United States
Environmental Protection
Agency
Office Of Air Quality
Planning And Standards
Research Triangle Park, NC 27711
EPA-452/R-02-005
September 2002\
Final Report for Proposal
Air
    Economic Impact Analysis of Metal Can
                   MACT Standards

-------

-------
Economic limpact Analysis of Metal Can
            MACT Standards
      U.S. Environmental Protection Agency
   Office of Air Quality Planning and Standards
Innovative Strategies and Economics Group, C339-01
        Research Triangle Park, NC 27711
          Prepared Under Contract By:

            i        RTI
     Health, Social, and Economics Research
       Research Triangle Park, NC 27711
              September 2002

-------
      This report has been reviewed by the Emission Standards Division of the Office of Air Qualit
Planning and Standards of the United States Environmental Protection Agency and approved for
publication. Mention of trade names or commercial products is not intended to constitute
endorsement or recommendation for use.  Copies of this report are available through the Library
Services (MD-35), U.S. Environmental Protection Agency, Research Triangle Park, NC 27711, or
from the National Technical Information Services 5285 Port Royal Road, Springfield, VA 22161.

-------
CONTENTS
Section
1 Intro
1.1
1.2
1.3
2 Indus
2.1
2.2
2.3
2.4
3 Engii
3.1
3.2
duction 	
, Agency Requirem
Scope and Purpos
Organization of th
.
stry Profile: Metal C
Production 	
2.1.1 Sheet Man
2.1.2 Can End N
2.1.3 .One- and'
2.1.4 Three-Piec
2.1.5 Coatings a
'2.1.6 Costs of P:
Uses, Consumers,
Industry Organiza
2.3.1 Market Str
2.3.2 Facilities .
2.3.3 Companie;
Market Data and r.
leering Costs 	
Methodology . . .
3.1.1 Monitoring,
3.1.2 Material Co
3.1.3 Add-On Co:
Engineering Cost
'
'
	
ents for Conducting an EIA 	
e 	 	 	 	 	
'
.
e Report 	
an Manufacturing . . 	 	

ufacturing 	
lanufacturing 	
^wo-Piece Can Body Manufacturing 	
e Can Body Manufacturing 	
nd Emissions 	
"eduction 	
and Substitutes 	 " 	
ion 	
ucture 	


"rends 	


Recordkeeping, and Reporting Costs 	 	
sts . .• 	 	 	
itrol Devices 	
Summary 	 	 	

Page
1-1
	 	 1-1
	 2-2
1-2
2-1
	 2-1
	 2-2
2-2
	 	 2-2
	 .2-3
2-3
2-7
	 2-9"
2-14
2-14
2-17
2-17
2-19
3-1
3-1
	 . . 	 3-1
	 3-2
	 3-2
	 ! 3-3


-------
4      Economic Impact Analysis:  Methods and Results	•	4-1

       4.1    Markets Affected jby the Proposed NESHAP	4-1

       4.2    Conceptual Approach  	4-2
             4.2.1  Supply  ..:	4-2
             4.2.2  Demand .|	'.". . "........ ".".'.".". ?  4-2
             4,2.3 . Foreign Trade	4-4
             4.2.4  Baseline and With-Regulation Market Equilibrium  	4-4
             4.2.5  Impacts for Facilities Excluded from the Market Model	4-6

       4.3   ' Economic Impact Results	4-6
             4.3.1  Market-Level Impacts 	4-7
             4.3.2  Industry-Level Impacts	:	4-7
             4.3.3  Closure Estimates	.4-11
             4.3.4  Employment  Impacts	4-11
             4.3.5  Social Costs	 4-11
             4.3.6  Sensitivity Analysis	4-12

       4.4    New Source Analysis	'.	4-13

       4.5    Energy Impact Analysis	4-13

.5      Small Business Analysis ;.;	5-1

       5.1    Identifying Small Businesses	5-1
                              !                              '              •
       5.2    Screening-Level Analysis  	:	•	5-2
             5.2.1  Results  .  '	•. .-	5-2

       5.3    Economic Analysis  ,	5-3
                              i                      •            '                  .
       5.4    Assessment. . . . :	•	5-4
                              i i

References  	J	:	R-l

Appendix A: Model Data Set and Specification	  A-l
                              !               '            '         '
Appendix B: Sensitivity Analysis	B'-l

-------
                                     LIST OF FIGURES
Number
       2-1

       2-2
       2-3
       2-4
       2-5
       2-6
       5-1
       4-1
Two-Piece Draw-and-Iron Aluminum Beverage Can Manufacturing
Process	;	:	
Three-Piece Can Sheet B.ase Coating Operation	 .	
Three-Piece Can Fabrication Process  	
Distribution of Metal Can Shipments by End Use: 1997	
Distribution of Soft Drink Packaging Mix by Type:  1997
Distribution of Metal'Can, Sheet, or End Manufacturing Facilities
by State ............. :. ...... ...... ...... .
Summary of Costs to Seven Industry Subcategories

Market Equilibrium without and with Regulation ..
.  2-4
.  2-5
.  2-6
2-11
2-12

2-18

.  3-4

.  4-5

-------
Number
                                     LIST OF TABLES
       2-1    Spot Prices for Steel and Aluminum Sheet and Plate: 1997-2001	2-8
       2-2    Historical Cost of Production Statistics for the Metal Cans Industry:
             1992-1997	r	2-10
       2-3    Metal Can Uses by Material and Type	2-11
       2-4    Measures of Market Concentration for the Metal Cans Industry .
             (NAICS 332431): 1997	2-15
       2-5   , Measures of Market Concentration for the Glass Containers Industry
             (NAICS 327213): 1997 	:	2-15
       2-6    Measures of Market Concentration for the Plastic Bottle Industry
             (NAICS 326160): 1997	'.	2-16
       2-7    Domestic Metal Can Shipments by Market:  1993-1999
             (million cans)	2-20
       2-8    Prices for Beverage Containers: 1993-2000 ($/l,000 cans or bottles)		2-21
       2-9    Apparent Consumption of Metal Cans (NAICS 332431):  1993-1999
             (million cans)	[.:	'	2-22
                                  i                .                   '               •      .
       3-1    Summary of Costs to Industry Subcategories/Segments . .,	3-4

       4-1    Market-Level Impacts of the Metal Can MACT: 1997	4-7
       4-2    National-Level Industry Impacts of the Metal Can MACT: 1997  	'	4-8
       4-3    Distributional Impacts Across Facilities of the Metal Can
             MACT:  1997	'•	4-9
       4-4    Impacts for Facilities Not Included in the Market Model:  1997	4-10
       4-5    Distribution of Social Costs for the Metal Can MACT:  1997 . . . .	4-12

       5-1    Summary Statistics for SBREFA Screening Analysis: 1997	5-3
       5-2    Small Business Impacts pf the Metal Can MACT:  1997	 5-4

-------
                                   i     SECTION 1
                                   ;  INTRODUCTION
                                   i
       The U.S. Environmental Protection Agency (referred to as EPA or the Agency) is developing an
air pollution regulation designed to reduce emissions generated by the metal can manufacturing industry.
In the baseline for this analysis, the U.Sj metal can manufacturing industry was comprised of 202
establishments, whiqh were owned by 3p domestic and foreign companies and employed more than
160,000 workers.1 Of these facilities, 1^2 are classified as major sources of hazardous air pollutant
(HAP) emissions,2 primarily due to emissions occurring during the coating process.  Under Section 112
of the 1990 Clean Air Act (CAA) Amendments, EPA is currently developing national emission
standards for hazardous air pollutants (If ESHAP) to limit these emissions. This report presents the
results of an economic impact analysis (iEIA) in which a market model was used to evaluate the
economic impacts associated with the proposed regulation.
                                   I         '                            '
1.1     Agency Requirements for Conducting an EIA
       Congress and the Executive Office have imposed statutory and administrative requirements for
conducting economic analyses to accompany regulatory actions. Section 317 of the  CAA specifically
requires estimation of the cost and economic impacts for specific regulations and standards proposed
under the authority of the Act.  In addition. Executive Order (EO) 12866 and the Unfunded Mandates
Reform Act (UMRA) require a more comprehensive analysis of benefits and costs for proposed
significant regulatory actions.3 Other statutory and administrative requirements include examination of
the composition and distribution of benefits and costs. For example, the Regulatory Flexibility Act
(RFA), as amended by the Small Business Regulatory Enforcement and Fairness Act of 1996
(SBREFA), requires,EPA to consider the economic impacts of regulatory actions'on small entities. The
Agency's Economic Analysis Resource Document provides detailed instructions and expectations for
economic analyses that support rulemaking (EPA, 1999).

1.2     Scope and Purpose           i
       The CAA's purpose .is to protect and enhance the quality of the nation's air resources (Section
101(b)). Section 112 of the CAA Amendments of 1990 establishes the authority to determine a
NESHAP. This report evaluates the economic impacts of pollution control requirements placed on
metal can manufacturing establishments! under these amendments. These control requirements are
designed to reduce releases of HAPs into the atmosphere.
1These establishments  include those  that produce steel  or aluminum  cans,  metal
   sheets used for can production,  and/or  can ends.   Metal  cans are  primarily
   used  in  packaging  foods  and| beverages.    They  are  also  .used  in  general
   packaging applications for products such as paint  and aerosol  cans.

2A major source  of HAP emissions  :is defined as a  facility-'that  emits, or has the
   potential  to emit,  10  or mor^  tons  of any  HAP  or 25  or more  tons  of any
   combination of  HAPs.
3Office of  Management  and Budget! (OMB)  guidance under EO  12866 stipulates that
   a full benef it.-cost.. analysis  jis required  when.the regulatory  action  has an
   annual effect  on the economy oJE  $100  million  or more.

                                   :         1-1

-------
       To reduce emissions of HAPs, the Agency establishes maximum achievable control technology
(MACT) standards.  The term "MACT floor" refers to the minimum control technology on which
MACT standards can be based. For existing major sources, the MACT floor is the average emissions  -
limitation achieved by the best performing 12 percent of sources (if there are 30 or more sources in the
category or subcategory). For new sources, the MACT floor must be no less stringent than the emissions
control achieved in practice ,by the best controlled similar source.  The MACT can also be chosen to be
more stringent than the floor, considering the costs and the health and environmental impacts. This
report analyzes the economic effects of the metal can manufacturing MACT floor on existing sources.

1.3    Organization of the Report
       The remainder of this report is divided into four sections that describe the metal can
manufacturing industry, present the methodology used for the analysis, and summarize the results of this
EIA:                          '     ;
       •  Section 2 provides a summary profile of the metal can manufacturing industry. It describes
          the affected production process, inputs, outputs, and costs of production. It also describes the
          market structure and the uses and consumers of metal cans.

       •  Section 3 reviews the regulatory control alternatives and the associated costs of compliance.
          This section is based on EPA's engineering analysis conducted in support of the proposed
          NESHAP.                ;

       •   Section 4 outlines the methodology for assessing the economic impacts of the proposed
          NESHAP and the results of this analysis,  including market, industry, and social welfare
          impacts.  In addition, this section describes the economic impacts specific to new sources in
          the metal can manufacturing industry and economic impacts on the energy sector.

       •  Section 5 addresses the proposed regulation's impact on small businesses.

In addition to these sections. Appendix A further details the economic model used to predict the
economic impacts of the NESHAP and Appendix B presents the results of sensitivity analyses performed
for the demand and supply  elasticities used in the economic model.
                                             1-2

-------
                   :                  I                                                    -

                                     i     SECTION 2
                   INDUSTRY PROFILE:  METAL CAN MANUFACTURING

       Cans are one: of the most widely] used containers in the world. Industry estimates that more than
 200 million cans are used each day in trie United States (Can Manufacturers Institute [CMI], 1999a).
 Consumers use metal  cans for a variety jof purposes, including the storage of food, beverages, and many
 other products (e.g., paint). During the production process, a variety of surface coatings are applied to
 these cans. Interior coatings prevent corrosion and protect the contents from being contaminated by the
 can. Exterior coatings are applied for decoration, to protect printed designs, or to facilitate handling by
 reducing friction. Traditional coatings used in this industry have a high concentration of solvents, which
 results in the emission of volatile organic compounds (VOCs) and HAPs. Currently, the U.S.
 Environmental Protection Agency (EPA1) is developing national emissions standards for these HAPs.
       This section provides an economic overview of the metal can industry. Section 2.1 describes the
 production processes with emphasis on surface coatings.  Section 2.2 identifies uses, consumers, and
 substitutes.  Section 2.3 summarizes the! organization of the U.S. metal can industry, including a
 description of the manufacturing facilities and the companies that own them. In addition, we identify
 small businesses potentially affected byjthe proposed rule. Finally, Section 2.4 presents market data for
 the industry, including U.S. production, jprices, foreign Trade data, and trends.
                                     i
                                     i                                       •
 2.1    Production                   I
       The can manufacturing process has changed dramatically since its beginnings in the early 19th
 century. Today's automated processes Have replaced the once labor-intensive process and produce an
 estimated  139 billion cans per year (CMJI, 2001a). Metal can manufacturers purchase two primary raw
 material inputs for the production of caris:  steel and aluminum.  In 1999, almost three-quarters of all
 metal cans produced were aluminum (CMI, 200la). These two raw material inputs are used to produce
 one-, two-, and three-piece can bodies and can ends. During the production process, the steel or
 aluminum (in the form of sheets or coil) j is shaped, coated, quality checked, and prepared for shipment to
 a variety of consumers across the United States and the world. The following sections describe
 individual manufacturing processes in greater detail. Much of the information in these sections was
 taken from EPA (1998).               |
                                     !
 2.1.1   Sheet Manufacturing          \
       The process of manufacturing metal sheets for use in metal can manufacturing begins by cutting
 a large coil of metal into pre-scrolled sheets. An inside protective coating is then placed on the sheets
 and cured.  At this point the sheets can b;e decorated. An over coat of varnish is placed on the decorated
 sheet and cured again.  A second inside protective coating is placed on the sheets and cured.  These pre-
 scrolled sheets are then cut into small scroll sheets which can be fed into the end or body making process
 (CMI,2001b).      !                  i

2.1.2   Can End Manufacturing       \
       The production of can ends varies by end use. Aluminum beverage can ends are made from
precoated coil that is stamped and scored to produce an oval pattern, and an end tab is attached.  This
end is attached to the1 can with a solvent-! or water-based compound, and the seal is allowed to dry. The
production process of ends for food cansj and other sheet-coated ends is similar to beverage cans with the
                                              2-1

-------
exception that food can and other sheet-coated ends are typically coated on metal sheets rather than coils.

2.1.3  One- and Two-Piece Can Body Manufacturing
       The one- and two-piece can manufacturing process, involves forming a can body, creating an end
(for the two-piece can), and applying Coatings to the open can and can top. Two fabrication processes
are used to produce these cans: the draw-redraw process and the draw-and-iron process. Manufacturers
of one-piece can bodies use the draw-and-iron process, while two-piece can manufacturers use both
processes.                            :
       During the draw-redraw process, aluminum or steel coil is fed into a processor called a cupper
that stamps shallow metal cups. The coil may be stamped one or two additional times to create a deeper
can.  This process typically uses pre-coated coils and if no additional coating steps are required, the cans
are tested and stored. However, some manufacturers use an uncoated coil and perform sheet coating
similar to the three-piece can body coating operation described in  Section 2.1.4.
       In contrast, the draw-and-iron process involves the following additional steps after the shallow
cup is created. Full-length can bodies are created from shallow cups through an extrusion process
(aluminum cans) or "ironing" process (steel cans). The can bodies are then trimmed, cleaned, and dried
in preparation for the application and curing of exterior base coats, printing inks, and protective
overvarnish coats (aluminum beverage cans) or corrosion-resistant wash coats (steel food cans). Once
the coatings are dry, the can necks are flanged (beverage) or beaded  (food cans). A leak tester applies air
pressure to each can and tests for any holes or cracks and rejects any inadequate cans. In addition, the
coating thickness may be tested by a random electrical resistance spot check. After passing these tests,
the finished cans are then stacked for storage or shipment., Figure 2-1 provides a detailed example of a
two-piece draw-and-iron aluminum beverage can production process.
                                     i                     '                        '
2.1.4  Three-Piece Can Body Manufacturing
       'Three-piece cans are typically made of steel sheets. The manufacturing process involves two
operations:  sheet coating and can fabrication (see Figures 2-2  and 2-3).  The sheet coating operation
includes the application of a base coat, inks, and overvarnish.  After application, the sheet passes through
an oven for curing and drying. The can fabrication begins with the processor slitting these coated sheets
and feeding them into a "body maker" where the seams are welded or cemented together. The seam
along the side of the can is welded or cemented and then coated in a process called "side seam stripe
application." This seam may be coated with an interior spray or an exterior spray, or on both sides.  The
side seam stripe protects exposed metal along the seam. At this stage of the production process, the cans
are flanged for proper can end assembly and the diameter of the wall may be reduced (necked-in)
according to end-use requirements.  In addition, if the can will be  used to store beverages, the can's
interior is sprayed with a protective coating and then baked or  cured. After curing, the end seamer
attaches one end to the can in a process called "double seaming" where end seal compounds are applied
and used as a gasket material to provide an airtight seal. Afterwards, the leak tester checks for leakage.
The finished can is stacked and prepared for shipment.
                                     i
2.7.5  Coatings and Emissions       \  '
       Coating is an integral part of the production processes of cans and can parts.- Without the
specialized interior coatings, cans could potentially contaminate their contents and render them
dangerous to consumers. Exterior coatings enhance the can's appearance, protect the can from
                                               2-2

-------
  0)
  to
  
-------
                                                                  e
                                                                  o
                                                                  •rl
                                                                  -P
                                                                  (C
                                                                  M
                                                                  0)
                                                                   Cl
                                                                  •H
                                                                  •P
                                                                   n)
                                                                   O
                                                                  O
                                                                   n)
                                                                  PQ
                                                                  -P
                                                                   d)
                                                                   nS
                                                                  U
                                                                   O
                                                                   Q)
                                                                  •H
                                                                  CM
                                                                   I
                                                                  CM
                                                                   I
                                                                  
-------
           sl
                                                              CO
                                                              CO

                                                              s
                                                              o
                                                              M
                                                              a
                                                              o
                                                             •H
                                                             -P
                                                              n>
                                                              o
                                                             •H
                                                              M


                                                             I
                                                             fa

                                                              a
                                                              fl!
                                                             u
 rs
 o
 o
 o
 CO

 a
 u
 0)
•H
CM
 I
 0)
 (1)
                                                             n
                                                              i
                                                             CM

                                                             Q)
2-5

-------
corrosion, and protect printed designs. However, the traditional coatings used in the metal can industry
have a high concentration of solvents, which results in the emission of VOCs and HAPs. Several types
of coating technologies exist:          ;
       •   Conventional solvent-borne coatings—Conventional coatings offer good abrasion resistance
           and ease of application. However, they have high concentrations of VOCs and HAPs.

       •   High-solid coatings—The most widely used high-solid coating is polyurethane.  These
           coatings are used as exterior bases, some interior sheet coatings, decorative inks, and end seal
           compounds.

       •   Waterborne coatings—These: coatings are used extensively in beverage can manufacturing.

       •   Ultraviolet radiation-cured (UV-cured) coatings—UV-cured coatings offer advantages of
           rapid curing, low process temperatures, and low VOC and HAP content as well as lower ,
           energy costs because drying ovens are eliminated. However, UV coatings are expensive and
           require specialized equipment.

       •   Powder coatings—These coatings offer excellent resistance to chemicals, abrasion resistance,
           and barrier qualities. The application process for these coatings is currently not fast enough
           for can coating line operating speeds, and only limited numbers of colors, finishes, and
           textures are available for can manufacturers (EPA, 1998).

Coatings are applied to both interior and exterior can bodies and ends. Emissions are generated during
coating application, during transportation to the oven (evaporation), and during curing.  However,
approximately 50 to 80 percent of emissions occur during the drying and curing process (EPA, 1998).

2.1.6  Costs of Production
       Raw material and energy costs account for the largest share' of the variable costs of metal can
production. In 1997, the cost of materials and energy totaled $8.6 million, or 72 percent of the metal can
industry's value of shipments. Steel and aluminum purchases totaled $8.1 million, or 94 percent of the
cost of materials.                     ,
       Recently, prices for steel and aluminum sheet, plate, and coil have fluctuated given the changes
in market conditions for these inputs. For 2001, Purchasing Online (2001) reported spot prices for a
cold-rolled steel sheet at $320 per ton. coiled-steel plate  at $288 per ton, and aluminum common alloy
sheet at $1,720 per ton (see Table 2-1).  The data show the price of steel has dropped significantly since
1997 as foreign steel imports have surged.  For September 1997, spot prices for cold-rolled steel sheet
and coiled steel plate were quite a bit higher than more recent levels at $480 and $390 per ton,
respectively.  In 1995. a shortage of aluminum led to significant raw material price increases, forcing
beverage canners, such as Coca-Cola and Pepsico, to increase the use of alternative packaging containers
such as plastic bottles (Sfiligoj, 1995). However, aluminum prices decreased significantly in 2001.
                                              2-6

-------
 Table  2-1.
 2001
Spiot  Prices  for Steel and Aluminum Sheet and  Plate:  1997-
                    Year
                                               1997
                                                         1998
                                                                  1999
                                                                            2000
                                                                                      2001
 Cold-rolled steel  sheet (Midwesti,
 $/ton)            ,             ,    j
                                    I
 Coiled steel plate  (Midwest,  $/tp'n)

 Aluminum (common alloy sheet  300J3,
    $/ton)         ;       '          |
                               $480
$410
$390
$380
$320
                               $390      $400      $300      $320      $288

                             $2,200    $1,920    $2,040    $2,240   $1,720
Source:      Purchasing Online,  September 15,  1998.  "Transaction Prices."  Purchasing
              Online.               !
       Purchasing Online.  September 16, 1999.  "Transaction Prices."   Purchasing Online

       Purchasing Online.  September 20, 2001.  "Transaction Prices."   Purchasing Online
       Labor is used throughout the production process as. well as during transportation of the product.
However, labor costs account for only ajsmall share of variable production costs in the metal cans
industry. In 1997, payroll represented only 10 percent of the value of shipments.
       In 1995, industry estimated that approximately 20 million gallons of coating materials were   •
consumed annually by two-piece beer and beverage can manufacturers (Sfiligoj. 1995). A more recent
estimate shows that two-piece beverage ^manufacturing facilities used 26 million gallons of coating in
1997 (Reeves, 1999). Using data on the; volume and value of coatings shipped to the metal coil coating
industry, the Agency estimates the average cost of coatings for 1997 at $15.60 per gallon (Bourguigon,
1999).  However, it is likely that some specialty coatings sell for substantially more—as high as $50 per
gallon.                             j
       The U.S. Bureau of the Census (Census Bureau) and U.S. Bureau of Labor Statistics (BLS)
publish historical statistics for costs of njtaterials (i.e., materials;fuels, electricity) and labor for the metal'
can industry using the following classification systems:
       •  North American Industrial Classification System (NAICS)—beginning with the 1997
          Economic Census, the metal cans industry was classified under NAICS code 332431, Metal
          Can Manufacturing.
          1987 Standard Industrial Classification (SIC) codes-
          was classified under SIC 34111, Metal Cans.
                                        -prior to 1997, the metal cans industry
As shown in Table 2-2, the cost of materials averaged 72 percent of the industry's value of shipments
between 1992 and 1997, while payroll represented roughly 10 percent of the value of shipments. Wages
for production workers ranged from $15i86 to $17.34 per hour during this period.
                                   •i
2.2     Uses, Consumers, and Substitutes
       Historically, steel cans were primarily used to store prepared raw food products.  During the
1970s and 1980s, the use of metal cans expanded to the beverage market, and aluminum cans
subsequently captured a significant shar? of the market (Hillstrom, 1994).  Today, it is estimated that
Americans use approximately 200 millitin cans each day.  Metal cans are used for a wide variety of
products, such as soft drinks, food products, and aerosol cans.  Table 2-3 lists selected end uses for metal
cans.                          .      j    	
                                             2-7

-------
      In 1997, the baseline year selected for this analysis based on data availability, more than 130
billion metal cans were shipped to three primary market segments—beverage, food. and. general
packaging (CMI, 1999b).  Figure 2-4 shows the distribution of shipments of metal cans by market for
1997. As shown, the beverage market accounts for the largest share of metal cans (73.4 percent),
followed by food (23.4 percent) and general packaging (3.2 percent).
Table  2-2.   Historical Cost  of Production Statistics for the Metal
Cans Industry:   1992-1997   ;
Year
1992
1993
1994
1995
1996
1997
Total/Avera
, ge
Value of Cost of
Shipments Materials
($106) ($106)
$12,
$11,
$11,
$12,
$12,
$12,
$71,
112
498
610
326
273
007
825
$8,
$8,
$8,
$9,
$8,
$8,
$51,
798
360
306
084
624
598
770
Cost of
Materials
Share (%)
72
72
' 71
'73
70
71
72
.6%
.7% "
.5% '
.7%
.3%
.6% '
.1%
Average
Payrol Earnings of
1 Payroll Production
($106) Share (%) Workers ($/hr)
$1,
$1,
$1,
$1,
$1,
• $1,
$7,
262.
212
256
183
194 .
183
485
10.4%
10.5%
10.8%
11.2%
9.6%
9.8%
10.1%
$15
$16
$16
$16
$16
$17
$16
.86
.23
.50
.74
.98
.34
.61
Source,s:     U.S.  Department of Commerce,  Bureau of the Census.  1999a.  1997 Census of
             Manufacturing    Industry    Series:       Metal    Can '   Manufacturing.
             .              •
      U.S.  Department of Commerce, Bureau of  the Census.   1998.   1996 Annual Survey of
      Manufactures     Statistics     for     Industry     Groups     and    Industries.
      .
      U.S.  Department of Commerce, Bureau of  the Census.   1997.   1995 Annual Survey of
      Manufactures     Statistics     for     Industry     Groups     and    Industries.
      .
      U.S.  Bureau of Labor Statistics.  National Employment,  Hours,  and Earnings—Metal
      Cans:   Series  ID eeu31341106.    .   As obtained  on August 27,
      1999.                       ;                  '        '      '•         ...,•..

                                  I

      CMI reports that nearly all beverage cans are made of aluminum. A recent survey conducted by
the aluminum beverage can industry identified characteristics of aluminum cans that consumers found
attractive compared to other packaging alternatives (CMI, 1999c). These include

      •   less spillage or breakage,    [

      •   ease of storage at home or when traveling,

      •   maintenance of soft drink carbonation, and

      «   ease of recycling.

      The ability to recycle aluminum'cans is one reason why they continue to dominate other
                                           2-8

-------
                               I
packaging alternatives in the carbonated soft drink (CSD) market, one of the largest segments of the
market. CMI estimated that in 1998, twjo out of every three manufactured aluminum beverage cans were
recycled as new cans, a process that takes approximately 60 days (CMI, 1999d).
Table 2-3.   Metal Can Uses|by Material and  Type
       Type
          Material
            Used
Products  Contained
Three-Piece Can   Steel
Body
Two-Piece Can
Body
   Draw-iron
   Draw/redraw
                      Food,  juices, spices,  aspirin,
                    I  paints,  glue, aerosols (includes
                    I  decorative tins)
                               I
        Aluminum      Beer,  carbonated beverages,
                    ]  juices
        Steel       j  Food,  other nonfood
        Steel,      , i  Food,  shoe polish, sterno, fuel,
        aluminum    ]  car wax,  other nonfood products
Source:
U.S.  .Environmental  IProtection  Agency.   1998.    "Preliminary  Industry
            Characterizati on:
                    ! Metal
                   A,
                                         Can
                                                Manufacturing—Surface
            .
                               i
                               Coating."
                               i       1997
                              l|35,468 Million Cans
                  Beverage
                   73.4%
                                                  Food
                                                  23.4%
                                           General
                                          Packaging
                                            3.2%  '
   Figure 2-4.   Distribution of  Metal  Can  Shipments by  End Use:
   1997                        i    •
                               i               •
   Source:     Can Manufacturers jlnstitute  (CMI).  "Domestic  Can .Shipment 1997."
               .    Obtained  August  31,
              1999c.           :    .         .. .. .     .
                                       2-9

-------
 In 1997, aluminum cans accounted for 75.7 percent of the soft drink packaging mix followed by plastic
(19.9 percent), glass (2.3 percent), and other (2.1 percent) (see Figure 2-5). Despite the current
dominance of aluminum beverage containers, the use of polyethylene terephthalate (PET) bottles has
recently experienced growth due to the widespread availability of the polymer and its low cost (O'Neill,
1998). Aluminum cost increases in the mid-1990s encouraged soft drink canners to substitute bottles
made of PET.  The glass CSD container share, on the other hand, is small and declining. For example,
the Census Bureau (1999a) reports shipments of glass bottles fell 14 percent from 1997 to 1998.
                                                        Plastic
                                                          9.9%
                  Metal Can
                    75.7%
                                                               Other
                                                               2.1%
Glass
2.3%
Figure 2-5.   Distribution  of  Soft Drink  Packaging  Mix by Type:   1997

Source:       Can Manufacturers Institute (CMI) .   "1997 Retail 'Sales Prove It's Better
              in Cans."  Canline 1(2).  .
              As obtained  on August 31,  1999a.
       Another important beverage segiinent is the beer market. Aluminum beer containers accounted
for approximately one-third of metal carj beverage shipments in 1999 (CMI, 200 la). Small aluminui.ii
cans (60 percent) and glass bottles (27 percent) dominate the beer market, with bulk packages such as
kegs accounting for the remaining 13 percent (Brody and Marsh, 1997). Recently, plastic containers
have entered the single-service beer market.
       A variety of alternative packaging methods in the food/general packaging containers market
exist. The primary factors in deciding \4hich type of material to use in packaging are temperature
control, counterpressure, and shelf-life, but in most cases plastic of glass can be substituted for metal
(Brody, 2001).  '
       Plastic containers have enjoyed widespread use since the 1970s, but this use has been
concentrated in the beverage market.  In 1998, only about 1 billion plastic containers were 'used in food
packaging versus 32 billion metal containers (Brody, 2001). Steel food can manufacturers have
primarily been affected by the increasing use of plastic in a limited number of food market segments as
they face increased competition from microwave and frozen food products using plastic packaging
(Hillstrom, 1994).  Plastic also has the advantage of being impact resistant, heat resistant, and
transparent. PET is often used as a glass replacement in both food and beverage bottles (Brody and
Lord, 2000).
       Glass is also used in food packaging. It is usually found in the form of wide mouth containers
(i.e., jars).  Approximately one half of glass containers are used for baby food.   Glass is much more
prevalent in the food packaging industry than is plastic (approximately nine times more glass containers
                                             2-10

-------
 are used) (Brody, 2001). Although consumers desire the transparency of glass, it might be less than
 desirable from the perspective of food preservation because light can accelerate reactions in the food.
 Although it can be substituted for metal jor plastic it is very heavy, breakable, and energy intensive to
 produce (Brody and Lord, 2000).      ;|  '*;         ;,
       Paper and paperboard are the mqst widely used package materials in the world.  However, in
 order to protect food from moisture, gasj, odors, or microorganisms, they must first be coated with
 plastic. For this reason, they are infrequently used as substitutes for glass, plastic, and metal in the food
 and beverage industry (Brody and Lordj 2000).
       Prices of raw materials can significantly influence beverage producers' choice of container
 material because containers represent a large share of the product's cost and because several substitute
 materials exist.4 For'example, aluminum can prices increased nearly 14 percent between 1994 and 1995,
 leading several manufacturers to consider expansion of plastic packaging methods (Sfiligoj, 1995).
       In addition to this anecdotal evidence, there is some quantitative data suggesting substitution
 between container materials based on relative prices. Aluminum can shipments in the beverage market
 declined, by 5 billion units, or 4.6 percent, from 1994 to 1995, as aluminum can prices rose relative to
 PET bottles. Since 1995, the price, of altiminum cans has fallen relative to PET, and shipments of
 aluminum cans have risen close to 1994|levels. A simple regression of the ratio of aluminum and PET
 prices on shipments.of aluminum cans provides an elasticity estimate of-0.6.5 In other words, a
 1 percent increase in the price of aluminum cans relative to PET bottles is estimated to reduce the
 quantity of aluminum cans demanded by 0.6 percent.
       Although the .cost of steel cans has remained constant over this period, sharp reductions in raw
 steel prices in 2000 and 2001 suggest lo^jver costs of steel cans in the future. However, in addition to
 declines in metal prices, plastic resin costs.have fallen since 1995, which makes plastic containers more
 attractive (O'Neill, 1998).  In fact, all of the major materials used in food and beverage packaging
 (aluminum, steel, plastic, and glass) have been declining in price over the last few years  in inflation-
 adjusted terms.                       I
                                    j                   -            •
 2.3    Industry Organization        j
       This section provides an overvie\|v of the market structure of the metal can manufacturing
 industry, including the facilities, the companies that own them, and the markets in which they compete.
^Economic  theory  suggests the elasticity of the derived demand  for an  input is
   a function of  the cost  share  of the  input .in  total  production  cost and the•
   elasticity of substitution between this  input and  other inputs in production
   (Hicks,  1966).   Because  the cost share of  containers  is relatively large and
   there  are  good; substitutes  available,'  we  may  infer an  elastic  demand  for
   aluminum   beverage  cans.    Containers  used  in  food  or  general  packaging
   applications  (e.g.,  steel  cans|  typically  have much smaller  cost shares  than
   those used for beverages  (because  the products  contained in  them often  have
   far higher values  than beverages)  and would be expected to face  less elastic
   demand curves.  .                 I           .           : .
5The model  estimated was lnQA
= j a  +
       b  In
                                                "PET
,  where  QA1  is  the quantity of
   aluminum cans;  PPET and PA1  are inflation-adjusted  price indices of PET bottles
   and aluminum cans,  respectively;  and a and b are parameters  to be estimated.
                                            2-11-

-------
2.3.1  Market Structure
                                                                                       n
                          3 ft£efctly%8mpetitive, then individual producers are not able to influence
the price of the output they sell or the inbuts they purchase. This condition is most likely to hold if the
        hds SM^peflfcmber of firms, the products sold and the inputs purchased are homogeneous, and
entry and(l^6of^finnsjre unrestricte^tntry and exit of firms Sfe8unrestricted for most irlffiistries
except. fcfr^X&fiple, in cases where get?^rjament regulates who iiatole to produce, where oasiirm holds
a patent on a product, wnere one iirm owns the entire stock of a critical input, or where a single firm is
Hbteto: Supply U^enfeB3fil^k£tJur- firm  concentration ratio.
                            infiM^Pf^ft5sL(eR^5ridlCR8, respectively) and
                                                               cfo^pflfivlhess of an   .
       ;. The
the most rec
bottle industry are reported in Tables 2-4, 2-5, and 2-6, respectively.

Table  2-4.   Measures  of  Market  Concentration  for  the  Metal Cans
Indus¥Kff-Crfi!SM§.

                                  I
Merger Guidelines. According to these ;criteria, industries with HHIs below 1,000 are considered
unconcentrated (i.e., more competitive), those with HHIs between 1,000 and 1,800 are considered
moderately concentrated (i.e., moderately competitive), and those with HHIs above 1,800 are considered
highly concentrated (i.e., less competitive). In general, firms in less-concentrated industries are more
likely to be price takers, while firms in more-concentrated industries are more likely to be able to
                                  |

Table 2-5.   Measures  of  Market  Concentration  for the-Glass Containers
Industry  (NAICS 3,27213) : 1997
  Value of Shipments
        ($10e)
         CR4
                               CR8
                                                     HHI
        $4,198
                                91%
                                                      98%
                                                                            2960
Notes:       CR4 denotes  four-firm concentration ratio.
       CR8 denotes eight  firm  concentration ratio.
       HHI denotes Herfindahl-Hirschmann index for 50  largest  companies.
Source:      U.S. Department  of
             Ratios in
Manufacturing
Commerce, Bureau  of  the Census.  2001.  Concentration
       .
                                           2-12

-------
influence market prices.              i
       In the metal can industry, the C$.4 was 58 percent, while the CR8 was 87 percent. The HHI for
this industry was 1,180. Based on the criteria above, the metal can industry can be classified as
moderately concentrated.             '   :          ;:   ;
       With only 11  companies, the glass container industry was concentrated with a CR4 of 91 percent
and a CR8 of 98 percent. The HHI for this industry implies that it was highly concentrated.
       In the plastic bottle industry, the;CR4 was 33 percent and the CR8 was 52 percent. With an HHI
of 425, the plastic bottle industry can be; classified as unconcentrated.
       Although the  metal can industry iappears to fall at the lower end of the moderately concentrated
range, the close substitutability of alternative materials such .as glass and plastic makes it likely that
metal can producers behave as price-takers.  Thus, based on the CR4, CR8, HHI, and the available
substitutes, an assumption of perfect competition for the metal can industry appears reasonable for
modeling purposes.                  j
                                    i
2.3.2  Facilities                    I
       In the baseline for this analysis, 202 potentially affected facilities manufactured metal cans,
sheets, or ends in the.United States.6 Th]ese facilities can be classified as one of two types of producers:
independent can manufacturers and captive can manufactures.  Independent can producers coat and
fabricate cans based on the customer's specified end use. Several of these plants manufacture cans
solely for one customer (EPA, 1998).  Cjaptive can producers coat and fabricate cans as part of the
vertical operations of a parent corporation.  The great majority of metal cans are produced by
independent can producers rather tnan for captive use (see Section 2.3.2 for more information).
       The size of can manufacturing plants varies depending  on the number and types of production
processes performed;  Some plants coat fmly the metal sheets,  while others may fabricate a particular
type of can body or end from the coated jsheets. Others both coat and fabricate the metal can.
       Metal can manufacturing facilities are generally located near sources of material supply (i.e.,
steel or aluminum plants) or near the customers based on the costs associated with transporting raw
materials and final products.  Figure 2-6J shows the distribution of these facilities across the United
States. California contains the most meljal can, sheet, or end manufacturing facilities (29), followed by
Ohio (19), Illinois (15), and Wisconsin (J13).
2.5.5  Companies                   \   '
       Thirty parent companies own the! 202 metal can manufacturing facilities.  These companies
report an average (median) annual sales bf $3.8 billion ($336 million). This figure includes revenue
from operations other than metal can manufacturing. The average (median) employment for these
companies was 17,400 (2,566) workers, j Three of the largest companies, based on annual sales, produce
containers as part of the company's vertical operations (i.e., Nestle S.A.—$52.1 billion, Con
Agra—$23.8 billion,  and H.J. Heinz Company—9.3 billion). However, these companies own a total of
only seven facilities, or 3.5 percent of thfc establishments.  Ward's Business Directory (Gale Research,
1999) identifies the top metal can manufacturing companies (i.e., those withNAICS 332431  as a
primary SIC) as Crown Cork and Seal Cjompany ($8.3  billion), Ball Corporation ($2.8 billion), and
American National Can Company ($2.4 |billion), all of which are independent metal can manufacturers.
These companies own 82 facilities, or 43^ percent of the total. Additionally, Silgan Holdings Company is
a major independent metal can manufacturer in this market:  they own 34  facilities (annual sales are $1.7
sThat  is,  there  were  202 facilities  classified in  the  metal can manufacturing
   industry.    However,  eight  of jthese  facilities are  classified  as  synthetic
   minor sources  and 52  as  area  Sources,  neither  of  which incur  any compliance
   costs under this  regulation.                      •
                                             2-13

-------
billion).
                                                    2-14

-------
                                                                         -P

                                                                         01
                                                                         •H
                                                                         -P
                                                                         •H
                                                                         H

                                                                         •3

                                                                         £
                                                                        •rf

                                                                         M
                                                                        ,


                                                                         g
                                                                        4J

                                                                        OS
                                                                        w


                                                                         cT
                                                                         rt
                                                                        -P

                                                                        2
                                                                        C
                                                                        o
                                                                        W

                                                                        •H

                                                                        Q
                                                                        I
                                                                        CN


                                                                       'a
                                                                       •rl

                                                                       fc

                                                                       r-
                                                                        g,
                                                                       •H
                                                                       •fa
2-15

-------
       Metal can coating companies can be classified as small or large businesses using Small Business
Administration (SBA) general size standard definitions for.NAICS codes. For NAICS 33243L the SBA
defines a business as small if it employs a,000 or fewer employees. Using this guideline and available
secondary data, the Agency identified 13 small businesses, or 43.3 percent of the metal can companies.
For these small businesses, the average (median) annual sales for companies reporting data were $27
($24) million, and the average (median) employment was 178 (175) employees. Appendix A lists
individual metal can companies and includes sales and employment data reported by secondary sources,
including Dun & Bradstreet (1999), Hoover's Inc. (1999), and company and industry websites.

2.4    Market Data and Trends
      'Growth in the metal can industry during the  1990s has slowed as a result of a mature domestic
market for aluminum and steel cans. As shown in Table 2-7, domestic shipments were reported at 137
billion cans in 1997 (baseline year), a small increase of 1.2 percent over 1996.  During the period 1993
to 1999, total metal can shipments increased at an average annual rate of 1 percent.
       There are a variety of metal can products, and prices vary by size and end-use application.  The
Agency conducted a search for can price data by type of can and found that this information is not
published in a statistical annual. However, an industry trade journal did report spot prices for aluminum
and steel beverage cans as well as plastic bottles for 1995 (Sfiligoj, 1995). Using these spot prices and
the producer price indexes published byithe BLS, the Agency computed a historical price time series for
these selected cans for the period 1993 through 2000. As shown in Table 2-8, the average prices per
1,000 units during this period were as follows: aluminum cans ($62.47), steel cans ($65.28), and plastic
bottles ($68.51).                     ;
Table 2-7.   Domestic  Metal ; Can  Shipments  by  Market:
 (million cans)
1993-1999
Year
1993
1994
1995
1996
1997
1998
1999
Beverage Food
97,
103,
98,
99,
100,
102,
102,
605 i. 30
119 I 31
116
136
680
789
253
: 31
31
31
; 31
32
,465
- 9°7
,313
, 971
,998
, 782
;349 ' .
Average Annual Growth
1993-1999
1%
t .1%
General
Packaging
4,
4,
4,
4,
4,
4,
4,
Rates
2%
072
228
275
361
375
404
457


Total
132,
'139,
.133,
;135,
.137,
138,
139,

1%
142
254
704
468 .._
137
975
059

 Source:      Can  Manufacturers   institute  (CMI).    "Historical   CMI  Can  Shipments."
              .  As obtained on December 6, 2001a.
                                            2-16

-------
Table  2-8.    Prices for  Beverage  Containers:    1993-2000  ($/l,000  cans
or bottles)
                                     I
         Year
Aluminum Clans
Steel Cans
PET Bottles
1993
1994
1995
1996
1997
1998
1999
•2000
Average:
$63.99
$61.01
$70.58
$63.02
$60.94
$61.01
$59.14
$60.04
$62.47
$64.78
$64.78
$65.66
$65.81
$65.76
$65.76
$65.30
$64.37
$65.28
NA
$65.23
$70.68
$68.57
$68.63
$67.73
$67.99
$70.75
$68.51
Sources: Sfiligoj, Eric. June 1995. "At What Prjce?" Beverage World.
       U.S. Bureau of Labor Statistics. Producer Price Index—Commodities:  Aluminum Cans—Series ID wpu!03103.
       . As obtained on Djecember 6, 2001 a.
       U.S. Bureau  of Labor Statistics.  Prbducer Price Index—Commodities:  Steel Cans—Series ID wpu!03102.
       . As obtained on December 6, 2001 b.
       U.S. Bureau of Labor Statistics. Producer Price Index Revision—Current Series: Plastic Bottles—Series ID pcu3085#.
       . As obtained on Dje. ,-mber 6, 200 Ic.

       Currently, foreign trade does not represent a significant share of metal can shipments. For 1996,
the value of imports and exports as a share of the total value of shipments for NAICS 332431 was less
than 1.5 percent. However, foreign interjest in the benefits of aluminum can packaging is growing and
this is expected to benefit U.S. producers of aluminum cans (Hillstrom, 1994). There has been growth in
exports since 1992, although exports peaked in 1995 and have generally been declining since then (see
Table 2-9). Similarly, imports (primarily from Canada) have risen between 1992 and 2000 but peaked in
1996 and have been on a downward trend. It is unclear why trade spiked in the mid-1990s and has since
been falling.  Even in the peak years, trade was a very small fraction of total production and
consumption of metal cans. Because imports and exports are such a small percentage of total shipments,
apparent consumption of metal cans in th!e U.S. does not differ greatly from total shipments by domestic
producers (see Table 2-9).             I

       In the domestic market, the aluminum container has become widely used because of its relative
advantages in price and weight as well as opportunities  consumers have to recycle it. The beverage
market grew rapidly during the 1980s and 1990s and began to dominate the entire can industry.
Aluminum has a 75 percent market shares in the beverage segment, experiencing rapid growth along with
the beverage industry.  As beverage industry growth has leveled off, so have sales of aluminum cans.
Although steel represents a declining share of the beverage market, steel cans still dominate the food and
consumer product markets.  However, they face increased competition' from food product packaging
using plastic materials.  Exports of both food and beverage products are anticipated to increase based on
                                             2-17

-------
Table  2-9.   Apparent  Consumption  of Metal  Cans  (NAICS  332431):
1993-1999  (million cans)    |
Shipments by Domestic
Year Manufacturers ;
1992
1993
1994
1995
1996
1997
1998
1999
2000

N/A ; ,
1 ,.
132,142 1
139', 254 !
133,704 j
135,468 ;
137,137 I
138,975
139,059 ;
N/A :
Average
1% 1
Imports
335
461
711
.559
1,454
627
334
691
634
Annual Growth
28%
Exports
395
568
1,390
2,196
'899
861
967
624
674
Rates
21%
Apparent
Consumption
N/A
132,035
138,575
132, 067
136, 023
136,903
138,342
139,126
, N/A
1%
Sources:     U.S.  International  Trade  Commission.    ITC. Trade Data  Web..  Version  2.4
             [     c    o    m   p [ : u    t    er           f    i    l:e    ]
             .    As  obtained' on  December  '7,
             2001.                ! !                     '                       -' ' .'     :
                                  i
      Can Manufacturers  Institute! (CMI) .   "Historical CM I Can Shipments'. •
      cancentral.com/>.   As obtained on December  6,  200la.
shttp://WWW.
trends established during the. 1990s. For example, between 1990 and 1992 soft drink and carbonated
 water exports increased 63 percent and fruit and vegetable exports increased approximately 32 percent
(Hillstrom, 1994). However, it is not clear that these trends will lead to increased exports of metal cans.
Because of the low value-to-weight ratio of metal cans, it appears unlikely that foreign trade in cans will
develop to a significant degree. On the Other hand, an increase in food and beverage exports may lead to
an increase in demand for metal cans sii^ce they may be used to package the exported items.
                                          2-18

-------
                                     j  '   SECTIONS
                                    ENGINEERING COSTS
                                     i
       This section presents the Agency,'s estimates of the compliance costs associated with the
regulatory alternatives developed to reduce HAP emissions during metal can coating operations. This
NESHAP will limit the amount of organic HAP emitted relative to the volume of coating applied. To
meet the requirements of this regulation,! most facilities will add control devices, with some facilities
substituting low- or no-HAP coatings for their current coatings. The tabular costs associated with
making these changes to the metal can production process were estimated for the 142 major source
facilities operating in the U.S. in the basfeline year. 1997. These costs are defined as the annual
recordkeeping and reporting, material, capital, and monitoring costs assuming no behavioral market
adjustment by producers or consumers. [The engineering costs will serve as an input to the economic
model, which incorporates behavioral adjustments, presented in Section 4.  An overview of the
methodology used to develop the engineering cost estimates is provided below. A more detailed
discussion of this methodology ana the assumptions used for the calculations can be found in Icenhour
(2002).                        .!....,.
                                     i                   .
3.1    Methodology                 j
       EPA identified three potential types of costs associated with pollution abatement:
(1) monitoring, recordkeeping, and reporting (MR&R) costs, (2) material costs, and (3) capital costs
related to the purchase and installation of add-on capture and control devices.  Each of the cost
components is briefly described below, j

3.1.1  Monitoring, Recordkeeping, and Reporting Costs
       MR&R costs are divided into six! types, including the cost of labor to track material usage and to
compile data for compliance reports; thejcost of buying and maintaining computer equipment to track
coating and solvent material usage; the cpst associated with buying and maintaining continuous
parameter monitoring systems for the adfl-on control devices; the cost of photocopying and mailing the
reports and notifications; the cost.of purchasing filing cabinets for recordkeeping purposes; and the cost
of hiring a contractor to conduct performance testing of the add-on control devices and monitoring
systems. The average annual total facility cost associated with MR&R activities is estimated to be
$52,700, for an industry total of $7.3 million.  Facilities that are subject to multiple subcategories have
this MR&R cost divided evenly among the subcategories such that their total facility cost is $52,700.

3.1.2 Material Costs                 \
       This cost component characterizes the of costs of substituting low- or no-HAP coatings for the
coatings currently being used. For this analysis, EPA assumed that facilities in well-controlled
subcategories such as, two-piece beverage cans, two-piece food cans, and sheetcoating operations will
                                              3-1

-------
meet HAP emission limits by installing a new regenerative thermal oxidizer (RTO) rather than incurring
material costs. In addition, three facilities that are within 10 percent of the organic HAP emission rate
for the well-controlled coating type segments were assumed to meet the limits by improving the existing
capture device. All other subcategories,; except for one-piece aerosol can facilities, are assumed to
reformulate the coatings to limit surfaceicoating HAP emissions.
       Because reformulation costs vary by type of coating, the Can Manufacturers Institute (CMI) was
consulted for accurate cost ranges.  Based on these data, an average cost was estimated for each specific
coating type segment. Costs were calculated using the assumption that each facility will use the same
amount of coatings that were consumed in the baseline year of 1997 and that there will be a greater cost
per gallon for low- or no-HAP coatings pompared to the cost per gallon for higher HAP-content
coatings. This incremental cost increase is assumed to be $2.00 per gallon for inside sprays and-$5.00
per gallon for side seam stripes, which are used in three-piece food can assembly and three-piece
nonfood can assembly subcategories, and $2.00 per gallon for non-aseptic end seal compounds, which
are used in the end lining operations subcategory. The total estimated impact for material costs is
estimated to be $4.1 million per year for the three impacted subcategories.

3.1.3   Add-On Control Devices
       In general, the two-piece beverage cans, two-piece food cans, and sheetcoating subcategories are
well-controlled in terms of air emissions.  Therefore, EPA assumed that all facilities in these
subcategories will require an RTO to meet the emission limit with two exception's. First, if the facility
has an organic HAP emission rate that is less than or equal to the organic HAP emission rate for the
coating type segment, the amount of control is considered  sufficient. Second,  if the facility has an
organic HAP emission rate that is less than 10 percent above the organic HAP emission rate for the
coating type segment, it is assumed that the facility can meet the limit by adding equipment to the
existing capture equipment.  The capital cost for this investment is estimated to be $400,000, which,
when annualized over 10 years at 7 percent, is an annualized cost of $98,000.  For all other major source
facilities, facility-specific capital equipment costs were estimated that include purchase, installation, and
operation of an RTO. Capital investment costs were annualized over a 10-year period with an interest
rate of 7 percent.  The total annualized capital cost for all facilities is estimated to be $44.8 million.

3.2    Engineering Cost Summary   '
       The Agency's facility level engineering cost estimates are summarized in Table 3-1 for each of
the 142 major sources and 8 synthetic minor sources in the metal can manufacturing industry. The
nationwide total cost is estimated at $56.2 million per year divided across  142 major source facilities.
This cost is divided among MR&R costs of $7.3 million, material costs of $4.1 million,- and capital costs
for add-on control devices of $44.8 million.
                                              3-2

-------





























)
1
1
il
1
1
i
'
)
I
>
n
)
>



i














1










'
_s
i %
c c
< CJ
3
0
r-
a:
I -
_, G
C
W5
"M
"a —
G ™
a o.
< g

V.

- 1
re w
s —
•5 ^
es
S
o
Ik;
5 •<
z:
v:
VI
S
es
"2
c


••= £
£ =
e «£
^

1
e
_o
u>






3^
0
Bt-
u



<»





c
!™T
«•*
T3

0







S



s






0



















o
>-










*5"
o
'—
1
To

P
H
o
to
o
aseptic sid

c
Z



-o

c^
fee







o



le






0
fee


















o
>~










^
c
1—
.1
w


1

aseptic sid


2:



r*
r*^

^o
o
fee





o
fee


S






o

^
*?
fee




























trt
^
O
C.
£T


1
T-I
aseptic ent


2



S


CM'
fee





o
fee


o
o
&e




0






































.E

eragc can <


CQ



O


S ±- J5 5 55 J5 J5 15 §
^£;^^^^ Swfee


•


oooooopooo
feefeefeefeetefee &efeefee


OC^Of^OO OOrn
o oo o v~t ^ ^ o o r*-
feefeefeefeefee^e &efee&e




OOC;O'!^CiC;OOO


•





-























t«
"H
1
; C


.E .E .E .S .E .E "H .E .E .E

UOOOCJO-^OOD
ooootioooeoooppofiboea
SS£££2"-2S£


CQODCQCQCQCQZCQCQCQ



O (N OC rj- C= OC —
V~, OO rn — — -^- -5}* ^ O OC
'
,


fee '.fee fee fee fee fee. "'..-'





OOOOOOOOQO
t~-t^-r^t^t~^t^-r*-r^ir^r^i
^^^^^^^£^£^£4


O VC O O O OO
c; o o o o —
fee fee fee le fee le




oooooco— — o

' '-' ^ °°r ™
S' s§ R
fee fee




cvj ___ —














s •&






^
"o
f 	 s ^ f 	 ,
"O U- "^3
: 1. 1 1
 oc
— Or— sO"OO

aseptic sid


Z



o»


v~»
&<=





in
fee


§






0























CO
tl










"H
Q
c.

p
"«
^tic end se;


<



CS

\^
^c
c*-





>o
fee


^






0

„
r—



























^
T3
^
U-
CJ
.£•
«


1

aseptic sid


Z



S


^
&





ir
&e


r —
00
—
fe«













-































1


K



g


-------




















73
01
d
•H
4J
C
O
CJ


0)
-P
c!
Ql .
w
w
0)
•H
M
O
S1
+J
id
y
,0
CO

PI

4J
to
3
•a
a
H
o

tn
4J
M
O
O
•8
L . Summary
i
n
o
rt

H
5 m
.S r-'

•5 if. *^
5 0
•< u
o o r^
^ m
d —
WM *2 rrr
*£ t/s *"
cs ^
3
a
e
S S S
fee %e *^
•— i •«•>
3 re
§ *£.
j« O O O
_ ®


s ,*«
•< J5
.5
*£


W
o
fe C
•O W
S =-
3 •<
*!-


"S .1
c: *3s
y s





o o
~ .E
r^ J£
*/*

>.
—
'i «
*» u c»
~ >• >-
e

•I.





-a
o
•— o
rt ^""*
sl-i
111
Subcalegory
Aseptic side seam s
Nonascptic side sea
Aseptic end seal coi
S

(£
•3
.E — —
5 — — 5
r-

OJ



^
u-}
r;
5


g
V5


0
•sO
•n
O)

««





































^»,

fe.
.¥
Aseptic side scam s





$
VC OC
NO fNr
"X exf

&


r- o
tn m
r^* vo
w fee


o o
fee fee


in so
O C1^
oc —
01 V.






































1 „

.li
i S
Nonascptic side sea
Nonaseptic end seal





£ a
R


S


0
m
vo
(N
t/J


^.
•=-.
OO
VO
V*

g











01

































Sheetcoatings





OJ
C3


5


o
fi
0^
fee


o
^
CO

o
fee












































a
Beverage can coatir





g
o
VJ


M


O
m
^
w


P



o






























trt

>-








M

3
O
CL
5
o
Nonaseptic end seal





g
0
o
o'

s


g
r^
«o



0
n
oc

0












































a
Beverage can coatir





OO
r-
m


01
fee


5
"n
r-
fee


0
sO
fee

o
fee












































a
Beverage can coatir





5
m


s


^
H~l
r^
w?


o
~°.
z

s













































Food can coatings





r-
vO
r^ r
so T
•n c
r-' c
— i



r- c
so c
in r
r- f
— u
fee fe
.

o r
r
c
fe

0 <



















*










v:
O









•8
C
|
5
0
Nonaseptic end seal




(•
r- v
so -
^ §
> K
•^ oc



i g

S «
e w


s o-
i •«,
D <^1
3" rO

i g












































a a
Beverage 'can coatir





o in
s


5


g





^

s

g












































a
Beverage can coatir





2
£


S


o
o

(N
oe


^
o
S
CO
fee

o
fee












































•a
Beverage can coatir





P
r--


w


^
V~l
^
&e


o



o
OO







































o
<8
6

.&
C3
General line side se





S
oc
of

fee


^
in
£Z
fee


o



2
&*r
TT

fee





































-a
c
o
o.
o
o
Nonaseptic end seal





01
in
«n
00

01
fee


^
m
«
fee


OO
^
1

o
fee













































Sheetcoatings





01
0
oc
oc

fee


in
r-

«
fee


o



o
l/~l

fee





































.
T3
O
tS
CI
O
C
CL
'fa
Aerosol side seam s





SO
r-^


&e


in
—
~
fee


o



g











































•o
1
Aseptic end seal coi





so
r-


fee


in
—
2
fee


o



g
r"i
O)
fee






































•a

o
o.
o
o
Nonaseptic end seal





SO
s

SO
fee


in
~-
™
•fee


so
in
in

g













































Sheetcoatings





S
0
in

O)
fee


CD
»n

so
ol
fee


o



g







































o

c
0

.S-
[/}
ra
General line side se





00
G

O)
fee


o
m

ol
fee


o



o
fee






























«

^>-












Sheetcoatings





OO
o
o

in
fee


o
o
r-
0)
in
fee


o



o
fee






























i«
4)













Sheetcoatings





3
3-4

-------















*«*•
•d
CD
•H
4J
C!
3


CO
4J
C
CD
&
CD
W
^*V^
CO
0)
•H
M
O
Ol
CD
4J
nj
0
W
h
M

CO
•§
C!
H
O
-P

CO
4J
CO
O
U
•>•>•
ca



o
« S W Wl V,
^ n o o o

>•» i**"
W3

i=
1 3
o
c
U.







Subcategory
2nd seal compounds
side scam stripe (Food
c. H- H.
w u a
O e/3 vi
a
u
f2
c
fiS ^ — —


g.O^^^JQjqc
, f--. [-^, [ — . — ~ ^ -^- (
S S 5 P 'S S S <




gcoo§!C|GlCi
s s £ s? « « « <
oo^^oooo.
o o '
f '
*O ^O '
: ZZ • ,

oooor-o-o*oc
fe^feefeefee-^-feeior^fe
•^r (Nr (N
O>' CN (S
rS S S




— tN
'

-


.


> '•
.




0
^-




3
'




'
1

•

T3
O
•i £
aa|l f
g 8 S -S ~§ g
il lit Iff]
M fee (N fee




m 04 — —
- .



1














OOU OOO OO O O








•

J -s J -s I J -s
e = " = = fc, c c
O 3Q3 ,^OV33
o g. la ,| Iji
'£ Si u 5 ° Sj £ §> S P" °
1 J III If j Iff lilt
:
-------
•
















I"""*
•8
0»
0
C
•H
^j
pj
O
CJ
0)
-p
c
a
a
CO

0)
CD
p
o
g1
•p
n)
u
•3
M
fr
(4
3
-a
C
H
O
M
4J
in
o
O

 •* «3 . °0 ' °
"S— wS£3t3Kw «« ' ««
= s
J-jL
t
^!Cgggggggg§gg!§; gg
•s. « i
•M
«£

w L
= «
b e
.S Ci M — • , » ^ ^
I/I • i


= £ S S3 8 • S
E £ > > > > >

, '


\

1
i
^» "o o "o f „
o |g £ eg. • -3
at t^ t±* c i , £

I .s. 1 .| 11 i
*^ ^« « w i • f •§•''!
K i g, g> g. .. g | g , .. «
S! -ii ^13" g El g
•8 8S 88g -S a • •?
! 1 1 1 I 1 1 -5 fr s 1 1 fr 2 .1
11 II! Ill !f!i;!f !
es7joo'c3t>uci-*o2iuOipa2'U
SHSS§SSS-='iss:'gi2
Zccn2£2:/!cae2< — Zw<^— Zw

•3 S ;
5 < :
^•^•oo^d-eNt^-r^-csooocDr o- CD o

o

o*
5
fee

o
»n
fN
fee


CD
fee




s
(N
fee

























S
^











Inside spray



CD

O

CN
oo'
>n
fee

o
fee


CD
fee




1
— '
fee


































f 	 ^
o
£

G.
Nonaseptic side seam stn



o



'n
£f
fee

^
fee


o
&e




g



























3
•*"









W
n
Aseptic end seal compou



r^

r- —

«n oc
cr-' oc'
-^ oc
fee in

^ ^
fee fee


fee o
t~~
n
fee rn
fee

|C §
ro C*f
fee le


CD —
fee o
fee



«n o
r-i fee
•o'
(N
fee








~~

























^_^
o
c
ti.

a.
Nonaseptic side seam str
Sheetcoatings



ro m



"^*r
CD


g
fee


•n
f-;
(N
(N
fee



o
fee










*"





























Sheetcoatings



0

r-

•n
r-
fee

•0
5


£




o



























0
^


•




^_^
1
c

Aerosol side seam stripe



CD

                                                 ^  -p
3-6

-------





















^"^
Q)
C
•H
4J
a
o
o

M
.p
(3
0)

W
•v^
w •
0)
•H
M
Q
tn
0)
rt
0
•§
w
fr
4J
W
•d
a
H
O
-p
D)
4J
CO
O
U
0
a

03
1
m
0)
I— (
•8
EH


CJ
.£


| S
c
r-
Ctf
if
c
c

'O *"
Id
is
S 'S.
S
0
"s U


— ™
-^ 3
«
§


c
fe a
Hk1
S <
«Z



05
GC 3
ca



(j

11

M

£>
™
|
|








>-,
Sllbcatcgor




1 *
~ fc.




Ef





5
r-'

s



§































o









._
seal compound
"2
o
.H
"H.
0


O


— 5e o

oo' of
O VI





r- o o
K CM'
— i/~>

•*• o o
"*• &e &e
i~^r
o

000
fee fee &e











(N —













CJ





ci










a
M) S) 5
.= .= "
"ra « R
o o 2P
o o £
"S S 
1
§
00



OC

f-
t-;
(N

fee



0
o^
f —
fee

s
o
o'

o
&e










































CO
.S
es
O
CJ
C
CC
CJ
CO
CO
£
CJ
>
CJ
ca



CT-.

£ £

O \C
rO OC

fee fe<



§ s
>n c-
fee fe^

o -^t
fee r-i
«=


•* o

<>
OC






































w-
o
Ci-
1 a
"B .£
CJ ^
•C! 8
5 =
--> S
1- §>
U3 d
U


CT-
00 r--
_ i_



C^ t — >O >o CN O (N ^" ^ \C^ ro  rn *n ^* 
-------

























?
g
c
o
o
**^

to
4J
C
0)

CQ
W
0)
•rl
O .
tn
0)
43
n)
i
co
s,
»
O
-a
C!
H

O
4J
IS
w
cS
o
*Vl
«
w
•
1
en
Q)
H
1

"3
=

s
2
5
^

0

j
_•
I
<

§

«
3
n
e
<

|
c
<

o
L.
tJ
S
3

J








^

w
*J
j?






















c*t^Tr--r-|n'Ofee
cs r-r o r-; oo CN oo
oe CN o CN — •' o -3-
r*^ — -3- — C* *n t--

1


00000000

* r\f r\f r«f r\T r-f rsf

*- fc4 fee fee fee fee fee fee
J3



£2 v. o. «i o — »n

— Le fee fee fee fee



JS5
»,oooooovgo
-««WW««V5W|£S«
*n
I§

^5
*"* — — —
cu
<



.j.
s
"1 S3
•s

en



O wi
S >
*=

.•?
1 #
O
~
M





•a
I" !•§
1 „ „ « « « 1 1
^ .f .= .1° .=* .S .=" -3 1
^* O O C3 O O O W "IS
C3 C3 BJ K S3 C5 -T3
oo oooo. He
e E s t g g i -|
iiiiilll
o
D
g
•s
.S CO ro r"" 1C
COc-» — — ootn — CN —

r- r-; i
CN' CN' r
VI l/"i *J
fee fee
«



O CD O C3 O <

CN CN i
•n  >• >-



tn
O








tn in
•O "O
o -oo
r ,ii
•« ^> q "ra
^ -S « S
-a g g -a
o .E i "2 o

Z 55 •<  t~~- CD C3 r*- r^- cvi >f~> ro *n c?-
•j — , — r^- ™ »/", u~i — — OC fN ro CN C^



-
^ CD CD C3 O O O O O CD CD CD O O

N' CN' CN? CN" CN' CN' cN1 CN' CN' so' so' so' \o' CN
n >n >n n "n m ,«n «n "OOOO O^ U^ U
cccncScr^ccScSc
Seocstoceearacceacsra
:o o oo. o o u o o.o u.o o
cscsoccettjj^csrarero^caj^TO
qjodjou^ajtijtijci^cj^flj
^,>,>.>.>.c^>-^>-c>-c;>
ucjajcJt)Oft3OSJi>OiuOai
acaca03CQZCQ03CQC3ZCQZ02



iSSpsSBSsl^.ISS
3-8

-------


,






















^^
Ti
0)
a
•H
-P
fi
O
u
w
-P
a
0)
&
0)
w
^*"»fc
n
01
•H
H
0
tn
(U

rt
O
•§
w
-P
W
jj
•d
c!
H
O
4J
•P
W
O
o
'H
O
«
|
w

•
1
n
0)
•8
EH
I

u ^^~f^i — — — o^C'^r^r-j-stj -— >n — occ
2 -, r^Mfeefee&efee:feefeemfee— m — -3- &4 — rn (Nfee<-n&e t--" r-~' r-' r-' r-* o' o ^d *o r--' r-* t--r n j ' oc — ^ ' r-r oo^ —

N r i "^" *Nt* C* *O. O ': -^ rn "5j* - OC C^

ec — fee fee feefeeifee feeoefee &=
it
,*• es
*ifeefeetNo4feefeetnfeefeefeefeeccfeecc!mfee — r^oofeefeefeefee fee c^ — &«
g \G co o* c^ col r*- t — c*' i "^ i
x ! !
>> :
1 -3 S 8 ' 2 8
, >- >- ' ^~ ?™ >"
3

H



^ ' ^ ^5" ! ^ ^5" ^ ^o"
o - c . o _ :c O'O o
£ £ £ | || £ | £ £ | f | £ '
| 1 III I I ° I | a I | £.11
w 5 — si is i: ^ Pi: ^Jr -i= ^ o^- —
s a I- P 1 : . S P s.§S§Sp8-g-8S
I|| is i i al S| 1| | 1 111
u"^ «o"2 ra^.S ^.H S ^->.5 yf^.E tJ ^.5 w,S ^.S"o y U.S
Si, u c. H- y c. "H. w H. w ^ H. K "HJ "H. K "H. H. "^ " « "H. ^ w "c. H. re
SJo^o o« o c  ^ii w ^ w - 0 « «0 c«u50wa Oyi.O™ trttrt O
o c, -^ J« c. "o 2 o S o "O ^;cj^c^cjc"K'a3c'(De"OJ ^- ** ^ o
>• o •— C CJ -j— ~4)£sjC^O"~O~™w5oCJ~*'~S
ffl<£z<42«zS£z«zz5zzSz5z'5
-------



















•"•V
-d
'(NO  o- oo oc o o
C3 O V"T of O O)



-^•viG^OOOOvOOOOOi:

V% M P) 01 K
*^ *^ *** *^





— 01 01 —







„











yi en v,
O O 5J






'

•u ~o
§o
«« . — . O OT ^"^
g u- ^ -g g, "g -g
= ^t* = o ^ = o
g..S- a &..!- a t,
E — E o is E o
S 1 8 f 1 8 f
Sial § a I a | a"
"H -o ™ "2 o —3 "^ *s ° 'B "
o'Slo^-g'ueo^S-il
Itiflllillili
e; c3 -a "{3 c.05 o ra cj*a CL~O
§§§§JSo|||||g.





Mro in-riinmc^ ooooc
^ — o — o — i— ^ — o o o-
"* O O f"i Vi OOO-sfOOOOOOvOO

n o- — — i— CD
f t— Tt — — OC
2 ^ Vj o^ O9 V3-





_ _ - __ ^t-'



'










;


1

> >







f"^ . V~*- ' x~^.
•0 T3 -0
i I 1 1 I' 1
li i| i I- i
5 « . o'« g. « 5
11 al 1 | a i 1
2 *a -.n ' *9 *o S *^ *a "2
a S i* 8 o "o 1" -S 8 *o c o =
B. a. S ^B.c-S ° SclSo-os
atuOS^4JOOoccQ>OciO
K KuOwwO-^UwOwO
2So"o22u^-'T32'u2o
5-OJ= oOO^ w§ 0^0^
^^ooti.Z2oo
-------














i
1
1
i
1
1
)



)
)


1
)
1
'


1
n






1
















i















5

^
3 a!
C "S
C O
< U
->
mnlizcd Annual MRJ
tal Costs Costs
c c.
< V
«
_ c

_
c |«
< s
,5
^

"o
(«
— C5

?- **
s •<



cr
K
c
"3?
cs
«
S


o
'•C x.
a c
"p- *~
>> *^
CC
^'
"3
ul
^_
c
—
iZ







(N — — oc — r

OO^^^OV.^^^O'C
o -^j- »n o -
•* fN 0 0 C
•rj- <^i r- ^o r~

i
c oc c !*^* o o o vi c o o c
o ,r^ OP *n
oc *^ — ' r-T
'.J .'O ~~~ O^
— i/^ v>






(N — — (N —



1















(

v. ' v. v. in
r>~ ' ^~ ^ ^" -



1

s—
1
^ ^
v. 'tn § 2 en

3 '5 <5 u 5
o o = .£• c
5= ^ "o" S f=
c o .£• «- 3
w o ecu
_= T3 _c ;-3 o r2 "O
§ o I o .if .If :§ = o .f1 .1? .=
§— "— ' — '— ' ^ •— • r^- '"" i— »J •- -^

c.
"o|'o1|88£|§'HS|
cSjSu.IZc/sSo'^OZc/StT^c?



=

C1^ >O •
— OOOfvC^D OCCCOCOCO'w'

-ooooooo«r>
D >^-
* rn
^ IV

•oooooooo
i r<
;ggg§SSS^
r t"-
o
'


oooooooo


























C/5 t/j U5 trt t/3 Wl t/1
O CJ U O O CJ O





U (5
>- )>~





^. ' ^.
1 ' 1
sr . ^" n
9 w § 2 «

O 5 Q O §
'§" n. -S '& E.
To P cj 'K c
n c .9- = 5
^ ° -i= 5 °
.12 -a S 3 -S
' g S .I1 -S s S .I1 .=
"~" ™ ^ crt ^^ ** 'ji ^

StnuSSS^wu
||§| = |8S
OZ^^OZwc/S



.

m m m o-i —
_ — _000ooo


-------
•d
 0)


 g
•ri
4J
 C
 O
O
 CO
4J
 a
 01
 Q)
CO
^s.
 w
 a

•a

 &
 0)
4J
 ID
 U
tn



S
 Cfl


I
H

 0
-P

 (0
4J
 II)
 O
u

«w
 0
CO
en

 Q)


I
1
S3

S "c?
S O
O
H
12
"S (j
t= —
S
K
•a "y.
.JS (j

3 S
g'E.
.<• cs
^ o

"K .2 s/;
is!
S _K y
fc a
.0 u
E -
_3 <
= 0
,c "—
^ a

^ =
s «~
&
O «
^ £




1
1
3




1 <


g'E B f: ! 3 s s p • S s 1 5 i
t-^f*1) r^ir^-fso'in r-r^.c>t^soor^
^& ^o ^ K |/j oSSa fe^»ni^-inocS('*i
fee _j ^e^feefeefeefee

, ;
£S ^iCgiS Sg§8§§§
n* n' m en  o

fee . r ^ fee fee fee fee fee
fee .


OO OOCJOO OO;-OOOOO
u-ite befeefeefeero feetefeefee&efee-fee
^t- r--'
ro &e
:
f
™_~ ^




1
S S S

J -S •§ ! ,
5 o — o ~ '
c ^ o .£• c
— ' »- C. k- C.
0 7 • P w P
.£- ^ 5 5 •
^ C O Co
ill 1 1 : :
S S -P S -a ;
'^W — ?/} V, V) IS) — t/l V3 V} VJ (A M
Ot> oeacjjco.* OiBijDOeocuoocij
•^3 ^ t> — ICC f- (jtl-CdCCCZ
1 ll §"111 ll 8 1 1 1 I 8 1
oScc^^^c^c^^^^1114*
w c>o o S r^ 5^ , o o o £3 •• $2 ££ Ji S2 _£;
•< ^ %5* 21 So 55 &o O ^ "£. Eo oo M w ^ co

(N 
*""


cc











tn
A
"o
H



                                                                          3-12

-------
                                     I     SECTION 4
                ECONOMIC IMPACJT ANALYSIS: METHODS AND RESULTS
                                     I •                       -
                                   •I
       The underlying objective of the El A is to evaluate the effect of the proposed regulation on the
welfare of affected stakeholders and society in general. Although the engineering cost analysis presented
in Section 3 represents an estimate of the resources required to comply with the proposed rule under
baseline economic conditions, that analysis does not account for the fact that the regulations may cause
the economic conditions to change.  For! instance, producers may elect to reduce output in response to •
cost increases or even discontinue production rather than comply, thereby reducing market supply.
Moreover, the control costs may be passed along to other parties through various economic exchanges.
The purpose of this section is to develop and apply an analytical structure for measuring and tracking
these effects as they are distributed across the stakeholders tied together through economic linkages.
                                     i
4.1    Markets Affected  by the Proposed NESHAP
       The determination of markets potentially affected by the rule requires identifying the products
produced  at the affected facilities and linking them to markets where they are exchanged. Based on the
Information Collection Request (ICR) and data provided by the Can Manufacturers Institute (CMI),
EPA divided the metal can market into three separate markets:
                   •                  i
       •   beverage1 cans,             ; ~         	

       •   food cans, and             ;

       •   general packaging containers.

       The economic impacts of the rule on the identified industries and related product markets are
examined in the following  sections usin^ both a conceptual approach and operational model.  The
conceptual approach is described in Sectjion 4.2, while Section 4.3 presents the economic impact results
based on the operational model.        j

4.2    Conceptual Approach         i
       The Agency developed three national partial equilibrium models to estimate the economic
impacts on society resulting from the proposed regulation.  The large number of metal can producers and
the close substitutability of alternative miaterials such as glass and plastic for metal cans in many
packaging applications lends support for|the notion that metal can producers will behave as if they
operate in perfectly competitive markets! As a result, we assume  that the number of buyers and sellers is
large enough that no individual buyer orjseller has market power (i.e., influence on market prices).
Under this condition,, producers and consumers take the market price as a given when making their
production and consumption choices.   I
                                     i

4.2.1  Supply                        j
       After critical review, the Agency] determined that the level of detail of facility  survey and .
compliance cost data is sufficient to support a facility-level characterization of supply. EPA assumed
each plant has some fixed factors of production (e.g., plant and equipment) that are augmented with
variable factors inputs (e.g., materials, labor) to produce metal cans. These fixed factors are the source
                                              4-1

-------
of diminishing marginal returns, hence, [increasing marginal costs. Therefore each supply segment
(beverage cans, food cans, general packaging containers) can be characterized by an upward-sloping
supply curve.
       An important measure of the magnitude of this response is the price elasticity of supply,
computed as the percentage change in quantity supplied divided by the percentage change in price.
Absent empirical estimates of the supply elasticity, we use assumed values of the supply elasticity in
each of the relevant markets and perform a sensitivity analysis on those assumptions. The supply
elasticity used to generate the primary impact estimates, which are presented in Section 4.3, is 1.0 for all
three markets modeled.  The sensitivity analysis presented in Appendix B examines the effects of
varying the supply elasticity between 0.5 and 2.0.
                                    1  ,                                            :              .
4.2.2  Demand                     '  '
       Consumption choices are a function of the price of the commodity, income, prices of related
goods, tastes, and expectations about the future, among other variables. In this analysis, we will
consider how purchases of metal cans change in response to higher prices resulting from regulation,
holding other variables constant.  The demand for metal cans is a derived demand, meaning that the
quantity of cans demanded is directly dependent on consumer demand for the final products metal cans
are used to produce. In this case, consumer demand for products such as beverages, food, and paint
influences the number of containers (e.g., metal, glass, or plastic) that will be purchased for packaging
those products. Nonetheless, the price of factors of production, such as metal cans, is still an important
determinant of the derived demand for that factor because of substitution possibilities among factors of
production. The economic model assumes a downward  sloping demand curve (i;e., the quantity
demanded for a good falls when price rises), consistent with the Law of Demand. Thus, ah increase in
the price of metal cans, as is expected to occur following regulation, is expected to result in a decrease in
the number of metal cans demanded by final product industries. The buyers of metal cans are likely to
switch to containers made from alternative materials (e.g., plastic, glass) to some degree and/or reduce
their total output in response to this increase in metal can costs.
       EPA modeled the demand for metal cans in each of the three markets defined above based on
using reasonable assumptions for the price elasticity of demand in each market.  The primary
consideration that will influence the choice of demand elasticity in each market is the availability of
substitutes for metal cans in that market.  Other things being equal, the more close substitutes are
available for a given product, the more elastic the demand for that product. The more elastic demand
arises because, with many close substitutes available, consumers can easily switch to alternative
products in response to a price increasej  As a results, manufacturers may have little ability to pass costs
onto consumers in the form of price increases. In contrast, firms in industries with few close substitutes
are likely to be able to pass a higher proportion of regulatory costs to consumers of their products.
       Based on information contained! in the metal cans industry profile, it appears that both metal food
cans and metal beverage cans have fairly strong substitutes available (primarily plastic bottles for
beverages and glass bottles for foods), ^hile there are fewer substitutes for metal general packing
containers in the markets where they are  generally used (e.g., paint cans). In addition, the demand for
aluminum beverage cans is likely to be jmore elastic than the demand for steel food cans because the cost
share of cans  in the beverage market is lower than in the food and general packaging-markets and plastic
bottles seem to be more generally substitutable for aluminum beverage cans than glass bottles for steel
food cans. Consistent with this notion, Palmer, Sigman, and Walls (1996) report demand elasticities of
-1.4 for aluminum beverage cans and -0.63 for steel cans (including both food cans and general
packaging containers).  EPA used these; elasticities as the primary elasticity values for the economic
analysis. However, because of the  inherent uncertainty involved in selecting point estimates of demand
                                    i-
                                              4-2

-------
elasticities, a sensitivity analysis was performed that examines the effects on the economic impact
estimates of different assumptions concejrning the demand elasticities. We examined a range of demand
elasticities from -0.5 to —2 for each of tljie three affected markets as part of the sensitivity analysis, the
results of which are presented in Appendix B.
                                     i
                                     i
4.2.3  Foreign Trade                 |
       A review of the international trad|e data shows that foreign trade is a very small share of the
domestic metal can market. Based on repent data, imports account for about 0.24 percent of 1998 U.S.
metal can consumption and exports account for about 0.71 percent of 1998 U.S. metal can production.
In addition, there is no information available to inform the allocation of imports and exports between the
three markets defined above for the analysis. As a result, we provide a qualitative description of the
foreign trade impacts rather than developing quantitative estimates.  For example, foreign imports may
become more attractive to U.S. consumers and U.S. exports may become less attractive to foreign
consumers as a result of the change in relative prices resulting from regulation in the U.S. In addition,
domestic facilities could potentially relocate to foreign countries with less stringent environmental
regulations if domestic production costs increase.7 However, the cost impacts are unlikely to be large
enough to cause significant trade impacts.

4.2.4  Baseline and With-Regulation Market Equilibrium
       A graphical representation  of the competitive model of price formation, as shown in
Figure 4-1 (a), posits that market prices and quantities are determined by the intersection of the market
supply and demand curves.  Under the baseline scenario, a market price and quantity (p,Q) are
determined by the downward-sloping ma|rket demand curve (DM) and the upward-sloping market supply
curve (SM) that reflects the sum of the do|mestic supply curves. EPA's model includes both affected and
unaffected domestic supply.            |                            .
       With the regulation, the costs of production increase for affected domestic suppliers. "The
imposition of these regulatory control cofets is represented as an upward shift in the affected facility
supply curve. As a result of the upward shift in this supply curve, the market supply curve for metal
cans will also shift upward as shown in Figure 4-l(b) to reflect the increased costs of production.
'However,  empirical  .studies  in  |the  literature  have  generally  found  little
   evidence  of  environmental   regulations  having  a  significant  influence  on
   industry  location decisions  (e.fj.,  Levinson,  1996).
                                             4-3

-------
         Affected Facilities
    P'


    P
            S'
   " a   ^a


Affected Facilities
                          +  p
                   =  p
                                                                        SM/
Unaffected Facilities
                            a)'Baseline Equilibrium
                      p'
                      P
                                 Unaffected Facilities
                        b) With-Regulation Equilibrium
                                                                         DM
                                                                      Q
                                                       Market
                                                                   Q'  Q
                               Market
Figure 4-1.   Market  Equilibrium without and with Regulation
                                      4-4

-------
       In baseline without the proposed Standards, the industry produces total output, Q, at price, p, with
domestic producers supplying the amount qa and imports accounting for Q minus qd, or qu. With the
regulation, the market price increases fro!m p to p', and market output (as determined from the market
demand curve, DM) declines from Q to QJ'. This reduction in market output is the net. result of reductions
in affected domestic supply and increases in unaffected supply.
                                     i
                                     i
4.2.5  Impacts for Facilities Excluded from the Market Model
       After review of the available data|, the Agency determined that 13 facilities manufactured unique
metal can commodities that did not fall within the market definitions above (e.g., commemorative tins).
However, the Agency concluded data linjiitations did not support the development of similar partial
equilibrium models for these commodities.  As a result, the Agency employed a simple nonbehavioral
financial analysis to estimate impacts, which takes the form of the ratio  of compliance costs to the value
of sales (cost-to-sales ratio or CSR). To icompute these ratios, EPA collected revenue data and
calculated a CSR for each of the firms asj follows:
                  CSR = Total Annual^zed Compliance Costs/Total Plant Revenue              (4.1)
       One drawback of this approach is! that it does not consider interactions between producers and
consumers in a market context. The analysis simply assesses the burden of the rule by assuming the
affected firms fully absorb the control costs, rather than at least partially passing them on to consumers
in the form of higher prices. Therefore, it likely overstates the impacts on facilities affected by the rule
and understates the impacts on consumeijs. However, the approach can  provide a quantitative measure
of the economic impacts for these facilities and has the advantages of simplicity and relatively limited
data requirements.

4.3
                                     I
       Economic Impact Results      I
       To develop quantitative estimates of these impacts, we developed a computer model using the
conceptual approach described above.8 Using this model, EPA characterized supply and demand of
three affected commodities for the baselihe year, 1997; introduced a policy "shock" into the model by
using control cost-induced shifts in the domestic supply functions of these markets; and used the market
model to determine a new with-regulation equilibrium in each metal cans market. We report the market,
industry, and societal impacts projected by the model below.           .
                                     I
4.3.1   Market-Leveilmpacts          |
       The increased, cost of production flue to the regulation is expected to increase the price of metal
cans and reduce production/consumption' from baseline levels. As shown in Table 4-1, the price
increases in all three metal can markets are similar in magnitude and are each less than 0.5 percent.
Domestic production of metal  cans is estimated to decline'by a total of 392 million cans, or 0.30 percent.
The beverage can market accounts for 8CJ percent of this decline, which is approximately proportionate
to its share of metal cans pro'duced.
8Appendix  A  includes  a  description   of  the 'model's  baseline  data  set  and
   specification. ;                 I
                                             4-5

-------
Table 4-1.   Market-Level Impacts  of  the  Metal  Can MACT:   1997

Tab 1 e 4 -fySv^etNa tional-L
JJgvyijnge
Price (S/can)


Revenues ($106/yr)
Costs ($10<7yr)
Compliance
Production
Pre-tax earnings ($10s/yr)
Plants {#)
Employees {#)
Total1
Price (S/can)
Quantity (10^
t

$0.061
r
Baseline
$10, 848 .12
$10, 030.25
i$0.00
$10, 030.25
$817.87
156
20, 846

$0.084
129,387
Absolute Relative
ti#y*e3S.lftfe>ets <
$0.061
With
Regulation
$10,849.63
$10,047.40
$48.50
$9> 998 .90
$802.22
156
20,670

$0.084
128.995
S^ngftie MetaShasgfia
$0.000 0.23%
Absolute v
Change
-$1.51
$17.16
$48.50
-$31.34
-$15.65 :
0
- 176
!
$0.000 0.31%
-392 -0.30%
MACT:

Relative
Change
. 0.01%
0.17%
NA
-0.31% :
-l.'9i%" \ I'
0.00%
-0.84%



* The prices reported for the total impacts on the metal can manufacturing industry are weighted averages of the prices in the
  three submarkets above.                ',
4.3.2   Industry-Level Impacts   -    \
       Revenue, costs, and profitability of the directly affected industry also change as prices and
production levels adjust to increased coists associated with compliance. For metal can producers, pre-tax
earnings are projected to decrease by a total of about $16 million across all three submarkets included in
the economic model (see Table 4-2).9 These losses are the net result of three effects: '
       •   Increases in reveriue ($1.51 million, or 0.01 percent)—based on the elasticities used in the
          model, revenue increases slightly because the average price of metal cans increases by a
          larger percentage than the quantity falls.

       •   Reductions in production cojsts as output declines ($31.3 million, or 0.31
          percent)—production  costs fall as firms reduce their output.10

       •   Increased control costs ($48,5 million)—we have assumed total annualized compliance costs
          vary with the ]evel of  output.  Therefore, the compliance costs being incurred with regulation
'Note  that there are  only  156 facilities  included  in the  market  model ! after
   excluding  the facilities  that; did  not fit. into the  three metal  can  markets
   modeled  and  allocating  costsi assigned  to  facilities  that  only  manufacture
   sheets  or  ends to their  sister facilities  that manufacture  the  cans.   This
   adjustment was made because  tjtie facilities producing only sheets or  ends do
   not compete  directly,  in the  [can market,  although  changes  in the  costs  of
   producing  these  inputs will  affect company-level can  output.

"Note that this  does not imply that  production costs per  unit are falling,  only
   that total production costs  will tend to fall as  less  output  is produced.   For'
   example, fewer raw materials are needed as  output declines.

                                            4-6

-------
           are smaller than the engineering compliance costs presented in Section 3 because the
           estimated reductions in output imply lower compliance costs.11
        The national-level results also highlight important distributional impacts of the rule across
 facilities, as shown in Table 4-3. Approximately one-third of the modeled facilities experience an
 increase in pre-tax earnings totaling aboUt $10.3 million as a result of increases in price that exceed their
• compliance costs per unit. In contrast, the remaining two-thirds of metal can facilities experience losses
 in pre-tax earnings totaling $26.0 million. As expected, facilities who are better off with regulation have
 relatively lower per-unit compliance costs than their competitors.
        The Agency also examined impacts on the 13 facilities not included in the market model.  By
 assumption, these producers experience reductions in profit equal to the total annualized compliance   '

 Table  4-3.    Distributional  Impacts  Across Facilities of the  Metal  Can
 MACTs   1997
                                    r

Pre-Tax Earnings
Loss Gain
Total
 Plants  (#)


 Baseline  Production
 Total  (units/yr)

 Average  (units/facility)


 Baseline  Compliance Costs
 Total  ($10s/yr)    ;       . '
 Average  ($/unit)
             '99
j  86,117,362,89

!               6
|    887,807,865
    $45,450,401'

       • $0.0005
                                57
35,843,620,63
             2

  607,518,994
   $4,167,867
       $0 .0001
                                                    156
121,960,983,52
   781,801,176
   $49,618,268
       $0.0004
Change in Pre-tax Earnings
($ios/yr)
Change in Employment (# employees)
. ,-$25.90
• -309
$10.25
133
-$15.65
-176
 costs estimated to fall on those facilities (j$4.5 million), an average of $350,000 per facility (see Table 4-
 4).  Revenues for these companies were estimated based on data collected from Dun & Bradstreet,
 Reference USA, Thomas Regional, and the Census Bureau.  References USA provides facility-level
 sales ranges, but this data was not available for all 13 facilities.  Therefore, we used Census estimates of
 Compliance  costs are expected to!be  lower,  on average,  as output  falls because
   many types of compliance  costs dre typically assumed to  vary with output.   For
   example,  as output  falls, some. Ifirms may  be able to meet pollution abatement
   requirements with smaller-,  less i expensive control equipment.
                                             4-7

-------
Table  4-4.
1997
Impacts  for  Facilities Not  Included  in  the  Market  Model;
Total Number of  Facilities       i
Total Annualized Compliance  Costs  (TACC)
Average (TACC)  per Facility ($10S)
                                                     $4.5
                                                    $0 .35
                                                          Number
                                                                Share
Facilities with Sales Data       ;
   Compliance costs are < 1%  of  sales
   Compliance costs are >_ 1% and; <  3%  of
sales                             I
   Compliance costs are >_ 3% of bales
                                                                 69%
                                                                 23%

                                                                  8%
 Compliance Cost-to-Sales Ratios
   Average
   Median
   Minimum
   Maximum
                                                1.34%

                                                0.43%

                                                0. 00%

                                               10.20%
the average revenue per metal can manufacturing establishment for the employment size category that
the facility falls into as an estimate of facility-level revenue for those facilities where Reference USA
data were not available. Because Reference USA provides fairly wide ranges in its sales estimates. EPA
chose to use a conservative estimate of facility revenue by using the minimum of:
       •   Total company sales (from Dun & Bradstreet- or Thomas Regional),
                                                              ,
       •   Midpoint of facility-level sales range reported by Reference USA, and

       •   Census estimates of the average revenues per establishment for the metal can industry for the
          state in which the facility is located.

This was done to ensure that we were not using facility-level sales that were greater than total company
sales and that the Reference USA estimate was not far out of line with the standard industry output for
an establishment with a given employment range. Relative to estimated baseline sales for these
facilities, nine facilities are impacted less .than one percent, three are impacted between 1 and 3 percent
of sales, and one facility is impacted at a level above 3 percent of sales.

4.3.3  Closure Estimates
       As shown, the economic model does not predict any facilities included in the market model will
close following regulation under the reference case elasticity assumptions. However, sensitivity analysis
shows that one facility may close under ;different supply and demand elasticity assumptions. In addition,
the cost-to-sales analysis for the 13 facilities not included in the economic model shows that one facility
has a CSR exceeding 10 percent. The U.S. Bureau of Census reports industry group financial ratios in
their Quarterly Financial Report for Manufacturing, Mining and Trade Corporations (U.S. Bureau of
the Census, 1998).  For 1997, the Census Bureau reports that income before income taxes (pre-tax
earnings) for SIC group 34 (Fabricated Metal Products) was approximately 7.6 percent of sales.  For
                                             4-8

-------
smaller firms (i.e., firms with assets under $25 million) this ratio is 6.9 percent12. Therefore, the Agency
believes the rule may^potentially result iik one to two premature plant closures.
                                    i
4.3.4  Employment Impacts          \
      Reduction in domestic production leads to changes in industry employment-. Facility-level
changes in employment were estimated tj>y multiplying the change in production by baseline
employment:                         j
                         .          I   AE^tAQ/QlEo            .                       (4.2)
Employment is projected to decline by 3\)9 employees at plants with profit losses and increase by 133
employees at facilities with profit gains. EPA estimates the net employment change resulting from the
rule is a reduction of 176 employees, or -j-0.8 percent.
                                    I    .
4.3.5  Social Costs                  |
      The value of a regulatory action ijs traditionally measured by the change in economic welfare that
it generates. The regulation's welfare impacts, or the social costs required to achieve environmental
Table 4-5.
1997
      Distribution  ofj Social  Costs  for the Metal Can MACT:
                          i

Change in Consumer Surplus
Beverage ;
Value ($10
. .. -, i .-.-...•.•-.• -, , • -$33.
•-, ..; -.-.•::• v. . • -$i3.
Food •••..,'•.•-_ -$10.
Packaging
Change in Producer. Surplus
• • • • . • -$8.
' -$20.
Market model ..! .'. ' '...-•" . ..-$15.
Not modeled
• •- '- •- '•' - -$4.
i
Total Social Cost •

, :•.. ••:. ,.: . • • •: -$53.
Vyr)
3
9°
8
5
2
6
5

5

improvements, will extend to consumersjand producers alike.  Consumers experience welfare impacts
due to changes in market prices and consumption levels associated with the rule. Producers experience
welfare impacts resulting from changes in profits corresponding with the changes in production levels
and market prices. However, it is important to emphasize that this measure does not include benefits
that occur outside the market, that is, thej value of reduced levels of air pollution with the regulation.
                                    i
      The economic analysis accounts  for behavioral responses by producers and consumers to the
regulation (i.e., shifting costs to other ecpnomic agents). This approach provides insights on how the
12In  the short run,  a plant would jbe  presumed to  continue to  operate as  long as
   variable  profits  are positive.]   The  Agency  considered  QFR's  income  before
   income
   profit
taxes
rate.
                 measure as  a
                                reasonable
approximation  of  plant-level  variable
                                            4-9

-------
regulatory burden is distributed across stakeholders.  As shown in Table 4-5, the economic model
estimates the total social cost of the rule! at $53.5 million. As a result of higher prices and lower
consumption levels, consumers (domestic and foreign) are projected to lose $33.3 million, or 60 percent
of the total social costs of the rule. Beverage market consumers experience over one-third of these
losses, or $13.9 million. Producer surplus declines by $20.2 million, or 40 percent of the total social
costs.               •           '     i  .
4.3.6  Sensitivity Analysis .          \  .
       As a result of uncertainty involved in selecting point estimates of supply and demand elasticities.
EPA also conducted sensitivity analysis to explore the effect of different elasticity values.  Detailed
results of this sensitivity analysis are presented in Appendix B. The social costs of the rule remain
essentially unchanged in the sensitivity analysis. As expected, changes in elasticities that make the
consumer more responsive to marginal changes in price relative to producers results in lower consumer
surplus losses and higher producer surplus losses. Conversely, changes in elasticities that make the
producer more responsive to marginal changes in price relative to consumers results in higher consumer
surplus losses and lower producer surplus losses.  Finally, closure estimates ranged from 0 to  1 facility
under all scenarios for those facilities included in the market model.
                                    i                                     ,
                                    i
4.4    New Source Analysis         |
       Potential new suppliers of metal cans have an investment decision concerning whether or not to
enter the market (or to build new facilities in the case of current market participants). Economic theory
tells us that investors are only expected to invest in projects that are expected to have a positive net
present value (NPV), that is, an internal rate or return higher than the opportunity  cost of capital.
Therefore, to the extent that the metal can manufacturing  NESHAP will result in a decrease in the
expected NPV of investing in new plants.' it could potentially reduce the number of new entrants.
However, EPA has estimated that there jwould most likely be no new entrants in the metal can
manufacturing industry over the next fe^v years even in the absence of this NESHAP.  Thus, EPA
concludes that there will be no impacts ^>n new sources as a result of this regulation.
                                    I                               '          •     .
4.5    Energy Impact Analysis      |
       Executive Order 13211, "Actions Concerning Regulations that Significantly Affect Energy
Supply, Distribution, or Use" (66 Fed. Reg. 28355, May 22, 2001), requires federal agencies to estimate
the energy impact of significant regulatory actions.  The proposed NESHAP will trigger both a small
increase in energy use due to the operation of new abatement equipment as well as a decrease  in energy
use due to a small decline in the production of metal cans. These impacts are discussed below.

       Based on information from the industry survey resporises, it is not expected that the substitution
of low HAP coatings and thinners for the materials currently used would result in any change in energy
usage. However, because many metal qan manufacturing facilities use add-on emission control devices
to meet existing limits, it is expected that these facilities would use additional add-on controls to comply
with the MACT standard. Facilities are| expected to add RTOs to reduce HAP emissions, which require
electricity and the combustion of natural gas to operate and maintain operating temperatures.  EPA
estimates that electricity consumption will increase by 36,730,000 kilowatt-hours  (kWh) per year and
fuel energy consumption resulting from! burning natural gas will increase by 1,197,000 million British
thermal units (MMBtu) per year, which roughly corresponds to 1.2 billion cubic.feet of natural gas. The
total electricity generation capacity in the U.S. in 1999 was 785,990 MW (DOE, 1999a). Thus, the
electricity requirements associated withjthe new abatement capital likely to be added to comply with, this

                                    i          4-10               '

-------
NESHAP represents a very small fraction of domestic generation capacity. Similarly, the natural gas
requirements associated with the NESHAP are very small relative to the 23,755 billion cubic feet of
natural gas produced in the U.S. in 1999 |x>OE, 19991?).'.
       In addition, as described in Section 4.3, the economic model predicts that increased compliance
costs will result in a reduction in annual output of 0.3 percent for the metal can manufacturing industry.
This small decline in production is expecjted to result in an approximately proportionate reduction in
energy consumption for this sector and w,ill partially offset the increased consumption to  operate add-on
control devices.            .           '
       Overall, both the increases and decreases in energy consumption expected to result from
implementation of the metal can manufacturing NESHAP are projected to be extremely small relative to
national energy markets (and will at leaslj partially offset each other).  Thus, it is extremely unlikely that
the proposed NESHAP will have any significant adverse impact on energy prices, distribution,
availability, or use.
                                             4-11

-------
                                         SECTIONS
                               SMALL BUSINESS ANALYSIS '

       This regulatory action will potentially affect the economic welfare of owners of metal can
manufacturers.  These individuals may be owners/operators who directly conduct the business of the
firm or, more commonly, investors or stockholders who employ others to conduct the business of the
firm on their behalf through privately held or publicly traded corporations.  The legal and financial
responsibility for compliance with a regulatory action ultimately rests with plant managers, but the
owners must bear the financial consequences of the decisions. Although environmental regulations can
affect all businesses, small businesses niay have special problems complying with such regulations.
       The Regulatory Flexibility Act (RFA) of 1980 requires that special consideration be given to
small entities affected by federal regulations. The RFA was amended in 1996 by the Small Business
Regulatory Enforcement Fairness Act (SBREFA) to strengthen its analytical and procedural
requirements. Under SBREFA, the Agency must perform a regulatory flexibility analysis for rules that
will have a significant impact on a substantial number of small entities.
       This section focuses on the compliance burden of the small businesses within the metal can
manufacturing industry and provides a screening analysis to determine whether this proposed rule is
likely to impose a significant impact on: a substantial number of the small entities (SISNOSE) within this
industry. The screening analysis employed here is a "sales test" that computes the annualized
compliance costs as a share  of sales for jsach company. In addition, it provides information about the
impacts on small businesses using a market analysis that accounts for behavioral responses to the
proposed rule and the resulting changes in market prices and output.

5.1    Identifying Small Businesses  |
       The Small Business  Administration (SBA) released guidelines effective October 2000 that
provide small business thresholds based on NAICS codes that replace the previous thresholds based on
SIC codes. Under these new guidelines., SBA establishes 1000 or fewer employees as the small business
threshold for Metal Can Manufacturing !(i.e.,,NAICS 332431). Using this .guideline and available
secondary data, theAgency  identified 13 small businesses, or 43.3 percent of the metal can companies.
For these small businesses, the average (median) annual sales for companies reporting data were $27
($24) million, and the average (median} employment was-178 (175) employees.
                       i
5.2    Screening-Level Analysis
       To assess the potential impact of this rule on small businesses, the Agency calculated the share of
annualized compliance costs relative  to baseline sales for each company. This type of analysis does not
consider interaction between producers jand consumers in a market context.  Therefore, it likely
overstates the impacts producer impacts' and understates the impacts on consumers. When a company
owns more than one affected facility, EPA combined the costs for each facility owned by that company
to generate the numerator of the cost-to;sales ratio.  Annualized compliance costs include total
annualized capital costs and operating and maintenance costs imposed on these companies.

5.2.1  Results
       Small businesses are expected to incur only 2 percent of the total industry compliance costs of
                                             5-1

-------
$56.2 million (see Table 5-1).13  The aveijage total annualized compliance cost is projected to be $90,000
per small company. The mean (median), &ost-to-sales ratio for the 13 small businesses is 1.10 (<0.001) "
percent, with a range of 0 to 10.20 percent. EPA estimates that 10 of the 13 small businesses experience
an impact less than 1 percent of total company sales, two small firms have CSRs between one and
3 percent, and one firm has a CSR greater than 3 percent of sales.
       Large businesses are expected to incur 98 percent of the total industry compliance costs of $56.2
million. The average total annualized compliance cost is projected to be $3.2 million per large company,
The mean (median) cost-to-sales ratio for the 17 large businesses is 0.27 (0.14) percent, with a range of 0
to 1.29 percent.  EPA estimates that 16 of the 17 large businesses experience an impact less than 1
percent of total company sales and one latge firm has a CSR between 1 and 3 percent.
                                    I
"This   disproportionately  small  limpact  is  primarily  due  to  the  fact  that
   relatively  few .small  businesses  in the  metal  can manufacturing  industry are
   major sources.  ;                 '

                                    1    '    5-2

-------
Table  5-1.   Summary Statistics  for  SBREFA Screening Analysis:   1997
                                           Small
                       Large
                          Total
Total Number of  Companies             13          17          30

Total Annualized Compliance Costs    $1.1       $55.1       $56.2
 (TACC)  ($10S)   '

Average  (TACC) per Company  ($10S)     $0.09       $3.24      $1.87

Numbe Share Numbe Shar Numbe Shar
r r ' e r e
 Companies with  Sales Data
   Compliance  costs  are <  1%
 sales
of
13

10
   Compliance  costs  are >.!%  and <
 3% of sales
100%

 77%


 15%
17

16
   Compliance  costs  are >. 3%
of
100%   30

 94%   26


 6%     3


 0%    'l
100%

 87%


 10%


 3%
 sales
Compliance Cost-to-Sales Ratios
Average
Median
i • l
'• 0
Minimum ' : 0
Maximum • ; 10

.10%
.00%
.00'% :
.20%

0
0
0
1

.27%
.14%
.00%
.29%

0
0
0
10

.63%
. 06%
.00% ,'
.20%
5.3   Economic Analysis          |
      The Agency also analyzed the economic impacts on small businesses who own operate facilities
included in the market model under with-regulation conditions expected to result from implementing the
NESHAP.  Unlike the screening analysis, this approach examines small business impacts in light of the
behavioral responses of producers and consumers to the regulation. As shown in Table 5-2, the
economic model projects pre-tax earnings to marginally increase by approximately $1.98 million, or
0.46 percent, for the eight small businesses14 included in the market model. As noted earlier, small firms
only bear 2 percent of the total annualized control costs and the per-unit costs of control are smaller
relative to other affected firms, leading [to an estimated increase in the level of pre-tax earnings.  This
increase is the net result of three effects:
"The  eight  small businesses  included in the market  model own a  total  of nine
  plants.
                                          5-3

-------
Table 5-2. Small Business Impacts of the Metal Can MACT After Market Adjustments:
1997a . .
Base
Revenues ($10s/yr)
Costs ($106/yr)
Compliance
Production-
Pre-tax Earnings
($106/yr)
Plants
Employment :
$56
$13
Line
0.87
2.06
$0 .
$13
$42

2.06
8.80

9
1,205
With
Regulation
$560.54
$129.75
$0.05
$129.70
$430.78

•9
1, 181
Absolute
Change
-$0.33
-$2.31
$0.05
-$2 .37
$1.98

0
' -24
Relative
Change
-0.06%
-1.75%
NA
-1.79%
0.46%

0.00%
-1. 98%
   This table  only presents results for those small  firms included in .the market  model.
   There  are  an  additional   six plants  owned  by  five  small  firms  that  manufacture
   speciality  products  and were, therefore  not included in the market model.
                                     i

                                .33 milli
Decrease in revenue ($0
This is offset to some degree
can is sold at a higher market
 llion, or -0.06 percent)—revenue declines as output declines.
?y increases in the market price of metal cans (i.e., each metal
price).
       •  Decrease in production costs ($2.37 million, or 1.8 percent)—production costs^decline as
          output falls.                         •    .

       •  Increased pollution control co[sts ($0.05 million)—these costs increase with the rule, although
          the estimated costs after allowing for behavioral adjustments are smaller than those estimated
          by the engineering cost analysis because these costs are assumed to vary with output.  Given
          that output declines, pollutiorjj control costs also decline relative to the costs estimated by the
          engineering analysis.

5.4    Assessment                   '
       After considering the economic impacts of the proposed rule on small entities. EPA certifies that
there will not be significant impacts on ajsubstantial number of small entities. We provide the following
factual basis for certification:          ;
       •  The screening analysis shows! only one of the 13 small firms is impacted greater than
          3 percent of total revenues.'  I

       •  Only one of the 15 facilities o|wned by small businesses is likely to prematurely close as a
          result of the rule using the bas|e elasticity assumptions. A second facility is estimated to close
          under some of the scenarios included in the sensitivity analysis.

       •  After taking into account behavioral responses of producers and consumers to the regulation,
          plants owned by small businesses included in the market model (nine total) experience a net
          increase in pre-tax earnings of $ 1.98 million.

       •  EPA does riot anticipate that small firms will be disproportionately affected relative to large
          firms. Small firms are only expected to incur approximately 2 percent of the total annualized
          costs of $56.2 million. In adcition, the average total annualized compliance costs are

                                              5-4

-------
          $90,000 per small firm compared to $3.2 million for large firms.  Finally, a comparison of the]
          cost-to-sales estimates shows small firms have a lower median CSR relative to large firms
          (<0.01 percent compared to!0.14 percent for the large firms, and 0.06 percent across all
          affected firms).

Although this proposed rule will not have a significant economic impact on a substantial number of
small entities, EPA continues to be interested in the potential impacts of the proposed rule on small
entities and welcome comments on issues related to such impacts.
                                             5-5

-------
                                    !   REFERENCES
                                    I                                     •  •

Bourguignon, Edward W. 1999. "Grown Accelerating for Coil Coating." Paint & Coatings Industry
       15(3):44-45.
Brody, Aaron L., and Kenneth S. Marsh'eds.  1997.  The Wiley Encyclopedia of Packaging
       Technology, Second Edition. Ne^v York: John Wiley & Sons, Inc.
Brody, Aaron L., and John B.  Lord (eds.j). 2000.  Developing New Food Products for a Changing
       Marketplace.  Lancaster: Technomic Publishing Co., Inc.
Brody, Aaron. November 27, 2001. Personal communication with Julia Wing, RTI.
Can Manufacturers Institute (CMI).  "1997 Retail Sales Prove It's Better in Cans." Canline 1(2).
       . As obtained on August 31, 1999a.
Can, Manufacturers Institute (CMI)r  "Consumers Vote Yes for Aluminum Cans." Canline 1(2).
       . As obtained on August 31, 1999b.
Can Manufacturers Institute (CMI).  "Domestic Can Shipment 1997." .  Obtained August 31,1999c.
Can Manufacturers Institute (CMI).  "History of the Can."  .
       As obtained on July 13, 1999d.  j
Can Manufacturers Institute (CMI).  "Historical CMI Can Shipments." .
       As obtained on December 6, 200Ha.
Can Manufacturers Institute (CMI).  "How Cans are Made." . As
       obtained on December  17, 2001 b.|
Chambers.  R.G. 1988.  Applied Production Analysis: A Dual Approach. Cambridge, UK: Cambridge
       University Press.              I
Dun & Bradstreet. Dun's Market Identifier Electronic Database. 1999.
Gale Research, Inc. 1999. Ward's Business Directory of U.S. Private and Public Companies. Detroit,
       MI: Gale Research, Inc.        j
Hicks, J.R.  1966. The Theory of Wages. \ 2nd Ed. New York: St.  Martin's Press.
Hillstrom, Kevin.  1994. Encyclopedia of American Industries.  Volume  1: Manufacturing Industries.
       Detroit, MI: Gale Research, Inc. j       .
Hoover's Incorporated.  1999.  Hoover's Company Profiles. Austin, TX:  Hoover's Incorporated.
       .
Icenhour, M., MRI. Memorandum to P. Almodovar, EPA and L. Pope, EPA. April 9, 2002. Tabular
       costs for metal can (surface coatirig) NESHAP.
Nicholson, W. 1998.  Microeconomic Theory:'Basic Principles and Extensions. 7th Ed. Fort Worth:
       The Dryden Press.             j
O'Neill, Martin.  1998.  "Polyethylene Terephmalate: In Packaging:  Low Price and High Growth
       Rates." Modern Plastics Jan:64.                                '
Palmer, K., H. Sigman, and M. Walls.  19,96. "The Cost of Reducing Municipal Solid Waste."
       Resources for  the Future Discussion Paper 96-35.
Purchasing Online.  September 15, 1998. "Transaction Prices."
Purchasing Online.  September 16, 1999. j "Transaction Prices."
Purchasing Online.  September 20, 2001. i "Transaction Prices."
Reeves, Dave.  "Metal Can (Surface Coating) MACT Floor Analysis." EPA Presentation by Dave
       Reeves of Midwest Research Institute, June 10, 1999.
Sfiligoj, Eric. 1995. "At What Price?" Beverage JForW June:46-50.
                                            R-l

-------
Thurman, W.N., T.J. Fox, and T.H. Bihgham. 2001. "Imposing Smoothness Priors in Applied Welfare
       Economics: An Application of the Information Contract Curve to Environmental Regulatory
       Analysis." The Review of Economics and Statistics 83(3):511 -522.
U.S. Bureau of Census. 1998. Quarterly Financial Report for Manufacturing, Mining and Trade
       Corporations. U.S. Bureau of the Census.
U.S. Bureau of the Census.  1999. 1997 Census of Manufacturing Industries: Metal Can
       Manufacturing. Core Business Statistics Series. EC97X-CS3.  Washington, DC • Government
       Printing Office.
U.S. Bureau of Labor Statistics.  National Employment, Hours, and Earnings—Metal Cans: Series ID
       eeu31341106. . As obtained on August 27, 1999.
U.S. Bureau of Labor Statistics.  Producer Price Index—Commodities: Aluminum Cans—Series ID
       wpul03103.  .  As obtained on December 6, 200la.                  .   .
U.S. Bureau of Labor Statistics.  Producer Price Index—Commodities: Steel Cans—Series ID
       wpulOS 102.  .  As obtained on December 6, 2001b.
U.S. Bureau of Labor Statistics.  Producer Price Index Revision—Current Series: Plastic
       Bottles—Series ID pcu3085#. {.  As obtained on December 6, 200Ic.
U.S. Department of Commerce, Burea|i of the Census.  1997. 1995 Annual Survey of Manufactures
       Statistics for Industry Groups and Industries, .
U.S. Department of Commerce, Bureau of the Census.  1998. 1996 Annual Survey of Manufactures
       Statistics for Industry Groups and Industries, .
U.S. Department of Commerce, Bureafa of the Census.  1999a. 1997 Census of Manufacturing Industry
       Series: Metal Can Manufacturing,  .
U.S. Department of Commerce, Bureau of the Census.  2001. Concentration Ratios in Manufacturing:
       .        '                           .
U.S. Department of Energy. 1999a. Electric Power Annual, Volume I. Table A2: Industry Capability by
       Fuel Source and Industry Sector, 1999 and 1998 (Megawatts).
       
                                  [

U.S. Department of Energy. 1999b. Natural Gas Annual. Table l:Summary Statistics  for Natural Gas in
       the United States, 1995 - 1999.; 

U.S. Environmental Protection Agency.  1993. Economic Impact and Regulatory Flexibility Analysis of
       Proposed Effluent Guidelines andNESHAPfor the Pulp, Paper, and Paperboard Industry.
       EPA-821-93-021. Washington, DC: Government Printing Office.
U.S. Environmental Protection Agency.  1998. "Preliminary Industry Characterization: Metal Can
       Manufacturing—Surface Coating."  .
U.S. Environmental Protection Agency.  1999. Economic Analysis Resource Document. RTP, NC:
       EPA.
U.S. International Trade Commissions ITC Trade Data Web. Version 2.4 [computer file]. ,
       . As obtained on December 7, 2001.
                                             R-2

-------
                         MODEL
   APPENDIX A
TA SET AND SPECIFICATION
       The primary purpose of the El A for the proposed metal can manufacturing MACT is to describe
       and quantify the economic impacjts associated with the rale.  The Agency used a basic framework
       that is consistent with economic analyses performed for other rules to develop estimates of these
       impacts.  This approach employs standard microeconomic concepts to model behavioral
       responses expected to occur with (regulation. This appendix describes the spreadsheet model in
       detail and discusses how the Agency
              •   collected the baseline 'data set for the model,

              •   characterized market ^upply and demand for three' submarkets of the metal can
                 industry — beverage cdns, food cans, and general packaging containers.
                                     i
              •   introduced a policy "shock" into the model by using control cost-induced shifts in the
                 facility-level supply functions, and

              •   used a solution algorithm to determine a new with-regulation equilibrium for each
                 market.              j
                                     i
A.I    Baseline Data Set              j
       EPA collected the following data {to characterize-the baseline year, 1997 (see Tables A-l and A-
   .    2):                            |
              •   Baseline Quantity—EPA collected facility-level production and mapped facilities to
                 appropriate markets using ICR survey responses. We estimated facility-level
                 production for plants Ajvithout ICR data using the following approach:
              i/     Collected secondary data on market-level output for each of the three categories of
                    metal cans modeled from a publicly available source provided by the CMI (see
                    Table 2-7).       !
                                             A-l

-------
Table  A-l.   Baseline Data  Set,  1997
             Market
Average Price
    ($/can)
 Domestic
Production
 (106 cans)
 Beverage
 Food
 Package
     $0.06

     $0.12

     $0.44
  100,680

   24,332

   4,375
Sources:     Sfiligoj,  Eric.-  June 1995.  "At What  Price?"   Beverage World June-.46-50.
      U.S.  Bureau of Labor Statistics.  Producer  Price Index—Commodities:   Aluminum
      Cans—Series ID wpu!03103.  [.  'As obtained on December 6, 2001a.
      Can Manufacturers  Institute (CMI).  "Historical CMI'Can Shipments."  .   As  obtained on December 6, 2001a.
Table  A-2.   Primary Supply and Demand  Elasticities  for  Metal Can
Market Models                 I
                Market
       Supply
  Demand
 Beverage
 Food.
 Package
          l
          1
          l
   -1.4 ;
  -0.63:
  -0.63
Sources:  Palmer, K., H. Sigman, and M. Walls. 1996.' "The Cost of Reducing Municipal Solid Waste." Resources for the
Future Discussion Paper 96-35.


             v^     Computed the difference between total market output for each of the three
                   categories modeled and total reported output calculated from summing ICR
                   responses for each market (i.e., total production-total reported ICR production =
                   total unknown production)
      Distributed unknown production across facilities that did not provide production data15 using
      ICR plant-level employment responses. Using this approach, the facility-level model is
      consistent with secondary market data.
             •   Baseline Prices—EPA computed 1997 baseline prices for the beverage can market
                using data from Sfiligoj (1995) and price  indexes from BLS (2001a). For the food
                can and general packaging container markets, the Agency employed -the following
                approach:
"These  are primarily area  sources.   In general  less information was  collected
   from area sources  than major sources because' major, sources are  the focus of
   the rule.   However, it is impojrtant to capture production from all sources to
   accurately  develop   the  baseline   and  estimate  post-regulation   market
   conditions.                    i
                                          A-2

-------
                     First, we estimated total revenue for the beverage can market using price16 and
                     total output.      [                  •
                     Next, we collected value of shipment data from the U.S. Census Bureau for Metal
                     Can Manufacturing (NAICS 332431) to obtain an estimate of total industry
                     revenue. We then subtracted revenue from the beverage market (as calculated
                                     l
                     above) from total revenue to approximate the total revenue in the food can and
                     general packaging container markets.
                     Using.census dataj CMI, and ICR data, we estimated the average revenue per
                     employee for the food can and general packaging container markets. We
                    • multiplied this valjue by total plant-level employment for each market to derive an
                     estimate of total revenue for each market.
                     Finally, we divided these two revenue estimates by their respective market
                    ; quantities to commute a market price. Using this approach, the facility-level
                     revenue totals are 'consistent with the value of shipments for the industry reported
                     by the Census Bureau (i.e., does not significantly understate or overstate total
                     industry revenues).
                 Domestic supply and demand elasticities—The primary demand elasticities used for
                 this analysis are drawn from Palmer, Sigman, and Walls (1996).  They report demand
                 elasticities of-1.4 for|aluminum beverage cans and -0.63 for steel cans. Because no
                 empirical estimates off the supply elasticity were identified, the primary supply
                 elasticity was assumed to be equal to 1.  Because of the inherent  uncertainty
                 associated with choosing point estimates of elasticities, a sensitivity analysis was
                 conducted where the supply elasticity was varied from 0.5 to 2 and the demand
                 elasticity was varied from -0.5 to -2.
1SEPA used the price  of  aluminuiji  cans   ($0.061/can)  for  the  beverage  market
   because  the overwhelming majority of beverage cans  are made  from aluminum.
                                             A-3

-------
A.2    Supply of Metal Cans
       The market supply of metal can? in each of the three defined submarkets (Qs) may be expressed
       as the sum of affected and unaffected producers, that is,
                                    i           Qs = qa + qu                                  (A.i)
where qa is the affected supply of a particular can type and qu is the unaffected supply.
                                    I  ~
A.2.1  Metal Can Facilities
       Producers of metal cans have some ability to vary output in the face of production cost changes.
       Production cost curves, coupled with data on market prices, can be used to determine the
      •facility's optimal production rate, including zero output (shut-down). EPA used the a
       Generalized Leontief profit function to characterize metal can facility supply curves.

A.2.1.1 Using the Generalized Leontief Profit Function to Derive Output Supply
       The specification of a facility's|profit function given by the generalized Leontief is as follows:17
                                                                                         (A.2)
Eq. (A.2) is an empirical model to estimate facilities' profit, where Pn is the net market price for product
       n manufactured by facility j, ^n is one variable proportion input (characterized by a cost index
       described below), P0, p,, and J32 pre model parameters, j indexes producers (i.e., affected
       facilities), and n represents the three commodities included in the market model. By applying
       Hotelling's lemma to the generalized Leontief profit function, the following general form of the
       product n supply function for facility j is obtained:
                                                       • 1_  '=m  '  —
                                                                                         (A.3)
where q,-,, is the quantity of product n produced by facility], Pn is the net market price for each product, ^n
       is the variable proportion input, ;yjn = P0 and Pn = P, are model parameters,] indexes producers
       (i.e.. affected facilities), and n represents the three markets. The theoretical restrictions on the
       model parameters that ensure upward-sloping supply curves are jjn > 0 and Pn < 0.  '
       Figure A-l illustrates the theorejtical supply function for product n represented by Eq. (A. 3). As
       shown, the upward-sloping supply curve is specified over a productive range with a lower bound
       of zero that corresponds with a shutdown price equal to  — -and an upper bound given by the
       productive capacity of qjM that is approximated by the supply parameter Yjn-  The curvature of the
"For additional details,
   form  (pages 172-173).

see Chambers  (1988)  for a discussion  of this functional
                                              A-4

-------
                                      I
       supply function is determined by (the Pn parameter.
       Supply function parameters: The 13 parameter is related to the facility j's supply elasticity for
       product n, which can be expressed as          .
                                        d
Taking the derivative of the facility supp
y function (Eq. [A.3]) with respect to price shows
                         P*.
                    Y2
                    '
                                                                            M
Multiplying this expression by Pn/qn results in the expression for the supply elasticity:
By rearranging terms, Pn can'be expressed as follows
                                                                                         (A.4}
                                                                                         (A.5)
                                                                                         (A. 6)
                                                                                          (A. 7}
                                      I
Values for the p parameter can be computed in two ways: econometric estimation using facility survey
       data18 or substitution of an econoraetrically estimated or assumed market supply elasticity for
 8For  a discussion, see  EPA  (1993)  and Thurman, Fox,  and Binghatn  (2001).
                                      !
                                              A-5

-------
       product n (£jn), the average annual production level of facilities (qjn), the variable production cost
       index (Ijn). and the market price! of the product n (Pn). Note that unlike the product-specific (3, the
       facility supply elasticity is not qonstant but varies with q, p, and I. For this analysis, we used the
       calibration approach because facility-level data available from the Information Collection
       Request (ICR) did not support econometric estimation. Using this approach, the remaining
       supply function parameter, Yjn> approximates the productive capacity and varies across products
       at each facility. This parameter does not influence the facility's production responsiveness to
       price changes as does the P parameter. Thus, the parameter Yjn is used to calibrate the model so
       that each facility's supply equation replicates the baseline production data.

       Variable production cost index: The cost-share weighted variable production cost index, IJ5 was
       constructed with the following data from the U.S. Bureau of Census:
              •   state-level wages paid by the metal can industry (NAICS 332431) divided by value of
                 shipments (w) and  |

              •   state-level materials:purchased by the metal can industry (NAICS 332431) divided by
                 the value of shipments (m).

Note, the Ij variable varies across facilities due to the two state-level variables (w, m).
       Before computing the cost-shar^ weighted index, the wage and materials variables were
       converted into indexes normalized to the average value of each variable. This conversion allows
       each variable to be measured in; terms of a relative index.  The state specific index was computed
       as follows:
                                *
                                    —        —   -              =                 ~      (A. 8).
where a is the national cost share of materials for the metal can industry (NAICS 332431) and 1-a is the
       national cost share of wages. Table A-3 summarizes the normalized cost index values computed
       for states with available data.
       Regulatory Response: The production decisions at these facilities are affected by the total
       annual compliance costs, Cj as provided by EPA's engineering analysis of capital costs, annual
       operating and maintenance costs, record keeping and reporting costs, and applicable monitoring
       costs required to comply with the metal can MACT.  The supply equation of
                                               A-6

-------
Table A-3.   Variable  Cost  Indexes,  1997
State
AL
•CA
CO
FL
GA
IL
IN
MO
NJ
NY
NC
OH
OK
PA
TN
TX
WA
WI
Labor
1.
, 1.
0 .
.' 1.
0.
1.
' 0.
1.
1 .
0 .
0.
1 .
. 0 .
' 0.
0.
0 .
1.
0.
Index1
01
02
89
00 '
94
35
63
08
42
94
97
15
93
93 . •
82
88
10
94
Materials Indexb
0
' - 1
0
' 1
0
0
0
1
0
1
--' •" 1
o
... - 1
'• 1
'- i
0
0
. ..-. ,]_
.71
.03
. 99
.14
.97
.96
.95
. 03
. 82
.06
.10 ' .
.97
.20
. 03
.07
• SB '
.95
.04 " " '
Variable
0
1
0
1
0
1
0
1
0
1
1
0
1
. • 1
1
0
0
1
Cost Index0
.74
.03
. 98
.12
. 97
.00
.92 '
.03
. 88
.05
.08
.99
.18
.02
.05
.97
.97
.03
a Computed as  follows:   (State  wages/State value  of  shipments)/(U.S.  wages/U.S.valueof
  shipments)                            •
b Computed as  follows:   (State  cos|t  of materials/State value  of  shipments)/(U.S.  cost  of
  materials/U.S.  value of shipments)I         -,,,.•
c Computed as follows:  0.90*Materials Index + 0.10*Labor  Index; shares were computed as follows:
  materials share =•• 0.90 = U.S. cost) of  materials/sum (U.S.  cost of materials+ U.S. wages) ' and
  labor share  = 1 -01 .90.            I     ,         -i   -.      '".".-•                    • ..
Source:      U.S.  Bureau of the Census.  1999.  1597 Census  of Manufacturing Industries:  Metal
             Can Manufacturing.   Cozje  Business.Statistics Series.  EC97X-CS3.  Washington, DC:
             Government Printing  Office.
each facility will be directly affected by the regulatory control costs, which enter as a net price change
       (i.e., PJ - Cj). Thus, the supply function presented in Eq. (A.3) becomes:

                                                        !;„
                                                                                         (A.9)

The total annual compliance costs per cap, cj5 are estimated given the annual production per facility and
       the regulatory cost estimates for each facility provided by the engineering analysis. Under this
       approach, we assume all regulatory costs vary to some degree with output.
       Closure Decisions: One of the most sensitive issues to consider in the EIA is the possibility that
                                             A-7

-------
       the regulation may induce a producer to shut down operations rather than comply with the
       regulation.  The data (i.e., direct observations of plant-level costs and profits) necessary to make
       definitive projections of these impacts are unavailable from the survey data.  Therefore, the
       Agency developed a method of identifying firm closure decisions using industry measures of
       profitability. The plant closure Criterion used for this analysis is:
                         ""  •"" • ' ••• ••HBaEBMBaii^^^^a^H • H -:»•      .(A,10)
where
              •   TR= Total Revenue !

              •   TVPC = Total Variable Production Costs (area under the supply function)

              •   TFPC = [(1-profit rate)*TR] - TVPC. This accounts for production costs that do not
                 vary with output (i.e., "fixed") and can be avoided by ceasing production.
                                    1
              •   TACC = Total Annual Compliance Costs.

Note that all of these variables are with-regulation values (i.e., they account for market adjustments).
       The U.S. Bureau of Census reports industry group financial ratios in their Quarterly Financial
       Report for Manufacturing, Mining and Trade Corporations (U.S. Bureau of the Census, 1998).
       For 1 997, the Census Bureau reports that income before income taxes (pre-tax earnings) for SIC
       group 34 (Fabricated Metal Products) was approximately 7.6 percent.19  For smaller firms (i.e.,
       firms with assets under $25 million) this ratio is 6.9 percent. Given the estimated 1997 values of
       revenue and variable production, costs, EPA developed an estimate of the total fixed production
       'costs so that the pre-tax profit rate for each facility exactly matches the rate reported by the
       Census.

A.3    Demand for Metal Cans
       Domestic demand for metal cans may be expressed by the following general formula for each
       product:                      j
where p is the market price for the product, r)d is the' domestic demand elasticity, and Bd is a
       multiplicative demand parameter that calibrates the demand equation for each product, given
       data on price and the domestic demand elasticity to replicate the observed 1 997 level of domestic
       consumption.

A.4    With Regulation Market Equilibrium Solution
       Producer responses and market adjustments can be conceptualized as an interactive feedback
       process. Plants facing increased production costs due to compliance are willing to supply
       smaller quantities at the baseline price.  This reduction in market supply leads to an increase in
       the market price that all producers and consumers face, which leads to further responses by
       producers and consumers and thjus new market prices, and so on. The new with-regulation
19In the short run, a plant would1 be presumed  to continue  .to operate as  long as
   variable  profits  are positive.   The Agency  considered  QFRs  income  before
   income  taxes  measure  as  a  reasonable approximation  of plant-level  variable
   profit rate.     -              \

                                             A-8

-------
equilibrium is the result of a sends of iterations in which price is adjusted and producers and
consumers respond, until a set of ptable market prices'arises where total market supply equals
market demand (i.e., Qs = QD).  Market price adjustment takes place based on a price revision
rule that adjusts price upward (downward) by a given percentage in response to excess demand
(excess supply).                I
The algorithm for determining wijth-regulation equilibria can be summarized by nine recursive
steps:                         |
       1.  Impose compliance cojsts.

       2.  Use supply functions to derive marginal responses given the base price.  ,
       3.  Check if TR>T.C (i.e.,
Eq. [A.7]); if not-set q^O.
       4.  Compare aggregate su Dply and demand.

       5.  Revise prices using the Walrasian auctioneer approach.

       6.  Use supply functions tb derive marginal responses given the revised price.

       7.  Check if TRXTC (i.e., JEq. [A.7]); if not set qj=0.

       8.  Compare aggregate supply and demand.

       9.  Go to Step #5 and continue until convergence is obtained (i.e., the difference between
          supply and demand is arbitrarily small).

-------
                                       APPENDIX B
                                 SENSITIVITY ANALYSIS

       As noted in Section 4, EPA's analysis is based on the best point estimates available of the
       responsiveness of supply and demand for metal cans to changes in their prices. This appendix
       examines the impact on the estimated results of varying these model parameters.  The key results
      .are discussed below:
             •   The social cost estimate remains essentially unchanged under all scenarios—As
                'shown in Table B-l and B-2, the social costs vary by 0.1  percent or less in each
                 scenario.

             •   The distribution of costs across producers and consumers depends on the relative
                 supply and demand elasticities—As consumers become more (less) responsive to
                 marginal changes in! price relative to producers, they will bear less (more) of the
                 regulatory burden. Similarly, as producers become more (less.) responsive to marginal
                 changes in price relative to consumers, they will bear less (more) of the regulatory •
                 burden. We can see !why these changes occur by examining a very simple
                 mathematical model of tax incidence:20
                                                                                     (B.laT
                                                                                      (B.lb)

"Derivation of  this result can  lie,found in intermediate microeconomic textbooks
   such as Nicholson (1998)      ,  ,

                                  i     '     B-l

-------
                                                                                      (B.lc)
             where             •    i
             dpD    = price paid by consumers
             dps    =price received by) suppliers
             dc     = per-unit control costs
             ss     = market elasticity of supply
             r|d     = market elasticity of demand
       For example, holding market elasticity of supply constant at one and varying the demand
       elasticity from -0.5 to -2.0 shows consumer losses fall as they become more responsive to price
       changes declining from (-$43.3.million to -$21.6 million) (see Table B-l).
             •  Closure projections slightly increase—one closure may occur in each market if we
                reduce the supply elasticity to 0.5 under all demand elasticity scenarios.

Table B-l.   Sensitivity  Analysis Result  Matrix
Supply

0.5



1.0



1.5



2.0

-

Elasticity

Change in consumer surplus
Change in producer surplus
Social cost
Plant closures
Change in consumer surplus
Change in producer surplus
Social cost
Plant closures
Change in consumer surplus
Change in producer surplus
Social cost,
Plant closures
Change in consumer surplus
Change in producer surplus
Social cost
Plant closures

-0.5
-$34.6
-$19.1
-$53.7.
-1
-$43.3
-$10.2
-$53.5
0
.-$47.7
-$5.7
-$53.3
0
-$50.3 '
-$2.9
-$53.2
0
Demand
-1.0
-$23.1
-$30.6
-$53.7
-1
-$32.5
-$21.0
-$53.5
0
-$38.1
-$15.2
-$53.3
0
-$41.8
-$11.3
-$53.2
0
Elasticity
-1.5
-$17.3
-$36.4
-$53.7
-1
' -$26.0
-$27.5
-$53.4
0
-$31.7
-$21.5
-$53.3
0
-$35.8
-$17.3
' -$53.1
0

-2.0
-$13.8
-$39.8
-$53.7
-1
-$21.6
-$31.8
-$53.4
0
-$27.2
' -$26.0
-$53.2
0
-$31.4
' -$21.7
-$53.1
0
                                            B-2

-------


B-3

-------


1. REPORT NO.
EPA-452/R-02-005
4. TITLE AND SUBTITLE
Economic Impact Analysis, of
TEC
(Please read
2.
Metal Can MA<
7. AUTHOR(S)
9. PERFORMING ORGANIZATION NAME AND ADDRESS
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standar<
Research Triangle Park, NC 2771 1
12. SPONSORING AGENCY NAME AND(ADDRESS ;
Director
Office of Air Quality Planning and Standan
Office of Air and Radiation
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
15. SUPPLEMENTARY NOTES

HNICAL REPORT DATA
Instmctions on reverse before completing)

3T Standards

Is
Is

3. RECIPIENT'S ACCESSION NO.
5. REPORT DATE
September 2002
6. PERFORMING ORGANIZATION CODE
8. PERFORMING ORGANIZATION REPORT NO.
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
None
13. TYPE OF REPORT AND PERIOD COVERED
Proposed regulation
14. SPONSORING AGENCY CODE
EPA/200/04

16. ABSTRACT
Pursuant to Section 112 of the Clean Air Act, the U.S. Environmental Protection Agency (EPA) is developing National Emissions
Standards for Hazardous Air Pollutants (NESHAP) to control emissions released from Metal Can Manufacturing. This report analyzes.
the economic impacts of the proposed rule. j
The estimated total annual cost for these facilities to comply with the proposed MACT standard is approximately $56.2 million.
Due to the total annual cost of compliance, an economic impact model estimates that production of metal cans will decline by 392
million cans, or 0.3 percent. The estimated price changejdue to the regulation ranges from 0.2 percent in the beverage can market to
0.4 percent in the general packaging market. The Agency! estimates pre-tax earnings for the companies owning the facilities in this
source category will decline by about 1.9 percent. In addition, EPA concludes that the rule may potentially result in one to two
premature plant closures. According to the Small Business Administration size standards, thirteen companies owing facilities in this
source category are considered small. Based on the resulb from the screening and market analysis, EPA certifies that there will not
be significant impacts on a substantial number of small entities.
17.
a. DESCRIPTORS
KEY WORDS AND DOCUMENT ANALYSIS
b. IDENTIFIERS/OPEN ENDED TERMS

18. DISTRIBUTION STATEMENT
Release Unlimited

air pollution control, environmental
regulation, economic impact analysis,
maximum achievable control technology,
metal cans
19. SECURITY CLASS (Report)
Unclassified
20. SECURITY CLASS (Page)
Unclassified

c. COSATI Field/Group

21. NO. OF PAGES
84
22. PRICE
EPA Form 2220-1 (Rev. 4-77)    PREVIOUS EDITION IS OBSOLETE

-------

-------