttempt to pass through the pollution control
costs, the U.S. products would become less competitive, forcing domestic and
foreign purchasers to buy semiconductor products in foreign markets that do
not have these added costs. As U.S. products became more expensive, imports
of semiconductor products into the U.S. would increase, exports would
decrease, and sales and profitability of U.S. plants would significantly
decrease as U.S. producer prices rose. Since U.S. semiconductor producers are
quite sensitive to these conditions, their most likely option would be to
absorb the regulatory costs in the short and intermediate period and continue
operations if possible.
The existence of differences in profits from one product line to another
within a firm implies the potential for price flexibility. That is, the most
mature product lines operating with narrow profit margins may have some of
their compliance cost burden borne by the most profitable newer products. For
this reason, there generally exists within the industry the capability for
absorbing the cost of the pollution control equipment.
In summary, although the market structure is not precisely determined, it
has been shown that the industry has both the incentive and the capability to
not raise prices in response to small and moderate levels of mandated
pollution control costs. For these reasons, it is assumed that there will be
no price effect resulting from the proposed regulations in the short and
intermediate time period.
The remainder of the economic impact analysis, therefore, focuses on the
extent that plants in the semiconductor industry can remain financially viable
II-9
-------
and competitive if they are required to install pollution control equipment.
The measures of financial viability that are considered relate to the plants'
profitability and their ability to raise the initial investment funds for
pollution control equipment in addition to other capital equipment needed for
expansion, growth, and competitiveness.
Having described the current industry structure, it was necessary to
determine if the key parameters would change significantly by 1984.
Projections of industry conditions begin with a demand forecast. The demand
in 1984 is estimated via trend analysis and market research analysis. An
examination was also made of the factors which might affect the real cost of
manufacturing semiconductors. No reason was found to expect the real price of
these products to increase between now and 1984. It was concluded from the
projections of industry conditions that only minor changes in market structure
would occur in the base case. For this reason, it was concluded that the
market structure previously described can be used to determine price changes
due to the regulation.
The post-compliance market price levels (i.e., zero price increases due
to the regulation) are used in a later step to assess the financial condition
of individual electronic crystals manufacturing facilities.
2.4 STEP 3: COST OF COMPLIANCE ESTIMATES
Investment and annual compliance costs for the recommended treatment
options 1, 2, 3, and 5 were estimated by EPA's Effluent Guidelines Division
for treatment systems of various selected sizes. Based on these cost
estimates, compliance cost curves were developed and then used to estimate
plant-specific compliance costs. A description of the control and treatment
technologies and the rationale behind these compliance cost estimates appear
in chapter 5.
2.5 STEP 4: PLANT LEVEL PROFITABILITY ANALYSIS
The basic measure of financial performance used to assess the impact of
the proposed regulations on the profitabilities of individual plants is return
11-10
-------
on sales (ROS). The use of this technique involves a comparison of the
measure with a critical return.
The return on sales is defined as the ratio of annual profits before
taxes to the total revenues of a plant. The principal virtue of this
technique is its simplicity and its common use in comparative analyses of
operating efficiencies of financial entities. Its principal shortcomings are
that it is based on accounting income rather than cash flow and that it fails
to account for the amount of invested capital and the timing of cash flows,
thereby ignoring the time value of money.
The profit impact assessment is determined by calculating the after-
compliance ROS for each plant and comparing them to a threshold value which
represents the lowest ROS which the industry typically considers to be
acceptable to remain in operation. Plants with after-compliance ROS below the
threshold value are considered potential plant closures. The underlying
assumption is that plants cannot continue to operate as viable concerns if
they are unable to generate a return on sales that is at least equal to the
threshold value. Equation (1) summarizes this technique:
(1) PM-~-£PM^. Significant Profit Impact
P
where
PM » the pre-compliance average annual profit margin (ROS) for the
industry
PM * the lowest ROS at which plants will remain open
L
ACC
— -° * the ratio of the total annual compliance costs (ACC ) to annual
p value of shipments (VS ) for each plant. ^
_ The lowest acceptable PM was determined to be 2.7 percent. This figure
was derived from the observation that it is the lowest annual profit margin
for the semiconductor industry over the 1971-1977 period. The precompliance
profit margin is curved to be 4.5 percent, which is the average for the 7-year
U.S. Department of Commerce, U.S. Semiconductor Industry, September 1979.
11-11
-------
period. Since 2.7 percent is the lowest ROS recorded by operating plants, it
is assumed to be a rather conservative threshold.
The values of shipments for the 21 sample plants were estimated by
multiplying the industry average value of shipments per production employee
by the number of production employees working at the specific sample plant.
The industry average value of shipments per production employee was derived
from the 1977 Census of Manufactures. The value of shipments data along with
the total annual compliance cost estimates are used to analyze the performance
of the plants after compliance.
2.6 STEP 5: CAPITAL REQUIREMENTS ANALYSIS
This step in the analysis assesses the ability of firms to make the
initial capital investment needed to construct and install the required
treatment systems. Some plants which are not initially identified as
potential closures in the profitability analysis may encounter problems
raising the amount of capital required to install the necessary treatment
equipment. The limit on a given firm's ability to raise capital to finance
investment expenditures is quite variable, depending upon factors such as the
firm's capital structure, profitability, future business prospects, the
industry's business climate, the characteristics of the financial markets and
the aggregate economy, and the firm management's relationships with the
financial community. The precise limit, considering all these factors, is
ultimately judgmental. Thus, even given firm-specific data, a limit on a
firm's ability (or willingness) to raise funds for capital investment would be
difficult to estimate. Because firm-specific data for this study is scarce
the problem becomes even more difficult. For these reasons, the analysis of
capital availability focused upon general indicators of the relative burden
created by the compliance expenditures for each semiconductor firm. That is,
to assess the firms' ability to commit the capital necessary to install the
specified pollution control systems, the following ratio is calculated for
each sample plant:
Compliance Capital Investment
Annual New Plant and Equipment Expenditures
11-12
-------
This.ratio reflects the typical level of capital expenditures of the
industry, and is used as a budget standard of the capital expenditure level
that firms in the industry generally hold. Although a specific threshold
value for this parameter is difficult to specify, a value that is useful for
comparison is the average annual value of this ratio for the entire industry.
Although this ratio provides a good indication of the relative burden
created by the compliance requirement, it does not precisely indicate whether
or not firms can afford to make the investments. If, for example, the same
investment requirements were placed on a firm which is already highly
leveraged (as indicated by a high debt/equity ratio) and a firm which is not
leveraged (as indicated by a debt/equity ratio of zero), the highly leveraged
firm is likely to experience the most significant impact. In addition, the
capital requirements must be evaluated together with other factors, such as
profitability. For example, a plant that is extremely profitable would
consider the risk of more leverage or increased cost of capital resulting from
investment more worthwhile than would a less profitable plant.
Because of the absence of plant-specific financial data, the annual
capital expenditures of the plants are inferred from the industry average new
21
capital expenditures per production employee ratios. These annual capital
expenditure estimates, along with capital investment cost for the pollution
control options, are used to indicate the significance of the capital impacts.
2.7 STEP 6: PLANT CLOSURE ANALYSIS
The plant-level analysis examined the individual production units in each
product group to determine the potential for plant closures and profitability
changes. The decision to close a plant, like most major investment decisions,
is ultimately judgmental. This is because the decision involves a wide
21
These ratios are derived from the 1977 Census of Manufactures.
11-13
-------
variety of considerations, many of which cannot be quantified or even
identified. Some of the most important factors are:
• Profitability before and after compliance
• Ability to raise capital
• Market and technological integration
• Market growth rate
• Other pending Federal, state, and local regulations
• Ease of entry into market
• Market share
• Foreign competition
• Substitutability of the product
• Existence of specialty markets.
Many of these factors are highly uncertain, even for the owners of the
plants. However, this analysis was structured to make quantitative estimates
of the first two factors, as described above, and to qualitatively consider
the importance of the others. In this analysis, the first two factors are
given the greatest amount of weight and the importance of the other factors
varies from plant to plant.
2.8 STEP 7: OTHER IMPACTS
"Other impacts" include economic impacts which flow from the basic price,
production, and plant-level profitability changes. These include impacts on
employment, communities, industry structure, and balance of trade.
The estimate of employment effects flows directly from the outputs of the
industry-level analysis and the plant closure analysis. The algorithms used
are:
direct » employment at + ( S ) (S/employee)
employment closed facilities
S * change in revenues at the remaining plants, which is derived from
the microeconomic and plant level analyses (in this case S * 0).
S/employee ** baseline revenue produced per employee„
11-14
-------
Employment estimates for the production facilities projected to close are
available from the EPA 308 Survey.
Community impacts result primarily from employment impacts. The critical
variable is the ratio of semiconductor industry unemployment to total employ-
ment in the community. Data on community employment are available through the
Bureau of the Census and the Bureau of Labor Statistics.
The assessment of industry structure changes is based on examination of
the following before and after compliance with the regulation:
• Numbers of firms and plants
• Industry concentration ratios
• Variance of average total production cost per unit among plants
• Effects of plant closures on specialty markets.
Decreases in the first two factors would indicate an increase in industry
concentration and may change the pricing behavior of the industry. Such
potential changes were qualitatively evaluated. An increase in variability of
average costs would indicate that some firms have become more competitive than
others, as a result of the regulation. The long-term implications of such
developments were examined.
Imports and exports are important determinants of pricing behavior in the
Electronic Crystals industry. The role of these variables is qualitatively
evaluated in Chapter 3 of this report. Basically, the threat of imports
appears to deter domestic producers from passing through compliance costs.
2.9 STEP 8: SOCIAL COST ANALYSIS
This analysis assesses the total social costs that can be associated with
the~EPA effluent regulations. The social costs measure the value of goods and
services lost by society due to a given regulatory action. These costs
generally include the use of resources needed to comply with a regulation, the
use of resources to implement and enforce a regulation, plus the value of the
output that is forgone because of a regulation..
11-15
-------
The partial equilibrium analytical framework is conceptually the most
practical means for estimating total social costs. This framework, in its
most sophisticated form, is based on an analysis of supply and demand rela-
tionships in the directly-affected markets. When an industry is regulated,
compliance requirements result in increased unit costs of production. This
increase, in turn, leads to an upward shift in the industry's supply curve.
The supply curve shift normally results in higher prices and a lower
production level. Compliance costs, production losses, and net welfare losses
incurred by producers and consumers due to decreased output are measurable
within this framework. There are other costs that are not measurable within
this framework. Costs of implementing and enforcing a regulation must be
added. Also, other social costs do not appear in this static analysis, such
as productivity effects, innovation impacts, and costs of reallocating
resources that become unemployed. Unfortunately, the data does not exist to
carry out such analysis at this time, and a compromise which captures the
major costs to society was performed.
For this analysis only the real resource costs are considered. This
provides a reasonable estimate of social costs, since EPA studies show that
95 percent or more of the social costs are directly related to compliance
expenditures by the regulated entities. Consequently, the present value of
social costs (PVSC) of regulations can be approximated by the following
equation:
PVSC - I + (OM/.l) r~n
where: PVSC = present value of social costs
I = investment cost
OM • annual operating and maintenance cost
r * real discount rate
n * number of years between now and the year the investment is
incurred (1984; n»2)
The above equation assumes that:
• -The regulations will be in effect in perpetuity
* Operating and maintenance costs will be incurred in the first year of
investment
11-16
-------
2.10 STEP 9: SMALL BUSINESS ANALYSIS
The Regulatory Flexibility Act (RFA) requires Federal regulatory agencies
to consider small entities throughout the regulatory process. This analysis
addresses these objectives by identifying and evaluating the economic impacts
that are likely to result from the promulgation of BPT, BCT, BATEA, NSPS, PSES
and PSNS regulations on small business in the semiconductor manufacturing
industry. The primary economic variables covered are those analyzed in the
general economic impact analysis such as plant financial performance, plant
closures, and unemployment and community impacts. Most of the information and
analytical techniques in the small business analysis are drawn from the
general economic impact analysis which is described above and in the remainder
of this report. The specific conditions of small firms are evaluated against
the background of general conditions in semiconductor markets.
A specific problem in the methodology was developing an acceptable
definition of small entities. The Small Business Administration (SBA) defines
small entities in SIC 3674 (Semiconductors and Related Devices) and SIC 3679
(Electronic Components, n.e.c.) as firms of fewer than 500 employees. The SBA
definition was found to be inappropriate as a basis for defining small
entities in the semiconductor manufacturing industry for purposes of
developing water pollution regulations. Instead, a definition was sought
which would account for firm size in comparison to total industry size and in
comparison to unit compliance costs (unit compliance costs increase
significantly in reverse proportion to plant size). Moreover, since the
available data on compliance cost and production were on a plant basis, the
individual production facility, rather than the firm, was used as the basis
for the analysis.
Alternative definitions for "small" semiconductor manufacturing plants
were selected for examination. Number of production employees was the primary
variable used to distinguish plant size. This is because plant level
employment data were considered most reliable.
The impacts on small plants under each definition were assessed by
examining the distribution of semiconductor plants by plant employee size,
wastewater volumes, compliance costs and potential closures from regulations.
11-17
-------
3. INDUSTRY DESCRIPTION
This chapter provides a summary of the economic and financial characteris-
tics of the semiconductor industry. It focuses on the determinants of demand
for semiconductors, industry structure, performance, and the financial condition
of plants and firms that manufacture semiconductor devices.
3d MARKET CHARACTERISTICS
3.1.1 Major Product Groups and Trends
In 1979, the value of shipments for all semiconductor devices in the U.S.
domestic industry was $6.9 billion (Table 3-1). This represented an increase
of 20 percent over the 1978 level of shipments. This rapid growth, however,
has not been the same for all classes of semiconductor devices over time.
During the 1970s, the semiconductor industry grew at a compound annual rate
of about 19 percent, which was substantially higher than the 12 percent annual
growth experienced during the 1960s. Moreover, the integrated circuit ship-
ments, which in 1979 were $5.1 billion, have been growing at about 28 percent
annually. This industry segment's rapid growth, however, has been at the
expense of other semiconductor devices, such as discrete resistors and capaci-
tators, which it has displaced in many types of equipment. The recent penetra-
tion of integrated circuits into the semiconductor market is reflected in the
fact that while integrated circuits accounted for 58 percent of the domestic
market share in 1974, they represented 73 percent by 1979. In contrast, the
discrete devices segment of the market has been increasing at a much slower
rate and has been losing its market share over the years. In 1974 the market
share of this industry segment was 42 percent, but by 1979 it dropped to 27
percent of total industry shipments. If new applications for discrete devices,
such as those used in automotive engines, had not been introduced, the decreases
in market share for this industry could have been even greater.
11-18
-------
TABLE 3-1. VALUE OF SHIPMENTS OF THE U.S. SEMICONDUCTOR INDUSTRY
1974 TO 1979
($ Millions)
YEAR
1974
1975
1976
1977
1978
1979
SEMICONDUCTOR PRODUCTS
DISCRETE DEVICES
1,519
1,320
1,532
1,823
1,880
1,900
INTEGRATED CIRCUITS
2,056
1,718
2,546
3,009
3,950
5,090
TOTAL SEMICONDUCTORS
3,575
3,038
4,078
4,832
5,830
6,990
SOURCE: U.S. Department of Commerce/Industry and Trade Administration, 1980,
U.S. Industrial Outlook for 200 Industries with Projections for 1984,
January 1980, p. 268.
11-19
-------
3.1.2 End-Use Markets
Since the invention of semiconductor devices in the late 1950s, their
.uses have pervaded almost every sector of our society. The principal end-uses
of semiconductor devices can be grouped into four major market categories:
• Computer
• Industrial
• Consumer
• Government, including military/space.
The growth of the semiconductor industry depends on the performance of each of
these end-use markets. The semiconductor industry end-use markets, with few
exceptions, have been outperforming the general economy in recent years. The
relative strength of these end-use markets is derived from the integrated and
pervasive role which electronic components (specifically semiconductors) play
in all sectors of the economy. These end-use markets are expected to continue
to grow in the future.
The relative growth of these four markets is shown in Table 3-2. Each of
these markets has experienced different historical development patterns due to
the relative trade-offs between performance, reliability, and cost of semicon-
ductors in product applications.^.' For example, component costs are far more
important in highly competitive consumer product industries than in other mar-
kets. Computer and industrial equipment manufacturers require high-performance
and reliable components, and price is of lesser importance than in consumer
products. The military/space market has very high performance and reliability
requirements and is willing to pay for the development of premium products.
Semiconductor devices used in military and space applications must be capable
of operating under extreme weather and other adverse conditions. This market
also requires specialized, high-performance semiconductors, such as infrared
£/ U.S. Department of Commerce, A Report of the U.S. Semiconductor Industry,
September 1979.
11-20
-------
TABLE 3-2. DISTRIBUTION OF SEMICONDUCTOR PRODUCTS BY MAJOR END-USERS
(1973 - 1975)
End-Users
Computer
Industrial
Consumer
Government , inc .
military/space
TOTAL
1973
C$mTT.
$590
605
510
378
$2,083
Z of
Total
)
28
29
25
18
Too"
1974
$855
775
575
420
$2,625
Z of
Total
32
30
22
16
Too
1975
C$mTT.
$1,005
905
715
450
$3,075
Z of
Total
33
29
23
15
100
Percent Change
72/73
66
54
72
26
~55"
73/74
45
28
13
11
~~26
74/75
18
17
24
7
"IT
SOURCE: Fairchild Camera and Instrument Corp., cited in "Coleman and Company
Electronics Letter," B. M. Rosen, March 8, 1974.
11-21
-------
detectors, radiation detection devices, and light-sensing devices, which have
very limited civilian applications. Because they are spread over a small pro-
duction volume, unit development and production costs are high.
3.2 INDUSTRY STRUCTURE
3.2.1 Plant Characteristics
The 1977 Census of Manufacturers reported 545 establishments whose primary
activity is the manufacture of semiconductors.£/ In that year, these plants
shipped over $5.3 billion worth of products. Table 3-3 illustrates the distri-
bution, the number of plants, and the value of shipments for plants in the semi-
conductor industry by plant size, in terms of employees. This table indicates
that the industry is composed of a disproportionately large number of very small
and medium-sized plants, and the number of these plants has been increasing at
a faster rate than large plants. The information shown in the table also indi-
cates that the plants in the semiconductor industry are highly specialized.
That is, approximately 94 percent of their product shipments were semiconductor
devices in 1977. Furthermore, the average coverage ratio for the industry is
high. In 1977, 545 plants in the industry manufactured 93 percent of the
products classified under SIC 3674.
As shown in Table 3-4, the plants of the U.S. domestic semiconductor indus-
try are concentrated in a few states. In 1977, approximately a third of the
plants were located in California, and over two-thirds were located in six
states: California, New York, Massachusetts, New Jersey, Pennsylvania, and
Texas. Most of the largest semiconductor companies have several domestic
plants (usually in the same region of the country) and several foreign plants.
la many cases, increases in the number of plants in a particular firm have been
the result of growth in the industry's market and limitations in adjacent land
needed to enlarge the size of existing plants. This phenomenon is especially
A significant number of these plants, especially the smaller ones, may be
R&D and other facilities that are not necessarily involved in wafer manufac-
turing and are not covered by the proposed regulation.
11-22
-------
TABLE 3-3. DISTRIBUTION OF SEMICONDUCTOR PLANTS AND VALUE OF SHIPMENTS
BY SIZE OF PLANT
(Selected Years)
PLANT SIZE
Establishments with:
1-99 employees1*
100 - 499 employees
500 - 999 employees
Over 1,000 employees
Total
Coverage Ratio0
Specialization Ratio**
SEMICONDUCTOR INDUSTRY (SIC 3674)
PLANTS
(No.)
1972 1977
225 414
65 90
17 19
18 22
325 545
a
PLANTS
(Percentage)
1972 1977
69.0 76.0
20.0 16.5
5.2 3.5
5.5 4.0
100.0 100.0
VALUE OF
SHIPMENTS
(Millions)
1972 1977
263.3
860.0
529.5
3696.8
5349.6
VALUE OF a
SHIPMENTS
(Percentage)
1972 1977
4.9 4.4
14.4 16.2
10.4 10.0
70.3 69.4
100.0 100.0
92.0 94.0
89.0 93.0
Percentage of total sampled plants within each employment category of the total
plants.
A number of these plants, especially the smaller ones, may be R&D and other
facilities that are not necessarily involved in wafer manufacturing.
Coverage ratio is the ratio of a given industry's primary product shipments to
total shipments of this product by all industries.
Specialization ratio is the ratio of primary product shipments to total
primary and secondary product shipments.
SOURCE: U.S. Department of Commerce, Bureau of Census, Census of Manufactures,
1977, Industry Series; Electronic Components and Accessories, MC77-1-36E
(Washington, D.C. 1980).
11-23
-------
TABLE 3-4. GEOGRAPHICAL DISTRIBUTION OF DOMESTIC SEMICONDUCTOR PLANTS IN 1977a
STATE
New England Division:
Maine
Vermont
Massachusetts
Rhode Island
Connecticut
Middle Atlantic Division:
New York
New Jersey
Pennsylvania
East North Central Division:
Ohio
Indiana
West North Central Division:
Minnesota
Missouri
South Atlantic Division:
Florida
East South Central Division:
Tennessee
West South Central Division:
Oklahoma
Texas
Mountain Division:
Idaho
Colorado
Arizona
Utah
Pacific Divison:
Oregon
California
Total U.S.
NUMBER OF
ESTABLISHMENTS
3
2
46
4
13
59
34
31
13
5
3
4
19
1
3
36
1
8
16
3
4
180
545
NUMBER OF
ESTABLISHMENTS
WITH 20 OR MORE
EMPLOYEES
1
1
23
2
6
22
11
16
5
2
1
1
9
1
1
16
1
3
9
1
2
80
219
•
«
aThe totals do not add up because of lack of state-specific information.
SOURCE: U.S. Bureau of the Census, Census of Manufactures, 1977, Industry Series;
Electronic Components and Accessories, MC77-1-36E (Wa8Jiin£tjon,_D..C..j 1980),
11-24
-------
true in the northern California area where the scarcity of large industrial
tracts has influenced companies to build many plants on different sites. In
some cases semiconductor companies have also constructed plants that specialize
in the production of specific products. For example, a semiconductor company
may have an integrated circuit plant and a separate transistor and diode plant,
or separate MOS and bipolar integrated circuit plants.
3.2.2 Industry Concentration
The semiconductor industry has shown a consistent level of market share
concentration over time. In 1972, the four largest manufacturers of semi-
conductors provided half of the industry's total value of shipments, only 1
percent lower than in 1957. Table 3-5 indicates that firms that produce
transistors are more highly concentrated than firms in other semiconductor
product groups. In 1972, the concentration ratio for the largest four com-
panies producing transistor products was 68 percent, as opposed to 57 percent
for firms in the integrated circuits, and 45 percent for those in the diodes
and rectifiers product groups.
3.3 FOREIGN TRADE
The market for semiconductors is worldwide. The United States is a
significant exporter and importer of semiconductor devices. It is important
to note that most large and many of the smaller U.S. semiconductor companies
conduct the design and wafer fabrication processes in the United States and
then ship the chips abroad to Southeast Asia or Mexico for wire bonding and
assembly. After assembly, the devices are imported back into the United
States subject to special-U.S. tariff provisions. These tariff provisions
specify that goods shipped abroad for assembly and then shipped back to the
United States shall be subject to duty only on the value added abroad.
11-25
-------
TABLE 3-5. CONCENTRATION RATIOS OF U.S. DOMESTIC SEMICONDUCTOR SHIPMENTS
1957, 1965, AND 1972
(PERCENT OF TOTAL U.S. DOMESTIC SHIPMENTS)8
PERCENT OF TOTAL U.S. SHIPMENTS
NUMBER OF COMPANIES 1957 1965 1972
All semiconductors
4 largest companies 51 50 50
8 largest companies 71 77 66
20 largest companies 97 90 81
50 largest companies 100 96 96
Integrated circuits
4 largest companies 69 57
8 largest companies 91 73
20 largest companies 99 91
50 largest companies 100 100
Transistors
4 largest companies 51 68
8 largest companies 83 84
20 largest companies 95 98
50 largest companies 100 100
Diodes and Rectifiers
4 largest companies 38 45
8 largest companies 56 61
20 largest companies 85 88
50 largest companies 97 100
a
Data include value of production in captive facilities.
SOURCE: Compiled from data collected in the Department of Commerce, Quarterly
Survey of Production Capabilities for Electronic Parts.
11-26
-------
Japan is one of our most important trading partners for finished semicon-
ductors, particularly for integrated circuits. Its growing equipment and com-
ponents industries have become a major market for U.S. semiconductors and also
a major supplier of semiconductor containing products to this country.
Table 3-6 lists exports, imports, and domestic shipments of the semicon-
ductor industry from 1967 to 1977 under these provisions. This table shows
that over this period the balance of trade in semiconductor products has been
positive, ranging from a low of $109 million in 1967 to $293 million in 1976.
From 1967 to 1977, the U.S. export of all semiconductor products increased
at a compound annual rate of nearly 26 percent. Meanwhile, imports grew at
a faster rate (41 percent), but still remained below the level of exports by
$151 million in 1977.
In recent times, the Japanese have made significant inroads into the high-
quality and technologically advanced markets of the semiconductor industry.
They have already captured 40 percent of the market for the 16K chip, and are
considered the leaders in the new 64K RAM market ..£/ Key factors that contri-
buted to the Japanese technological advancements in these areas, which were
traditionally dominated by U.S. firms, are the Japanese's high level of manu-
facturing discipline, team effort, and long-range planning capabilities. If
these trends continue, the balance of trade position for semiconductor products
will deteriorate as U.S. electronic manufacturers use increasingly more of the
Japanese low-priced, high-quality semiconductor products.
3.4 PRICING BEHAVIOR AND TRENDS
Since their introduction in the marketplace, there has been a downward
tjrend in the prices of semiconductor devices. This downward trend is illus-
trated in Table 3-7, which shows price indexes for major groups of semicon-
ductor products over the 1975-1980 period. This table indicates that the
*J Fortune, Japan's Ominous Chip Victory, December 14, 1981, pp. 52-57.
11-27
-------
ON
r-t
I
ON
CO
OCi
o
§
Z ^x
O ca
O hi
M CO
IE i-i
*3 i-l
CO O
o
fc
Q CO
2S §
H -^
t-t
Z -4
fc
CO
vO
en
W
•J
a
p-«.
s
SO
1*^
s
m
f»^
Os
-
^
^K
ON
en
PS*
a\
cs
r+*
2
1-1
P»»
1-1
o
r»
Ov
i— i
Ov
NO
ON
i— i
00
vO
ON
^
NO
ON
W
— 1
^
I***
CN
A
1-4
4J
C
01
hJ
CO
a
CL
e
o
•H
^ i
a
e
CO
C
O
u
r-4
en
CN
en
m
en
^.
en
(^
CN
m
CN
m
CN
^
CN
r-4
CN
^
i— i
i_1
r-4
»•<
CO
CO
CO
4-1
hi
o
a
X
w
CO
4J
e
O)
e
a
CO
U-l
o
ON
CN
00
CN
00
CN
ON
CN
i— I
CN
ON
r-4
^
(— 1
r-4
r—4
^
sO
en
O
K
CO
ce
CO
4J
hi
O
a
£3
M
C
o
•H
4j ±j
C D.
o> e
hi 3
CO CO
a c
a o
-
•-) O O
e o
0) C •
oo •!-( r-»
O
0) 6000
t3 C
3 -H -o
i— 1 CO C
U CO CO
C 0)
•r-4 y o
O en
CO hi •
1 1 f\n ^^
CO O
^0 ^ 00
0)
0) J= CO
•o 4J e
CO hi 01
hi 3 4J
e hi •<
60 O CO
0) >xCO
hi 1-1 H
O .O.
M-l g hi
oi a>
" 0) 00 "O
.C co C
H « 3
CO
C
0
hi
u
1— 1
w
hi r-°
O r-
u-i ON
fmmt
3 a>
•a co
o
hi A
P-t CO
4J
U-l hi
0 0
ex
>v 0>
0* P£
>
3 CO
CO •-<
hi
CJ 4J
O CO
CQ 3
T>
•• C
CO M
4-1
3 J->
O C
iJ ^
C hi
•H hi
hi 3
(L, U
co «
4-J 3
CO CO
O 0)
hi
CU 3
T3 po
CO
hi CO
H 3
CO
c c
60 01
•f«4 ^J
3 CO
CO 4-1
C hi
V CO
U Pu
-*•-
w
O
oci
a
o
CO
11-28
-------
TABLE 3-7. PRICE INDEXES FOR SEMICONDUCTOR DEVICES
(1975 - 1980)
YEAR
1975
1976
1977
1978
1979
1980
INTEGRATED
MICRO CIRCUITS
(36741)3
100.0
93.3
82.5
69.3
65.8
71.1
TRANSISTORS
(36742)a
100.0
97.8
94.0
91.7
90.8
98.2
DIODES AND
RECTIFIERS
(36743)a
100.0
102.3
103.4
101.6
101.4
102.2
OTHER SEMI-
CONDUCTOR
DEVICES
(36749)a
b
100.0
90.9
85.8
85.8
86.1
1975 » base year
>
1976 » base year
SOURCE: Bureau of Labor Statistics, Producers Price Indexes (selected years)
11-29
-------
prices of three of the four product groups declined over this period. The
only price increases are in the discrete segment of the industry (diodes and
rectifiers) which represents a much slower growing and mature segment of the
semiconductor industry.
There are several reasons for the decline in prices of semiconductor
devices over the years. Some of the major factors contributing to this
phenomenon are the following:
• The life cycle of most semiconductor products is short.
The products pass from sales of prototypes through market
expansion to a technologically mature product in only a
few years. Price reductions are generally very rapid
during the market expansion phase. The prices tend to
fall most rapidly on those new devices for which there has
been new entry by several firms or for new devices which
can be easily imitated.
• Competition and trade in both the domestic and international
markets are strong. As a result, high profits on new products
generally lead to new entry by firms that, in turn, cause
prices to fall. Semiconductor firms also recognize that large
price drops may slow or prevent new firms from entering the
industry, and they use price cuts to retain or increase their
market share.
• In addition, the learning or experience curve appears to
have a strong influence on the downward trend in prices
of this industry. The experience curve theory implies
that as the cumulative output increases by a fixed per-
centage, prices also tend to fall by some fixed percentage.
An example of the learning curve for integrated circuits
from 1964 through 1975 is depicted in Figure 3-1. This
figure indicates that for each doubling of cumulative unit
output, the selling price in constant dollars has decreased
by about 27 percent.
- It should be noted that these factors may impede the industry's ability or
inclination to pass through the costs of pollution control to their customers
by rarsing prices. The extent to which some of the costs of pollution control
could be passed through to customers would be observed in the form of slower
declines in the prices of semiconductor products.
11-30
-------
5100.00
JS
C. learmnj Cjne 5!oo«
From 125* tfsrouzn 19?:. eaca
doubling of cumulative units
result! m i 27.3 * dsciine in
constant dollar a»«ra{! pries.)
._ .1
100
Cumulative Unit Volume
(milliuns of circuits)
1,000
10,000
Integrated Circuit Learning Curve Data
Millions of I.e. Units
Year
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
Annual
2
10
29
69
133
253
299
362
602
1,060
1,477
1,375'
Cumulative
2
12
41
110
243
496
795
1,156
1,759
2,819
4,296
5,571 -
Annual
Value
(Saillior.s)
41
79
149
223
303
413
433
443
608
1,049
1,359
1,100
Average Selling Price
Current
Dollars
•513.
8.
5.
3.
2.
1.
1
1.23
0.
0.
0.
50
33
05
32
28
63
45
23
01
99
a?
Constant '72
Dollars
324.54
11.21
,58
.20
,75
,88
Scurce:
"Morran Scar.ley Ileccrc:
3er.;a=ir. M. P.cser., Zdit
r.ics Letter" October 15, 1976,
6.
4.
2.
1.
1.59
1.23
1.01
0.94
0.79
0.52
FIGURE 3-1. INTEGRATED CIRCUIT LEARNING CURVE
11-31
-------
4. BASE CASE PROJECTIONS
This section provides projections of the growth and performance of the U.S.
semiconductor industry over the regulatory impact analysis period in the absence
of water pollution control requirements resulting from the Clean Water Act. The
baseline projections in this report provide a general point of reference for the
analysis and are not intended to be a comprehensive, authoritative forecast of
future industry conditions. When the impact estimates (based on the estimated
compliance costs) are compared against the baseline, the difference represents
the estimated impact resulting from these regulations. The primary variables of
interest are the projected sales, employment, and capital requirements.
4.1 U.S. SEMICONDUCTOR SALES PROJECTIONS
The world demand for semiconductors quadrupled during the 1970's and
industry sources have projected world demand to increase fivefold during the
1980's reaching $60 billion in sales by 1990.JJ/ U.S. shipments of semicon-
ductors have grown from $1.7 billion in 1970 to $6.6 billion in 1979, an
average of 14 percent annually. U.S. shipments are expected to continue to
grow rapidly and are projected to be $15.7 billion by 1983.27 However, the
U.S. share of the world semiconductor market has been declining steadily,
down from 88 percent in 1963 to 63 percent in 1977. Assuming that this trend
continues, the U.S. share of the world market is estimated to be 46 percent
in 1990, representing $27.6 billion in sales.
Table 4-1 summarizes the forecasts of U.S. semiconductor shipments. It
indicates that U.S. shipments will increase rapidly through 1983 and then
average about 8 percent annual growth for the rest of the decade.
Circuit - Semiconductor Industry Association Newsletter, April Issue 1980.
I/Ibid.
11-32
-------
TABLE 4-1 U.S. SEMICONDUCTOR SHIPMENT FORECASTS ($ MILLIONS)
YEAR
1979
1980a
1981a
1982a
1983a
1990b
VALUE OF
SHIPMENTS
6,628
8,532
9,742
12,237
15,690
27,600
AVERAGE ANNUAL
GROWTH RATE (%)
29
14
26
28
8
aSIA projections from the SIA Circuit. Semiconductor Industry Association
Newsletter, April Issue 1980.
demand in 1990 is projected to reach $60 billion by the Semiconductor
Industry Association and U.S. market share is estimated by SIA to be 46
percent.
11-33
-------
4.2 CAPITAL EXPENDITURES PROJECTIONS
Capital investment requirements are generally high in the U.S. semicon-
ductor industry compared to other industries. During the 1973-1977 historical
period, the capital expenditures in this industry increased approximately
eightfold from $52.4 to $409 million, and ranged from 6 percent to 12 percent
of the industry's value of shipments. The high level of capital expenditures
over this period is due mainly to rapid changes in both product and production
technology which results in a high rate of equipment obsolescence. In order
to maintain market share and growth semiconductor firms have had to invest
continually in new equipment and technology. This trend is expected to con-
tinue over the regulatory impact period.
Capital expenditures are projected based on the average 1963-1977 his-
torical capital expenditures/value of shipment ratio which equals .083. Using
this historical relationship the estimate for new capital expenditures require-
ments would be $1.3 billion in 1983, and $2.2 billion, in 1990 (Table 4-2).
4.3 EMPLOYMENT PROJECTIONS
In the past, employment in the semiconductor industry has been increasing
at a slower rate than the value of shipments. Table 4-3 shows that between
1972 and 1977, employment increased by 17 percent, while value of shipments
nearly doubled. The lower employment growth rate was due to improved produc-
tivity which was a result of a combination of factors, including technological
advancement and increasing transfer of assembly and other production activities
to U.S.-owned offshore facilities to take advantage of lower labor costs. It
is expected that these productivity increases, measured in terms of value of
shipment per employee, will continue as demonstrated in the past. The ratio
of value of shipments to employees was projected over the regulatory impact
period based on the historical trend of these variables. Using the forecasts
of value of shipments (calculated above) and the time trend projections of
11-34
-------
TABLE 4-2. NEW CAPITAL EXPENDITURES FOR THE SEMICONDUCTOR
INDUSTRY, HISTORICAL3 AND FORECASTb VALUES
YEAR
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1980
1983
1990
EXPENDITURES
($ MILLIONS)
52.4
43.4
61.3
123.6
131.1
118.4
143.6
98.4
94.3
170.3
334.6
479.2
282.9
362.3
409.0
708.2
1,302.3
2,290.8
PERCENT OF VALUE OF
INDUSTRY SHIPMENTS
7.6
6.0
6.7
11.0
11.5
9.0
9.1
6.3
5.9
6.3
9.2
11.1
8.6
8.1
8.3
8.3
8.3
8.3
Historical data on capital expenditures obtained from the U.S. Department
of Commerce, Annual Survey of Manufacturers, 1977.
^Projection based on the average historical capital expenditure to value of
shipments relationship.
11-35
-------
TABLE 4-3. SEMICONDUCTOR INDUSTRY EMPLOYMENT FORECASTS
YEAR
1972
1973
1974
1975
1976
1977
1980P
1983P
1990P
EMPLOYMENT
(000)
97.6
120.0
133.1
96.7
102.7
114.0
149.0
228.1
287.2
VALUE OF
SHIPMENTS
($ Millions)
2,704.8
3,647.7
4,305.1
3,276.9
4,473.8
5,322.6
8,532
15,690
27,600
VALUE OF SHIPMENTS/
EMPLOYMENT
($000/Employee)
27.7
30.4
32.3
33.9
43.6
46.7
57. 2a
68. 8a
96. la
P * Projected
aProjected based on 1972-1977 historical trend
SOURCE: U.S. Department of Commerce, 1977 Preliminary Census of Manufactures
11-36
-------
the employee/value of shipment ratios, the level of employment in the semicon-
ductor industry was calculated for this 1980-1990 period. These projections
are also shown in Table 2-3. Total employment is projected to be 228,100 in
1983 and as high as 287,000 by 1990.
11-37
-------
5. COST OF COMPLIANCE
5.I OVERVIEW
The water treatment control systems, costs, and effluent limitations
recommended for the semiconductor industry are enumerated in the Development
Document for Effluent Limitation Guidelines and Standards of Performance for
the Electrical and Electronic Components Point Source Category. That document
identifies various characteristics of the industry including manufacturing
processes, products manufactured, volume of output, raw waste characteristics,
supply, volume and discharge destination of water used in the production pro-
cesses, sources of waste and wastewaters, and the constituents of wastewaters.
Using that data, pollutant parameters requiring effluent limitations or stand-
ards of performance were selected by EPA, and the costs of the treatment con-
trol systems to achieve certain standards of performance were estimated.
The EPA Development Document identifies and assesses the range of con-
trol and treatment technologies which apply to Semiconductor industry wastes.
This assessment involved an evaluation of both in-plant and end-of-pipe tech-
nologies that could be designed for this subcategory. This information was
then evaluated for existing direct industrial dischargers to determine the
effluent limitations achievable based on the "best practicable control tech-
nology currently available" (BPT), and the "best available technology economi-
cally achievable" (BAT). Similar evaluations were performed for new direct
dischargers to develop new source performance standards (NSPS). Finally,
pretreatment standards for- existing sources (PSES) and pretreatment standards
for new sources (PSNS) were developed for dischargers to publicly owned treat-
ment works (POTW). Each of the technologies identified was analyzed to calcu-
late cost and peformance. Cost data was expressed in terms of investment,
operating and maintenance costs, depreciation, and interest expense. Pollution
characteristics were expressed in terms of median and mean concentration levels
(per liter of water).
11-38
-------
5.2 POLLUTANT PARAMETERS
5.2.1 Pollution Parameters Analyzed
The selection of pollution parameters to be considered for regulation was
based primarily on laboratory analyses of wastewaters sampled from several semi-
conductor plants and responses to a mail survey of semiconductor manufacturers.
The specific approach to selecting the pollutant parameters is presented in
Sections 5 and 6 of the Development Document. The chemical pollutants analyzed
fall into three groups:
• Conventional
• Toxics
• Nonconventional.
Conventional pollutants are those generally treatable by secondary municipal
waste water treatment facilities. The toxic pollutants are comprised of 129
chemicals that are identified as priority pollutants. The nonconventional
pollutants are those which are neither conventional nor are toxic pollutants.
5.2.2 Pollutants to be Regulated
The specific pollutants selected for regulation in the semiconductor sub-
category are pH, fluoride, and total toxic organics (TTO). These pollutants
are commonly generated during the manufacture of semiconductor products.
The first pollutant parameter to be regulated, pH, is a measure of acid-
ity or alkalinity of a wastewater stream. The term pH is used to describe
the hydronium ion balance in water. During semiconductor manufacture, alka-
line wastes result from alkaline cleaning solutions. Acid wastes occur from
the use of acids for cleaning and etching operations. The pH in the raw waste
can range from 1.1 to 11.9 from these operations.
11-39
-------
Fluoride wastes result from the use of hydrofluoric acid as an etchant
in providing proper surface texture for application of other materials and
creating depressions for dopants in device manufacture. Fluoride concentra-
tions have been observed as high as 50,000 milligrams per liter in raw wastes
from semiconductor manufacture. Although not identified as a "toxic pollu-
tant," fluorides in sufficient quantifies are toxic and can be lethal to
humans. Fluorides may also be harmful in certain industries, particularly
those involved in the production of food, beverages, Pharmaceuticals, and
medicines. Moreover, fluorides found in irrigation waters in high concentra-
tions have caused damage to certain plants exposed to these waters and chronic
fluoride poisoning of livestock has been observed.
Toxic organic pollutants were frequently found in wastewaters from
semiconductor facilities. The sources of these organics are solvent cleaning
operations and possibly oil contaminants. The high concentrations observed
indicate probable dumping of solvent cleaning baths.
Because of the wide variety of solvents used in the manufacture of semi-
conductors and electronic crystals, and the subsequent large number of toxic
organics found in process wastewaters, the Agency is proposing that total
toxic organics (TTO) be used as the pollutant parameter for discharge limita-
tions and that TTO be defined as the sum of all toxic organics found at
greater than 0.01 milligrams per liter. This recommendation is based on
the fact that solvent discharges can be reduced to a minimum with good house-
keeping practices and solvent management techniques.
5.3 RECOMMENDED TREATMENT TECHNOLOGIES
_ Based on the analysis of the potential pollution parameters and treatment-
in-place in the semiconductor industry, EPA identified six treatment technolo-
gies that are most applicable for the reduction of the selected pollutants.
These treatment technologies are described in detail in Section 7 of the Draft
Development Document and are listed below.
11-40
-------
• Option 1:
• Option 2:
• Option 3:
• Option 4:
• Option 5:
• Option 6:
Segregation and collection of spent solvents
containing toxic organics for reuse, resale,
or contract hauling (referred to as solvent
management), plus end-of-pipe treatment for
pH control.
Option 1 plus end-of-pipe precipitation/
clarification for control of fluoride, arsenic,
and suspended solids.
Option 1 plus precipitation/clarification of
concentrated fluoride wastes for control of
fluoride.
Option 2 plus 75% recycle of treated effluent
to further reduce pollutant discharges.
Option 2 plus filtration to further reduce
pollutant discharges.
Option 5 plus activated carbon to further
reduce toxic organics.
The Agency's evaluation of treatment options 4 and 6 concluded that these
technologies would not be technically feasible for all semiconductor plants.
For this reason, the economic impact analysis concentrated on treatment options
1, 2, 3, and 5 only.
5.4 TREATMENT COST ESTIMATES
The treatment costs were estimated by EPA's Effluent Guidelines Division
based on model plants which were structured to resemble the types and capaci-
ties of waste treatment facilities needed for each product subcategory. Data
are based on plant visits and contacts with industry to verify treatment prac-
tices and to obtain data on size, wastewater flow, and solid waste disposal
s-ystems. The model plants characteristics were selected to reflect differences
in treatment capacities and the existing range of flows and concentrations of
pollutants.
11-41
-------
Although the model plants are believed to be typical of plants of their
type, the costs which will actually be incurred by a specific plant may vary
from that presented in the cost estimate, because of differences in pollutant
concentrations and specific site conditions from one plant to another.
Examples of specific conditions which may differ are piping lengths, climate,
land availability, water and power supply, the location of the point of final
discharge, and solid waste disposal options.
In developing the compliance cost estimates for recommended treatment
options, the following major assumptions were made by EPA:
• All costs are expressed in end-of-year 1979 dollars.
• The treatment facilities were assumed to operate 8 hours
per day, 260 days per year for plants with discharge
less than 60,000 gpd; 24 hours per day, 260 days per
year for plants with 60,000 gpd to 200,000 gpd; and
24 hours per day, 350 days per year for plants with
more than 200,000 gpd.
• Labor costs are based on an hourly rate of $20, including
fringe benefits and plant overhead.
• The cost of land is valued at $12,000 per acre.
• Energy costs are based on $306 per horsepower.
• Sludge disosal costs are included. Available informa-
tion indicates that the sludge to be disposed is not
defined as hazardous waste by RCRA.
• Capital costs are amortized at 5 years and 13 percent
interest.
Costs of compliance are estimated by EPA for treatment options 2, 3,
and 5. Treatment costs of option 1 are assumed to be negligible and are not
estimated by EPA, because information available indicates that most electronic
crystal plants are practicing good solvent management techniques to control
toxic organics. Moreover, many of these plants are also controlling the pH
of discharges by end-of-pipe neutralization.
11-42
-------
5.4.1 Existing Sources
Based on the above assumptions, capital investment costs and annual
compliance costs at several selected flow rates were estimated for treatment
options 2, 3, and 5. Figures 9-1 to 9-3 in the Draft Development Document
present the cost curves for these three options. These cost curves were used
to estimate the compliance costs for 21 semiconductor plants. Key operating
characteristics for the 21 plants are presented in Table 5-1, and the average
investment and annual costs for pollution control are presented in Table 5-2.
The five model plant size groupings were based on the plant sizes defined in
Table 3-2 and wastewater flow rates and other data developed from an EPA
survey of the industry.
It is estimated that 257 semiconductor plants will be affected by the
EPA effluent regulations if option 2 or 5 is required. Of this total, 77
plants are direct and 180 plants are indirect dischargers.
EPA has determined that 25 of the 77 direct dischargers already have
option 3 treatment systems in place. Therefore, for this treatment option
there are only 52 direct dischargers that will incur costs as a result of
the option 3 regulation. Using the average costs for the various pollution
control options by employee size classification in Table 5-2, the compliance
costs for the total industry were calculated and are shown in Table 5-3.J1/
The information from this table indicates that option 3 is the least costly,
and option 5 the most expensive treatment technology.
JL'The average investment and annual compliance were multiplied by the total
number of plants in each size classification to determine the total com-
pliance cost by size classifications. The results were then summed to
determine the total industry compliance costs.
11-43
-------
TABLE 5-1. SEMICONDUCTOR PLANT OPERATING STATISTICS
PLANT
ID #
A
B
C
D
E
F
G
H
I
J
K
L
M
N
0
P
Q
R
S
T
U
NUMBER OF
PRODUCTION
EMPLOYEES
4,650
2,395
40
353
600
3,600
2,100
733
1,800
2,500
800
60
300
150
50
300
1,150
33
12,450
60
595
TOTAL
FLOW RATE (GPD)
(000 's)
2,494
540
43
360
290
2,013
408
280
432
3,026
663
50
324
2
3
14
199
151
792
35
41
FLOW RATE FOR
CONCENTRATED FLUORIDE
ACID WASTE (GPD)
6,290
1,340
120
912
700
5,090
1,030
696
1,080
7,560
1,660
120
65
24
4
24
504
384
1,990
720
98
SOURCE: The EPA Effluent Guidelines Division.
11-44
-------
CO
H
CO
o
o
W /-s
U CD
z wi
o
O Os
r-»
H Os
5^
J U-l
CU O
J a
U T3
O C
Q «!
Z co
O
- -• j:
CM U
I -^
in
Ed
J
03
<
H
0
o
o
^
H
5J
jj
0«t
Cti
U
04
H
CO
o
W
cj
z
JF
0
g
M
€
in
Z
o
H^
H
O
cn
Z
O
l-t
H
P-
O
CM
Z
0
H
O
z
o
M
H
Bu
O
_3 CO
•< H
p CO
Z O
Z 0
J W
^ S
g_| ^H
M CO
0U (x3
•rf ^
U Z
M
J CO
< H
p co
Z O
Z U
H
Z
^J S
H H
M CO
0, W
U Z
M
J CO
^ ^H
p co
Z 0
Z U
gH
z
hJ W
H H
l-l CO
< >
0 Z
M
p co
Z 0
Z U
H
Z
^C £
^ §^
M CO
PL> U
cj z
M
W
W H *-»
0 < Q
K O
W S s-'
OS J CO
W Pii H
BP ?1 2>*
I m 3
o
CO
H
Z
(X
fa
0
U
N
l-l
CO
00
•
CM
OS
i— 1
CM
•
00
p-
CM
°:
en
m
in
•
1-1
SO
00
•
in
^»
**
CO
CM
O
o
CM
r*.
^,
A
sO
in
m
^^
1 'a
N»X d)
^,
fH 0
(0 CX.
E e
CO W
en
•
^\
CM
m
ps.
•
\£)
If}
*&
m
^^
•st
in
O
•
vO
^0
c^
•
v~4
0
en
^-*
•
en
O
^*
o
0
o
in
^^
«
,^
^*»
"*
i— i
f*i
(Q ^N
CO O^ 01
1 Os 0)
e s
3 1 0
• ^ 0 -H
•a o a
o> ^ e
X^ td
so
e
CM
CM
«^
00
•
*— t
oo
m
oo
*
m
m
^-4
sO
«
cn
f^
CM
Os
m
•d-
CM
O
m
o
0
m
CM
cn
ao
i— i
cn
**
QJ
w
03 CO
_3 Os 01
1 Os 01
B Os >^
9 1 O
• i-l O i— '
*^ O Cu
oi m g
X ^ w
m
•
Os
in
r-»
O
o
o
CM
B
0
"*^
sO
•
,^
f—4
<-4
«^
OS
SO
sO
OS
CO
00
o
o
-vf
CM
CM
00
cn
CM
1-4
00
^
0) *~*
> w
o o>
^
01 O O
too *-*
U O D.
(Q ^ S
J ^ W
CD
01
U
O
CO
(0
•O
01
a,
o
W
o
oi
O
CO
11-45
-------
co
H
CO
O
CJ
u
o
§
M ,
J
t!
o
o
o
•o
g
H ON
co r-
D ON
^
03
O C
H ce
oo
3
. O
en j:
i u
m •^^
U
I
H
co
0
O
W
z
CM
33
O
U
*-4
J CO
*D CO
Z O
z u
O Z
*""^
J CO
< H
D CO
Z O
Z O
•^
H
Z
^4 S
£•4 f^
M CO
0«4 Sd
<1 >
U Z
M
J CO
<: H
D CO
z o
<• °
E™^
z
^C 3u
E-i E-^
M CO
CU (t]
cj z
M
O fa
u o
H CO
fH- « §
c^ 3G *J
CO S (X)
M Z
CO
- H
5
i
fa
o
w
N
H
co
f*» fH 00
O CM CM
ON en CM
A A A
r» oo NO
fH CM
NO r^ en
O CM en
-* «* CO
A A A
fH vO f*»
fH CM en
«
NO CM en
fH CM 00
Q\ £*$ Jm4
«t 9*
en .*
CO
fH en .
fH CM en
CO
r- ^ f- 1
in oo -^
r- m en
A A
CM en
CO
CM fH en
CM en in
^H QO ON
A A A
1-1 en »*
en oo i-i
en r»- i-i
in .* o
A A A
r- r>> m
— 4 CM
ON en CM
f^ oo NO
O en -^
B* A ft
O en en
fH CM en
O O 0
o o o
in oo en
CM m oo
iJ fH
U CO
4J Of 4J
0 H O
OJ 'H JJ
Id "O iA
•H C 3
Q M CO
fH
I— 1
CO .—.
B oo
CO ON 0)
1 ON 3)
e sr x
3 1 O
•H O fH
•a o a.
41 fH g
X^ w
en en \o
fH ON O
fH
CM in r*.
O^ ^ co
O so r-
^^ in ""s^
A A A
CM 1^ O
""*
CO
00 NO -*
NO CM ON
fH 1^ 00
CO
fH r^ oo
CM in r^
CM ON fH
A
fH
o in in
NO ON in
ON O O
A A A
fH m r-
CM f» ON
r» oo m
m NO CM
A A A
CM VO O>
O O O
O O O
m en oo
fH fH
W rH
U CO
±J 4) 4J
U I-I O
01 -H U
I-I -O XI
•H C 3
O M CO
0)
00
H s"^
CO CO
>J ON 01
1 ON OJ
B ON X
3 1 O
•H O fH
•O O O.
oi m B
53 •— ' W
t*- * o m
rH rH
0 O 0
O O 0
0 O 0
NO «tf O
rH CM
CO
CM NO 00
in NO rH
>* CM f-*
A A
rH rH
CO
en in oo
r^ o r*
in NO rH
A A
rH CM
^ f^» 1— 4
NO rH 00
1-1 r«, oo
A A A
•^ ON en
i— i
oo 00
O Oi
**^ V
01 0 O ""
60O rH
H o a,
co « B
rJ rH W
0 r*. r^
rH ,3- tn
oo m en
A A A
CM en NO
CM m r^
-* CM NO
en oo fH
r^. sf CM
A A A
i— 1 -- en <->
in CM ao
A A A
en CM m
r-l i— 1
00 ON t-«
NO ON NO
00 ON 00
A A A
O 00 00
CM •>-; NO
rH ON O
Q\ Q\ Q\
ON VO NO
A A A
r>. in en
CM NO ON
O O O
o o o
r- O f~
r^ oo in
i-l CM
u
4J 0)
U H rH
O» -H CO
M "O *J
•H c 0
fi M H
_
^
H
o
H
in
II
in
oi
oo
en
41
60
• H
T3
01
CO
S
ID •
I 4J
B C
3 4)
•H H
•O 0)
r-
CM to
• H
-< C
CO O
09 W
3
.0
CO H
C
-------
5.4.2 New Sources
/•
The basis for the new source standards (NSPS and PSNS) is the best avail-
able demonstrated technology (BDT), including in-plant controls and end-of-
pipe treatment technologies that provide the maximum pollution reductions
feasible. For the Semiconductor subcategory, the proposed NSPS and PSNS
discharge limitations are the same as those proposed for BAT and PSES. Since
new source costs are defined as incremental costs from BAT and PSES, costs of
NSPS and PSNS are zero.
11-47
-------
6. ECONOMIC IMPACTS
This chapter provides an assessment of the economic impacts that are
likely to occur as a result of the costs of the effluent control/treatment
technologies described in Chapter 3. The analysis is based upon the examina-
tion of the estimated compliance costs, other economic and financial char-
acteristics of 21 sample plants, and the analytical methodology described in
Chapter 2. The primary economic impact areas discussed include the effect
of the pollution control costs on semiconductor prices, the profitability
of semiconductor manufacturing, plant closures, employment, changes in imports
and exports, and industry structure. The analysis is not intended to be plant-
specific; instead, the sample plants serve only as a basis for deriving
indications of potential industrywide impacts.
6.1 PRICE AND QUANTITY IMPACTS
The price and quantity impacts of the compliance costs are determined
through an examination of industry market structure and pricing behavior.
Economic theory predicts different pricing behavior for different market
structures (e.g., competitive, oligopoly, monopoly). As described in Chapters
2 and 3, the market structure for this industry is not precisely determined
in this study. Nevertheless, there is evidence that the semiconductor industry
has both the incentive and the capability to completely absorb all small and
moderate levels of mandated pollution control costs. The incentive not to
raise product prices is in the form of significant international competition
for world market share; and the capability is implied by the variations in
profit rates among various semiconductor products within multi-product firms.
For these reasons it is assumed that there will be no price changes resulting
from the proposed regulation in the short and intermediate time period.
11-48
-------
6.2 PLANT LEVEL PROFIT IMPACT ANALYSIS
The basic profitability measure used in this study is before-tax return
on sales, or profit margin (PM). The profitability of a plant is considered
to be significantly impacted if the total annual compliance costs reduce the
return on sales of a plant to an extent that it is equal to or less than the
lowest average industry PM. As described in Chapter 2, The lowest acceptable
PM was determined to be 2.7 percent. This is based on the performance of the
semiconductor industry over the 1971-1977 period,2J
Table 6-1 presents data on the actual number of production employees and
the estimated value of shipments for each sample plant. The value of shipments
data, along with the total annual compliance cost estimates for each of the
recommended treatment options for the semiconductor plants, are used to
analyze the performance of the plants after compliance.
Table 6-2 presents the percent of annual revenues that is estimated to be
spent annually to operate, maintain, and finance the proposed pollution control
technologies for each of the options for the sample plants. This table was
developed from the value of shipments data in Table 6-1 and the compliance
cost data from EPA. Compliance costs for option 1 are estimated to be zero,
since most plants already use the option 1 technology. Option 3 costs are
also relatively insignificant, since there is only one plant that has treatment
costs greater than 1 percent of the estimated annual revenues of the plant.
The post-compliance PM for this plant is not below the threshold value. The
annual compliance costs associated with options 2 and 5 are much more signifi-
cant. Eight of the plants examined may incur costs that range from 1-12 per-
cent of their annual revenues. Four of the eight sample plants have PMs below
the threshold value, all of which are small plants.
2f U.S. Department of Commerce, U.S. Semiconductor Industry, September 1979.
11-49
-------
TABLE 6-1. SALIENT STATISTICS FOR THE SEMICONDUCTOR SAMPLE PLANTS
PLANT
ID CODE
A
B
C
D
E
F
G
H
I
J
K
L
M
N
0
P
Q
R
S
T
U
PRODUCTION3
EMPLOYEES
4650
2395
40
353
6OO
3600
21OO
733
1 8OO
25OO
800 »
6O
300
150
50
30O
1150
33
12450
60
595
VALUE OFb
SHIPMENTS
3897651S6
200750O26
3352S19
29588626
5Q292282
301753692
176022987
61440405
150876846
209551175
67.056.376
5029228
25146141
12573071
4191024
25146141
96393541
2766O76
K>43.564.352
5O29228
49873 ISO
NEW CAPITAL13
EXPENDITURES
29950393. 7 O
15426062.99
257637.80
2273653.54
3864566.93
23137401.57
13525984.25
4721 21 2. 6O
115937O0.79
161 02362. 2O
5152755.91
386456.69
1932283.46
9661^ 1.73
322047.24
1932283.46
74O7086.61
212551. IS
80189763.78
386456.69
3832362.20
aSOURCE: EPA, Effluent Guidelines Division
^Estimates derived from average industry ratios obtained from the Census
Bureau.
11-50
-------
TABLE 6-2. SEMICONDUCTORS - ANNUAL COMPLIANCE COSTS
AS A PERCENTAGE OF REVENUES
11-51
PLANT
ID CODE
A
B
C
D
E
F
G
H
I
J
K
L
M
N
0
P
Q
R
S
T
U
TREATMENT3
OPTION 1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
TREATMENT
OPTION 2
.27
.26
5.23
1.50
.80
.31
.27
.65
.32
.54
.89
3.73
1.68
.36
1.30
1.17
.35
10.96
.06
3.16
.34
TREATMENT0
OPTION 3
.03
.04
.89
.21
.12
.04
.04
.09
.04
.07
.12
.60
.22
.24
.71
.12
.05
1.66
.01
.ooi/
.06
TREATMENT
OPTION 5
.28
.34
5.53
1.63
.87
.33
.29
.71
.35
.57
.94
4.13
1.83
.41
1.48
1.28
.39
12.02
.06
3.51
.38
aThe annual compliance cost for this option is estimated to be zero.
"Information on concentrated fluoride acid waste was not available
for this plant.
11-51
-------
6.3 CAPITAL REQUIREMENTS ANALYSIS
As described in Chapter 2, the assessment of the firm's ability to raise
the required capital is based on ratios of compliance capital investment to
average annual compliance expenditures. Because of the absence of plant-
specific financial data, the annual capital expenditures of the plants are
calculated from the industry average new capital expenditures per production
employee ratios reported in the 1977 Census of Manufactures. The estimated
annual capital expenditures from Table 6-1, along with capital investment
costs for each of the pollution control options, are used to indicate the
significance of the capital impacts.
Table 6-3 presents the compliance capital investment to plant revenue
ratios for the 21 sample plants. These ratios reveal that the proposed pollu-
tion control costs would range from zero to 29 percent of the average annual
capital expenditures for options 1 and 3. Investment costs of these magni-
tudes are not by themselves enough to cause plant closures. The pollution
control investment costs are more significant for options 2 and 5. Four of
the plants, all of which are small, would have to invest more than 50 percent
of their annual capital expenditures in pollution control to comply with the
proposed regulation. These results will be used below, together with other
information, to assess the potential for plant closure.
6.4 POTENTIAL PLANT CLOSURES
As described in Chapter 2, potential plant closures are estimated by simul-
taneously examining the results of the profit and capital impact assessments
for the sample plants, and qualitatively assessing other nonquantified factors.
Using these results, possible closures are identified, and then used as the
basis for extrapolating industry closures.
11-52
-------
TABLE 6-3. SEMICONDUCTORS - POLLUTION CONTROL INVESTMENT
COSTS AS A PERCENT OF ANNUAL CAPITAL EXPENDITURES
PLANT
ID CODE
A
B
C
D
E
F
G
H
I
J
K
L
M
N
0
P
Q
R
S
T
U
TREATMENT3
OPTION 1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
TREATMENT
OPTION 2
4.35
4.49
94.97
25.79
13.87
5.14
4.56
11.20
5.45
8.76
14.63
67.47
29.05
7.02
25.32
20.56
6.20
192.71
1.01
57.74
6.24
TREATMENT
OPTION 3
.56
.57
23.87
3.33
1.77
.66
.59
1.43
.70
1.13
1.88
15.91
3.39
6.37
19.10
3.18
.83
28.93
.13
.00b
1.60
TREATMENT
OPTION 5
4.8
5.0
108.4
29.1
15.6
5.7
5.1
12.6
6.1
9.8
16.4
76.9
32.8
8.1
29.2
23.3
7.0
218.7
1.1
65.9
7.1
aThe pollution control investment cost for this option is estimated to
be zero.
"Information on the concentrate fluoride acid wastes was not available
for this plant.
11-53
-------
It is assumed that a plant would close if the costs of pollution control
caused its profit margin to fall below 2.7 percent, and/or if a plant had to
spend more than 100 percent of its annual capital expenditures on pollution
control equipment.
The pollution control costs associated with options 1 and 3 are not great
enough to cause plant closures. However, the analysis of the pollution control
costs associated with treatment options 2 and 5 indicate that the annual profit
margins for four of the plants examined may be reduced below 2.7 percent, and
that these four plants will have to invest significant portions of an average
year's capital expenditures in pollution control equipment. Under these circum-
stances, four, or 19%, of the sample plants are identified as likely candidates
for closure, if they are required to install treatment option 2 or 5. All four
plants are small, with employment ranging from 33 to 60 employees. The results
of the sample plant analysis are used as the basis for estimating plant closures
at the industry level. These estimates are shown in Table 6-4. It is assumed
that the percentage of industry plant closures will be the same as that indi-
cated for the sample plants. Using this approach, it is estimated that 19
percent or 48 of the plants in the semiconductor industry are likely to close
under options 2 or 5. Of these 48 plants, it is estimated that 14 are likely
to be direct and 34 indirect dischargers. The number of direct versus indirect
dischargers that are likely to close is estimated using the pre-compliance
ratio of direct to indirect dischargers for the total industry.i?_'
i£'An important qualification to these extrapolations is in order. First, 136,
or 53 percent, of the 257 plants in the industry are small (employ fewer
than 100 production employees). However, only 5, or 24 percent, of the
plants in the EPA sample were small plants, indicating that the economic
_ impact estimates based on these data may not fully represent the range of
- likely impacts among small plants. For example, if 80 percent of small
plants closed (4 out of 5 small plants in the sample), then the costs of
options 2 or 5 could result in (0.8)(136) • 109 plant closures, instead of
the 48 estimated above.
11-54
-------
CO
SE
3
OH
06
0
E3
a
a
z
o
CJ
M
JgJ
W
CO
06
£
CO
§
i
M
H
Z
JM
0
OH
J£
w
Q
Z
CO
w
as
D
CO
O
CJ
£
•<
OH
j
<
H
Z
W
g
•
1
\o
w
5
m
z
0
H
OH
O
CO
Z
0
M
H
PH
O
CN
z
0
M
H
OH
O
r- 1
Z
0
H
OH
O
0
H
j£
M
H
CO
W
S5 |
5J
M >•
O O
Z >J
II
,j
< CO
H^ £—1 [xJ
E~* 2 w
z < »
W -3 CO
H OH O
O t-t
OH CJ
H
Z Z
M W
3J
w >•*
< fe
CJ W
< CO
M H W
III
H OH O
OH CJ
H
Z Z
Mg
W >-
Is
33 2J
O W
j
^J C/D
w H W
H Z OS
U i-9 CO
H OH O
OH CJ
Z Z
w P
z S
< OH
S w
J
< CO
i-i H W
gSg
ta i-3 co
H OH O
O -3
OH CJ
W
CO O
g^
^ EC en
^« HM WJ
a. co H
H4 -rf
Cn O H
O CO
9*
* "
O CO
H
fW| Jg;
N •<
M iJ
CO OH
vO vO CS
ao vo m
vO SO CO
« M
i— I CM
sfr -* 00
<-H CO «d"
OOO
OOO
\C vO CM
oo vo m
SO vO CO
•k A
i-l CM
-* vfr 00
f— ( CO *^"
OOO
000
0 <0
0 t- O
w -a 43
•H C 3
O M CO
1— 1
1— 1
(Q
E
CO
OOO
000
OOO
OOO
OOO
OOO
OOO
000
o H
0) « Q
X J J
^D sO ^J
00 vO i^
\O sO CO
M M
i-l CM
-* -* 00
v** en "^
OOO
OOO
sD sO CM
\O ^O CO
t-l CM
vj- -* oo
i-l CO
-------
6.5 EMPLOYMENT IMPACTS
In general, a regulation can impact on employment in an industry through
job losses that are associated with closures, and from price increases that
cause output and employment to fall. Since no plant closures or price changes
are expected for either options 1 or 3 the regulation is not expected to have
any impact on the level of unemployment for these options. On the other hand,
48 plants are likely to close if either treatment option 2 or 5 technologies
are required. Since all of the potential closures are small plants (i.e.,
having between 1 and 99 employees), and the average number of employees in the
small plants is 49, it is estimated that 2,352 employees (i.e., 48 plants mul-
tiplied by 49 employees) may lose their jobs if either of these two pollution
control technologies is required.
6.6 FOREIGN TRADE IMPACTS
The impact of the regulation on the levels of imports and exports depends
on (1) the extent that semiconductor product prices in the domestic market rise
faster than prices in the rest-of-world market, and (2) the extent to which the
domestic production losses are replaced by imports from foreign countries.
Because the domestic producers are not expected to increase the price of semi-
conductor products as a result of this regulation they should continue to be
competitive with foreign producers. Therefore, there will be no foreign trade
impact from this factor. The increase in imports of semiconductor products
because of domestic capacity reductions (i.e., plant closures) is difficult to
quantify precisely. For this analysis, it is assumed that the production
losses will be made up by U.S. producers only. Therefore, no balance of pay-
ment impacts are predicted.
67? NEW SOURCE IMPACTS
As indicated in Section 5, costs of NSPS and PSNS are zero. Consequently,
there will be no economic impact from NSPS and PSNS requirements., Also, because
11-56
-------
there is no difference in compliance costs between similar existing and new
sources, the proposed regulation would not foster any competitive advantages
or disadvantages between new and existing sources.
6.8 LONG-TERM IMPLICATIONS AND OTHER IMPACTS
In spite of the impressive growth of the semiconductor industry, a pros-
perous future is not guaranteed. Technological obsolescence and foreign compe-
tition are likely to continue to be of great concern to the industry. Because
of this, capital expenditures for R&D and new equipment are very high. Failure
to keep pace on a year-to-year basis may mean the loss of important market
segments to competitors, which in many cases are European or Japanese companies.
The costs associated with EPA regulations will add to these problems. To the
extent that this added burden could be significant, it may force some firms
into a noncompetitive position and could lead to potential plant closures in
the long run, or it could lead to a firm's relocating their wafer fabrication
operations overseas. Most companies already have moved their labor intensive
assembly operations overseas, and it would be quite possible for them to
relocate wafer fabrication as well. Such a move would result in a loss of
employment, wages, and a substantial adverse shift in the current balance of
trade.
6.9 SMALL BUSINESS ANALYSIS
The Regulatory Flexibility Act (RFA) of 1980 (P.L. 96-354), which amends
the Administrative Procedures Act (APA), requires Federal regulatory agencies
to consider "small entities" throughout the regulatory process. The RFA requires
analyses to be performed which determine if a substantial number of small
entities will be significantly impacted. If so, regulatory alternatives that
eliminate or mitigate the impact must be considered. This analysis addresses
these objectives by identifying and evaluating the degree of economic impacts
of the aforementioned regulations on small entities (the primary small entity
affected by these regulations is the small semiconductor manufacturer). As
11-57
-------
described in Chapter 2 the small business analysis was developed as an integral
part of the general economic impact analysis described above in this report.
6.9.1 Definitions of Small Business
As described in Chapter 2, the definition of small for purposes of regula-
tory analysis is not precise or universal. The small business defintions used
by other Federal agencies, which were developed for purposes other than regula-
tory analysis was found to be inappropriate for this study. Instead, informa-
tion on both the amount of effluent discharged and the number of employees
working in the plants was used to determine an appropriate cutoff point that
identifies small entities. The criterion selected consider plants with fewer
than 100 employees as small entities because at this cutoff point the plants
within this subcategory have some special and distinct properties. Table 6-5
illustrates the pollution discharge characteristics of plants in the industry
by employment size categories. The information presented demonstrates that
the small plants (i.e., plants with less than 100 employees) discharge an
average of 57 thousand gallons of wastewater per day. Furthermore, although
the small plants make up more than 50 percent of the plants in the industry
they discharge only 15 percent of the wastewater.
6.9.2 Impacts on Small Entities
Table 6-6 shows the average ratios of compliance costs to revenues and the
pollution control investment costs to annual new capital expenditures which are
associated with plants of various sizes in the semiconductor industry. The
information in this table indicates that the smaller plants (i.e., the plants
with less than 100 employees) would be more severely affected than the larger
plants. In fact, the impact variables examined are at least 3 times greater
for the small plants than for larger plants. In addition, all the projected
plant .closures are small plants if either treatment option 2 or 5 are required.
11-58
-------
TABLE 6-5. SEMICONDUCTOR - AVERAGE PLANT AND TOTAL INDUSTRY FLOW RATES
Size of Plant
Small
(1-99 Employees)
Medium-Small
(100-499 Employees)
Medium-Large
(500-999 Employees)
Large
(Over 1,000
employees)
Total
DISCHARGE
STATUS
DIRECT
41
25
5
6
77
INDIRECT
95
58
13
14
180
TOTAL
# OF
PLANTS
136
83
18
20
257
AVERAGE
FLOW RATE
(GPD)
56,772
174,450
318,325
1,238,224
205,038
Total Flow Rates
(GPD)
Total %
7,720,992 15
14,479,350 27
5,729,850 11
24,764,480 47
52,694,672 100
SOURCE: Compiled by JRB Associates.
11-59
-------
i
M
i
M
ft
CO
CO
.3
o
w
tvj
M
CO
U
!
a,
x
co
CO
u
CO
5
e
H 4)
co y
< ^
It
i
M
Oi
cl
1
Zj
S"
O N*^
H
CO
Ws
JJ6|
•3 5n
Ed W
P ^s
SO W
z «!
H H
M Q^
U>
<
3
CO
^
«5
w
p»
i
6*3
0
H
H
CO
o
u
1
3
Q
o§
<
«
o
H
Oi
o
en
2
O
M
H
CU
O
01
2
0
E
o
m
55
o
M
H
Pu
O
en
2
0
M
B
o
CM
2
O
B
o
0
M
H
QJ
M
CO
CO
^3
u
w
N
CO
00
ON
ON
NO
in
«
F*^
iH
sr
c^
en
en
^O
(^
f*«*
•
CO
00
•*
/-s
(0
4)
4)
O
^H
- B-
4)
ft ON
**H ^^
a |
en
en
CN
^
O
*
^f
vD
O
CM
ON
<%|
«
iH
ON
rM
»
00
iH
iH
^^
H
Ic>
1 ^^
a •*
9 1
•H 0
•0 0
ON
CM
1^
vO
«
fm^
00
fl
1^^
•
ON
O
vO
sJ
i-4 ON
1 ON
H3 1
•H O
•O O
IP
CM
vO
in
m
NO
o
in
en
en
en
O
•
O
en
/\
V o
MO
>-. O
09
0)
JJ
dt
•H
O
O
n
o>
a-
o
O
CO
11-60
-------
Although the small plants will experience greater profit reductions than
larger plants, there are no plant closures projected under the selected option
(options 1 and 3).
6.10 SOCIAL COSTS
This section assesses the total social costs that can be associated with
the EPA effluent regulation. The social costs measures the value of goods
and services lost by society due to a given regulatory action. These costs
generally include the use of resources needed to comply with a regulation,
the use of resources to implement and enforce a regulation, plus the value
of the output that is forgone because of a regulation.
The total social costs of the regulatory options are shown in Table 6-7.
The total social costs for treatment options 1, 2, 3, and 5 are $0, $435 mil-
lion, $59 million, and $469 million, respectively.
11-61
-------
TABLE 6-7. SEMICONDUCTORS - SOCIAL COSTS FOR POLLUTION CONTROL
($ Million)
TREATMENT OPTION
Option 1
Option 2
Option 3
Option 5
DISCHARGE STATUS
DIRECT
0
$130
$ 13
$140
INDIRECT
0
$305
$ 46
$329
TOTAL
0
$435
$ 59
$469
SOURCE: Compiled by JRB Associates.
11-62
-------
7. LIMITATIONS OF THE ANALYSIS . .
This section discusses Che major limitations of the economic impact
analysis. It focuses on the limitations of the data, methodology, assumptions,
and estimations made in this report. Information pertaining to the estimation
of the compliance costs for specific plants and the limitations of the method
are outlined in Section 9 of the EPA Development Document.
The economic impacts assessed are the result of the proposed water efflu-
ent requirement only. The assessment does not include the economic impacts
from such items as air pollution control, OSHA requirements, solid waste
requirements, and costs resulting from other regulations.
7.1 DATA LIMITATIONS
The major assumptions and estimates that have the greatest impact on the
accuracy of the conclusions of this report relate to the data used in the
analysis. It is important to realize that (1) no economic survey was con-
ducted to collect plant-specific financial and economic data, and (2) these
types of data are not in the public domain for this industry. As a result,
most of the plant-specific data used are estimates derived from average
industry operating and financial ratios. The plant-specific financial data
must, therefore, be considered order-of-magnitude estimates, and not the
actual data for specific plants.
The value of shipments and new capital expenditures were estimated by
multiplying the number of production employees at each plant by the appro-
priate average industry performance ratio obtained from the Census of Manu-
factures. This methodology relies on the assumptions that a fixed relation-
ship exists between the key performance variables for all plants,. and that
this relationship remains constant over time. These assumptions are probably
11-63
-------
invalid in the real world because of the following factors:
• Worker productivity varies significantly among plants
• The type, scale, and age of technologies vary widely
among plants and greatly affect the value added by
each production employee
• The type of products produced greatly affects the value
added per employee or unit of raw material used; and
product mix varies from one plant to another.
Ignoring these factors could result in overestimating the impacts on
some plants while underestimating those of other plants. The reason for
these potential errors is that the intra-industry variation in baseline
conditions is underestimated. That is, baseline conditions for lower-profit
plants are probably overstated and those of high-profit plants understated.
7.2 METHODOLOGY
7.2.1 Price Increases
The analysis assumes that the semiconductor industry will not increase
the price of their products and pass through any portion of the compliance
costs to their customers. While this is a reasonable assumption for some
semiconductor products it may not be true for all the product segments in
the semiconductor industry. It is possible that the price pass-through
percentage will vary among the various semiconductor products. Adequate
information on these segments is not available. Therefore, no further
analysis is carried out to measure the extent of a price increase which may
occur as the production costs of the plants increase with the installation
of pollution control.
The estimation error which might result from the price change assumptions
depends primarily on the cross elasticity of demand with respect-to the ratio
11-64
-------
of domestic to foreign prices. That is, if increased domestic prices shift
demand production overseas, there would be a drop in aggregate quantities pro-
duced by domestic producers and, consequently, increased unemployment and
possibly other impacts. Thus, the remaining domestic semiconductor producers
would have healthy profits, but account for a smaller portion of world market
share. On the other hand, if such shifts were small, then domestic producers
could raise prices to cover compliance expenditures without losing market
share.
If the latter situation prevailed in the real world, the economic impacts
estimated in this report are overstated. If the former situation prevailed,
the impacts would take a different form, but would probably not exceed the
levels estimated in Chapter 6 of this report.
7.2.2 Sensitivity Analysis
In addition to the biases that may be generated by the method used to
estimate revenues and new capital expenditures of individual plants, the
results of the analysis can be influenced by the pollution control cost data.
To alleviate some of the problems associated with these data, as well as
analyzing the effect of revenues and capital expenditure estimates, sensi-
tivity analyses were performed. For this analysis, the estimated compliance
costs were varied by -20 percent and +20 percent. Tables 7-1 and 7-2 present
the results of the sensitivity analysis on the profitability and capital
availability assessments for all the recommended treatment options. The
tables show that the changes in the pollution control costs (or equivalently,
changes in revenue and capital expenditure estimates) would not significantly
change the post-compliance financial ratios and the study's results for treat-
ment option 3. For treatment options 2 and 5 reductions in the pollution
control costs by 20 percent would not cause changes to the previous results.
Alternatively, increases in the pollution control costs by 20 percent would
cause 2 additional plants to close (i.e., six of the sample plants instead
of four plants, as predicted in the impact analysis in Chapter 6, are likely
11-65
-------
TABLE 7-1. SEMICONDUCTORS SENSITIVITY ANALYSIS -
ANNUAL COMPLIANCE COSTS AS A PERCENTAGE OF REVENUES
PLANT
CODE
A
B
C
D
E
F
G
H
I
J
K
L
M
N
0
P
Q
R
s
T
U
OPTION 2
-20% +20%
.21 .32
.21 .32
4.18 6.27
1.20 1.80
.64 .96
.25 .37
.21 .32
,52 .78
.25 .38
.43 .65
.71 1.07
2.98 4.48
1.35 2.02
.29 .43
1.04 1.56
.93 1.40
.28 .43
8.77 13.15
.05 .07
2.53 3.79
.27 .41
OPTION 3
-20% ^ +20%
.03 .04
.03 .04
.72 1.07
.17 .26
.09 .14
.03 .05
.03 .04
.07 .11
.04 .05
.05 .08
.09 .14
.48 .72
.18 .27
.19 .28
.57 .85
.09 .14
.04 .06
1.33 2.00
.01 .01
.00a .00a
.05 .07
OPTION 5
-20% +20%
.23 .34
.27 .40
4.43 6.64
1.31 1.96
.70 1.05
.27 .40
.23 .35
.56 .85
.28 .42
.46 .69
.75 1.12
3.30 4.96
1.47 2.20
.33 .49
1.18 1.77
1.02 1.54
.31 .47
9.61 14.42
.05 .08
2.81 4.22
.30 .46
Option 3 cost data were not available for this plant.
SOURCE: Compiled by JRB Associates
11-66
-------
TABLE 7-2. SEMICONDUCTORS SENSITIVITY ANALYSIS -
POLLUTION CONTROL INVESTMENT COSTS
AS A PERCENTAGE OF NEW CAPITAL EXPENDITURES
PLANT
CODE
A
B
C
D
E
F
G
H
I
J
K
L
M
N
0
P
Q
R
S
T
U
OPTION 2
-20% +20%
3.48 5.22
3.59 5.39
75.98 113.97
20.63 30.94
11.09 16.64
4.11 6.17
3.65 5.48
8.96 13.44
4.36 6.54
7.00 10.51
11.71 17.56
53.98 80.96
23.24 34.86
5.62 8.42
20.25 30.38
16.45 24.67
4.96 7.44
154.17 231.26
.81 1.21
46.19 69.29
4.99 7.48
OPTION 3
-20% +20%
.45 .67
.46 .69
19.10 28.64
2.66 3.99
1.42 2.13
.53 .79
.47 .71
1.15 1.72
.56 .84
.90 1.35
1.50 2.25
12.73 19.10
2.72 4.07
5.09 7.64
15.28 22.92
2.55 3.82
.66 1.00
23.15 34.72
.10 .16
.00a .00a
1.28 1.93
OPTION 5
-20% +20%
3.90 5.84
4.05 6.08
86.74 30.11
23.32 34.98
12.55 18.83
4.61 6.92
4.12 6.19
10.14 15.20
4.93 7.39
7.84 11.76
13.20 19.79
61.57 92.36
26.28 39.42
6.50 9.76
23.41 35.12
18.67 28.01
5.62 8.43
174.97 262.46
.91 1.37
52.79 79.18
5.70 8.55
aOption 3 cost data were not available for this plant.
SOURCE: Compiled by JRB Associates
11-67
-------
to close). Therefore, incorrect costs for these options (options 2 and 5)
could significantly affect the findings of the study. On the other hand,
incorrect costs for options 1 and 3 are not likely to alter the findings of
the study.
7.3 PLANT CLOSURES
The criteria for plant closure is not all-inclusive. It does not address
the long-term financial condition of the plants, captive operations, the time
value of money, the amount of invested capital, or other factors which affect
closure decisions. A more detailed closure analysis would require more pre-
cise data on the long-term capital structure, cash flow, and the profitability
of plants in the semiconductor industry. However, the costs of the recommended
treatment technology are small, and a more detailed analysis will not necessarily
yield different results.
7.4 SAMPLING
The statistical significance of the 21-plant sample is skewed in such
a way that the large plants are adequately covered. However, most of the
plants in the semiconductor industry are small, and these plants are repre-
sented by only a few plants in the sample.
7.5 SUMMARY OF LIMITATIONS
Although a number of assumptions and data limitations may significantly
bias the economic impact conclusions, the potential changes to the conclu-
sions resulting from the elimination of these biases are probably small.
Consequently, the above methodology appears to have provided a reasonable
industry-wide assessment of potential impacts.
11-68
------- |