-------
TJ
C
C
-H
(C
a
0)
f-t
•£*
4-1
C
•H
W
4->
C
0)
1
in
• H
r-t
XI
(0
JJ
w
0)
Dl
c
-H
w
3
U
U
(0
I4_l
3
C
(C
e
4-1
O
c
o
-H
U
(0
U
0
J
•
1-^
in
00
/•o
\ N
U
M
w
••
r~
00
1
M
4J
I/I
3
73
C
•H
U]
JJ
U
3
*O
O
V^
a
-o
0)
•H
1— 1
l-(
(0
a>
w
(0
01
i— i
a>
cc
,
w
0
-H
U
01
W
4J
U
O
a
01
a
*,
T-l
(D
3
r^
O
>
.
uses
to
c
0)
0
U
-H
E
o
c
o
U
U
r-
oo
o
rH
0)
U
k-l
1)
s
0
U
i-H
&\
i-i
O
±
U
r-
OD
O
S
0>
1— 1
-H
M-l
[/)
JJ
c
*
U
o
§ ,
M
0)
>J
C 3
0) C
3
D)
(0 >w
a o
0)
W
C
0)
U
01
U
M
o
t/1
1-36
-------
displays the location of manufacturing establishments in the
paint and allied products industry by state.75 California has
the greatest number, 201, followed by Illinois with 118.
Paint manufacturing is fairly well represented in most states
east of the Mississippi River.
Over the past decade, consolidation has been a major
trend in the paint and allied products industry. The maturity
of the industry and increased technology requirements are
factors contributing to the restructuring. A large number of
mergers and acquisitions took place in response to pressure
from the higher cost of paint ingredients, intense industry
competition, compliance with government regulations, and low
profit margins.76 Other companies are divesting their paint
and coating operations to focus on other businesses or as an
alternative to making the capital and research and development
(R&D) commitments required to remain competitive. The number
of coating manufacturers and the number of establishments
operated by these manufacturers have decreased over the past
two decades. As indicated in Table 1-12, from 1972 to 1991,
the number of companies decreased by 12 percent, and the
number of manufacturing establishments decreased by over 20
percent.77
On average, 35 to 40 mergers or acquisitions have taken
place each year in the coatings industry for the past few
years.78 A transaction involves the transfer of production
capacity from one company to another but does not necessarily
indicate the dissolution of the company making the transfer.
The selling company could sell only a division or product line
and remain in business. Some of the large recent acquisitions
reported in trade journals, by the press, and in companies'
annual reports are listed in Table 1-13.79
Most of the larger companies produce architectural, OEM,
and special purpose coatings. Several of the largest coatings
producers are chemical corporations; however, paint manufac-
turing represents only a small part of their overall busi-
ness.80 In 1991, merger activity slowed down and left the
1-37
-------
TABLE 1-12. NUMBER OF COMPANIES AND ESTABLISHMENTS IN THE
COATINGS INDUSTRY, SELECTED YEARS, 1972-1991
Year
1972
1977
1982
1987
1991
% change 1972-1991
Number of
establishments
1,599
1,579
1,441
1,426
l,400a
-12.4%
Number of companies
1,317
1,288
1,170
1,123
1, 030a
-21.8%
* 1991 figures are from Finishers' Management. The U.S. Paint and Coatings
Industry. pp. 23-25. April 1991.
Source: U.S. Department of Commerce. 1987 Census of Manufactures,
Industry Series: Paints and Allied Products. Washington, DC, U.S.
Government Printing Office. 1990.
TABLE 1-13. RECENT ACQUISITIONS IN THE COATINGS INDUSTRY
Selling company Acquiring company
Division sold
DeSoto
Whittaker Corp.
Azko Coatings
Inc.
DeSoto
Clorox Co.
Sherwin Williams
Morton International
Reliance Universal
Inc.
Valspar
PPG Industries, Inc.
Consumer Paint Operation
Specialty Chemicals
Operation
Buyout
Coil Coatings Operation
Olympic and Lucite
finishes
Source: Loesel, Andrew. Coatings Industry Faces New Mix. In Chemical
Marketing Reporter. 238(18):SR3-SRB. New York, Schnell Publishing
Co. 1990.
1-38
-------
industry basically divided into two groups: a few, well-
financed and highly diversified multinationals and a large
number of regional paint companies.81
1-39
-------
SECTION 2
COST AND ECONOMIC IMPACTS OF PROPOSED REGULATION
This section assesses the costs that must be incurred to
comply with the proposed AIM coatings regulation and examines
the economic impacts of these costs as they are absorbed by
producers and consumers of the regulated products through
market processes. Although the effects of this regulation may
be felt outside of the markets in which the regulated products
are traded (e.g., in markets for coating substitutes and
coating input markets [including labor markets]), the analysis
focuses entirely on the AIM coatings product markets
themselves.
2.1 BACKGROUND
The EPA plans to control VOC emissions from AIM coatings
using a combined regulatory approach: (1) product category-
specific VOC content limits and (2) an option for producers of
products over the proposed content limits to pay a fee on the
VOC content in excess of the limit. Using reformulation cost
estimates and a proposed fee rate, the potential impacts of
the proposed regulation are analyzed, first in static analyses
of the reformulation, fee, and withdrawal options and second
with a dynamic market analysis that estimates changes in
prices and quantities and welfare costs.
2-1
-------
2.2 OVERVIEW OF RESPONSE OPTIONS
The regulation to reduce the VOC content of AIM coatings
will affect both production decisions for the suppliers of the
coatings and consumption decisions on the demand side. Before
developing a formal economic model to analyze these
regulations, the scope of responses available to producers and
consumers are briefly characterized.
2.2.1 Supply
The EPA is proposing a set of limits for the VOC content
in specific product categories to be met by 1997. Firms that
produce products exceeding the proposed VOC limits essentially
have three potential options: reformulate the products so
that they comply with the standard, potentially paying a fee
on the excess VOC content over the standard, or remove the
product from the market.
The firm will presumably choose the option that maximizes
net benefits, as measured by the expected (discounted) value
of the profits generated under each option, less the
(discounted) cost of each option. However, in the short run,
firms may be unable to reformulate their coatings to meet the
content limits requirement because of the timing of the
regulation or constraints on the resources that can be
allocated to reformulation. They may remove the product from
the market temporarily until the VOC standard can be achieved
or potential pay the exceedance fee. If technological and
market conditions allow, they may increase their production of
within-limit products to compensate for the reduction in over-
limit products until the firm makes long-run adjustments to
adapt to the product-line restrictions. This strategy may be
temporary until the over-limit product(s) can be successfully
reformulated, or it may reflect a permanent response to the
regulation.
The compliance strategy decision is likely to be
complicated by issues other than cost that relate to the
2-2
-------
profitability of reformulation. If a product serves a narrow
market niche, reformulation may fundamentally alter the
product's attributes and erode the niche position. In such a
case, the producer may not find choosing reformulation
profitable. Although concerns regarding the regulation's
constraints on product differentiability are undoubtedly real
in some cases, this complexity is not explicitly addressed in
the quantitative analysis that follows, primarily because of
the difficulty in observing both levels of and changes in
product differentiation. Moreover, the proposed limits have
been determined technologically feasible, subject to Best
Available Controls, which should imply that the reformulation
will not fundamentally alter the functions that the product
can perform.
Often, product reformulation involves an investment in
research and development (R&D) to develop a compliant product.
The extent of the reformulation necessary to bring a product
into compliance can vary from product to product. In some
cases, compliance can be achieved for a particular product
without large R&D investments because the product is similar
enough to an existing formula or other product undergoing
reformulation. A major reformulation, as discussed here,
typically requires a significant resource and time commitment.
The process can take several years and is divided into a
number of different stages. Figure 2-1 identifies the basic
reformulation stages for a prototype AIM paint (other coatings
such as varnishes may have fewer stages) .82 The firm may need
to alter its capital equipment to produce the reformulated
product, but these physical capital adjustments are usually
small compared to developing the intellectual capital to
devise the new formula.3
aOne member of the Regulatory Negotiation Committee raised the point
that substituting away from solvents in the new formulations introduces the
problem of disposing of old solvent storage containers. Another member
mentioned disposing of obsolete labels as a potential problem as well.83
2-3
-------
Formulate a White Paint
Conduct Field Testing and Color Formulation
Formulate the Bases
(that will be tinted)
Determine Prescription for Retail
Stores
Develop Information
(technical and safety sheets, marketing
information, labels)
Scale Up Production. Introduce, and Distribute
Source:
Estimated Effort 2 '° 3 person-years
Estimated Elapsed Time 1 5 to 5 5 years
Figure 2-1. Basic stages of AIM coating reformulation
(prototype firm and product).
AIM Coatings Regulatory Negotiation Committee meeting, July 28-
30, 1993, Washington, DC. Meeting Summary.
2.2.2 Demand
The regulations can be expected to induce changes in the
prices of the affected products. A consumer may alter his/her
selection of coatings based on the relative prices of coating
products and on the relative prices of coating versus
noncoating alternatives. For example, consumers may opt for a
waterborne coating rather than its solventborne alternative if
the regulation-induced change in prices increases the relative
price of the solventborne product. Moreover, a potential user
of a high-VOC coating product facing reformulation may even
opt for a noncoating alternative if the price rises too much.
The reformulated products may also possess different
characteristics that affect their demand. For instance, VOC
content reduction in a typically high-VOC product may change .
2-4
-------
consumers' perceptions of the product's performance, dura-
bility, and ease of application. The lower VOC content may
also work as a signaling device for the "green" consumer in
pursuit of products deemed more friendly to the environment.13
These factors collectively affect the benefit consumers derive
from using the product and thus affect their willingness to
pay for the reformulated product versus other product
alternatives.
2.3 COST ANALYSIS
The regulations will cause firms who make products that
exceed the allowable VOC content specified in the Table of
Standards (TOS) to either reformulate, pull the product from
the market or pay an exceedance fee. This section evaluates
the costs imposed on manufacturers to reformulate non-
compliant products and the option to pay an exceedance fee.
Section 2.4 incorporates the option of withdrawing products
from the market into the decision process to evaluate national
market efforts.
2.3.1 Reformulation Costs
The estimated national cost for the regulation is based
on information developed by industry representatives during
the regulatory negotiation. The assumption in estimating
these costs was that coating technologies would need to be
researched and developed in the laboratories of resin
manufacturers/suppliers and paint manufacturers in order to
meet VOC requirements. Although the proposal is significantly
less stringent than the three-phases of requirements discussed
during negotiations, the EPA has relied on these same
reformulation cost estimates for calculating the national cost
of the proposed rule. Given that the rule has similar VOC
content requirements to State rules which have been enforced
fcSome manufacturers currently produce zero-VOC-content coatings that
are marketed as "clean air" coatings.
2-5
-------
since 1990, the EPA believes the reformulation estimates used
may be overstated. Since the proposed rule is implementing
available resin technologies, the cost to comply for those
manufacturers who have or will reformulate their higher VOC
coatings is expected to be partially reduced through the
assistance of resin manufacturers/suppliers. Upon request,
most resin suppliers are willing to share information and
sample low VOC coating formulations with interested paint
manufacturers, both large and small. In addition, another
limitation in the cost data is that no distinction for
reformulation cost is made between categories (i.e., the
reformulation cost in one category is the same as the
reformulation cost in any other category), or in relation to
the required VOC content reduction (i.e., it does not
distinguish between coatings at different VOC levels above the
limit). In the preamble for this proposed rule, the EPA
requests comment and technical information on previous (since
1990) or potential reformulation costs specific to each
category and VOC content level change. In addition,
information is requested on any changes in variable (e.g., raw
material) costs or disposal costs associated with
manufacturing coatings to meet the proposed VOC levels.
2.3.1.1 Product-Level Reformulation Cost Estimates.
Whether reformulation is a feasible response for the proposed
VOC content limit regulations is an empirical issue that
varies by firm and product-specific factors as well as by
timing specifications of the regulation. Since there is
insufficient data for this study to observe these firm and
product-specific factors, the analysis that follows discusses
costs involved for a representative firm facing a typical
reformulation in response to the regulations.
The new formula represents a different mix of the four
coating components: resins, solvents, pigments, and addi-
tives. For solventborne products, a new formula may involve
increasing the ratio of solids (resins) to solvents to reduce
2-6
-------
the solvent's contribution to VOC emissions. As explained
previously, a reformulation may be "major," involving
significant investment in R&D or it may require relatively
little investment cost. The discussion of reformulation costs
that follows refers to major reformulations.
A presentation to the AIM Regulatory Negotiation
Committee indicated the level of effort required for develop-
ing a new coating product. Costs were estimated based on this
level of effort.84 These cost estimates are for developing a
new coating product, which may be more difficult and more
expensive than reformulating an existing product. However,
these data provide the only available estimates for reformula-
tion costs in this study. Some of the upward bias implied by
using new product development costs as a proxy for
reformulation costs may be offset by downward biases resulting
from omission of certain non-R&D costs in the reformulation
effort (e.g., test materials, non-R&D labor). Raw materials
costs also may increase because reformulated products often
have a higher solids-to-solvent ratio, and higher solid resins
are more expensive than solvents.85 The various uncertainties
and potential biases surrounding the reformulation costs esti-
mates are presented in Table 2-1. Since these biases cannot
generally be quantified, the costs derived in this section are
used as a qualified best estimate of reformulation costs.
The levels of effort for developing a new product
indicated by the Regulatory Negotiation Committee presentation
are
• 2 to 3 scientist years and
• 1.5 to 5.5 years elapsed time.
Given the above time ranges, a midpoint estimate of the level
of effort is 2.5 scientist years over a 3-year period. Esti-
mating the cost of a scientist year at $100,000 gives the
total cost of a major reformulation effort as $250,000,
assuming these costs are incurred evenly over the 3-year
2-7
-------
TABLE 2-1 REFORMULATION COST ESTIMATION UNCERTAINTIES
Assumptions
• Reformulation effort: 2.5 scientist years over 3-year
period
Cost per scientist-year (1993 $) = $100,000
• Discount rate: 7 percent
• Initial lump-sum cost annualized on perpetual "capital
rental" basis (discount rate • reformulation cost)
Potential upward bias factors
• Cost estimate is for new product; reformulated product
may be lower.
• Cost estimate is for "major" reformulation; minor
adjustments may be feasible.
Major reformulation effort is based on expectations of
the originally proposed three-phase Table of Standards,
with each phase progressively more stringent. The
reformulation effort may be less intensive under the
current single phase proposal.
• Costs may fall over time as new technology is developed,
disseminated.
• Costs may be partly borne by material suppliers.
Potential downward bias factors
• Perpetual capital rental basis for cost annualization
assumes reformulation benefits accrue forever. Finite-
lived benefits would produce higher annualized costs.
Ignores nonresearch costs of reformulation (e.g.,
materials).
• Multiple products may be lumped together as one in the
survey. Therefore, multiple reformulations may be
necessary in some cases where a single reformulation is
proj ected.
Potential factors with unknown bias
• Estimate is for a white paint; other products may
differ.
• Reformulation may positively or negatively affect
variable production costs (e.g., materials).
Effects on product quality (i.e., "cost" of reduced
quality).
Costs may rise/fall based on amount of "excess VOC" to
reduce.
period ($83,333 per year). Since the effects of the
regulation are annual reductions in VOC emissions, placing the
corresponding costs on an annualized basis is appropriate.
2-8
-------
First, the cost of the initial reformulation is valued in the
same year in which the benefit (emissions reduction) stream
begins (i.e., at the end of the 3-year reformulation period).
Thus the future value of the 3-year expenditure is computed at
the end of the third year, when the reformulation is complete
and the benefits of the reformulation commence. Here and
throughout the analysis a 7 percent discount rate is used.
Valuing each annual expenditure at the end of the year gives
the future value of
FV = 83,333 • (1.07)2 + 83,333 • (1.07) + 83,333
= $267,908 . (2-1)
To compute the annualized value, we took the product of the
discount rate and the initial cost
A = r • FV = (.07) (267,907) = $18,754 per year . (2-2)
This annualization formula is based on the notion that the
reformulation produces a perpetual stream of emissions
reduction benefits. Thus the annualized cost of the
investment can be viewed as the foregone perpetual annual
yield of the funds applied to alternative uses.
In the cost and market analyses that follow, all dollar
values are expressed in their 1991 equivalents, to match the
price data that were obtained for the market analysis. The
annualized cost figure in Equation (2-2) is based on a 1993
cost estimate. This value is deflated to 1991 dollars using
the GDP price deflator and derive an annualized cost of
$17,772 per reformulation.
As a point of comparison, estimates of the cost of AIM
coatings reformulation are provided in a study conducted for
the South Coast Air Quality Management District (SCAQMD) to
address economic impacts of VOC content regulations in
California.86 This study identified costs associated with
2-9
-------
product reformulation and temporary and permanent product
sales losses. Reformulation costs varied depending on the
extent of the reformulation necessary. Most of the small
firms surveyed indicated that they did not have full-time R&D
employees. Costs for additional research and development due
to the regulation ranged from $1,000 to $5,000 annually for
firms with few products affected by Rule 1113 and more than
$50,000 for firms with many affected products and little or no
research staff.
The SCAQMD study also indicated other compliance costs
not related to R&D. Rough estimates of the cost of equipment
adjustments necessary to accommodate reformulation ranged from
$5,000 to $35,000 per firm. Costs attributed to temporarily
or permanently discontinued products ranged from zero to
$3,000 for firms with few affected products to more than
$75,000 for firms with many affected products. Employment
changes for the surveyed firms in the SCAQMD study were
expected to be minimal, affecting only the possible addition
of R&D chemists.
Because the timing, number of reformulated products, cost
components, and regulatory structure associated with each
SCAQMD cost estimate is not apparent from the report, they
cannot be combined with the Regulatory Negotiation Committee
estimates given above in any meaningful fashion to provide a
better estimate of regulatory costs. Therefore, the estimates
provided by the Committee are used in the analysis performed
here as the best estimate of reformulation costs, recognizing
the limitations discussed above and outlined in Table 2-1. To
evaluate the sensitivity of the resulting economic impacts of
the regulations, the analysis that follows also evaluates
alternative reformulation and exceedance fee values.
2.3.1.2 Total Reformulation Costs and VOC Reductions.
At the industry level, some producers will respond to the
regulations by reformulating products that exceed the VOC
content limits, some will pay an exceedance fee, and others
2-10
-------
will remove noncompliant products from the market. The
analysis begins with the conservative assumption that only the
first option, reformulation, is allowed. Then, the cost of
reformulation is recomputed assuming all products over the TOS
limits will reformulate. In this section, aggregate
reformulation costs are for the products reported in the
Architectural and Industrial Maintenance Surface Coatings VOC
Emissions Inventory Survey (the survey).87 The survey
population represents roughly three-fourths of total industry
output. In Section 2.4, the analysis is extended to the
national industry level to calculate market effects of the
regulations.
To estimate reformulation costs for the entire survey
population, a determination of the number of AIM products that
will need reformulation to comply with the standards is made.
This number depends on the number of AIM products with a VOC
content exceeding the proposed standards for the respective
product categories. Table 2-2 provides the proposed TOS.
Next, the number of products in the survey that exceed
the limits imposed by the TOS is determined. The survey
reports the number of products, sales volume, and average VOC
content for specific VOC content ranges (e.g., 0 to 50 g/L, 51
to 100, 101 to 150) within specific product groups (e.g.,
exterior flat waterborne, exterior flat solventborne, interior
flat waterborne). Knowing the limits imposed by the TOS, the
number, volume, and average VOC content of products over the
limit can be derived using the survey data. These data can be
used to generate estimates of the expected cost of
reformulating products subject to the TOS, as well as the
associated reduction in emissions accomplished by the
reformulations. Using these cost and emission reduction
estimates, the cost-effectiveness per metric ton (Mg) of VOC
reductions is analyzed.
2-11
-------
TABLE 2-2. TABLE OF STANDARDS
VOC content limit
AIM coating (g/L)
Antenna coatings 500
Antifouling coatings 450
Antigraffiti coatings 600
Bituminous coatings and mastics 500
Bond breakers 600
Chalkboard resurfacers 450
Concrete curing compounds 350
Concrete protective coatings 400
Dry fog coatings 400
Extreme high durability coatings 800
Fire-retardant/resistive coatings
Clear 850
Opaque 450
Flat coatings, N.O.S.
Exterior 250
Interior 250
Floor coatings 400
Flow coatings 650
Form release compounds 450
Graphic arts coatings (sign paints) 500
Heat reactive coatings 420
High-temperature coatings 650
Impacted immersion coatings 780
Industrial maintenance coatings 450
Lacquers (including lacquer sanding 680
sealers)
Magnesite cement coatings 600
Mastic texture coatings 300
Metallic pigmented coatings 500
Multicolor coatings 580
(continued)
2-12
-------
TABLE 2-2. TABLE OF STANDARDS (CONTINUED)
VOC content limit
AIM coating (g/L)
Nonferrous ornamental metal lacquers 870
Nonflat coatings, N.O.S.
Exterior 380
Interior 380
Nuclear power plant coatings 450
Pretreatment wash primers 780
Primers and undercoaters, N.O.S. 350
Quick dry coatings
Enamels 450
Primers, sealers, and undercoaters 450
Repair and maintenance thermoplastic
coatings 650
Roof coatings 250
Rust preventive coatings 400
Sanding sealers 550
Sealers 400
Shellacs
Clear 650
Opaque 550
Stains
Opaque 350
Clear and semitransparent 550
Waterborne low solids 120
Swimming pool coatings 600
Thermoplastic rubber coatings and mastics 550
Traffic marking paints 150
Varnishes 450
Waterproofing sealers and treatments
Clear 600
Opaque 400
Wood preservatives
Below ground 550
Clear and semitransparent 550
Opaque 300
N.O.S. = Not otherwise specified.
2-13
-------
Table 2-3 presents the results of the analysis for the
TOS.88 Table 2-3 reports reformulation costs and emissions
reduction summed across all survey products. A breakdown of
costs and emissions reduction by product category is provided
in Appendix C. Aggregate cost estimates for the survey
population were derived by multiplying the number of products
in the survey facing reformulation by the annualized cost of
reformulation. As Table 2-3 indicates, 1,729 products from
the survey exceed the 1997 limits, which is 36 percent of the
total number of products in the survey (4,846).c The same
presentation to the Regulatory Negotiation Committee that
serves as the basis for reformulation cost estimates also
indicated that roughly one in three products that exceed the
proposed limits would not need a major reformulation,
primarily because the product lines are similar to others that
will be reformulated. Thus, the costs are assessed for the
remaining two-thirds of products over the limit to computing
the aggregate cost estimate. After reducing the number of
products, the estimated number of reformulations is 1,153,
yielding an aggregate cost of reformulation of $20.5 million
dollars (1991 dollars) for the survey population.
To compute the average cost per Mg of emissions
reduction, an estimate of the expected emissions reduction for
the survey population is needed. Total baseline emissions for
the survey population is 344,059 Mg. The 1997 TOS is
projected to reduce aggregate VOC emissions by approximately
20 percent.89 Applying these percentages to the baseline
emissions figure for the survey population yields estimated
reduction emissions of 68,812 Mg.
cThe actual survey total number of products is 4,920. However,
throughout Section 2 we use 4,846 as the total number (and the
corresponding quantity and emissions) because product-level data were
unavailable for 74 products in the survey.
2-14
-------
s
o
M
£-1
J
ID
(J-i
O
CU
>H
W
>
&
D
CO
-— '
CO
^
£-*
o
M
EH
U
B
w
CH
CJ
o
>
Q
S
rtj
CO
EH
CO
O
U
§
M
EH
<
j
h-j
2
CM
0
Cn
W
f^j
Pi
B
m
l
05
TABLE
j_j
w —
O D
U Dl 2
2 --.
Q) -H
O> l-i cn
(0 Q) en
M Q.
«} O —
Jj 4-1
O QJ
EH H
M— J
o
tn1 §
X -H
0 £*
u * ^
T3 -H en
QJ g en
M C rH
2 oi2
^ ^
2 QJ
1 *
*^
O 0
JJ
kl W -H
QJ JJ E
JJ. U -H
Sr-< ^
tJ rH
3 -O
2 O
^
a
CO
en
CN
(N
CO
CO
VO
if
(N
m
in
CO
•a1
o
(N
CN
r~
r^
r~
t-i
ro
in
<— i
T-H
en
(M
[-
rH
C
O
-H
JJ
03
' '
D
E
l-i
0
4-1
OJ
1-1
i-l
o
•r— i
(0
E
03
T3
Q)
Q)
C
jj
O
c
o
T3
jj
-H
E
•H
i— 1
QJ
JS
JJ
M
QJ
>
o
w
jj
u
3
•a
0
S-l
a
4-(
0
•a
'H
-H
x;
jj
i
QJ
C
0
jj
CO
x;
jj
c
o
-H
JJ
a
E
3
W
tn
oo
Q)
x; •
jj —
jj
c x
O 0)
Jj
TJ
0) QJ
W QJ
03 W
CQ —
ID
>l
J3
T3
Q)
-H
rH
a
-H
JJ
rH
3
E
„
M
TD
VH
CO
T3
C
05
JJ
W
4-1
0
Q)
rH
XI
03
E-
r^
CTl
cn
rH
QJ
x:
Jj
VH •
0 >i
4-1 QJ
>
dP In
0 3
CM W
4-1 QJ
o x;
jj
w
Q) £
JJ O
03 l-i
Dl E4-i
C -H
•H JJ Dl
-U MS
C Q)
3 en
O Cm
lH O 0
-H
O JJ ^
JJ U •*
3 n
01 T3
3 QJ 4-1
T3 l-i O
4-1 VI U)
4-1 C C
O O O
-H -H
0) cn cn
X! cn cn
•H -H
>i E E
co 0)0)
c i
Q) ^l
l-l O
(Xi JJ
03
r-l
• 3
>i Dl
0) QJ
> OS
l-l
3 2
CO M
<
>1
l-l QJ
o x;
jj jj
c
Q) x:
> jj
C -H
M 3
W C
C O
O -H
•H JJ
W 03
cn 1-1
•H QJ
E a
W 0
o
0 U •
O C*O
> C 0>
•rH CA
W iH
Dl C
C 0
•rH -rH
JJ JJ JJ
CO CO l-l
0 -H 0
cj u a
O QJ
QJ cn oi
u cn
03 < JJ
4-1 4-1
VH cn co
3 D> ^
W C Q
•H
2 -U rH
M CO CO
,r-t ^
Industr;
Nationa
Industr;
Q)
u
lH
3
O
W
2-15
-------
Combining the cost and emissions reduction estimates for
the survey population, an estimate of the cost per Mg of VOCs
reduced is derived as $298/Mg.
2.3.2 Exceedance Fee Option
As an alternative to strict adherence to the limits
imposed by the TOS, the EPA is taking comment on an option to
provide AIM coatings producers the alternative of paying a fee
per unit of output for products that exceed the limit. The
per-unit fee will be computed as follows:
fee = (VOC content - VOC limit) • rate . (2-3)
VOC content is measured in grams per liter (less water and
exempt compounds) and the fee rate is paid on the grams per
liter in excess of the limit. The proposed fee rate is $0.005
per excess gram per liter (1995 dollars) with annual adjust-
ments based on the gross domestic product (GDP) price defla-
tor. Total fee payment per product simply equals the per-
liter fee times total liters of production.
2.3.3 Reformulation versus Exceedance Fee: Least-Cost
Analysis
In this section, an expansion of the producers'
compliance options is considered by allowing for the fee
option. The premise of the analysis conducted in this section
is that AIM coatings producers will choose the less costly of
the reformulation and exceedance fee options as a compliance
strategy.d The choice is based largely on two product-
specific factors: quantity of output produced and the
"excess" VOC per unit.
The diagram in Figure 2-2 helps explain the effect that
output quantity has on the choice between reformulating the
dThe coatings manufacturers may undertake the reformulation activity
themselves, though, in some cases, manufacturers who allocate few resources
to R&D may rely on material suppliers to provide compliant formulations.
Although the reformulation activity may be performed by another firm, the
costs incurred will presumably be borne by the manufacturers facing the
regulatory constraints.
2-16
-------
Average Cost of
Reformulation
QT'
QT
Output Level
Favors Fee Favors Reformulation
Figure 2-2. Fee versus reformulation.
product and paying an exceedance fee. Since the cost of
reformulation is a fixed cost (i.e., it is independent of the
level of output) the average reformulation cost per unit of
output falls as output levels increase. This situation is
represented by the downward-sloping line in Figure 2-2.
However, the exceedance fee per unit of output is constant
with respect to the output levels. Let F be the fee per unit
of output; the flat line extending from F on the vertical axis
indicates that the fee rate is constant. In Figure 2-2, for
all output levels less than QT the average cost of reformula-
tion is higher than the per-unit fee, and for all output
levels above QT, the average cost is below the fee. This
relationship indicates that the fee is the less costly
alternative when output is below QT and reformulation is the
less costly alternative when output is above QT. Thus the fee
is more likely to be chosen by small volume producers, all
else equal. As Figure 2-2 illustrates, the existence of a fee
2-17
-------
places an upper limit on the per-unit costs of complying with
the regulation.
Figure 2-2 also allows us to see the effect of different
fee rates on the "threshold point" of quantity, below which
the fee is the preferred option. If the fee were F' instead
of F, reflecting either a higher assessment rate per Mg of
emissions or a higher amount of excess VOC per unit, the
threshold point would be lower. Thus for higher excess VOC
categories and for higher assessment rates we would expect to
see fewer producers selecting the fee option, all else equal.
Because the fee will be more cost-effective only for lower
volume products and lower excess VOC categories, allowing the
fee option should have a relatively small impact on variation
from the aggregate emissions reduction targets as long as the
fee assessment rate is not set at an extremely low level.
Sensitivity to the fee rate is also analyzed below.
The least-cost analysis presented here determines which
option (fee or reformulation) would impose a lower cost to the
manufacturer for specific products within a certain VOC
content range from the survey. For the purpose of this
analysis, a product stratum is defined as all products
existing in a specific VOC content range for a specific
product category. An example of a stratum would be all
exterior flat waterborne products in the 101 to 150 g/L VOC
content range. Thus, for the TOS, all strata in the survey
are examined to determine those that exceed limits for their
respective product categories. As indicated above, the survey
includes data on the number of products, sales volume, and
baseline VOC emissions for each stratum. These data are used
to compute average sales volume per product for all strata
exceeding the TOS limits. Then, the average (per liter) cost
of reformulation is computed for each stratum by dividing the
annualized reformulation cost per product by the average
volume per products. Suppose the average sales volume per
product for one stratum is 100,000 L/yr. Given the annualized
2-18
-------
reformulation cost estimates from the previous section
($17,772 per product, 1991 dollars), the average reformulation
cost for this stratum is $0.17/L.
To determine the exceedance fee for each stratum, the
midpoint of the VOC content range is used as an estimate of
average VOC for the stratum. This measure was used to compute
excess VOC content because it is consistent with the regula-
tory definition of VOC content (grams per liter less water and
exempt compounds) and is available for each stratum. By
contrast, the "actual VOC" data in the survey are not consis-
tent with the regulatory definition of VOC content and are not
available for each stratum in the survey.
First the fee rate is adjusted to 1991 dollars by
multiplying the proposed fee rate (in 1995 dollars) of 0.005
by the ratio of GDP price deflators for 1991 over the 1995
(projected). The resulting rate is 0.0045. Suppose the
midpoint of the stratum is 75 g/L above the proposed limit.
The associated fee per unit would be 75 • $0.0045 = $0.3375/L.
This fee is higher than the average reformulation cost per
liter ($0.17/L). Under these conditions, it is assumed that
products in this stratum would reformulate rather than pay the
exceedance fee.6 This decision would be reversed if, for
instance, the stratum exceedance were 25 g/L, in which case
the fee would be $0.1125/L, which is less than the average
reformulation cost ($0.17/L), or if average sales volume were
25,000 liters per product instead, in which case average
reformulation cost would be $0.68/L.
These average reformulation cost and fee per-liter
calculations are performed for each stratum above the proposed
eBy conducting the fee-versus-reformulation decision at the stratum
level, and basing the decision on average cost and fee for each stratum, it
is implied that all products within the stratum are identical to the mean
values. In reality, there will be some variation around the mean so that
some products may find one alternative less costly while others find the
other alternative less costly. This analysis is unable to capture this
heterogeneity with the available data, but presumably these effects are
smoothed out as the analysis compares means across the hundreds of strata
in the survey.
2-19
-------
TOS limits to determine the relative frequency of
reformulation/fee selections and their impact on costs and
emissions reduction. Results are presented in Table 2-4 and
can be summarized as follows: (1) the fee option is chosen
for a sizable minority of products exceeding the limit, and as
a result, (2) there is a substantial reduction in aggregate
compliance costs; however, (3) overall product volume for
products selecting fee is very small.91 As explained shortly,
the combined result of factors (1), (2), and (3) is that the
fee option provides a means to reduce compliance costs with
little negative effect on overall emissions reduction results.
Under the standard fee rate of $0.0045 (1991 dollars), we
estimate that the fee is the preferred alternative for 323 of
the 1,153 products (28 percent) facing the reformulation
versus fee decision/ However, these products only account
for 26.1 million liters of output, about 1 percent of total
coating volume for the survey population. Total fee payments
for those products is just under $4.0 million (average:
$0.15/L); however, the estimated avoided reformulation costs
for the 323 products choosing the fee is over $5.7 million
($0.22/L) for a net aggregate savings to producers of about
$1.8 million. Moreover, because the fee payment is simply a
transfer from one sector of society (AIM coatings producers)
to another (the government), the net social cost savings are
the full $5.7 million reformulation cost savings, less any
costs of administering the fee.
One obvious policy-relevant question is how much will the
selection of the fee option undercut VOC emission reduction
goals. Since products will be allowed to exceed the regula-
tory standard if they pay the fee, the associated excess VOCs
from these products can be viewed as the foregone emissions
£Note that 1,153 products represent two-thirds of the total number
exceeding the limits because the other one-third were assumed to
reformulate without incurring the "major" reformulation cost.
2-20
-------
CO
H
CO
>H
iJ
<
*3,
EH
CO
O
U
1
r .
^
CO
< ^
S§
EFORMULATION : :
SURVEY POPULATI
: VERSUS R
FOR THE
w
W
h
,
^
i
CN
W
J
PQ
<;
£
01
c
§1 5
*> Ul~
^ c
° § w
^ •-, c
rn W 0
^ CO -H
2 •" jj
m 6 U
p
tO E — -
3 ^
c 3 -u
c ,R w
<^o
l-l
0)
AJ ,
^ *>
10 (Tl
^S
^si
cui2:
fc
ro co r~
CO O 0
co ro CN
CN CN
in o CN
CN CN CN
in U5 U3
CN CO O
r~ r- ro
cyi o en
m m a~\
LD i^D U3
rH m CN
OJ CN O^
in rH CN
r- in vn
O CTi ^3"
UD -^ «ai
CN ^ ^r
m in CN
CN UD in
m in in
CN CN "S1
r- r^ -si1
r- r~ in
r- r~ in
r-H rH n
in CM in
•^ CN ^r
0 O 0
o o o
.
o o o
JJ
-H
E
•H
rH
w
r .
fcH
Q)
£
JJ
lH
Q)
>
O
ie excess VOCs (g/L
r^
JJ
>1
£1
T3
OJ
-H
, — <
pi
>r^
f^
^3
0)
JJ
ro
^
i
0) >H
w o
a, jj
ro
rH
• 3
>, 01
CD m
> a
i-j
3 S
W H
<
>i
V-i 0)
0 .C
^-
CD £.
>ji
-1— '
C -H
M 3
ings VOC Emissions
ion in cooperation
1993.
urface Coat
gs Associat
raft Report
co C Q
•r-t
2 AJ rH
M (0 K3
< O C
O -H
T-
UM
• T3
U) C
JJ (0
X! W
Cn JJ 3
•H C O
U) -H 3
C ID rtJ
M O. U
>irH X
!H ro M
JJ C AJ
WOW
3 -H 3
T3 JJ T3
C (0 C
M S M
CD
tj
M
3
O
W
2-21
-------
reduction resulting from the fee. This quantity is estimated
as follows:
Ej = eve, • Qi (2-4)
where evcL is excess VOC content per liter for stratum I and qi
is stratum I product volume in liters. EiF is summed across
all strata selecting the fee option to compute aggregate
foregone emissions reduction as a result of the fee selection.
In Table 2-4 we see that foregone emissions reduction in
1997 amounts to roughly 883 Mg per year for the survey popula-
tion. This is approximately one-quarter of a percent of base-
line emissions, so the aggregate effect is quite small.
Dividing the avoided reformulation costs ($5.7 million) by the
foregone emissions (883 Mg) indicates that the fee option
precludes reformulation efforts that cost approximately $6,500
per Mg of emissions reduction. This is more than 20 times the
average reformulation cost per Mg indicated in the "reformula-
tion-only" scenario in Table 2-3. These results suggest that
the fee option provides a cost-effectiveness "backstop" by
avoiding reformulations that generate relatively little in the
way of emissions reduction benefits. By their existence,
however, these fees do provide an incentive for marginal
reductions in VOC content down to the limit, since the fee is
paid on excess VOC content per unit volume.
To analyze sensitivity of this decision to the fee rate,
we also examined the effect of halving the fee from the
proposed rate of $0.0045/L to $0.0022/L (1991 dollars). The
lower rate leads to a significant increase in the number of
products selecting the fee in 1997: 565 compared to 323 under
the higher rate. Interestingly, total fee revenue rises with
a rate reduction to $5.1 million. Still, only a small portion
of coating volume is affected and the foregone emissions
reduction is now over 2,308 Mg, which is almost three times as
large as with the higher fee rate, but still just over
2-22
-------
one-half percent of total baseline emissions. Avoided
reformulation costs are approximately $10.0 million,
indicating that the cost per Mg of these 2,308 Mg of emissions
foregone by paying the fee is about $4,350 per Mg.
Table 2-4 also examines the sensitivity of the results to
the cost of reformulation by doubling the assumed costs (to
$35,544 per year).9 The results are similar to those derived
when halving the fee in terms of the number of products now
opting for the fee option and the foregone emissions
reduction, but the fee revenue effect under the double-
reformulation-cost scenario is higher since the fee rate is
not reduced.
2.4 MARKET ANALYSIS
In this section, market effects of the regulatory action
are analyzed by presenting a model of the firm's decision to
reformulate, pay an exceedance fee, or withdraw the product
from the market, followed by a model of how the outcome of
this decision collectively affects aggregate supply conditions
and market outcomes. Then, operationalizing the model using
baseline market data and regulatory costs is discussed to
analyze the welfare effects of these market outcomes.
2.4.1 Potential Firm Responses and Market Effects with
Product Reformulation
In this section, the current "least-cost" model of the
reformulation/fee decision is extended to include the
possibility of product withdrawal if the cost of the least-
cost option exceeds the profits generated by the product.
Since the exit option completes the set of producer options
considered here, this discussion can be viewed as evolving
9The increase in annualized costs could reflect either an increase in
the initial cost of reformulation or a reduction in the time horizon of
benefits (from infinite to finite).
2-23
-------
from a "least-cost" to a "best-response" criterion for
decisionmaking.
A simple model is presented of a firm's decision to
either reformulate or pay the exceedance fee and remain in the
market or to do neither and exit the market. The potential
for some firms to withdraw products from the market and for
the effect of the unit fees on the marginal cost of fee-paying
producers is incorporated into the discussion of potential
market effects.
2.4.1.1 Firm-Level Model Extension. Up to this point,
the analysis has focused on firms responding to the regulation
by choosing the least-cost regulatory option. However, this
view of a producer's likely response is incomplete because the
cost of the regulatory response must be weighed against the
benefits of the action to the firm. Here the analysis equates
regulatory compliance with the decision to pay the costs and
remain in the market. Thus the benefits of the compliance
action are the net returns (revenues minus variable costs)
obtained from continuing to produce the product. The net
payoff of compliance for a particular AIM coating exceeding
the limit can be expressed as follows:
nR = P • q - c(q) - r* . (2-5)
To ease the notational burden, all terms are expressed in
their annualized form: P is product price, q is annual
output, c(q) is the cost function with respect to annual
output, and r* is the annualized cost of the least-cost option
among regulatory responses (i.e., reformulation or fee). In
other words, r* gives the cost of the solution to the least-
cost decision discussed in the previous section.
Assuming that the market for the AIM coatings product is
competitive, the price-taking firm maximizes profits by
2-24
-------
maximizing function nR with respect to its choice of q. This
is indicated by the first-order condition
dnVdq = 0 : P = dc/dq + 3r*/6q . (2-6)
Here the profit-maximizing solution equates product price with
marginal production cost (dc/dq) plus the marginal compliance
costs (dr*/dq) . Note that, if reformulation is chosen as the
least-cost compliance option, then dr*/5q = 0, because the
cost of reformulation is independent of product output level.
However, if the exceedance fee is the least-cost option, then
<3r*/6q = F, where F is the per-unit fee rate.
The solution to Equation (2-6) generates (first-order)
optimal output level and profits, qR* and nR", respectively.
However, the firm will only operate in this market if it can
cover its production costs and compliance costs; that is, if
the following condition is met:
nR" (qR* r') > 0 . (2-7)
If the condition in Equation (2-7) is not met, then the firm's
best response is to withdraw the product, produce no output
(qR* = 0) , and generate zero profits for the product (nR* = 0) .
The absolute change in the output of the regulated
producer can be approximated as follows:
Aq = qR" - q° = (6q/3p) (AP - dr*/dq*) (remains in the market)
= -q° (leaves the market) (2-8)
where (6q/dp) is the firm's usual supply response with respect
to a price change (i.e., the inverse of the marginal
production cost function) and (AP - 8r*/dq*) is the change in
"net" price for the firm, reflecting a change in the market
price after market adjustments take place less the marginal
compliance cost. Again the marginal compliance cost term is
2-25
-------
zero when reformulation is chosen and equals the per-unit fee
under the fee option. For products with VOC content below the
proposed limits (i.e., unconstrained by the regulation), the
supply response can be characterized as a special case of
Equation (2-8) with the marginal compliance costs (<3r*/dq")
equal to zero. Net changes for each producer remaining in the
market depend on whether the price increase exceeds the
marginal compliance cost. When it does, as is the case for
unconstrained producers, the output effect is positive. When
the price increase is less than the marginal compliance cost
(e.g., a relatively high per-unit fee), the output effect is
negative.
The change in market price depends on the aggregate
effects of the supply responses of the individual producers.
Product exits will shift the aggregate supply function inward,
and marginal cost effects, such as the per-unit fee, will
shift the function upward. This change can be expected to
raise the post-regulatory market price as the new equilibrium
is attained.
Appendix B describes the methodology for incorporating
the reformulation/fee/withdrawal effects into a linked
multiple-market model framework. Appendix B also presents the
methodology for measuring the social welfare effects (e.g.,
producer and consumer surplus) of the changes in market
equilibrium, which is effected by the proposed regulation.
2.4.2 Model Execution and Results
To estimate the effect of VOC content limits on AIM
coatings markets, a baseline characterization of affected
markets is constructed, empirically estimated the shifts in
market supply and demand as a result of the regulations are
computed, and the market equilibrium model is applied to the
data to generate changes in prices and quantities in each
market.
2.4.2.1 Baseline. The coatings categories are grouped
into market segments, as defined in Table 2-5.92'93 The price
2-26
-------
TABLE 2-5. AIM COATINGS MARKET SEGMENTS BASELINE DATA
No.
1
2
3
4
5
6
7
8
9
10
11
12
13
Market segment"
Exterior & high performance
solventborne coatings
Exterior & high performance
waterborne coatings
Interior solventborne
coatings
Interior waterborne
coatings
Solventborne primers &
undercoaters
Waterborne primers &
undercoaters
Solventborne clear
coatings, sealers, & stains
Waterborne clear coatings &
stains
Architectural lacquers
Wood preservativesc
Traffic marking paints
Special purpose coatings
Industrial maintenance
coatings
Totals /averages
Quantity
produced
(kL)b
162,
468,
94,
833,
61,
75,
134,
120,
40,
27,
91,
34,
231,
2,375,
937
345
935
434
298
212
678
738
Oil
449
067
568
261
933
Value
($103)
540,
1,046,
302,
1,747,
171,
160,
412,
266,
83,
493,
132,
141,
797,
6,296,
511
383
264
341
583
960
743
174
320
965
358
633
006
241
Average
price
($/L)
3
2
3
2
2
2
3
2
2
I
1
4
3
2
.32
.23
.18
.10
.80
.14
.06
.20
.08
.45
.45
.10
.45
.65
a See Appendix A for an explanation of products included in each
market segment.
b The quantities and values are taken from Census data except the
quantity for wood preservatives, which is taken from the survey.
c For wood preservatives the quantity is taken from the survey, but
the price is taken from the Census data.
Sources: U.S. Department of Commerce. Current Industrial Reports:
Paints and Allied Products, 1991. Washington, DC, U.S.
Government Printing Office. 1992.
Industry Insights. Architectural and Industrial Maintenance
Surface Coatings VOC Emissions Inventory Survey. Prepared
for the National Paint and Coatings Association in
Cooperation with the AIM Regulatory Negotiation Industry
Caucus. Final draft report. 1993.
2-27
-------
and quantity data necessary to analyze market effects are not
provided in the survey conducted for this study but are
available from the U.S. Census Bureau Current Industrial
Reports publications.94 Because the Census Bureau categorizes
AIM coatings products differently than they are classified for
this study, the market segments were constructed so that data
can be used from both sources and provide the necessary level
of resolution for market analysis. This process resulted in
the 13 market segments presented in Table 2-5. Appendix A
provides the details of this product/market cross-referencing
scheme.
Table 2-5 lists quantities and value of shipments for
each market segment. From these data, the average price for
each market is imputed. Because the market segment price is
an average value, it may obscure heterogeneity of products
within each group. Although the model aggregates different
products together to construct individual market segments, the
objective in aggregating to the market segments in Table 2-5
is to provide a level of resolution that both highlights
differences in the end use of the product (e.g., exterior
coatings versus interior coatings) and distinguishes between
groups that will be affected differently by the VOC content
regulation (e.g., solventborne versus waterborne). Eight of
the 13 segments consist of four pairs of related product
groups; one in each pair represents solventborne products and
the other represents waterborne products (e.g., interior
coatings) . Although the products in each of the paired market
segments possess different attributes, they perform similar
functions, thereby suggesting a high degree of product
substitutability in demand. Demand elasticities are estimated
using procedures outlined in Appendix A. Supply elasticities
could not be econometrically estimated because of data
limitations; therefore, the aggregate supply elasticity for
each market segment was assumed to be unitary (1.0) .
2-28
-------
2.4.2.2 Quantifying Market Shocks. To simulate the
reformulation/fee/exit decision, per-unit profits are
estimated to compare with unit costs for each stratum, and
computed as follows:
nu = P • m , (2-9)
where, P is output price and m is the profit margin. For each
product category analyzed, the average market price in Table
2-5 for the market in which the product category belongs was
used. The model derives the returns-to-fixed-factors (RFF)
profit margin as follows:
m = 1 - (variable cost/revenues) . (2-10)
The ratio of variable cost to revenue can be computed using
values provided by the NPCA.95 The variable cost component in
the numerator includes cost of goods sold plus variable
selling and storage costs. These variable costs comprise 81.7
percent of revenues for the mean producer surveyed by NPCA, so
our estimate of the RFF profit margin is 0.183.
The least-cost regulatory option for each stratum in the
survey exceeding the TOS limits is computed in the previous
section. For the market analysis, the least-cost solution
obtained previously is compared to an estimate of per-unit
profits. If the cost term exceeds the profit term, that
stratum is identified as a "withdrawal" stratum. If the
profit term exceeds the cost term and the least-cost option is
reformulation, the stratum is identified as a "reformulation"
stratum. If the profit term exceeds the cost term and the
least-cost option is the fee, the stratum is identified as a
"fee" stratum. The model computes the total quantity share of
the withdrawal strata by summing the total quantity from these
strata (Qsx) and dividing by the total baseline quantity from
all strata for that market segment in the survey (QST) - This
share was then multiplied by two-thirds to compute the market.
2-29
-------
quantity subject to the withdrawal option, which is denoted as
the term Rx.h
Rx = (QSX/QST) • (2/3) . (2-11)
Similarly, the model computes the total quantity shares for
the reformulation (R superscript) and the fee strata (F
superscript), respectively:
RR = (QSR/QST) • (2/3) (2-12)
RF = (QSF/QST) * (2/3) . (2-13)
Finally, all quantities not allocated to the exit, reformula-
tion, or fee actions can be viewed as the unconstrained share:
Ru = 1 - Rx - RR - RF . (2-14)
To perform the market and welfare effects calculations,
the initial baseline market-level values for the exiting,
reformulating, fee-paying, and unconstrained sectors are
obtained for reasons explained in the methodology description
in Appendix B. The model derives baseline quantities by
multiplying the quantity shares derived from the survey data
by the initial baseline market quantity, Q0.
Qx = Rx • Q0 (2-15)
QR = RR • Q0 (2-16)
QF = RF • Q0 (2-17)
"Multiplication by two-thirds incorporates the previously discussed
assumption that one-third of all products exceeding the limit can be
costlessly reformulated (and thus would not be withdrawn).
2-30
-------
Qu = Ru • Q0 . (2-18)
To quantify the supply effects of the per-unit fee from
the fee-paying sector, as indicated in the equilibrium model
discussion in Appendix B, the model computes a value for the
unit fee. Since the fee actually depends on the VOC content
of the product, the model computes an average fee for the fee-
paying producers in each market segment by taking a quantity-
weighted average of the stratum-specific fee for each fee-
paying stratum per market segment.
NF
F = £ F. • (QsiF/QsF) (2-19)
where Fx is the fee for fee-paying stratum I, QsiF is stratum
I's quantity, and N is the number of fee strata in the market.
Finally, note that the measure of producer surplus losses
requires an estimate of market-wide reformulation costs. The
model estimates this cost by first taking the number of
products in the reformulating strata from the survey and
multiplying it by the annualized cost of reformulation. This
calculation gives an estimate of annualized reformulation
costs for the survey population, which is a subset of total
market quantity. The model then computes the market-wide
estimate by multiplying the survey population reformulation
cost estimate by the ratio of survey quantity for that market
segment by the total market quantity indicated by the census
data. This calculation may provide an underestimate of
market-wide costs if the ratio of the number of total products
to survey products exceeds the ratio of total product volume
to survey volume. For instance, in some markets many of the
nonsurveyed products are produced by small producers and are
low volume products. In these cases, the ratio of total
products to surveyed products will likely exceed the ratio of
total volume to survey volume. The magnitude of the effect on.
2-31
-------
the cost estimates is unknown, since there are no data on the
nonsurveyed products (such as the number of products over the
limit, VOC content, and other data necessary to determine how
many products face reformulation, how many will opt for the
fee, how many will be withdrawn, and how many will be
unaffected). However, the possibility that these cost
estimates are biased downward should be considered along with
the other downward and upward biases discussed previously.
2.4.3 Market and Welfare Effects Results
To assess the potential effects of the proposed regula-
tions we considered four separate scenarios:
1. "Standard": reformulation (at $17,772 product/year
reformulation cost) or exit
2. Reformulation plus fee option:
reformulation cost = $17,772/product/year
fee rate = $0.0045/L (1991 dollars, equivalent to
$0.005 in 1995 dollars)
3. Low fee rate (one-half): $0.0025/L (with standard
reformulation cost)
4. High reformulation cost (double):
$35,554/product/year (with standard fee rate).
The first scenario reflects the impacts of the TOS and
reformulation costs, while the fee rate considered in the
regulation. The third scenario considers the impact of
setting a lower fee rate. The fourth scenario is generated to
see whether a substantial increase in the cost estimate would
have a large impact on the estimated market effects. This
scenario was modeled because, as will be seen below, the
market effects are quite small under the other scenarios.
Each of the four scenarios was modeled for the proposed TOS
limits, leading to a total of four alternatives. Tables 2-6
through 2-9 report the associated market and welfare results.
A separate estimate of welfare costs that increases the number
2-32
-------
Q
OS
j*
<•
Q
2
EH
CO
rH
O
M
OS
g
w
u
CO
1
CO
EH
.1
t— t
*3
*—>
CO
w
OS
i-3
H
D
O
s
EH
W
^M
PS
s
VD
1
CN
w
j
m
t/j.
(i) —
z
c "*•' •"""
.is c w o
g 3 -
S § a-
§ * ss
ri O w
u 01 —
•S|SS
0) 5 r*
£ | S£
5 g g S
u u
•n
il 2 ~
o -S ° o
c 2 Si «
D a $ ~
w
*T w
£ g = » -
. V- -,H U o
« 0^3;;
;2 «w * TJ v>
0) -H O
g « = &
"3. en „
5 .5 2
" i?§-o
n ^ 3 ,— I
m 7" "0
S S 2 ~
•o u a
o
a „£
C 0) *~"
.5 •- So
*J 3 rH
* X 0 W
? w °
re a
u
fl) "o
J-1 rH
O -H -v
Z >-" v>
Q. —
>,
ti j
S.fH -1
„ XJ
0) 7- "
z g o
" rH
&~
c
•.H — .
QJ t— (
Oj U cr\
DI -H en
C ^ ^
re a v>
6 ~
>i
« "3
W -rH «-J
ecu
<0 -H C o
u 1 =
Jj
c
0)
1
01
tn
ij
o>
x
re
£
O
Z
o
0
p"
m
CC
m
1
O
O
o
o\
T^
in
r~
0
m
in
CT>
CN
O
CC
v£i
n-i
0
O
O
ri
p-
U3
P-
1
CN
in
_,
U3
U3
m
i
•^T
CO
m
m
v>
n
U3
CO
pj"
U3
1-1
^
0
o
o
v>
m
p-
X
tn
!c i o>
o> e
•Jl O M
c o
^ re ja
O £ *J
-H W C
WOO)
0* «u >
X 01 O
U O. «
*-*
i^ tn ^-* ui tn co
1-1 (N m — f o o
^ fN V3 rg o co
'-O T-H f^i CC tN O
i (J\ 1 CC 1 CO
mm f-i
o o o o o o
o o o o o o
o ooo o o
\o m CN vo I-H o
vo r- P- m *» r-
o OD TT ^n cr^ o\
p- m co •& vc ^
i i i m i pj
CO P- O P- PJ 00
O i-t P- m ro cr>
in o -w 04 cr\ ^»
fN iH
rH CO O -^ •-* MS
cr\ p* 1 U3
mm m
O OOO O O
O OOO O 0
o ooo o o
»J ff\ C\ i-H "O- CTi
\o oo pj m vo xo
in m pj m p- o
1 rH (N ^T 1 O
m o o co p- co
in in CN r- m m
^o m oo o 0*1 co
r- «* *? ^ m
CD m in
mm m
•o1 m r- p- o m
*» ^ \D "^ «-H 1C
m oo 01 o ^ ^D
CM rH O 00 i-H O
(N m fN CN CN m
V> t/> V> V> V> V>
(N *? r- o m o
U5 PJ rH ^ (N CD
m o-> ^T rH fN m
CD •"» m rH m ^i
u) o~i m vo p- m
•^T CO rH
fN ^ rH \O ON CT\
o o o in o -M
0 OOO O O
o ooo o o
o ooo o o
W V> V> V> Vr tO-
CO (N oo oo m OD
rH rH t-t in T-t 0^
1 1 rH 1
t
X m
w - c
• o)0> >EO> 01 E ro -H
xure co cm one -
UD.3 •-<« w3 Uia3 3O.3 wuw
CN m ^ m vo p-
r-> p-
•W T
rs; ^?
r-i ys
( ^-
O O
O O
o o
a\ o
m o
rj o
m
a\ o
in o
PJ 0
O P-
o *»
o *»
^D
-w
o o
o o
o o
m O
T 0
rvj O
rH
'
VO P-
O *»
•^ ^D
<»
O *J
m fN
O CD
(N O
fN fN
Vr V>
rH v>
P- O
1
to
CT <-*
c re
-H U
0) iJ 3
C (0 *-"
u o n u n
50 c a> w
•H 4J 01
Ul r4 (C -H 3
01 * *J X &
m r-* n a
S u J> < •-<
CO (T>
CO CN P- P-
rH ** r- *r
P. in mm
p- •<* rH p-
m m p- in
--H \£>
O ,-, O O
0 g 0 0
o 0' oo
in PJ * co
^- CO rH 0»
m O m «-*
VD fN CO O
rH rH m m
rH
in
p. rH PJ rH
O Cft 00
rH
(N P- ID ON
IH m -
CO *» CM CO
en m m PJ
o en PJ «)
p- m p- o
(N P) in O
rH \O
1 1
0 0 O O
o o o o
o o o o
O) \O iH \D
CO rH rH O
^" m m p-
^ m o m
i i rH m
m O rH CT\
p- ^ ^D rH
m ^ o fN
rH fN m p*
P4 PJ m PJ
i •• i
rH m
i i
*» P- mo
fN ^" CO (N,
m m o in
*f *f rH ^
rH rH T m
v> v> w v>
fN U"l rH PJ
en PI oj r-
m o mo
P- rH ^ rH
PJ cn mm
PJ
O m I-H \o
\£> rH rH IO
O O rH O
0 O O O
O O O O
w w
in •*» ^f co
1 1 1 r-t
t
Oi Oi
C tn
« -H O
0) M Q.
> w u
•* «J 3 rM
we a c
A -H
> o »H n
M -H co re AJ
Q) l*-( AJ -H 0)
•O n «w c o 3
I §, S I S I
O »H fN m
rH -H »-( rH
«
(J
c
<0
c
0)
JJ
C
re
£
2-33
-------
w
H
Cn
CO
J
P-i
0
H
EH
*3<
t_3
O
fc,
w
ex;
CN
O
H
OS
1^
u
CJ
CO
1
CO
EH
*""j
CO
w
w
Q
0
S
EH
W
&
s
•
r-
i
CN
W
m
j^
EH
| B °
CD 0) ^
at ^
2
5 £ 2 2
£
c £ & »
5 |S?
c ~
•~ jj w o
Q) § r^
Ci 5 D. M
c J2 u <*
, •g
C u 2 ~
5 a
u
(0 o
o «
e-> ^-
aj
3 u J
S1S
>^
j .t< J
2 C "o
&~
a) "*
*° *H 'u 2
o a £
^,
01 u ~
O> -H J
C C U „
« -< c ^>
cr
4J
C
0
0
CO
iJ
0
V.
£
o
Z
VC
T4
s
7
m
oi
o
*~4
O
CTs
'
O
«-•
o
rH
O
CN
n
in
o
ft
m
o
r-
VO
7
CD
*-«
^
m
1
cJ
^
ro
v>
O
r-
"
^^
fN
O
O
o
r-
r-<
£
•H
O
X
u
r-(
eo S^cn S """ ° r-oorMvocN
1-1 cT\ino r- oo o "»vommm
^> OOJ.-I --( CO 1 VOVO'JVOoj
7 ^ ^ ' ' ^ ' „•
'
p- vomrM vo m m otrvo^oo
o in 1-1 f4 CD r-t m o^om^
r-i in m in r-J CN rn O WocNr-i
S 5 S S 5
r-t 04
CN m^tm m co <*-i otrifN^
o cNoor- CN m ty> or-cooo
vo i^rncn m o vo omovooo
r-t fs)
1 1
vo ^r-o m m CN ovo«HoOf-i
vo r-mtn ^ (N vo ocor-co *AV>V> W V> V> V>
CN mmm cri ^ m r-immmm
m CO^»GD nj m r- o-^oinoo
^> CD r-l r-t CN
m cNrHrH n in n ovoonvo
O rlOr- (N CN O O^r-(COr-t
0 O 0 r-l O O O OOOO.H
o ooo o o o ooooo
o ooo o o o ooooo
v> vt vt vt- v> v* v> v> v> v* «• w
oo mooin oo n in o*»
01 - c « « -S o
cue 01 C CrJ i^c-HjJ -H X4 '"m 3^-iu
co cc o coutJJCt«jJO °m'~H U£
aw was ^w i-t s was sa3 wu« su^i «t--i sa E-Q. w we
O «H CN m
fN f-i-wm vo r- oo c^rnr-i-Hr-t
p*
en
ac
7
o
f-1
r-
Cfi
"
^
CN
p,'
CC
in
^
m
CC
VO
*H
*
^^
VO
r-
m
r-
i
CN
DC
m
r-
1
VD
in
m
•^
CO
*•*
r-
r*
r^
-------
w
EH
3
w
w
fa
1
S:
0
ro
O
H
n
3
H
3
CO
lucer
u
o
a
c
o>
01
\
u
v _—
*"" O
to n ._,
^ U r-t
01 0) a,
•*-> (U t/>
0) v
2
.5 c w o
1 3 ^
o> 5 .-H
o> r a •"•
C r. u °^
* > 3 2
u g, w */>
c ~~
« §|:
c !2 u
, 7 w
c c g ~
c 2 jjj 2
M w
u)
E C 01 _~
S - 3 2
«-< (0 T)
0) ^ O _
a: 3 v.
a
c w
••I *"* — .
0 | °
?••§ «
s £ -
w
Cn w
C 0| -
'U 3 °
-^ TJ rx
X O _
a
11 "o
o Z>
H _
01 ~
S u J
CD -H ^
2 W v>
C. —
3 £ 3
2 C "o
s. ~
&
g .C .H en
— ^ t-<
u a JJ
C ^
* *J
?•?.—•
u °
c
s
i
«
in
jj
01
i
0
2
r- rv rffMfM m ,- r^ om/*i^in
in VD ^- TT a> r- en o Tfcn^r-4cri
CD oo vom— i CN i*i«-^fn •^ co *"^ rjcocof^f^
CO ^»H in_ *** C»ff»
*-« (N
2 S 5§l; S S S SSSSS
^J1 i-l OJOJCO i-l ro 1 rMI«Hi-tO
I 1 1 1 1 1 1 1 II-
T*
^
m S SS2 2 S S 3?^S^
ro ro O) ^ «-< ro IH 9 CO in ^O in y-j
^ N
fN i« OJ_tn -, O r-) OO1C°i-\^
r- CD ojinoj ^f r~ co minorjro
VD OJ(NVDr-»CTi .HOJiHi-HVO
1(11 II
o in incor- oo ^ o\ t^-r^'Hoin
f-> OJ ,H 0 CO r^ 0 o ONr^tH^ro
VD VD ^S^ ^ ^ ^J roojchmro
oj o oaOro ^-t 01 o r-riiHinm
O O OOt-l O O O OOOOrH
O O OOO O O O OOOOO
O O OOO O O O OOOOO
OJ VO COVD^ CPl «H CD OO^fNi-4
oo vo en co «*^ 01-HO
j: i o> x; i a> mm tnaiCffi c t) at ^ a
OIC 0) C C^ ki c -^ *J -1 *-< >*-• ^ (U
co cc o coji-ucuj-uonJ c * u *J e D,QC
o e 4J o £ o o^j oo jjwo ono a.) u w o u c a> i-« > u <-< ^ c
•HtiC-HUjQ-HC-Hja CWUJ3V4U C VJJ3 -H4JCI W-HOT (0 4-* (H
u w o t-< k< o> wt-i a) Q) u< Lifuv^ ai^ai s-( u *o -H 3 u "4-< iJ -H tn AJ
0>*4-i> Q)vwQ) (D> 4)4) >6dl £QJ > (0 •-* flJftjAJ XCT'OW <0 CO C
-------
o
H
EH
d
^*j
S
OH
o
w
EH
O
CJ
1
ffi
o
H
ffi
f t
"HI
o
ss
w
u
CO
1
CO
EH
^
D
CO
w
K
i_5
W
Q
O
S
EH
s
K
^C
•
CTi
1
CN
W
CQ
EH
*-* "o
c ^
• H JJf U) O
§3 ^
v "o, i~l
C « ti "*
£ o 5 S
(_> 0 V>
n
' r- CM *» CD
^ "" or^rsj r- m -H fN(N^i-HrM
o o rn o\ -q> o -T o fNoor-or-
^ r< or^iro ri oc i fia\oco^
in rH O'O i in insc^rHi^
Tf ' kO " " ' ' " '
''ii *? T
** r^ ^mrg \o o m ooomtN
O «H mr^ico »H CN m o *r Q *r \o
- -" «
o in ccmr* p- in QD pocfiin^in
V£> rH COVOO O O O COr-rHCOfN
cr^ *o r** m r^* in CN Of-r^rHO
1^1 ft oi^-rN m «> rn o% r-iinm
rH rH r-
rH m oonm (N£ CN fN ° ° ^ M P P
fl in rH ^ fN CJ
1 1 1 1 II
S 2 * 15 £ 0 ro ^ O^gSS
O 1 fN 1 fN 1 O 1 1 rH O>
rH ^| frj
rH eo o-iinc-, m co o a\(NfNcor-
in CN OO^HT, o m o ujvoi^cnCD
ri f-i rnr^tN in CD o rn'riorHr-'
«-< i i m ro m i «-H i r-
i
O r- fjcNin o os oo *cr^f-^^
^ « "^^^ °! °. ^ "^^^"^
S ^ *SS o £ N ®^£St2
in VO (N in rH CN
1 1 1 1 1
•W 0 CDrHO^ in fN ^D OOflMJO
(*! (N rHOCO rH O fN rH^^rHtf
f-t fN r-ifNfN fN rn pj oirHrH-^r^
V> V>V>V> « V)- V> V)-V>V>V>VV
2 S «^^; S « s 5SSis
fN CD W m" o' in W o (* r-" o V m
v> v>*/>v> v> v> v> ww>v>v>v>-
r- «> p. r» fH oi o in r-^McftJJ
^ rH rH «H O*l OO JJ01O OWO u O tfl O U C 0)h > O «-^ rM> 0)*WOI 0)> OlOi >EO) OtgOl >«r-t (UfljAJ *£"r'SS *B^H GJ -O^
Xft»O XOlfO CO C<0 OlnC «J*HC Or-fO) (0<-< IntO Oti tlS ^ '"c
O "-< (N ft
rH (N m*»m «3 t- CO <^rHrHrHt-t
2-36
-------
of produces projected to face reformulation is provided in
Appendix D.
In general, the price, quantity, and welfare effects are
quite small relative to baseline values. Under each scenario,
the average price increase and quantity reduction across all
market segments are less than one-tenth of a percent of the
industry average baseline price and total baseline industry
output. Estimated quantity reductions, across all AIM
coatings markets range from approximately 670,000 L/yr under
scenario 1: Standard, to over 3.6 million L/yr under the
"high-cost" reformulation scenario. Still, this latter figure
is less than two-tenths of a percent of the industry baseline
quantity. Price increases are typically well below 1 percent
of baseline price, with the exception of the solventborne
primers and undercoaters market segment, where the projected
price increase is as high as $0.034/L (1.1 percent) under the
high-cost scenario.
In some of the waterborne market segments there is a
projected increase in market quantities, as consumers
substitute away from the solventborne counterparts because of
the regulation-induced supply contraction and price increases
in those segments. While noteworthy, these increases are
quite small empirically.
The method for estimating changes in consumer and
producer welfare effects are demonstrated in Appendix B. To
summarize, the net welfare effect (social cost) of the
regulation equals the sum of consumer surplus and producer
surplus effects. Changes in consumer surplus measure losses
to consumers from higher prices and foregone consumption. The
total change in producer surplus for each scenario equals the
sum of the change in producer surplus for the exiting
producers, fee-paying producers, reformulating producers, and
unconstrained producers. Losses to exiting producers reflect
the foregone profits they would have received had they stayed
on the market. Losses to fee-paying producers measures the
2-37
-------
net effect of fee payments plus the partial offset of these
losses by the rise in price caused by the regulation. Thus,
the producer surplus loss estimates for fee-paying producers
in Table 2-6 through 2-9 are not as large as the fee payments
these producers will make. For instance, the producer surplus
loss to fee-paying producers in Table 2-7 is approximately
-$4.7 million. Actual fee payments under this scenario are
estimated at $5.2 million. The difference between these
numbers ($0.5 million) reflects the offsetting gains to these
producers through the rise in market prices. In a similar
vein, losses to reformulating producers measures the net
effect of the reformulation cost outlays and the offsetting
gains from price increases. In the standard scenario modeled
in Table 2-6, the producer losses for reformulating producers
is estimated at -$23.4 million. Actual projected
reformulation costs are $24.1 million; the balance
($0.7 million) is due to offsetting price gains accruing to
the reformulating producers.
Note that in all cases presented in Tables 2-6 through
2-9, the producer surplus effects for unconstrained producers
is positive, reflecting the fact that these producers gain the
benefits of the regulation-induced rise in price, without any
change in their cost structure caused by the regulation.
Please note, however, that the welfare gains accruing to the
unconstrained producers are transfers from coating consumers
and, as such, should not be viewed as a net welfare gain of
the regulation.
Costs tend to be distributed across parties in such a way
that reformulating, fee-paying, and exiting producers
experience welfare losses by incurring the regulatory costs
(or ceasing operation) and consumers bear welfare costs
through higher prices. Net welfare cost estimates range from
approximately $13 million under the low-fee scenario to $28
million under the high-cost scenario. Welfare gains accrue to
unconstrained producers (through higher prices) and the
2-38
-------
recipient of exceedance fee revenues, identified here as the
government sector. However, the government may redistribute
these revenues back to any of the parties affected directly by
the regulations or back to the citizenry via the Federal
Treasury. From society's perspective, the net welfare effects
of the current transfer method (AIM producers to the
government) or alternative distributions (e.g., back to AIM
producers) are zero.
Approximately 1.4 percent of all products are projected
to withdraw under scenario 2: reformulated plus fee; these
products account for less than one-tenth of 1 percent of
baseline industry output.1 Not surprisingly, the very small
products are the ones most likely to exit in light of the
relatively higher unit costs of reformulation. When a fee is
introduced, about 5.8 percent of products choose to pay the
fee instead of incurring reformulation costs or exiting, but
these products only account for about 1 percent of industry
volume. Again, the attractiveness of the fee option seems
restricted to the smaller volume products. Approximately 16.4
percent of products are projected for reformulation and the
remaining 76.4 percent of products bear no costs with the
regulation.
To focus more on the relative merits of the fee option,
model results under the first scenario are compared with the
results under the second scenario. The fee reduces the net
welfare costs of the regulation by about $6.4 million (25
percent), primarily through avoided reformulation costs and
product withdrawals. As indicated in the previous section,
these cost savings are attained with very little effect on the
'Note that the analysis is performed under the assumption that any
product exceeding the category standard whose annual profits exceed
annualized reformulation costs will reformulate when the Table of Standards
takes effect. It is possible, however, that some products may satisfy this
condition but are not (immediately) reformulated because of technological,
product quality, or capital constraints. The product withdrawals estimated
here may therefore be viewed as low-bound estimate of true short-run
product withdrawals in response to the regulation.
2-39
-------
attainment of aggregate emissions reduction targets. However,
a greater portion of the total welfare costs (29 percent) is
borne by consumers when the fee is allowed than when it is not
(12.2 percent) because of the fee's upward effect on product
prices.
2.5 SOCIAL COST-EFFECTIVENESS
The results of the market analysis can be used to compute
measures of the social cost-effectiveness of the proposed
regulation. The distinction of "social" cost-effectiveness is
made to illuminate the fact that the costs that are evaluated
are the net costs imposed on society (i.e., the net welfare
costs estimated using the market model). Earlier, when cost
per Mg of emissions reduction was evaluated, a static view was
taken, one with no product withdrawals, aggregate supply
shifts, and market price and quantity effects to consider. By
using the net welfare costs of regulatory scenarios, these
expected market effects are incorporated into the assessment
of costs in a manner consistent with microeconomic theory.
The measure of social cost-effectiveness is computed as
follows:
SCE = AWF/AE . (2-20)
AWF is the aggregate net change in welfare (i.e., social
welfare costs), summed across all markets, and AE is the
change in aggregate emissions. The social welfare cost
estimate is produced by the market model. The change in
emissions estimate needs some elaboration here. In Table 2-3,
an estimate is presented for the emissions reduction target,
but this estimate was for the survey population, a subset of
the industry. To correspond with the welfare cost estimates
generated by the market model, a national estimate of
emissions reduction must be used. The baseline estimate of
2-40
-------
national VOC emissions from regulated AIM coatings products is
approximately 481,000 Mg.3 Given the targeted reduction of
20 percent in 1997, the aggregate emissions reduction target
is approximately 96,200 Mg, which is AET. However, the
emissions target must be adjusted by two factors: foregone
emissions reduction due to selecting the fee option and
changes in emissions due to regulation-induced changes in
industry output.
The first adjustment is computed by taking the ratio of
foregone emissions reduction to emissions reduction target in
the least-cost analysis performed above. This ratio is
(AESFR/AEST) , with the numerator indicating the foregone
emissions reduction from the survey and the denominator is the
emissions reduction target from the survey. This ratio is
based on the survey population but is then multiplied by the
national emission reduction target to provide a proxy of
foregone emissions of the different fee options at the
national level.
AEFR = (AESFR/AEST) • AET. (2-21)
The second adjustment is computed by taking the ratio of
change in industry output to baseline industry output and
multiplying by baseline industry emissions:
AEQ = (AQ/Q0) • E0 . (2-22)
AQ is the change in industry output, which is the sum of
market-level changes, Q is baseline industry output (2.375
billion liters), and E0 is baseline emissions (480,816 Mg
indicated above).
3This estimates is based on a national baseline emissions estimate
provided by Radian Corporation96 of 530,000 tons, which we convert to Mg by
multiplying by the ratio of tons/Mg = 0.9072. The result is a national
estimate of 480,816 Mg.
2-41
-------
Thus, the net change in emissions reduction is computed
as follows:
AE = AET + AEFR + AEQ . (2-23)
The net change equals the sum of the targeted change, foregone
emissions change (due to fee) , and emission changes due to
changes in industry output via regulation-induced market
interactions .
The analysis focuses on computing social cost-
effectiveness measures for the four market model scenarios
outlined above. Table 2-10 presents the results.97 Social
cost-effectiveness estimates range from $139 /Mg for the low-
fee scenario to $299 /Mg for the high-cost scenario. Social
cost-effectiveness in the standard scenario is $260/Mg, while
offering a fee lowers the cost-effectiveness by $64/Mg,
reinforcing the notion previously discussed in the least-cost
analysis that introducing the fee provides significant
improvement in cost-effectiveness, yet, further improvements
are possible with a reduction in the fee rate (Scenario 3) .
As a normative policy issue, the socially preferred
outcome cannot be inferred without knowledge of the social
benefits of marginal reductions in emissions. Moving from the
suggested fee rate to the lower fee rate reduces social costs
by about $5.7 million but forgoes about 2,000 Mg of emissions
reduction. Dividing the cost savings by foregone reductions
gives the marginal social cost of actually achieving the
foregone reductions. This figure is about $2,900/Mg.
Although this is much higher than the average social cost-
effectiveness numbers reported in Table 2-10, the normative
issue is whether this action generates marginal social
benefits that exceed $2,900. If so, then society loses by
forgoing these reductions by allowing the fee option; if the
benefits are lower, society receives a net gain by forgoing
the reductions.
2-42
-------
CO
W
EH
S
^.TJ
II
EH
CO
W
CO
CO
W
£2
W
>
H
r .
U
W
fl4
fe
W
1
EH
CO
o
CJ
-]
v~j
^4
H
U
O
CO
O
rH
1
CN
W
PQ
rtj
EH
a
Z
T3
0)
E
•H
m
W
T3
d)
4J
1
•H
4J
to
W
National
1J
W
0
r— (
(0
''H
U
O
C/1
£
o
-H
w
en
•rH
£
0)
j-J
CJ
•£
»*D
E
(D
C
O
Dl
rH
O
14-H
targeted
01
rH
fO
14-1
rH
0)
4J
CU
2
0)
v
j_i
rrf
2
m
w
0)
C
CD en
> -S
•H \
U •—
CU
,
**H
CU
•D
CU
DI c;
c |P
£ —
o
„
Dl
4-1 S
D ^
a
4J 0
D '
0 -o
to
W Ol
C 2
o ~"
1/3 Dl
E its
(1) _£•
u
emissions
hanged (Mgi
u
^^
o *-'
1 ^
u t2
o
r-H '["J
o> ::
•o 2
0 £
CJ
en
o «3 a* en
i£> ci m c^
CN rH rH CN
CT* VO CTt rH
CN rH CM CO
vc in ^o ro
CTl CTt CT* O*\
1 1 1 1
l£> OO CO O
rH CN CN f~
1 1 1 1
O CM CM O
V « *
rH m n
p") po m ro
VO VD VO VC
CTi CTl CTi CTt
1 1 1 1
0 O O O
o f- r-- VD
CTI M1 in oo
kD CJ^ rH «3
«a- vo r- m
O VD CTi O
in CD CM CD
CM rH rH CM
1 1 1 1
rH CN rn ^J1
CD
^r
o
CM
O3
^^
0
O
o
CM
Dl
S
O
4->
JJ
C
01
4J
in
3
•51
10
03
CU
•H
4-J
in
C
o
"
o
o
o
o
m
in
in
C
O
• H
in
in
• H
E
0)
0)
of baselin
01
4J
10
E
• H
4J
01
01
rH
18
C
O
•H
4-1
IB
C
^^
in
o\
o\
r-t
in
*c
CU
W
CO
3
,
01
c
o
-H
in
in
-rl
E
W
U
O
T3
18
01
01
rH
10
W
0)
c
0)
en
10
n
in
D!
C
-H
jj
O
CJ
rH
18
3
4J
CJ
CU
-H
• jr,
— u
t~- M
Ol <
c«
uction in 1
mination of
'O IM
CO 0)
r-l 4-1
01
01 Q
C
o
•H
01 C
en o
•H -rH
E 4J
01 10
IH
T3 O
0) Q
4J rJ
ifl O
e cj
•H
•U C
01 10
El3 >rH
— T3
o p£
,
o ••
01
oi u
01 ,4
e 3
•r| 0
OJ W
01
c
0
•H
en
01
-rH
E
~
CO
Cb
o
o
o.
CO
D
c
0)
rH
rH
U]
*
^
0)
u
c
3
Q
O
4J
C
o
• H
(8
r4
O in
r-l CTl
O rH
CJ
Rob, Radian
November 22
* 'D
n eu
0 4J
en 18
•n Q
^
ifl C
tc o
-H
E
UH -H
Q
3 01
rD 'O
C rH
Ifl Ifl
M -D
o c
E ifl
0> 4J
S M
C
0
-H
4-J
Q
3
TD
01
tH
01
C
0
•H
en
en
-H
E
Or*
TJ
0)
01
Dl
rH
18
4J
O
4J
C
0
-H
4J
U
3
TD
0)
rH
in
C
0
>rH
01
01
• H
E
(1)
01
C
O
Ol
01
rH
o
1-f
o
0
times rati
4J
0)
D)
rH
IB
4J
0
-H
4-1
U
•o
01
V4
en
C
0
•H
01
01
-rH
E
0)
rH
Ifl
C
o
4J
18
z
JD
10
10
•o
>,
01
>
rH
3
en
O
01
-H
en
>,
rH
18
ifl
4J
01
O
U
1
u
01
10
0>
rH
E
O
V4
14-1
4J
3
O.
4J
O
4J
en
3
C
-rH
01
C
-H
rH
0)
in
18
Q
0
4J
0
•H
4J
U
3
•d
0)
u
4J
c
18
D1
4J
eu
ndustry mar
•H
14H
O
0
.rH
4J
Ifl
)H
01
01
-H
4J
01
C
0
•H
01
01
-H
E
0)
0)
c
•H
rH
01
01
18
u
4J
C
4J
in
3
•r-t
•a
Ifl
^
4J
-H
4J
c
Ifl
cr
01
Dl
c
10
0
01
c
0
Dl
01
rH
0
H-
CU
Dl
18
j:
CJ
TD
0)
4J
0)
Dl
M
10
4J
II
01
Dl
ifl
rC
u
01
C
O
-H
m
m
• H
g
Ol
u
0)
2
•D
2-43
-------
2.6 EMPLOYMENT IMPACTS
Regulation-induced reductions in industry output may lead
to corresponding reductions in AIM coatings employment.
Employment impacts are estimated by multiplying the baseline
industry employment level (L0) by the proportional change in
industry output from its baseline level:
AL = (AQ/Q0) • L0 . (2-24)
This assumes a fixed relationship between output and
employment, at least for the marginal changes considered here.
Q Q Q O
Table 2-11 presents the employment impacts results. • '
Total employment for SIC 2581 is 51,100 employees.100 The AIM
coatings sector is a subset of SIC 2581, so the AIM coatings
TABLE 2-11. ESTIMATED EMPLOYMENT EFFECTS
Regulatory
scenario
1
2
3
4
AIM coatings
output change
(103 L)
-670
-1,177
-1,225
-3,659
AIM coatings
output change3
(% of baseline)
-0.028
-0.050
-0.052
-0.154
AIM imputed
employment change3
(no. employees)
-7.4
-12.9
-13.4
-40.2
Baseline quantity and employment computations are as follows:
Industry output Industry output
from Census from Census
Sector (103 gal) (103 L) Industry employment
SIC 2581
AIM model
1,229,800
627,723
4,654,793
2,375,932
51,100 from Census
26,083 imputed from
output share
Sources: U.S. Department of Commerce. Current Industrial Reports: Paints and
Allied Products, 1991. Washington, DC, U.S. Government Printing
Office. 1992.
U.S. Department of Commerce. 1991 Annual Survey of Manufactures:
Statistics for Industry Groups and Industries. Washington, DC, U.S.
Government Printing Office. 1992.
2-44
-------
employment is computed by taking the ratio of AIM coatings
output to SIC 2581 output (same source as above) and
multiplying this by SIC 2581 employment. This produced an
estimate of approximately 26,100 employed in the AIM coatings
sector.
The proportional change in AIM coatings output is
computed by taking the ratio of the change in output from the
market model (summed across all market segments) over baseline
AIM coatings output. This computation is performed for all
four scenarios of the market model.
Given that the output change estimates in the market
model are relatively small, it follows that the estimated
employment impacts are also small. Under the standard
scenario, there is an estimated loss of approximately 7 jobs
nationwide, a 0.03 percent reduction. Employment losses range
from approximately 7 to 40 jobs under the different scenarios.
The largest effects are associated with the high-cost
scenarios, which has the largest corresponding reduction in
output.
2.7 SUMMARY
The proposed regulations impose a set of standards for
VOC content for individual AIM coatings products. Products
that exceed the limits imposed by these standards must either
be brought into compliance with the limits, have an exceedance
fee assessed on the product's VOC content above the limit, or
be withdrawn from the market. These compliance actions must
be taken by the producers of the violating products. This
leads to a reallocation of resources toward these efforts,
which imposes opportunity costs directly on the producers and
indirectly on other members of society as producers act,
markets respond, and prices and output change. The purpose of
the preceding section of this report is to characterize the
reallocation of resources and quantify them in dollar-
2-45
-------
denominated terms to provide an assessment of costs and
economic impacts of the proposed regulations.
Initially, the regulatory impacts are viewed in a very
restrictive light, assuming that reformulation down to the
proposed standards is the only option available to producers.
The aggregate costs of this restrictive option are then
computed for a sizable subset of all AIM producers comprising
the survey for this study to provide a sense of the empirical
magnitude of the proposed regulations.
The analysis is expanded by progressively shedding the
restrictive assumptions of forced reformulation. First, the
exceedance fee option is incorporated, taking into account
that producers may choose to pay an exceedance fee rather than
reformulate if this is a less costly alternative for them.
Then, the least-cost compliance option (fee or reformulation)
is compared with benefit streams (net revenues) to determine
if the least-cost option is also profitable. If the value of
the benefit stream is less than the cost of compliance, firms
are assumed to remove the products from the market as a best-
response strategy. The collective effect of some producers
removing unprofitable products and some producers bearing a
per-unit fee on output will contract the aggregate supply
function and lead to changes in market prices and quantities.
The optimal best-response actions and resulting market
outcomes will determine how the welfare costs of the policy
are distributed across producer groups, consumers, and the
government sector.
Several scenarios are modeled for the proposed standards.
In general, market model results indicate very small change in
baseline market conditions as a result of the proposed
regulations. This derives from the empirical expectation that
aggregate costs of the regulation are a small share of
aggregate industry costs. However, because there is a high
degree of producer heterogeneity within the AIM coatings
sector, the costs for some producers may be empirically large.
2-46
-------
Of particular concern is the potential impact of the
regulations on small producers, which is the topic of
Section 3.
The analysis does demonstrate the potential for substan-
tial cost savings due to adopting the fee alternative and how
this cost savings is likely to accrue especially to producers
of small volume products. Moreover, this cost savings is not
expected to have a significant impact on undercutting aggre-
gate emission reduction targets. Two fee options were
considered, $0.0045 (1991 dollars) per g/L over the limit and
a rate half that amount ($0.0022). The lower fee is projected
to cause a substantial increase in the adoption of the fee
rate alternative with significant cost savings to the affected
producers and a negligible increase in foregone emissions
reduction. The relatively large cost savings with small
attendant effect on emissions reduction targets may make the
lower fee rate warrant further consideration by EPA.
2-47
-------
SECTION 3
REGULATORY FLEXIBILITY ANALYSIS
3.1 BACKGROUND, AFFECTED ENTITIES, AND REGULATORY REQUIREMENTS
The Regulatory Flexibility Act (RFA) of 1980 requires
special consideration of the effect of federal regulations on
small entities. The RFA requires that federal agencies consider
whether the regulations they develop will affect small entities
including small governmental jurisdictions, small businesses, and
small nonprofit organizations.101 Under the 1992 revised EPA
guidelines for implementing the RFA, an initial regulatory
flexibility analysis (IRFA) and a final regulatory flexibility
analysis (FRFA) must be performed for every rule subject to the
Act that will have any economic impact, however small, on any
small entities that are subject to the rule, however few, even
though EPA may not be legally required to do so. The severity of
the rule on small entities may be measured once the small
entities are defined.
Small entities may be defined using the criteria prescribed
in the RFA or some other criteria identified by EPA. The SBA's
general size standard definitions for Standard Industrial
Classification (SIC) codes is one way to define small businesses.
These size standards are presented either by number of employees
or by annual receipt levels, depending on the SIC code. For SIC
2851, Paint and Allied Products, the SBA defines small business
as fewer than 500 employees. Because the coatings manufacturing
3-1
-------
industry is not labor intensive, use of this SBA definition would
result in, almost all firms in the AIM coatings industry being
classified as small. Alternatively, based on input from the
regulatory negotiation process, EPA has defined small businesses
as having less than $10 million in annual AIM coating sales and
less than $50 million in total annual sales of all products.
U.S. industries are composed of businesses ranging in size
from small independently owned single-facility firms to large
corporations. Environmental regulations affect all businesses,
large and small, but small businesses may have special problems
in complying with such regulations. Therefore, this analysis
specifically addresses the RFA requirements by measuring the
impacts on small entities of regulating AIM coatings
manufacturers. This analysis focuses on small company impacts.
3.1.1 Potentially Affected Entities
A regulatory action to reduce VOC emissions from AIM
coatings products will potentially affect the business entities
that own the regulated facilities. Facilities, or
establishments, comprise a site of land with plant and equipment
that combine inputs (raw materials, energy, and labor) to produce
outputs (AIM coatings). Firms, or companies, that own these
facilities are legal business entities that have the capacity to
conduct business transactions and make business decisions that
affect the facility. Figure 3-1 shows the chain of ownership may
be as simple as one facility owned by one company or as complex
as multiple facilities owned by subsidiary companies.
Potentially affected firms include entities that own
facilities that manufacture AIM coatings. Determining the total
number of facilities and firms that will be affected by the
regulation is difficult because most of the available Census data
are reported at the four-digit SIC code, and AIM coatings
manufacturers, for whom this regulation applies, are a subset of
the entire coatings industry represented by SIC 2851. The 1987
Census of Manufactures, Industry Series: Paint and Allied
3-2
-------
Parent Company
Parent Company
i
Parent Company
(Direct Owner)
Other Companies
or Legal Entities
I
Subsidiary
Company
(Direct Owner)
1
Subsidiary
Company
(Direct Owner)
Facility
I
Facility
Facility
ABC
Figure 3-1. Chain of ownership.
Products identified 530 companies with shipments of $100,000 or
more that manufacture architectural and special purpose
coatings.9'102 Data from the Architectural and Industrial Surface
Coatings VOC Emissions Inventory Survey (the survey) conducted
for this study were provided by 116 firms, 36 of which identified
themselves as having under $10 million in annual net sales.b'103
While small businesses represent about 31 percent of the firms in
the survey, a larger share of nonsurveyed firms appear to fall in
"These are the two Census categories within SIC 2851 where most of the
AIM coatings products are represented, and this figure includes companies that
produce AIM products, whether or not it is their primary product.
bTwelve survey respondents did not indicate company size.
3-3
-------
the small business category.0
3.1.2 Regulatory Requirements
The proposed regulatory structure imposes a TOS establishing
VOC content limits for AIM coating product categories. The
standards must be met by all products produced after 1997 .
As discussed in Section 2, the proposed regulation
constrains firms that produce AIM coatings products over the
content limits in one of three ways:
forces them to produce products with VOC content under
the established set of limits,
• forces them to withdraw the product from the market, or
• may provide an option to pay a fee on each unit of
product they produce that exceeds the limits established
in the regulation.
Thus, firms with a heavy (baseline) concentration of products
above the limit for their respective product categories are more
tightly constrained by the regulation than those with a lighter
concentration of above-limit products, all else equal.
3.2 ANALYSIS
*
3.2.1 Baseline Market Presence of Small AIM Coatings Producers
Small business presence in specific coatings markets
indicates one dimension of how small firms may be affected by the
regulation. For certain product markets, small businesses
predominate and thus may be disproportionately affected if limits
are particularly restrictive on those categories. Table 3-1
cThe 116 survey respondents comprise about one-fifth of the firms making
AIM coatings products but account for about three-fourths of industry output.
Thus the nonsurveyed firms are relatively numerous but produce relatively
little volume.
3-4
-------
Ifl
w
a!
ff.
*?
**
cn
CD
s
H
P_4
fi
O
U
s
h- i
X
EH
s
H
U
s
w
cn
cu
tn
W
3
M
cn
CQ
,J
^•<
^•^
CO
•
J—^
1
ro
W
•J
§
£•*
-
ij
QJ
"D
C
O
QJ
1-*
n
—
cn
U]
0)
c
• »H
W
J3
_,
2
0"j
01
E ^j
.E O
n 3 --
2 T3 J
? 0
11 0,"°
HI j_, —
< a
° rH ^
(D ^ ~ —
i« O X — '
nj ^j ^
u) a
a "
^3 ifc-i -^
^ ^ O
2 ti
a
"*" *
*s s ^
4J ?• —1
3 ? v>
0- 5,
M £"0
"^
c
e
O _j "
- c
^
ir- C r' ' — '
5 i
C
tr.
IL-,
^ 0
tfi —
c
0
(C W
o c
a a
^ ^
3
rH
i-H rH
(0 —
O
E-1
&i o3
10 5 >J 3
h 3 (D Ti-
ll 7! a o o
> £ ^ "!
< •* c. —
Ul
U XJ
(D U
6 O "D
2 *-<
a
o
T3 w
j-i 3 7-, ~*
-3 C c^
Q, 11 C ^
E > -1
M 1* *^
U
J
U]
11^
r-1 O
IB —I
t/3 —
0 >,
jj n
(0 M
rH d)
QJ (J
•*-> u
Q) ID _^
ID D 2
S W ~
<
Z
0
0
0
rH
rH
rt;
-*.
2
0
o
o
OS
2
2
rH
,«£
2
a
2
U)
i-l
-V
fO
0)
,Q
-o
c
o
ca
CN
rH
O OCN rHf-- r^i^ cn
O fNCN U3O On UD
mm m rn fN ro cn
CT>
o r-^1 r-^r >x>o o
O OV£> <>DrH rHCN in
CM rH CN m CN rH CN
0
o rHt"- fNin r^co o
co co cn m m p- CN
o mcN ^ro CNO vo
rH cn n ^D co co ^
m X> CO rH UD CO O
rH in
o cn co rH in co m
*C i— (O r-* ro fNfN CN
"~" fN
,—, ^CN oin On m
rH o rN UD co o cn
£-. CNCO ror^ r^'.D
^ * cn
<3< CN CO ^ CN CN rH
*^> co r~ rH m r- cn
^ ^co cccn r^irH co
« CN o^ ^r o r~- ^ *-^
<~lin LTlm rHrH f"1
rH rH
fN rHCn ^3*^ ^*fN O
m in rn ry H3 v£)
< fNr- cno ^ro m
~^. ocn cncn rnrN co
2 mo r~»cN oco ^
no n n co co ^
r^ o r^ co vo I-H CN
cn r- T n fN ro (N
co VD ** in o
cn co ^o o in CN in
*3- r- r- r- co <-H n
^r tn
Di O^D COCN Ch- ^
2 ^rH COLO OCN cn
^T n rH rH «H Cn CO
V)
Di V-* ^
coo n
•H -H -H t>
JJ V-i VH (L)C
03 0) (U rH U -H
Q) O AJ AJ CO C AJ
AJ(JC -MX ->H -H(OWfO
• H -H AJO
in
o
in
^"
*3<
0
in
CN
rn
fN
n
CO
**
CO
cn
rH
rH
in
O
cn
I**
-o
U] IQ
D
O W
C Dl W
-H C U
E -H -«H
3 AJ AJ
iJ (0 W
•H O CO
co o e
fN
rH
^r in
ri n
n
CO CO
CO CO
CO rH
rH rH
rH n
0 O
\& rH
^
rH rH
rH O
o °^
\£> ^
co n.
co r^
in r^
tN n
in in
0 CN
CN rH
r-i ^O
rH <3"
rH l/l
^ CN
rH rH
r- co
0 rH
rn r-
U"l ^f
0)
1
d a
m O
P
- (TJ
W V«| Ul
c AJ q
-H .H -H
•3 g rO
JJ QJ AJ
X CO
r- r*
o r- rN
Cn fN U3
\£J CN M3
fN
c** r^ n
^ cn CN
CM rH CN
rH ^T CD
fN rH ^T
tN ^O fN
cn ^» in
rH CO ^3-
CO CO
O Cn rH
fN O CN
rH rH
cn r-
^ r-
\O CO rH
» i-H »
in n n
vo r** CD
t-H VD rH
in CN CN
in in rH
in rH in
CO r- rH
CN
in ^ cn
co ^o ty
cn ^ vo
r** cn cn
vo cn ^o
\D n : cai
JH -H rH
(D UH O
E O W
O -
^ CO CO
—
T,.
C
j^
c
0-
C
O
o
.
j*
*»
s
^4
Q
•o
c
4J
Q
n
Q
o
C
&
HI
CO
3-5
-------
Q
W
J3
EH
O
U
«
W
si
W
o
H
EH
0
u
s
M
^t
W
K
EH
a
H
W
O
£5
W
W
0,
C/}
C/}
r^
M
r«
D
m
Jq
*i
cn
rH
i
PO
U
j
m
EH
w
^
•c
c
o
HI
W
VC
"~*
U)
H
41
C
W
,-,
10
£
t/j-
c
0
•rl i
Jj —
10 M
rH *J
3 c
a -I
QJ W
> U
3
0) ^>
'H
rH »~<
TO -"
U
o
u
£|*3
10 g O
v "o a°
* " £-
a
i" in
O ,_, u
« u
S * ^*
o £ o ~
£ 1-
w a
o u
J2 >*j -
e o -o
3 0
2 ti
a
•a ™ £
If*
M P 0
W rH
0)
IfcJ w
0 In v
E
<" 3 «
^ (—1
O
^
~
vn *~^
a —
^i jj
HI O
"e 'o -o
3 O
2 ^
a
0 •"*
•o w S
d) 11) l— ,
3 c v>
Q, 0)
e >-
tH 0) 0
—
W ^
rH O
10 rH
c
o >
u O
10 Dl
rH 0)
3 J->
Di (0
01 U
a.
jj r" ^
1 if
tc QJ r:
S w G
co m (N en o
r~ v^> [•-- co o
T rn rn
rH "31 O O O
CT\ CN o en o
rn r-* CN
in (N IH
CO TJ CO CN O
en VD rH r-i
[^ m rH en
en c^
co CO m m o
o o in CN o
r-* CN
O ^j
CTi CO
m in m o o
«g" -in
CN iH ^T m
«H ^" cn co r~\
r*~ o vo in o
rH rH fN rH O
rH
co VD in o in
CN CO ^D O rH
rH rH rH
CO ^" O CD f-i
^D r^ cn en ^3"
CO O3 O CO CO
O CO VD O -H
O"i ^ n ^3" OO
m f*i rH in ^
vo ro in r- ^>
rH (N m m CD
CN cn r^- tn in
to
0)
U TJ-H
T) C O J->
U Q) W 03 Ul O CQ
-HiJDi £ v-t w 3>
rH C C M 0) l-f ^
rHClt-H |Q 3 AJ O
W r-l - CO O -
o c ^ u a v-i
y-i -H ,y a) v-i MO)
AJ U 6 Q> OJrH
>, CC -H -H T3 AJ (0
Ui O 3 H C CD 0)
Q U O Q43 3 W
« * "
1-1 in r-
o in r^ o
o cn m o
CN
o m co in
0 ^ m vli
rH IN rH
O
O t^ O T
rH rH
O
O CN ^ O
U) ff\ c^
rH O 00
ai
O CTi C^ 'J
O O 0 T
o r-
O VO UD O
V£> - 00
(N CM
CvJ O ^3" IN
co in ri in
CN rH rH
(N m fN \Tt
&\ ro O rn
O rH ro fN
cn m PJ m
rv] c^ rs) o
rH rH OJ IN
IN \o m o
o cn *r a\
-
V
m ^
ft) W 3
> M Jj
a -H ai x
ui u u m >
c ra 10 w ui ^
10 > o ai •o «
H t4 U O C rH
U 0) IH -H .H .* QJ
•HT30J 01 4JJJ UE
EOO) T3 wiO -nra
OJOM c 100 3c
W3ft S SO OH)
0 "1 m ro
1-1 in ^
o o
o o
o o
0 0
0 0
o o
CD O
o o
0 0
0 0
o o
0 0
c^ r-
t~ rH
u-i r-
rH CO
r-H
r- r-j
o m
OJ *!
rH
r~ ^J
CA O\
-
r4
10
HI 01
rH 11) 0)
0 C T! >
U O'rt
O O 4J
M ma s ro
O 3 JJ >
10 crc ai u
rH 10 H> 3 i1
rH a > CT01
0) OrH 10 01
ff. O Q< ^
cn j; w o a
o
—
QJ
C
TJ
0
u
V
n
a
UJ
Q
•o
c
10
10
^J
o
c
(U
01
U)
3-6
-------
____ ^
Q
W
^
2
M
EH
O
0
tH
W
H^
^g^
CO
0
n
in
to
0>
c
U)
3
JJ
,
e
c
O
1C w
"3 c
0 T3
O
>, O.
01 jjl
5) *o
'"'
10 —
o
E-
QJ - J-* —
tT> = U J
t i-i D
^ ,£ CJ T3-
O - C. O °
> ^ U «~<
< a
'— CO
c ^ t*
^ /i 3 -j
S ° 0
r- J-1 ^
OQ a
U)
3 u
ja iu 3
£ 0 Tl
3 0
2 ^
Q.
T3 W cr!
4) 2 m
^ C *H
E "•
M " 0
Ul
t! -S
m £ .
r^ " O
M 0
fl ""*
^ o
1C rt
0) —
& ^ U J
1C § ^ 3
k< _- 41 tJ-
0) "^ D. 0 0
> ^ 1-4 *"*
< a -"
W
VJ 4J
QJ U
n y_j "j
1 O 73
3 O
2 *-<
DJ
" rt
0) 4| ^
3 C '^
cx ^
M 41*0
^1 »H
tn ^
OJ-
r-( 0
CO rH
CO —
C
0 >,
•u o
1C Dl
.-< OJ
W 10
0) U
en
u
III
oD fM i— ( O O i-Hin
rH rH u3fNLn cNr-r^
o o o o o o o
coor—i OO^^TO o\ mvc
CN T-I
oo^c^o ^ "^
CN 2 o in <^> *H
< „ ^ ^ OO Or*")r'"iCN >c -i 01 o i- us r- u; r~ rt
*-DmvjD *& *f CQ \D r^ m in CM
?3fN S^SSS S CM"^
CN rH rH rH
rH W 01
0) W iJWOTJO) O-H
W iJ CV^O C> 01 OJ->
O WCDWtO EE O'*-1 3 (rjtw
rH W Q> *D rH *O U} J-) - *-iC
-H OJ -H -H V4 C O O £ -H H E -H J 0) 0)-H t-f O»*H
t3rH iJjJ U'Hfi E£X Q*'*-' AJ£ C*-" O'OW .C Q< J-* 3 *«H ^J
CfO rHfQ C^nE *H£ CO(D Qjy) -
1— t
CQ
i| j
o
•D
C
0>
1C
m
41
O
C
01
01
en
3-7
-------
Q
W
O
U
H
«
OS
CO
§
O
u
g
H
H
EC
EH
U
a
w
&
CO
CO
J
•J
CO
m
W
J
I
c
10
X!
JJ
tn
in
0)
rH
M
O
W
JJ
o
3
T3
o
JH
ft
in
Dl
c
•H
JJ
(0
o
u
52
M
<
c
-rj
c
O
• H
t — 1
rH
•H
E
O
rH
c
<0
£
JJ
in
tn
0)
rH
DI
g
•rH
O
3
T>
O
u
a
tn
(0
73
0>
c
•rH
4H
0
•0
4)
In
(0
tn
(U
in
tn
0)
c
•H
in
3
.5
rH
rH
C
W
ro
U
0)
rH
10
tn
rH
to
jj
O
jj
c
•H
c
o
•H
rH
rH
•H
O
in
C
•H
C
o
•rH
rH
rH
•rH
6
0
rH
M
0)
13
3
Dl
C
•H
>
(0
XI
tn
(0
tn 01
O) N
> -H
rH t/1
01
in >,
E c
01 (0
xi a
jj e
o
13 U
0)
•H JJ
MH tH
•rl O
•u a
C 0
0) tH
T)
•H JJ
o
Oi C
in
O 13
rG -H
JJ 15
m in
O JJ
C
vo 0)
m "D
C
T) O
c a
iO tn
0>
tn tn
JJ
C >i
01 0)
•0 >
C M
O 3
ft in
in
0) 0)
tH >
rH
vo 01
rH %
rH &
tn
(0 •
XI tn
0)
>rH
01 (0
> tn
>H
3 rH
in 10
3
0) C
XI C
EH (0
.a
01
XI
JJ
c
4)
x:
3
T3
0)
tn
3
tn
•rH
0)
Dl
10
tH
0)
>
(0
•a
01
jj
Xi
D)
•H
0)
3
<
CN
C
O
• H
JJ
U
0)
w
E
0
tH
V4H
c
Q)
J4
10
JJ
in
0)
u
-rH
rH
a
O)
Dl
(0
tH
0)
>
(0
Dl
C
•rl
in
3
13
01
JJ
3
I
•rH
Q)
tH
(0
in
0)
3
0)
>
0)
Oi
u
in
jj
C
0)
O>
0)
in
jj
QJ
^
tH
(0
E
O
3
jj
C
•rH
in
D)
c
o
rH
4)
X)
rT
O
Dl
41
JJ
<0
U
jj
U
3
13
O
tH
ft
0)
XI
01 JJ
in tn =
•H (0 C C •
•H 3 tn
- XI O -H
j u i; c in
•H O) .* >
r- x; tn c rH
r- 3 3 3 m
vo = C
- - > 10
in in jJ in
m rH -n O JJ
O) JJ 01
- E C = X
01 10 10 JJ M
IH c 3 a 10
4) 0) D1 E B
XI 0)
>, >, X 41
0) >H O) 01 X!
U 13 > = JJ
C in
0) ,X 3 in tH
tH U W 0) O
01 -H XI tn
H 0) C
m C D) 3 "->
in -H
00 JJ >i JJ JJ
- (0 rH (0 O
rH O JJ XI C
u x; jj
in DI a>
•H JJ -H >i >H
C rH QJ 0)
>i 0) in > 3
0) E tH
> 01 in 3 ~
tH U -H W 0)
3 C
tn 01 0) 0) tn
jj tn x: o
oi -H o) jj xi
x: in x: M
JJ 0) CO)
C "O -H JJ
C Dl 0) iO
•rH (0 jj tn 3
E -H 01
T! O -H tn
0 13 ft JJ O
JJ C 0) --H
tH (0 tH JJ 4)
o c c
a tn >,
rH Xi tT 10 rH
o u o
> 13 4) 0 W
c xi xi
rH O EH O
iO XI in JJ
JJ -H
o o • tn JJ
JJ JJ T! >i U
0 rH 0
TH TJ jj
•H
T!
c
•H
<4~l
o
0
tH
3
in
o
rH
u
in
• H
TJ
o
JJ
0
T3
T3
0
jj
tH
O
a
0
tH
jj
o
z
II
OS
s
0
rH
XI
(0
rH
•rH
10
>
10
JJ
0
z
ii
<
\
s
3
0
rG
JJ
W rC
C JJ
O -H
-H 3
in
in C
•rH O
E -H
W JJ
(0
U tH
O 0
> a
0
in o
Dl U
C
•H C
JJ - rH
C (0
rH (0 C
(0 -H
• rH JJ Itr
tH C
Jj -H
in (0 •
3 Cu in
13 3
C rH U
H (0 3
C 10
13 0 U
C -H
(0 JJ >i
(0 tH
rH 2 JJ
(0 in
tH tH 3
3 O t?
JJ MH C
U H
0 T)
JJ 0 C
•rH tH O
rG (0 -H
U ft JJ
tH 0 (0
< tH -H
cm JJ
o
Dl
in • 0
JJ >i 2
j- 0)
01 > >i
•H M tH
W P O
c cn jj
H (0
>T-I
>1 tH 3
tH O Ol
JJ JJ 0
w C OS
3 0
13 > S
C C H
H H <
0
U
tH
3
o
W
-8
-------
lists the coatings product categories provided in the survey.104
The survey data represent producers that account for
approximately three-quarters total industry product volume.d
Small producers produce more than 20 percent of the products
in the survey, but these products account for just 3.6 percent of
total coatings volume and 3.7 percent of total revenue. The
average price per product in the small business segment is
$2.52/L, compared to $2.44/L for the industry. The largest
volume category for small producers is roof coatings, at 19.9
million L/yr. Small producers comprise just over 22 percent of
the volume in that category. Small businesses produce over 95
percent of the total volume of antigraffiti coatings, but the
volume is quite low, with six products totaling about 40,060 L.
Other categories in which small producers comprise more than
20 percent of the market volume are lacquers, mastic texture
coatings, graphic arts coatings, bond breakers, and
appurtenances. In addition to roof coatings, small producers
collectively produce over 4 million L in the following
categories: traffic marking paints, exterior nonflats,
bituminous coatings, lacquers, and interior flats.
The extent to which small businesses are affected by the
regulation to reduce VOC emissions from AIM coatings will depend
partly on the average VOC content of small business products
relative to the industry average. Table 3-2 presents the average
baseline VOC content for products manufactured by small busi-
nesses as compared with those manufactured by the industry as a
whole.105 Small business products generate approximately 6.2
percent of total VOC emissions in the survey, which is substan-
tially greater than their output share. The average VOC content
for small business products, 325 g/L, is almost 75 percent higher
is based on the ratio of Census product volume (part of the total
SIC 2851 volume) to the survey product volume.
3-9
-------
TABLE 3-2. BASELINE VOC CONTENT
Size
category*
All products
Small business
products
VOC emissions
(Mg)
344,059
21,431
Sales
(kL)
1,853, 623
65,914
Average VOC
content
(g/L)
186
325
a The survey has 116 respondents and 36 of those identified themselves as
having under $10 million in annual sales. Twelve survey respondents
did not report company size.
Source: Industry Insights. Architectural and Industrial Maintenance
Surface Coatings VOC Emissions Inventory Survey. Prepared for
National Paint and Coatings Association in cooperation with the
AIM Regulatory Negotiation Industry Caucus. Final Draft Report.
1993.
than the average VOC content for the industry, 186 g/L. This
suggests the potential for disproportionate impacts on small
businesses unless the regulatory approach were designed to
mitigate these impacts.
Small business products have a higher VOC content than the
industry average for two possible reasons. First, small
businesses specialize in products that tend to be higher in VOCs
because of fundamental performance requirements of the products.
Second, small businesses tend to produce higher VOC-content
products regardless of the product category. The first reason
can be termed a specialization effect and the second reason a
.technology effect.
The observed difference in average VOC content of small
businesses and all products as separated into the specialization
and technology effects using a simple procedure. First, a
measure of the projected average VOC content of small business
products is computed based entirely on the distribution of small
3-10
-------
business products among product groups with different average VOC
contents. This is a measure of its specialization-based VOC
content:
N
Vs = £ V^ • SiB . (3-D
i=l
Here, V^ is the average VOC content for all products in product
category I, StB is the share of total small business product
quantity attributable to product category I, and N is the total
number of product categories.6 The separation of the average VOC
content difference into the two component effects derives from
the following equation:
(VB - V1) = (VB - Vs) + (Vs - V1) (3-2)
Difference = Technology + Specialization
in Average Effect Effect
Content
Vs and V1 are, respectively, the small business and industry-wide
VOC content averages. The specialization term quantifies the
difference between the specialization-adjusted average for small
businesses and the overall industry average; the technology term
quantifies the difference between the specialization-adjusted
average and the actual average VOC content for small businesses.
Table 3-3 yields the computation of the Vs measure for the
small business products in the survey.106 The computed Vs value
is 261, meaning that one would expect an average VOC content of
261 g/L for the small business sector, based purely on the way
their products are distributed among product groups (i.e., their
specialization). Placing this value into Equation (3-2), along
with the values for VB and V1 given above (325 and 186), the
breakdown is computed as follows:
estB is not the small business share of total production in category I,
but rather the contribution of category I to total small business production.
3-11
-------
TABLE 3-3.
SPECIALIZATION-BASED AVERAGE VOC CONTENT:
SMALL BUSINESS PRODUCTS'"
Market
segment
number
12
12
1,2
11
1,2
1,2
9
3,4
1,2
7,8
3,4
5,6
13
13
12
7,8
7,8
7,8
3
12
7
13
1,2
Regulation
category
Bond breakers
Concrete curing
compounds
Roof coatings
Traffic marking
paints
Nonf lat,
exterior
Bituminous
coatings and
mastics
Lacquers
Flat, interior
Flat, exterior
Varnishes
Nonf lat ,
interior
Primers
Mastic texture
coatings
Industrial
maintenance
coatings
Metallic
pigmented
coatings
Stains ,
semi transparent
Sealers
Waterproofing
sealers, clear
Quick dry
enamels
Graphic arts
coatings
Shellacs, clear
Sc opaque
solventborne
Apurtenances
High performance
All
products
average VOC
(g/L)
N/A
621
239
369
173
23
657
52
79
474
134
172
146
374
459
475
312
632
461
366
539
411
335
Share of total
small business
volume
N/A
N/A
0.3025
0.0857
0.0723
0.0675
0.0665
0.0639
0.0504
0.0482
0.0425
0.0422
0.0400
0.0395
0.0363
0.0091
0.0053
0.0048
0.0042
0.0038
0.0032
0.0030
0.0022
Share -weighted
content factor
(q/L)
N/A
N/A
72.20
31.66
12.49
1.54
43.72
3.30
3.99
22.84
5.71
7.23
5.85
14.78
16.66
4.34
1.66
3.05
1.96
1.40
1.72
1.25
0.74
(continued)
3-12
-------
TABLE 3-3. SPECIALIZATION-BASED AVERAGE VOC CONTENT:
SMALL BUSINESS PRODUCTS3 (CONTINUED)
Market
segment
number
12
13
5,6
12
12
7,8
7,8
12
13
10
10
10
10
12
12
13
13
5,6
Regulation
category
Swimming pool
coatings
Sanding sealers
Undercoaters
Dry fog coatings
Antigraf f iti
coatings
Stains, opaque
Waterproofing
sealers , opaque
Pre treatment
wash primers
High-
temperature
coatings
Eelow ground
wood
preservatives
Clear wood
preservatives
Opaque wood
preservatives
Semi transparent
wood
preservatives
Form release
compounds
Multicolor
coatings
Fire-res is tent/r
etardant
coatings
Magnesite cement
coatings
Quick dry
primers ,
undercoaters
Sums / averages
All
products
average VOC
(g/L)
552
525
206
300
397
257
239
706
561
541
419
362
548
599
321
16
N/A
439
Share of total
small business
volume
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
1.
0019
0012
0010
0010
0006
0006
0003
0002
0001
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
Share -weigh ted
content factor
(g/L)
1.06
0.64
0.21
0.29
0.24
0.15
0.06
0.12
0.04
0.00
0.00
OfOO
0.00
0.00
0.00
0.00
N/A
0.00
260. 87b
° Small businesses are defined as producing less than $10 million in AIM coatings
products or less than $50 million in total sales.
Specialized average VOC content equals the sum of share-weighted content
factors.
N/A = Not available
Source: Industry Insights. Architectural and Industrial Maintenance Surface
Coatings VOC Emissions Inventory Survey. Prepared for National Paint
and Coatings Association in cooperation with the AIM Regulatory
Negotiation Industry Caucus. Final Draft Report. 1993.
3-13
-------
(VB - V1) = (VB - Vs) + (Vs - V1)
(325-186) = (325-261) + (261-186)
139 =64 +75
Approximately 54 percent of the 139 g/L difference between
the small business sector VOC content average and the industry-
wide average can be attributed to greater specialization in high
VOC product categories (specialization effect), and the remaining
46 percent can be attributed to the disproportionate presence of
small business producers in the higher-VOC products of the
respective product categories (technology effect).
This finding has implications for the feasibility of
designing a TOS to minimize small business impacts. Since small
business producers are somewhat concentrated in the higher VOC
categories, as indicated by the empirically sizable
specialization effect, the TOS can be designed to be somewhat
less restrictive in categories with high small business presence.
However, the effectiveness of such an approach in mitigating
small business impacts will be limited by the fact that small
business producers are also concentrated in the high-VOC range of
each product category.
In 1993, the National Paint and Coatings Association
performed an analysis of the then-current 1997 Table of Standards
(TOS) and found that the projected emissions reduction from the
small business sector would be 19.65 percent of baseline
emissions, compared to a projected 25 percent reduction for the
industry.107 This estimate provides some evidence of relief for
small business products under the standards proposed at the time.
Unfortunately, data were not available to recompute these
estimates based on the current proposed TOS to see whether the
proportional reduction from the small business sector is still
less than the overall target of 20 percent.
3-14
-------
3.2.2 Costs Associated With Regulatory Compliance
As discussed in Section 2, compliance options include
product reformulation and the optional payment of an exceedance
fee. Depending on the technical difficulty to reformulate the
current formula, the reformulation may be considered "major" or
"minor." A major reformulation requires the following level of
effort:
2 to 3 scientist years and
1.5 to 5.5 years elapsed time
at an estimated cost of $250,000 per reformulation over 3
years.108 This level of effort was converted to an annualized
cost of $17,772 (in 1991 dollars)/ Fee payments made for
products that exceed the limit is an alternative compliance
mechanism under consideration for the proposed rule. The per-
unit fee is computed as follows:
fee = (VOC content - VOC limit) • rate. (3-3)
VOC content is measured in grams per liter, and the fee rate is
paid on the grams per liter in excess of the limit. The proposed
fee rate is $0.005 per excess g/L (1995 dollars). Total fee
payment per product simply equals the per-liter fee multiplied by
total liters of production.
3.2.3 Reformulation Cost Impact Estimates
Given the data from the survey and the 1997 TOS limits, the
number of products produced by small businesses that exceed the
proposed VOC limits are identified. The number of potential
reformulations is estimated by applying the TOS limits from
Section 2 to the number of products reported by category and VOC
content in the survey to determine the number exceeding the limit
for each category. Results are reported in Table 3-4. An
fDetails of the derivation of these estimates are presented in Section 2
of this report.
3-15
-------
It!
P-
rH
-
CO
o
o
s
o
H
EH
J
B
2
o
W
a
CO
CO
u
M
CO
m
t "i
f^i
BE!
CO
*tf
1
m
W
J
f*4
<
EH
-- M
o e c "
~ -H C "^
U3 "4-1 ^ ^ ~—
O (U J~ ?
O l-< -j ^,
U-l 1 ~-.
o E c
^i O '~~
AJ O-H tO
O 01 ra
U U--H
3
«>" i CTC 2
N o E C-H 7;
•H 5 .H ra Si ~
rH j_) Q jj ^| •* vy
nj (o IM fl JJ o *~*
3 O 0>rH W g
i°*3§a
< 0
"S^o
'ra^'o ~
M ™ n*
In O ^^
C-O '
5 ° *
O,^
•0
O C ^j
* ra 3
^ ij'D
E in 2
28
O -U -H
U 3 ^-H
0) T3 > f-l
"E ^t ® o
3 a o
0 3
3 jr v>
£ ">
Kl
£
Oi CD -
"-H ^
0) W V>
> a~-
<
CO
01 —
ra ^"
CO
0 >
•H It
*J O
(8 Oi
r-l CD
3 4J
CD O
4J aJ lj
JSp
11 =
Z yj C
CO
o
O O
m
o on
CO
CN
rH
in
CN
o o o
in
0 0 m
0 0 rn
^*
0 00
o o cn
o o rn
rH
O
O
**• ** \o
*7 7 ^*
*• ^0
0
in
0 O rH
rH rH 1/1
•* ** CN
cn
vo
^. -^ co
Z 2m
cn
rH
Cn
•H Ul
U] tH O>
kn 3 C
0) U -H
X. W aJ
(^ 0*0 *fl
01 J-J C O
w a; 3 u
£) ^ O
-o u a •"
C C E 0
O 0 O O
m o u ct
rN (N M
cn TT M
CN r^ o
rH
&
u
fl
cn
r-
m
0
in
^
in
CN
O
in
rH
[^
O
rH
O
cn
m
rH
cn
o
m
rH
m
cn
rH
CN
CN
in
0
**~
rH
CN
^P
^
O
•H
i-l
0)
4J
C
•H
4J
(0
rH
[lj
*»
m
rH m OP*
CN CN rH m
rH r-H ^H p-i
O *»
r*- n O «-H
in rH \£> £~*
O O -
PO - - ^D
PO in cn rH
in r- o co
m co in cn
oo 0*1 o cn
CN o co in
o cn o m
»x> co cn TT
«3" m cn co
CN ^T rH m
o o o
n CN t**- ro
(N CN rH in
^ rn vo o\
mm fN r-
r«- vD CTl ^
CO U? VD O
m rH rH CN
r-i in cn co
m ^" rH f"-
CO CO VJ C tJ) -H jJ O>
CJC UOJC rnlCC
•H -H OJ JJ -H i— ( Q) -«H
jj jj 3 C J-> rtl E *->
W(0 *D'H(0 jJ&iHJ
(DO CtOO (U-HO
zu M£U SE&U
mm (N
rH rH (H
^D CO
m <£>
fN
P"t
CO O
m
rH
r- cn
(N pn
^J CN
rH
9 CO
«7 -?
in
CN rH
rH CN
o m
v0 cn
r-,
vO VD
^O VO
CN CN
co *r
r- m
cn ^
-H 0
o in
vo m
jj
C
ft)
Wi
a
01
c
- (0 01
W rH lH
C JJ 0)
-------
^^
Q
P
2
E-
2
0
Si
10
r-
cn
1 '
**
CO
EH
CO
0
CJ
2
O
H
<
cJ
5
§
o
(X
CO
CO
w
*2
H
CO
D
PQ
1
CO
•
^^*
1
m
W
«
^H
° M 2~
jJ O "— ^ dP
0 £ ~ ^
— ^i
y-i . ~-
DEC
w, O ~
JJ O -H w I Oi C 2
H •*•* C « *—•
•H jj Q jj ^j J ^/J.
3 o "a) ^ 1r O
C U U 3 C £
< 8
*I? w
*C , AJ
-H
M 3 bl -H
0) TJ S^-H
|So8
"81
AJ 3 —
E ^*
1-1 2
'"o
g1 00
P~
a
•H J
« —
0>
xJ O
<0 Dl
3 JJ
DI (0
o o
IX
» §1
S * C
vo o « ^.HC ^^^(N ^^ tn o o
(N M cri cocTio ^r.-ir-t^ cnm CN o o
i— i i— i n~i r^jtN ^ino
•H
^ co co ocoo in *? r*i o ^CN o o o
f*"l C*~l f*l r*lO~iO i-HlTliHrO CO^> CN O O
O O O OOO OOrHO rHrH ^" O O
ro m in or* co^coco cove oj
*<& ^ r^ (Ni-i_ co m o co OVD m _ o
vo VD ^ cr\oi »H in «— t «-^ rH m r*-
o o cr* in^* «Hrnr*«H r*-c^ ^*
»H r-< rH
co r*~ o r*coo rovDrHr-t ^ r** r*" o o
CN \XJ 0 ^0*3-0 C0cdr-*t-i VOVO U3 0 0
'•O VD LO OrnCC Or-^rN)^1 r-t^" i— ( PI O
en en co ompj o»Hf*ivo I-HVC c^ *d* o
CN O \D OO^* Ot^^^D CO VD rH ^ •— I
O ID O Or"l*? r^-COCOVD *f lT\ Cn t-* (N
o o in i-HVDco oinoo (NCO ^r (N ^H
»x> o o m t-t n co^ojcn ^"(N ^ *X) in
CO CD O v£)^f"i inr>J t— ( r*~ in in r-* (N co vo VD c^ r^ CD
o oo n co «3" en m rH **J «H o VD p* CN PO
oo en rH oo^1 vDo*iD m o co VD »-H ^
rH r^ in t-Ho^1 r^covD vo ^ rn *H t-n
m rg fN CN (N rH .-H
0) m V
In U W O> , 10 -DC**-- WJJU*M OO EEa
hW *OO) UDl U3a-> C OJ CD) DI O Ol (OD> •» i-iW RJ^-t
a^ rH-HC focrc CD a-nc c u o v^c w 0.1-1 aja-uc
1) ^H U E a •iJ rH G< J> V-< JZ E*J t3 U -r^ •*"* "*^ I' "^ *^ *G £ ^J
J-J (0 *H nj (0 <0 QJ O •— < 3 D^ -H (Q C *O ^i A-1 (0 fO JJ (0 QJ tfl &m
too 3c MO x: o a-H 30 to c M co jj (0
-------
II
•Sic 8
0 £ 0 3
11 W , — • — •
"•251-
S2-*
U " 3 Ul
II
11 J
"o EC:
II Ui O '~~
H Oi >AJ AJ —
II o 01 ro
| O UlrH
Q =
W Mr, "°
r^ T3 0) „,
t£ oiu-i i oic rj
\2 1 N O E C -rl
fj 11 ~-" U-rl «J " —
~ [I rHAJOAJUl^;
F-i ™
10 u ^ u
f- ?• o S,
^ -3-1
2 3s2*
rH M °
£3
II T3
- 5 ° m
S aw
E-1
CO „ -o
O •" <£ 01
u °5£
^2 =
2 -o^o
0 eg 2
H 2§D
P II f I
E-1 u
<
J M-l 0) AJ
M O AJ -n
B II u ui E
§ U •§ £ rt
ffi jg o S
o E r. ° u
^ 3 a o
ClH Z >
N
&
•o S
CO v%
_ -. (1 i) J ^^
c/) 3 crx
S l>~
* *£
M ^
cn
r-. I)
1—* II «
m 8," 3
, * u J
^ »'£«
iJ > a-
rf! [1 ^
W
11
l| W
*3" tn ~
II
ro
w
"J
9 1 §^
<3 || -H Ul
^ 1 *&
r-l 01
1 3 ^
I Ol ID
1 «• °
a
AJ e ^
II *•* C (1
0) OJ ,
It Ci '
(0 01
S W
oooo ooo o o
oooo ooo o o
oooo ooo o o
oooo ooo o o
oooo ooo o o
oooo ooo o o
oooo ooo o o
oooo ooo o o
oooo oo oo o
oooo oo oo o
oooo oo oo o
mi/i intn oo mm ^
TJ^ ry^r fH r-i ^" ^ <*
oooo oo oo o
AJ
jj ^ C
C AJ QJ
W W W (U y) C E M
•COJOJOJ^OJat tO Q) W
C>>T3>*D>W AJ U -0
3 -H *D -H O -H a -H (0 ^ W C
OAJ OAJ OAJ COAJ OJU) O -HAJ (U >i3
^ m om 3m c ic --co «-• w wcw AJW u^
& > S> > S > OJC 0 0» 0) at v^ai 3d) AJ
!-H
CJ
in
(N
CD
C\)
^
^«
rH
a\
m
V£>
C
O
•r-*
(0
^H 3
ro E
o AJ o
4J (1) «4-t
^. OJ
C Ul Ul
E Ui Ui
c o o
O O •!-> -n
-n 3 10 IB
rH AJ E 6
•n C O -O
E 'rl Ol 0)
Ul N
O 01 01 -H
in Ol 'O rH
c c ro
O 3 3
C rH C
ro oi >t c
£ J2 rH 10
AJ r-l
>* ID 01
01 ui -H s:
Ol O AJ AJ
01 Oi C
rH 01 01 >*
AJ AJ X3
Ui ID O
o u a TJ
01
Ul AJ AJ -H
AJ U 10 rH
u 3 JH a
3 -O AJ -H
•DO AJ
O Ul C rH
ui a o 3
a -r. e
QJ AJ
01 £ Ul C
Oi AJ O O
c a -r<
-rH C AJ
AJ 11 0) 10
TO £ JS rH
O 3 AJ 3
0 E
•O AJ Ul
E 11 C O
M 01 01 y-i
< 3 03 1)
0) Ui
C 0) Ul
•H -H P< CT
QJ C
C QJ lu -H
O 01 U
•T-< (Q 0 <3
r-l !M xJ «AJ •
•— i o; —
•H > U) W fN
E ro *0 AJ
U U C
O *O -«-t 3 O
rH QJ X! *O'-»
AJ AJ O AJ
.C i JH 0
C Di O Q-OJ
10 -H ? W
J^ QJ U H-t
AJ 3 0 C
>i -^
W fQ JJ *-<
W • Q) W
QJ *O *O *O J2 •"*
rH C 01 01 E-H
to -H T? 3 ro
Ol rH C C AJ
c - as 01
•H OJ -rH O 01 *D
3 C rH AJ
•O O 3 Ol M
O -H 6 Ui ui r^
Ui AJ (0 -H r-
a o AJ
0) -n oi c r-
W 03 6rH OrH
ID -H ja -H v>
E r-l ID AJ
•O O AJ (0 >*-!
01 Ul 01 rH O
C IAJ J3 C 3
-H *J -H E AJ
U-l C K U
o> oi ri 01 o g
businesses are d
ge prices are tak
nts.
r of products ove
mulation. Number
lized cost of ref
estimate per proc
rH 01 ID 01 0) M (0
r-l 01 Ul g J3 O 3 AJ
(OrHOlOlE'A-'COl
ElO>OI30ICO
WOlriiOlSZ^I^U
, c u -a
•o
UrH
ui ro
ro c
01 U.
Ul
0.
03
• 3
>iO
01 3
> 10
W C_J
tn >,
Ul
>.AJ
Ui 0}
O 3
AJT3
c c
01 rH
>
c c
M O
• rl
0) AJ
C ifl
O-H
•rl AJ
W Ol
•H 01
gz
u
>,
U Ui
o o
> AJ
iO
0] rH
Ol 3
c 01
•rl 01
AJ 0£
ro
O .E
CJ UH
<
01
U D
ro z.
IM AJ
Ui
3 C.
W AJ
01 3
u
c c
ro o
C *ri
01 AJ
AJ 10
C Ul
•H 01
S §•
0
rH O
ro
•rl C
Ul-H
4J
01 C
g o
rrj .,_(
C AJ
H (0
T3 U
C O
ro 01
U)
r-l «3*
f—1 «,
ID
Ul 0]
3 Ol
AJ C
CJ— 1
0) AJ
AJ (TJ
-H O
•C U
L)
< c
ID •
' AJ CV
05 C fj,
AJ-H rH
£ a
>t available.
Industry Insig
for National P
Draft Report.
W
Z
CJ
II U
tH
< 3
^ 0
Z t/3
3-18
-------
estimated 421 small business products in the survey (42 percent)
exceed the 1997 TOS. This figure is slightly higher than the
proportion of all survey products that exceed the limit (36
percent). As established in Section 2, approximately one-third
of products over the proposed limit can costlessly comply with
the regulation because of their similarity to the remaining over-
the-limit products that are being reformulated. The remaining
over-the-limit products are referred to as "constrained" by the
regulation and the sum of the costless compliance products and
under-the-limit products as "unconstrained" by the regulation.
Less than 10 percent of the small business products in the
sanding sealers, mastic texture coatings, and bituminous
categories will be constrained by the regulation. Swimming pool
coatings, shellacs, and high-temperature coatings produced by the
small business sector will require no reformulations. Traffic
paints, roof coatings, and varnishes are all relatively high-
volume categories in which over 40 percent of the surveyed small
business products are constrained by the 1997 TOS limits.
3.2.3.1 Small Business Costs Under "Reformulation-Only"
Option. In this section, total and per-unit annualized
compliance costs for small producers in each product category
with reformulation as the only compliance option is estimated.
The annualized $17,772 estimate of the cost per reformulation (in
1991 dollars) is multiplied by the number of products constrained
by the regulation (i.e., all products over the limit less the
one-third that can costlessly comply). Table 3-4 provides the
cost estimates. To contrast these costs with product revenue,
the analysis uses average price per liter for each category from
Section 2 for the market segment in which the category is
classified.9 The cost of reformulation as a percentage of
revenues is computed using the estimated cost of reformulation
'Where a coating category could not be separated into waterborne and
solventborne market segments (categories in market segments 1 through 8) a
weighted average of the two prices was used.
3-19
-------
divided by the imputed revenues for each product category.
The data presented in Tables 3-1 and 3-4109 illustrate a
number of scenarios pertaining to potential small business
impacts of the regulation under a reformulation-only response
scenario. Some of the phenomena indicated by the data are
examined below.
Roof coatings is the largest quantity and highest revenue
category for small businesses. For small business roof coatings,
43 percent of the individual products will be constrained;
however, the cost of reformulation as a percentage of sales is
relatively small, less than 1 percent.
In the opaque waterproofing sealers category, small
producers comprise a very small share of the market--less than 1
percent-both relative to all producers in this market and
relative to small business shares in other markets. However,
only two products are manufactured by small businesses in this
segment, and the data indicate that both of them exceed the TOS
limit. Reformulation costs would exceed 50 percent of total
small business sales for the category. This increased cost may
well lead to a compliance strategy other than reformulation
(i.e., fee or withdrawal).
Antigraffiti coatings present quite a different small
business impact scenario. Small businesses represent almost the
entire market but produce small quantities in relation to other
coating categories and generate lower revenues. Only one product
requires reformulation under the TOS limits, but the cost
represents over 7 percent of revenues in the category.
For the small business segment of the AIM coatings market,
42 percent of their products are over the TOS limits, and 28
percent are expected to undergo a major reformulation, pay a fee,
or exit. The total annualized cost for the sample of small
businesses in the survey under the reformulation-only option is
$5.0 million. Table 3-5 presents small firm and industry
3-20
-------
TABLE 3-5. AVERAGE REGULATORY IMPACT BY FIRM SIZE3
Industry Small firm
average average
38,990,002 4,614,010
($1991)
Number of products'1 42.4 27.5
Number of products facing 9.9 7.8
major reformulationc
Annualized reformulation cost"3 176,648 138,720
($1991)
Ratio of annualized reformulation 0.45 3.01
cost to revenues (percent)
" The survey has 116 respondents and 36 of those identified themselves as
having under $10 million in annual sales. Twelve survey respondents
did not report company size.
b Data for revenues and products per firm were based on data reported in
Table 3-1. The number of products per firm is based on the total
number of products for which quantity data are available.
c This number represents two-thirds of the products over the 1997 TOS.
Industry experts estimate that approximately two-thirds of the products
with VOC contents exceeding the TOS limits will incur the estimated
$250,000 per product "major" reformulation cost.
Annualized cost of reformulation is the number of major reformulations
multiplied by the annualized reformulation cost estimate per product of
$17,772 (1991 dollars).
Source: Industry Insights. Architectural and Industrial Maintenance
Surface Coatings VOC Emissions Inventory Survey. Prepared for
National Paint and Coatings Association in cooperation with the
AIM Regulatory Negotiation Industry Caucus. Final Draft Report.
1993.
averages for revenues, number of products, and reformulation
costs.110 Small businesses on average manufacture approximately
one-third fewer products than the industry average. On average,
small firms have fewer constrained products than the industry
average, but they comprise a slightly larger percentage of total
number of products, 28 percent, as compared to 23 percent for the
industry. Similarly, small business reformulation costs as a
percentage of revenues are higher at 3 percent than the industry
at roughly 0.5 percent.
3-21
-------
3.2.4 Potential Mitigating Effects. Fee Option and Small
Product Exemption
Fee Option. As discussed in Section 2, a product's output
level affects the choice between reformulating the product and
paying an exceedance fee. Since the cost of reformulation is a
fixed cost (i.e., it is independent of output level) the average
reformulation cost per unit of output falls as output levels
increase. However, the exceedance fee per unit of output is
constant with respect to the output levels. Thus, the fee is
more likely to be chosen by small volume producers, all else
equal. Because the fee will be more cost-effective only for
lower volume products and lower excess VOC categories, allowing
the fee option should have a relatively small impact on variation
from the aggregate emissions reduction targets as long as the fee
assessment rate is not set at an inappropriately low level.
Therefore, the fee option provides increased flexibility for
small businesses by placing an upper limit on the per-unit costs
of complying with the regulation, without significantly
jeopardizing VOC emission reduction targets.
The model is unable to conduct a least-cost analysis of the
fee/reformulation decisions (see Section 2) for the small
business segment of the survey because of insufficient VOC
stratum-specific data on small businesses. Therefore, the
analysis cannot directly estimate the mitigating effect of the
fee option on the costs borne by small businesses. However, the
results of the least-cost analysis in Section 2 can shed some
light on small business impacts.
In Section 2, the proposed fee rate of $0.005 per g/L over
the limit (1995 dollars, adjusted to $0.0045 in 1991 dollars)
leads to selecting the fee over reformulation for 323 of the
1,153 survey products constrained by the proposed content limits.
Note that the average sales quantity for the 323 fee-selected
products is approximately the same as the average product volume
for all small business products in the survey, suggesting a
3-22
-------
strong correlation between small business producers and selecting
the fee. To provide a measure of the maximum potential benefit
of the fee option for small producers, assume all of the products
opting for the fee are produced by small producers. Recall from
Section 2 that the fee option saves surveyed producers from
spending over $5.7 million in reformulation costs in return for
fee payments of just under $4.0 million, for a net savings to
surveyed producers of $1.8 million. Deducting this cost savings
from the total reformulation cost for small business producers in
the survey (Table 3-4) leads to a 36 percent cost reduction if
these savings accrue entirely to small business producers.
Low Volume Exemption. As discussed in the previous section,
the fixed cost of reformulation places a substantial burden on
low volume producers that exceed the applicable VOC limit. As an
alternative to a fee option which also can be costly for low
volume producers, the EPA can establish a low volume exemption.
This exemption would be designed to relieve low volume producers
from reformulation costs that can be difficult to recover from
the small amount of revenue generated by a low volume product.
Both the exceedance fee alternative and the low volume exemption
are compliance options aimed at addressing the potential issue of
"niche markets" in which low volume products exist for which it
may not be cost-effective for either the manufacturer or resin
supplier to develop a lower VOC formulation.
The EPA lacks data to evaluate an appropriate sales volume
cut-off for a low volume exemption. In the architectural coating
rule proposal, the EPA requests comment on a low-volume exemption
in the range between 1,000 and 5,000 gallons per year. Through
proposal of the architectural coating rule, the EPA is seeking
detailed information from manufacturers who produce low volume
specialty niche products which cannot be cost-effectively
reformulated. Specifically, the preamble requests the following
information for each identified product: sales volume, VOC
content, a detailed cost estimate for reformulation, and whether
3-23
-------
alternative products exist in the market which are able to meet
the table 1. VOC content standards. Once this data is gathered,
the EPA can evaluate factors such as reformulation cost versus
affordability to small manufacturers. In order to determine
affordability, one could compare the estimated reformulation
cost/unit with the current production cost/unit. Further, the
size of the small volume cut-off would affect the overall
cost/effectiveness of the rule. In addition, the EPA plans to
work with the Small Business Administration Office of Advocacy to
evaluate available data on lower volume sales which was collected
in a survey conducted by the National Paint and Coatings
Association in 1992. Under either the exceedance fee approach or
the low volume exemption, the EPA would consider the expected
magnitude of foregone emission reductions as compared to the
expected cost to achieve these reductions.
While seeking ways to mitigate the impacts of the regulation
for small manufacturers, EPA also recognizes that the two
different approaches discussed here, the fee option and small
product exemption, have different implications for the marginal
incentives for VOC reductions. Although the fee option continues
to provide incentive to reformulate the small niche products
because marginal reductions in VOC content will reduce the per
unit fee paid, a low volume exemption would provide no such
incentive. Thus, when evaluating these options, EPA must
consider the tradeoff between the level of relief for small
products versus the incentives to achieve VOC reductions.
3.2.5 Market Effects
Data were not available to estimate market-level consumer
and producer surplus effects for the small business segment of
the AIM coatings industry. Table 3-6 summarizes the small
business share of total volume and industry-wide producer surplus
effects for the 13 market segments defined in the market model
described in Section 2.111 The survey data did not allow
separation into waterborne and solventborne segments for the
3-24
-------
small business data; therefore, market segments 1 through 8
combine the waterborne and solventborne portions. In the eight
categories that can be divided into solventborne and waterborne
market segments small businesses produce more products in the
solventborne segment than in the waterborne segment, except for
interior coatings. Although it is not reflected in Table 3-6,
the majority of reformulations in each sector occur with
solventborne products.
In the architectural lacquers market segment, small
businesses produce over 25 percent of total volume, but that
market has relatively small adverse impacts on producer surplus,
a $464,000 reduction. Small businesses produce 12 percent of the
total quantity in the traffic marking paints market, where there
is a $1.2 million decrease in producer surplus due to the
regulation. Industrial maintenance coatings experience the
largest producer surplus impacts, a $3.7 million decrease, and
small businesses produce slightly more than 5 percent of total
volume.
The data in Table 3-6 can be used to compute a rough proxy
for small business producer surplus effects that is a share-
weighted average of total change in producer surplus in each
market (APSJ :
APSSB = 2 PS1 • Si58 (3-4)
where s^8 is the small business market share in market I.h The
standard case scenario for the 1997 limits produces an estimate
for APSSB of approximately $840,000 in producer surplus losses
for small business producers. This is roughly 4.7 percent of
This estimate is a very rough measure of producer surplus effects for
small businesses since it merely apportions the market-wide producer surplus
based on small business market share, rather than on direct tracking of small
business withdrawals, fee payments, and gains by unconstrained producers, for
example. However, it does underscore the notion that the static losses
(aggregate reformulation costs) without accounting for price effects and
withdrawals will overstate the regulatory burden on producers. This holds for
all producing sectors, not just small businesses.
3-25
-------
TABLE 3-6. MARKET-LEVEL IMPACTS AND
SMALL BUSINESS PRESENCE3'D
Market
segment
number
1
3
5
7
& 2
& 4
& 6
& 8
9
10
11
12
13
Small All
business businesses
share of producer
total surplus
sales volume effects
Market segment
Exterior & high-
performance coatings
Interior coatings
Primers &
under coaters
Clear coatings,
sealers, & stains
Architectural
lacquers
Wood preservatives
Traffic marking
paints
Special purpose
coatings
Industrial
maintenance coatings
(
6
0
2
2
25
0
12
6
5
3
%)
.16
.96
.18
.44
.27
.00
.05
.78
.39
.56
(103 $1991)
-3,279
-3,468
-2,257
-3,336
-464
-245
-1,225
-182
-3,699
-18,156
Share-weighted
small business
producer
surplus
effects
(103 $1991)
-202
-33
-49
-81
-117
0
-147
-12
-199
-842
.1
.1
.2
.4
.4
.0
.6
.4
.2
.2
a The emissions survey data did not allow separation into waterborne and
solventborne segments for the small business data; therefore, segments
1-8 combine the waterborne and solventborne portions.
b The survey has 116 respondents and 36 of those identified themselves as
having under $10 million in annual sales. Twelve survey respondents
did not report company size.
Source: Industry Insights. Architectural and Industrial Maintenance
Surface Coatings VOC Emissions Inventory Survey. Prepared for
National Paint and Coatings Association in cooperation with the
AIM Regulatory Negotiation Industry Caucus. Final Draft Report.
1993.
See Appendix A for construction of the market segments.
3-26
-------
total producer surplus losses, a proportion that is higher than
small business share of industry output (3.6 percent) but lower
than small business share of industry emissions (6.2 percent).
3.3 SUMMARY
The potential for significant impacts on small businesses of
the proposed regulation arise from two primary sources:
• Products made by small producers, on average, have
higher VOC content than the industry average.
The costs of reformulating products to comply with the
regulation are fixed and thereby impose higher average
costs on small volume coatings.
The first problem is related to small producers' tendency to
specialize in coating categories that are naturally higher in VOC
content and to their tendency to concentrate in the "high-VOC"
end of the distribution of products within a given category.
Thus the potential for disproportionate impacts of VOC reduction
regulation on small businesses follows partly from the fact that
small businesses contribute a disproportionate amount of the
aggregate VOC emissions that are targeted for reduction.
The second problem follows from the nature of reformulation
costs. A coating's formula is the product of an intellectual
capital investment, much like the development of a drug or a
computer software product. The cost of the investment follows
directly from the level of effort necessary to revise the formula
to meet both the VOC standards imposed by the regulation and
performance standards imposed by the marketplace. This level of
effort is essentially independent of the quantity of the product
that is eventually sold. Therefore, the relative impacts on
smaller volume products is, by definition, greater.
The data used in this analysis suggest that these two
primary factors are empirically relevant in the case of small AIM
3-27
-------
coatings producers. The average VOC content of the products made
by the small business producers in the survey is 75 percent
higher than the average VOC content of all products combined. A
little over half of the difference in the averages is attributed
to the specialization of small producers in high-VOC content
product categories, with the remainder attributed to the tendency
for small businesses to produce higher VOC products within each
product group. Moreover, the average product volume of products
made by small businesses is less than 20 percent of the average
product volume for the entire survey population, implying much
larger average reformulation costs. Thus, without mitigating
factors, the impacts on small businesses are potentially
significant.
The proposed regulation has been designed to mitigate small
business impacts. Despite their inherently higher VOC content,
the proportion of small businesses products exceeding the
regulatory standards proposed for 1997 is not much higher than
the corresponding proportion for the survey population at large
(42 percent vs. 36 percent). This suggests content limits have
been assigned among categories in the TOS with sensitivity to
small business presence. In addition, the availability of the
exceedance fee option is beneficial to small business producers
because it places an upper bound on the per-unit costs of
compliance. Data analyzed in this study indicate that producer
costs are reduced by nearly 40 percent in the cases where the fee
option is chosen. It is likely that much of this savings in cost
.will accrue to small producers.
3-28
-------
SECTION 4
REFERENCES
1. Stigliani, William M. Chemical Emissions from the
Processing and Use of Materials: the Need for an
Integrated Emissions Accounting System. Ecological
Economics 2.(4) :325-341. 1990. (Adapted from Figure 2).
2. Industry Insights. Architectural and Industrial
Maintenance Surface Coatings VOC Emissions Inventory
Survey. Prepared for the National Paint and Coatings
Association in cooperation with the AIM Regulatory
Negotiation Industry Caucus. Final Draft Report. 1993.
3. U.S. Department of Commerce. 1987 Census of
Manufactures, Industry Series: Paints and Allied
Products. Washington, DC, U.S. Government Printing
Office. 1990.
4. Ref. 2.
5. Ref. 2.
6. Johnson, Duane. The Best Paint for the Job. Family
Handyman. 42.: 18. June 1992.
7. U.S. Department of Commerce. Current Industrial Reports:
Paints and Allied Products, 1991. Washington, DC, U.S.
Government Printing Office. 1992.
8. Industrial Economics, Inc. Use of Economic Incentives to
Reduce Consumer and Industrial Consumption of Solvents.
Draft Report. 1989.
4-1
-------
9.
10,
11,
12
13,
14
15,
16
17
18
19
20,
21
22
23
24
25
Radian Corporation. Regulation of Architectural
Coatings. (Appendix E--U.S. Environmental Protection
Agency Non-flat Architectural Coatings). June 26, 1987.
Consumer Reports Magazine. House Paints and Stains.
52{6) -.365-374.
Consumer Reports Magazine. High-Gloss Enamels. p. 173.
March 1988.
Consumer Reports Magazine. Interior Semigloss Paints.
p. 317. May 1989.
Consumer Reports Magazine. Paints for Finishing Touches
p. 619. September 1990.
Consumer Reports Magazine. Interior Latex Paints,
p. 333. May 1991.
Ref. 10.
Ref. 13.
Ref. 14.
Brand, Benson G., and Guy E. Weismantel. Exterior
Coatings for Wood. In: Paint Handbook, Guy E.
Weismantel (ed.). New York, McGraw-Hill. Pp. 9-1 to
9-35. 1981.
Sunset Magazine. What You Really Need to Know About
House Paint and Painting, 179:80. September 1987.
Ref. 19.
Ref. 10.
Ref. 11.
Ref. 12.
Ref. 13.
Consumer Reports Magazine,
p. 568. September 1988.
Interior Latex Paints,
4-2
-------
26. Ref. 17.
27. National Paint Coatings Association. The Household Paint
Selector. New York, Barnes and Noble Books. 1975.
28. Bakke, Timothy 0. Clean Air Paints. Popular Science.
211:85. August 1990.
29. Ref. 9.
30. Ref. 6.
31. SRI International. U.S. Paint Industry Data Base. Menlo
Park, CA. 1990.
32. U.S. Department of Commerce. The 1982 Benchmark Input-
Output Accounts of the United States. Washington, DC,
U.S. Government Printing Office. 1991.
33. Whittington, Trevellyan V. Paint Fundamentals. In Paint
Handbook, Guy E. Weismantel (ed.). New York, McGraw-
Hill. Pp. 1-1 to 1-23. 1981.
34. Beno, J., W. Brown, and P.P. Obst. Formulating and Using
Water Based Thermoplastic Resins for Wood Finishing. In
Proceedings of the Nineteenth Water-Borne, Higher-Solids,
and Powder Coatings Symposium. Robson F. Storey and
Shelby F. Thames (eds.). University of Southern
Mississippi, Department of Polymer Science. Pp. 626-638.
1992.
35. Singer, Elias. Raw Materials. In Paint Handbook, Guy E.
Weismantel (ed.). New York, McGraw-Hill. Pp. 3-1 to
3-22. 1981.
36. Rauch Associates, Inc. The Rauch Guide to the U.S. Paint
Industry. Bridgewater, NJ, Rauch Associates, Inc. 1990.
37. Ref. 35.
38. Kemezis, Paul. Wait-and-See Stance Taken on Zero-VOC
Architectural Paints. Chemical Week. Pp. 52-53.
October 1992.
39. Ref. 34.
4-3
-------
40. Ref. 36.
41. Ref. 33.
42. Ref. 33. (Adapted from Figure 1.4).
43. Ref. 36.
44. Ref. 28.
45. D'Amico, Esther. Waterborne Systems Gaining Nitche By
Nitche. Chemical Marketing Reporter. 238 (18) :SR20-SR28.
1990.
46. Ref. 28.
47. Ref. 6.
48. Ref. 28.
49. Ref. 28.
50. Ref. 28.
51. Ref. 38.
52. Ref. 38.
53. Mullin, Rick. Reformulation Continues as the VOC Target
Broadens. Chemical Week, p. 37. October 1992.
54. U.S. Department of Commerce. Current Industrial Reports:
Paints and Allied Products, 1982. Washington, DC, U.S.
Government Printing Office. 1983.
55. U.S. Department of Commerce. Current Industrial Reports
Paints and Allied Products, 1983. Washington, DC, U.S.
Government Printing Office. 1984.
56. U.S. Department of Commerce. Current Industrial Reports
Paints and Allied Products, 1984. Washington, DC, U.S.
Government Printing Office. 1985.
4-4
-------
57. U.S. Department of Commerce. Current Industrial Reports:
Paints and Allied Products, 1985. Washington, DC, U.S.
Government Printing Office. 1986.
58. U.S. Department of Commerce. Current Industrial Reports:
Paints and Allied Products, 1986. Washington, DC, U.S.
Government Printing Office. 1987.
59. U.S. Department of Commerce. Current Industrial Reports:
Paints and Allied Products, 1987. Washington, DC, U.S.
Government Printing Office. 1988.
60. U.S. Department of Commerce. Current Industrial Reports:
Paints and Allied Products, 1988. Washington, DC, U.S.
Government Printing Office. 1989.
61. U.S. Department of Commerce. Current Industrial Reports:
Paints and Allied Products, 1989. Washington, DC, U.S.
Government Printing Office. 1990.
62. U.S. Department of Commerce. Current Industrial Reports:
Paints and Allied Products, 1990. Washington, DC, U.S.
Government Printing Office. 1991.
63. Ref. 7.
64. O'Reilly, Richard. Product Development Drives Market
Growth. In: Standard and Poor's Industry Surveys:
Chemicals Current Analysis. Pp. C40-C43. New York.
Standard and Poor's Corporation. 1991.
65. Ref. 64.
66. Ref. 64.
67. Ref. 6.
68. Ref. 14.
69. Ref. 27.
70. Ref. 3.
71. Ref. 32.
4-5
-------
72. U.S. Department of Commerce. Census of Manufactures,
Subject Series: Type of Organization. Washington, DC,
U.S. Government Printing Office, February 1991.
73. Source for Percentage Accounted for Data: U.S.
Department of Commerce. 1987 Census of Manufactures,
Concentration Ratios of Manufacturers. Washington, DC,
U.S. Government Printing Office. 1992.
74. Telecon. Lindsay, Sam. Small Business Administration,
with Dempsey, Jenny. Research Triangle Institute, July
6, 1993.
75. U.S. Department of Commerce. 1987 Economic Censuses.
Volume 1, Report Series, Release ID. Census of
Manufactures: Location of Manufacturing Plants. file
MC87LMCO. 1991.
76. Ref. 36.
77. Ref. 3.
78. Loesel, Andrew. Coatings Industry Faces New Mix. In
Chemical Marketing Reporter. 231(18):SR3-SR8. 1990.
79. Ref. 78.
80. Finishers' Management. The U.S. Paint and Coatings
Industry. Pp. 23-25. April 1991.
81. U.S. Department of Commerce. U.S. Industrial Outlook
'92, Business Forecasts for 350 Industries. Washington,
DC, U.S. Government Printing Office. 1992.
82. AIM Coatings Regulatory Negotiation Committee meeting,
Washington, DC. Meeting Summary. July 28-30, 1993.
83. Ref. 82.
84. Ref. 82.
85. Brady, Catherine. VOC Regs Put a Strain on Structural
Coatings. Chemical Week. Pp. 47-48. October 1989
4-6
-------
86. ICF Consulting Associates, Incorporated. Small Business
Economic Impact Study. Prepared for the South Coast Air
Quality Management District. Final Report. June 17,
1988.
87. Ref. 2.
88. Ref. 2
89. Fax from Sarsomy, Chris, Radian Corporation, to Murray,
Brian, Research Triangle Institute. January 16, 1995.
90. Ref. 2.
91. Ref. 2.
92. Ref. 7.
93. Ref. 2.
94. Ref. 7.
95. Letter from Nelson, Robert N, National Paint and Coatings
Association, to Ducey, Ellen, EPA/QAQPS.
April 2, 1991.
96. Radian Corporation. Determination of Architectural
Coatings Baseline Sales and VOC Emissions. Memorandum
from Harrison, Rob, Radian Corporation, to Duncey, Ellen,
U.S. EPA, OAQPS, Emissions Standards Division, Dated
November 22, 1995.
97. Ref. 89.
98. Ref. 7.
99. U.S. Department of Commerce. 1991 Annual Survey of
Manufactures: Statistics for Industry Groups and
Industries. Washington, DC, U.S. Government Printing
Office. 1992.
100. Ref. 99.
101. U.S. Small Business Administration. The Regulatory
Flexibility Act. October 1982.
4-7
-------
102. Ref. 61.
103. Ref. 2.
104. Ref. 2.
105. Ref. 2.
106. Ref. 2.
107. Memorandum from Nelson, Robert, National Paint &
Coatings Association, to Madariaga, Bruce, EPA/OAQPS
October 14, 1993.
108. Ref. 82.
109. Ref. 2.
110. Ref. 2.
111. Ref. 2.
4-8
-------
APPENDIX A
MARKET DEFINITION,
DEMAND ESTIMATION, AND DATA
-------
A.I PRODUCT/MARKET CROSS-REFERENCE METHOD
Data on coating prices, quantities, average VOC contents,
and proposed VOC content limits are necessary to estimate the
effect of VOC content limits on AIM coatings products. Price
and quantity data were taken from the 1991 Current Industrial
Reports: Paint and Allied Products.1 The Architectural and
Industrial Maintenance Surface Coatings Survey (the survey)2
provided the sales-weighted average VOC emissions, which
represent VOC content. VOC content limits were from the 1997
and 2000 TOS developed by EPA.
Census data are organized according to product codes,
which define product categories; however, these Census product
categories differ from the product categories in the survey.
Furthermore, the TOS (see Table 2-2) gives VOC content limits
for product categories that differ slightly from those
categories for which data are provided in the survey. Data
from all three sources are necessary to conduct the economic
impact analysis. Therefore, a fourth product categorization
was constructed, which is called market segments, that
aggregates the categories so that data may be used from all
three sources to provide the necessary level of resolution for
market analysis. Table A-l3'4 illustrates the individual
product categories represented by each data source and how
they map into the market segments used in the analysis.
A-l
-------
W
tt
i
EH
O
B
o
cr;
n ki
o c
0 0
u A
u
n
•o B
•O E
C -H
a -,
* 1
X! C
a o>
|M kl
X
inventor
tsgory
_ «
0 °
•^ >
2*
e -
® n
u
O
" »
3 y 01
S| o
01 P U
u 0
X
kl
0
AJ
0
u
B
3
B
C
U
AJ
kl
O
£
a
ki
AJ
m
|
AJ
•
3
e
«
B
AJ
ki
JE
0>
ki T) k-
O C AJ U
O 1C W C
- 30
u m ki o
01 .
c - - K
01 •»* 0) 31 O
• U AJ u U ~«
• ki c a c c ki
B o n o ic • « • oi
Bi ••«. E u E oi e » AJ
C kl k. n > A. > X
04 01 OB a • o ~< o -* oi
a em ti o ^
«j AJ £0 3 * £ k, .c i. IA«
o a DIU AJ E en a. a o, c
a; u. i B x z z
. AJ
S £ 2
no > e
0 XI rt «
2 £ . g S
C 01 £ I 0
S> u i a
-1 kl (J 01 1
-H O « C Dl i
O B * C kl
B • > -« 0
1 • U ** AJ -H
1 kl C O IB kv
B o « B o «
C C k- i X
~< 41 O B W 01
« X K C 0
O 01 Oi •*« C AJ
U 0. AJ -i 1C
O B
a « o> 3 -* AJ
c o c
AJ C AJ « • •«
g S S U V, S
> XI > 0 0
*-» iH O •* «
O Di O *H ki 3
0 C 0 *4 V O
kl T3 kl C V ^4
0 3 0 -1 « E
-, — _!, 3
01 C 01 X « -H
AJ -^ AJ . £ XI
K X AJ
U U O
01
c
kl
3
AJ
c
V
o
n
u
o
V
AJ
"
a>
k, T, £
O C AJ u
o « m c
-< 3 O
IM tn u u
> Oi
k, . c
• -H U AJ U U
* • ki C * C C
o ma IQ o IQ«IO«
•^ OlAJ E O EVEOI
kl CX k. « kl>kl>
01 -1 01 OBB< O -i O -I
••H « E B 01 1
E K z 3 3 x x
9 h
C « «
II -5
ki a £, i
9 ki O *
AJ 9 ( kl 0
o AJ i a DV
2 C k. k, ^C
U 1 ki C 41 10
o B oi n k o
••4 Ol AJ 01 E 1 O
kl C X C ki 1
01 ~« 0) kl o B «
AJ AJ O »w Ol 3
X 10 - XI kl C O
0) O AJ M 01 ~* C
0 « 01 Q AJ -g
a o c * DI o AJ
^ o o ~« -<
U. K Z X «
*-i (N in
w ^ in
in tn in
as as as
(N CN n
ki
O
O
T>
C
n
01 X
B U
a a>
XI -O
C B 0
^ AJ -4
AJ C kl
C -H 01 B
•H 10 AJ Q>
AJ a c c
? oT 3
« O ki O
kt O 0
0 -H
AJ "O kl kl
C C 9 9
iH « AJ AJ
Oi C 9 )
ki « kt ki
9 £* 9 O
5 c 5 S
* 5 * 0 S
O 3 O *9 9
ki U ki q ki
9 C 9 C 9
*J -rt AJ 9 £
u u o
9
C
o
A
kl
9
«
ki
O
ki
9
AJ
"
m
kl 0
O -^
• — « 9
k. kj £
O 4) H)
k. C 9
AJ >t
C - kl
AJ *^ U
5 1 g
9
C
U
O i
0 kl 0
i O —
o » c e
ki COD
4) -H J3
AJ AJ >^
C - C ki
C >
AJ VI O U
« C « -H
£20
ra in ^
^3 ^ H
OD 00 OD
r< n o*
-o
c
c
C 0
AJ V C
C E Of
-H « >
AJ C ^
« 0
« k> C
0 AJ
AJ J£ V
a u AJ B
-« — c
« 3 -£ ^-
— < *o *0 (5 «
^H C C -C 0)
c c « n a
0 * Oi J2
AJ B AJ 9
C j> C Oi
> »'•«"•«
— a 0 AJ
O AJ 0 C
0 C AJ ~* t-»
9 C 01 AJ
AJ > • -H
a — > E TJ
- 0 -• 9 C
0
O 0 O ki O AJ
£55^22.
C C C
9
c
jg
*J
c
9
O
0
b
O
Ll
V
c
t-H
kl
o
kl kt
0 9
ki C
AJ
C »
a
CO C
*H O
u. z
9
c
waterbori
r--
i 0
O V
ki C C
01 •* k
5 J "!
«o <
• rH AJ
AJ *J *
10 C *
•-< m
OD 00
in tn
00 00
(N (N
ki
0 9
9 £
0 AJ
C «
^H
AJ C
C -H
-H AJ
AJ 10
0
•o
c -
AJ £
C n «
ifl O* C
a.,,
9 0 *0
AJ 0 C
« O *
3g5
a
35 &
C C
M M
C
kl
9
AJ
J
kl
O
kl
9
AJ
C
0
« kl
m 9
9
9
c
w
9 AJ 0 <
C C U >
W 9 9 — i
X) ^ "3 •
AJ O 9
C 0 0 '
5 . .S
O kl k. AJ
0 9 9 «
AJ E O
- « -* 0
IR O kl kl
w u a 9
9 ki *C
E 9 • (
ki C •
Ox O O
tn •-(
rH r*
in tn
eo ac
Ol OI
0 a
9 9
E E
ki W
a a
* 0
0 0
kl kl
9 ffl
AJ AJ
* c
O O
0 0
9 9
11
AJ AJ
C C
9 9
<-H .-I
O O
o o
-H «4
AJ AJ
B
01 B
E kl
a •
o
« u
C kl
kl «
.81
AJ 3 a
C 3
0) *O O
> C k,
~! « 0
0
M
4J
C
o
o
XI
IB
TJ
a>
w
(1)
4J
0
0)
4)
A-2
-------
O
U
u
En
W
cn
I
CO
CO
o
tf
o
EH
0
g
o
OS
a
3
O
•o
Oi >
W ki
o o
C. O>
0 01
kl JJ
a 0
u
V)
TJ n
kl 4J
10 -«H
•a E
c •••«
0 *H
JJ
O J.
«
3 i
H !
><
O
4J H.
c £7
i|
c •
s »
« i
§
J^J
3 *
"8 £
kl
o
S1
•J:
u
«
3
C
a>
o
Jj
kl
o
a>
kl
JJ
to
*g
4J
s
kt
kl
3
jj
&
V
JJ
01
kl
X
n
"w o>
o> m
JJ 0)
0 w
0
u
k< n
o> ki
TJ 0> •
C E Q
3 1-1 h
kt V
*0 k\ JJ
C 0
0 NO
u u
n T) t
E U i
£ I3
c
O * ki
J3 k< 4
0) kl 01 4J
C 01 » 0
i t B 1
ki a
01 - - k.
jj a a o
0 W kl JJ
> 01 01 0
jj E O
- 0 -rt 0
a o k< fc*
w o a o
E 8 • }
•** Tl Q I
ki C •
O. D O
9 «>
9 CO
•H «-*
r* ft
in in
00 CD
(N (N
m a
0> 0)
E E
a a
TJ TJ
C B
n a
01 0)
0 0
o o
O 0
0) 01
TJ TJ
C C
3 3
kl kl
0) Ot
4J 4J
§ S
o o
kl kl
01 01
X C
bl M
B a
kl kl
01 01
E <•>
-5 0
ki O
a u
ki
01 01
£ c
ki TJ 0
01 C kl
JJ 0 OI
0
X
01
£ &
0) «
- a
u o
n M kl
ai a> oi
U) U) U
0>
c
kl
c
a>
*o
m
• tn
n o
u a
01 ~r
0) JZ
W M
m m r- r-
n r*- r*\ r-
r4 rl r+ *+
.H *-4 f* >-l
in in tn m
CD CD CD CD
to a
a oi
0 0
0) 0>
n 09
T) TJ
c c
0 c
n n
a> a*
n »
c c
n n
u u c c
* 01 0 "0
u u a a
jj jj jj AJ
£c c c
V « 9
o o o o
n « • n
O O O 0
"u *u "u "u
v v a* v
X C X C
w M « £
kl -
a kt
Ol 01
u <
a>
a) ta
c _c
jS c 2
jj ^ ta
C JJ
Ol « TJ
> 0 C
^ U *3
o
w
•
10
u
n m
a, ai
"« 1o TJ ^
01 01 C 1
m m "Q * !
» & 3 kl I
c c cr * a
*j *>-i « a -< c
§•0 n o u a
0) O 0) ki
W 3 k! £ - - JJ
& a & - c c i
X 3 > W W
O
c
kt
I O JJ JJ
I C C C
n os u • oi
oi c a> ja -* *
** v * jj o a
0 > *s c n n
0) ^ 0) JJ U C
n o ta c > - *
n ai -•< « ki a>
a 01 > o 3 jj c
c -c^no»-^ki
-H JJ — O « E 0
IM c *j m -aoijQ
o 01 o « o a jj
o E o • » c
Cli*^Q'BB01W>
kt&k.4J-^CC^
2 3 > « CO
ki
i-i
U
n n
0) 0>
•0 « ^ -
c JJ « n
a c o> 01
• ai n m
0) - ki
&U fi Ot 01
« Q. C c
a -i c *j iu
O U 0 O . O
. ki o 01 O
- a > tj u 3 k*
5^:^ § s as
w w co X 2
01
c
O JJ i
i- oi tn n
jj 9 n v oi
« c Q, ^H n ^
> b a 0 V 0
O C « JJ 0)
- ,Q 0 n 0 n
0) ki kt S
& JJ — 0) C - C
0 0 E C -H JJ —
0 « O 0 Oi O
j2 o e o
m ki m at Q, ^ t^
caiCJ-'kTQikiOi
w w w 2 2
Ch 00
^ 00
*-» <-4
in m
00 OD
0 n
Oi pi
n C
TJ JJ
C (0
0 0
o
n
c u
•«H Q)
0 JJ
n 2
a> o
> 0)
JJ «
kf e ki
-H <-H
01 01 01
JJ £ Q
w O
n
c
0
ki 01
0) TJ
** C
U 0
01 01
ki C
o -H a
A JJ 3
ki 0 O
0) O ki
JJ U Ol
2
0)
T
(
01
• o
tn — i
0)
«c .
•si
5S
01
kl
o
01
*J
;
n
n
c
>
0)
kl
<
i
i
n
ki
0
I
J
s
IT
CO
CJ
kl
&
U
rH
10
kl
3
jj
U
0)
4J
U
kl
JJ
kl
0)
&*
u
i - Ol
0 <0 &
-» 0) 0
* -H a
ja u o
tn n C tn
SOi 01 0>
_^ _> ki >
jj JJ U jJ
0 0 B 0
> > C >
kl kl 0 kl
0) 01 k) 0>
H} a jj n
a> T) a -« oi
ki C U E M
a s a « o.
1 I'S^ 1
i i . £
i a>
,2 J g1
oi > a
J3 - 0
u •a
a o
a > o n
0) U > 01
> 01 >
-H B U 01 -.
u 0) C 01 u
V kl O t» 0
> a ki -H >
kl 0 4J kl
oi T a 0 «
» o a > a
* TJ O C ki 01
kl C 3 0 V kl
as ki a a
0 kl *J 01
TJ ki 0 -H ki 13
o 01 01 E a o
o — * » o
X u in 3
CT*
tn
en
(M
a
TJ TJ
0 C
- O 3
K J C
jj ^
3 k. O
TJ 3 O
O a
ki a oi
a oi c
u a N
ceo
•« O -"
0 C Ol
a
- T)
TJ B C
01 kl 0
~* 01
-i C -
-< 0 >•
0 01 U
-t U
m u 3
3 a
O £
0) a -
c 3 a
« k, 01
-i .0 _>
1 CT u
u j: 0
-~ TJ k.
65 S .
kl (J 01 U
01 C ki u
c •* a 01
o
a
j;
^j
«
4)
0)
h. a
a 3
Q
•O k,
0 Oi
tn
c
I
c
u
I
U
•^
•u
ki
I-
n
c
i
u
lu
ki
t-
S
tH
in
CO
n
C
a
c
i
u
IW
0
ki
t-
0>
c
Jt
w
a
•i! c
IW «]
kl
•o
0)
c
• H
XJ
O
u
*"*"
aJ
i— i
(0
u
«-i
o
T3
C
10
V)
(U
4J
0
c
-------
Q
W
P
^H
FERENCE (CONTI
W
K
t
W
W
o
ry
H*
U
EH
CO
t*S
K
JEJ
^^
EH
O
»-^
i— >
Q
O
OS
0,
i
T"H
«:
u
ij
CQ
g
•o
Table of standards propose
regulation limits category
>
VOC emission inventor
survey category
Hi
f 0 ?,
o |°
X
^
o
0.
4)
u
«
u
B)
a
i
u
w
o
d
«i
14
r«
*
1,
4.
n
—
c
u
c
•
IH
x
4,
C
i
i
i
4
4
A
I
l
a
in w
tji D>
c n c n
•^ -O -H V-
jj C jJ 41
RJ W 3 W iQ En
O W UOT3«O -HOI
U 0) Ol 01 d C Dl O IMC
pi C u£3C n D. -^
TJC--^ flOO-HCOi *-»
4) -H *J >MUQ.UOC£iQ
OT 4J *J * U E «•**—• « O
O\ C Q O 3DiOO«4J(BU
c o» o u tn c u u i-i c S
••-leu w 01 -^ *o ^*n
4J& -^bWM«a)EU4-*O&
« ^. Cn 4J flJ 3«Jjg COC
o cxc^tXTJU Qt-i-^ki at n -^
u -K *j * M oa o E *J
{j ^^ **-• fti « ® •"• *O -^ *J W «
&^3*UO^fU4>O«CO
O<-4O U £ 13 0))M-4iJU *» -H U
tM^ucn JC »- JS O -^ n E
iO ^ — « T3 -« U E & «J J-> *J E S
> *j *J i-> C * C >- « Ci — 4) -^ O
W«CCO£OOU£3U}^
Dr-2ia3uou.o-a:a.coti.
m
? n n
C 01 w)
•r* TJ ^*
4J C 0)
« 3 • E a
O 01 o t3 n — » o>
u cn a C en ^ c
C E 3 C 81 ft-H
TJ —^ O O -H O> *J
nu m e « -H B o
O>C O t?i O 0 AJ «J U
co) u c u u o a
^ E -H O -H
*j tji -^ ki m U4JO
(0 -^ 4J 3 AJ CO
O O. ••-< U kJ ^ 4> O.
u >w a o E
tl *M 0 "H -^ *J Ol
Cn-H ,AJ AJ C * -.It-H
tn 41 C OOV4 31x3
ax 3. u u. o z ix w
o
0
r>
s
J<
0)
c
0)
O)
c
-r4
AJ
01
> c c *
-r* w< v4 Ct
*J iJ *J C
a a e -*
•* o o «J
: ° ° §
U « *J O
•-*. u C
AJ 3 0 «
C 4J E w
« a *> 3
Tl fc< U AJ
i* u ai x
c w a * *
W C E *J AJ
« •** 0) •**
h< AJ AJ n o
i « di -H
a o £ c AJ
u u a> o> m
••* -^ « «
b. XXX
in
o
n
in
OD
r<
•o
c
«
c
o
o
3
14
AJ
C
0
(J
J
0)
Q c n
AJ ^H AJ
C «J C
a b a
f± u Q
9 3 V
u -a u
£55
C C
« U 0)
WO*-1
C — t C
~H k« -«
S * S
E AJ E
X
u
V
u
c
(0
c
0)
u a
C 3
-H O
« W
1 &
n
ai
c
Sanding sealers
Thermoplastic rubber coati
Sanding sealers
and maitici
Industrial maintenance
Industrial maintenance
coatings
Heat reactive coat ings
coatings
Appurtenances
j=
AJ
o
£
u
ft)
TJ
C
3
TJ
O
JJ
to
n
•r4
u
u
3
TJ
O
l-i
a
A)
j=
^
c
o
JJ
fi
JJ
c
o>
n
ft)
u
U
W
o
•u
.,
£
>
Oi
n
^-t
n
a)
c
u
£
«
jj
c
»
•0
C
a
4J
>
p-t
0
n
£
U
O
J3
»4
«M
iJ
^4
^
ft)
C
o
>.
•-4
g
O
.
p
f*
ft)
£
*J
c
M
•
. Government Printing Office
u
Q
^
c
O
roducts, 1991 . Washingt
&
•o
0)
<
TJ
c
«
y
c
Kj
04
8)
0
a
V
{£
.-t
«
b4
iJ
n
3
T)
C
M
jJ
c
9
W
V4
3
U
0
U
iM
e
1
5
*0
jj
c
ft)
E
u
w
a
Q,
^
o
en
5
••
w
U
W
3
O
n
Prepared for the National
»ft report. 1993.
>. £
«TJ
W «^
3 a
CO C
VOC Emissions Inventory
tion Industry Caucus. F
n c
o> -^
c ^
— • o
AJ en
n o
o z
U
>,
0) U
o o
c u
*J C
w ^-*
3 3
V) O
o
o) o:
u
c x
•3 IH
C <
C £
•^ 4J
££
4J
rH — ^
« S
W C
U 0
n — •
3 *J
TJ «
C ^
»-< ft)
a
TJ O
C 0
« u
«-» c
« -H
1-
3 C
JJ O
O -~
a> *->
«j «
u o
tx B
»> C «
rH *-t Ow
A-4
-------
The mapping in Table A-l proceeds from the most
aggregated category to the least aggregated category. In some
cases, however, the survey provides more detail than the TOS.
Where possible, the market segments were paired as solvent-
borne and waterborne coating categories. Separate market
segments could not be created for flat and nonflat coatings in
the interior and exterior segments because the Census data do
not differentiate between exterior flats and nonflats.
The necessary data were developed for each of the 13
market segments using the mapping scheme presented in Table
A-l. Data for individual Census product codes were summed
where necessary to compute prices and quantities.
A. 2 ESTIMATING DEMAND ELASTICITIES FOR COATINGS
To perform the market analysis, own- and cross-price
elasticities of demand were estimated for four broad coating
categories: exterior solventborne and interior solventborne
and their two respective substitutes, exterior waterborne and
interior waterborne. The variables used in estimation are
domestic consumption quantity; real value of domestic
consumption; real consumption price; national income; a
housing variable; and the real price of alkyd resins, acrylic
resins, and titanium dioxide. Complete data for these
variables were collected for the years 1981 through 1991.
Justification of these variables and their data sources is
given below.
A.2.1 Estimation Procedure and Results
Econometric estimation of the interrelated demand system
for interior solventborne, interior waterborne, exterior
A-5
-------
solventborne, and exterior waterborne architectural coatings
generated estimates of own-price demand elasticities for each
of the four groups and cross-price demand elasticities between
the solventborne and waterborne segments of each interior
(exterior) pair.
The quantity demanded of a commodity is a function of its
price, the price of any substitutes and other factors, such as
income, that affect aggregate demand. Estimating the demand
function, however, is more complicated than just running
regressions of observed market quantities on observed market
prices and other demand variables. One must account for the
fact that the observed prices and quantities are equilibrium
values, which are simultaneously determined by both demand and
supply factors.
Variables that are determined within a system (such as
prices and quantities in a market equilibrium system) are
endogenous to that system, whereas those variables determined
outside of the particular system (e.g., income, housing
activity) are termed exogenous. In simultaneous equations
models, endogenous variables are correlated with the error
terms through solution of the system. As a result of the
interdependence of the endogenous variables and the error
terms, the application of standard regression techniques is
modified to estimate the effect of an endogenous right-hand
side variable (i.e., equilibrium price) on the endogenous
left-hand dependent variable (equilibrium quantity) . In
general, ordinary least squares estimation of the individual
demand equations leads to biased and inconsistent parameter
estimates when a regressor is endogenous.
A-6
-------
Endogeneity bias is corrected by applying the two-stage
least squares (2SLS) regression procedure for each estimated
equation (see, for example, Reference 5). In the first stage
of the 2SLS method, the price observations were regressed
against all exogenous demand and supply variables in the
system. This regression produced fitted (predicted) values
for the price variables that are, by definition, highly
correlated with the true endogenous variable (the observed
equilibrium price) and uncorrelated with the error term. In
the second stage, these fitted values were employed as
observations of the right-hand side price variables in the
demand equations. This procedure can also be used to estimate
the underlying structural supply equations; however, because
of the poor performance of various specifications in the
supply estimations, only demand estimates are reported here.
The 2SLS procedure was used to estimate the four demand
functions. Both linear and double-log regressions were
estimated. The double-log specifications are presented here
because of slightly better statistical fit and because the
parameter estimates are directly interpretable as point
elasticities.
For the two exterior categories, housing completions are
included as an exogenous demand determinant. Exogenous supply
factors incorporated into the first-stage regressions include
the prices of various raw material inputs and a price index
for substitute outputs, which captures the effect of non-
exterior coatings prices on the supply of exterior coatings.
For the two interior categories, U.S. domestic GNP is included
as a proxy for the exogenous effect of aggregate income on the
demand for interior coatings. Exogenous supply factors
A-7
-------
incorporated into the first-stage regressions also include the
prices of various raw material inputs and a price index for
substitute outputs, which in this case captures the effect of
noninterior coatings prices on the supply of interior
coatings. The results of the demand estimations are shown in
Table A-2.
Unfortunately, sufficient data to estimate the demand
parameters for the other market segments were unavailable.
For the other two solvent/water-paired segments—clear coatings
and primers/undercoaters—the mean of the respective own- and
cross-price elasticities from the interior and exterior
estimation process were used as proxies for the elasticities.
The other five segments—special purpose, industrial
maintenance group, traffic marking paints, lacquers, and wood
preservatives—are specialty groups whose demand is assumed to
be fairly inelastic and not dependent on prices in the other
segments. Therefore, a value of -0.5 for the own-price
demand elasticity and zero for all cross-price elasticities
were assigned to each of these categories. Table A-3 provides
the matrix of own- and cross-price elasticities for all 13
market segments.
A.2.2 Data Used in Demand Estimation
Domestic consumption quantities and values were
calculated using data from U.S. Department of Commerce
publications Current Industrial Reports: Paint and Allied
Products6"14 and U.S. Exports Schedule B Commodity by
Country.15"22 Domestic quantity and value of shipments figures
were used, which include exports. Exports were then
subtracted to estimate domestic consumption (AIM coatings
imports are negligible and are not included in the
A-8
-------
TABLE A-2. DEMAND CURVE ESTIMATES
Variable
Exterior solventborne
demand
Log-housing
completions
Log exterior
solventborne price
Log exterior
waterborne price
Exterior waterborne
demand
Log-housing
completions
Adjustable Elasticity
R2 F-value estimate
0.94 50.52
0.17
-1.43
0.20
0.92 39.36
-0.05
t-statistic
3.30
-1.89
0.36
-0.62
Log exterior 0.51 0.42
solventborne price
Log exterior -1.89 -2.17
waterborne price
Interior solventborne 0.69 8.49
demand
Log GNP
Log interior
solventborne price
Log interior
waterborne price
Interior waterborne 0.99
demand
Log GNP
Log interior
solventborne price
Log interior
waterborne price
1
-1
1
588.90
1
0
-1
.01
.50
.43
.00
.36
.39
1
-1
1
5
1
-3
.67
.74
.28
.07
.28
.80
A-9
-------
w
EH
U
O
W
W
W
a!
•?
g
c/3
0
s
M
EH
O
U
y
H
rf»
c:
o
XJ
H
E-J
s
^4
r .
H
o
r
rrt
Li-<
^
w
D
2
&
m
D
,
ro
^*<
TABLE J
4J
U
1
a
o
01
o
•H
a
c
•H
01
D)
«
J=
u
01
DI
-------
consumption variable). Consumer price indexes from the U.S.
Department of Labor's Handbook of Labor Statistics23 and the
U.S. Department of Commerce's Survey of Current Business24"26
are used to adjust the current figures to real values. Real
consumption price was imputed for each product by dividing
real value of domestic consumption by the quantity of domestic
consumption.
The GNP in constant 1987 dollars from 1981 through 1991
was used as an aggregate income measure.a'27"29 Housing
completions for 1981 through 1991 were obtained from the U.S.
Department of Commerce's Current Construction Reports.30
Prices for alkyd and acrylic resins are obtained from the U.S.
International Trade Commission publication Synthetic Organic
Chemicals, U.S. Production and Sales.31"40 Prices for titanium
dioxide were imputed using quantity and value of shipment data
for U.S. production from the Current Industrial Reports,
Inorganic Chemicals.41 Real prices for these raw materials
were calculated by deflating normal values using CPIs. Alkyd
and acrylic resins were used to represent raw materials for
the nonvolatile vehicle portion of the coatings, which are
found mainly in solventborne and waterborne coatings,
respectively. Titanium dioxide was used to represent a raw
material in the pigment portion of the coating, which is found
in both types of coatings. A Laspeyres price index was
constructed to incorporate the price of substitute outputs as
a supply-side effect in the first stage regressions of the
2SLS procedure. Let the price and quantity of commodity n in
aAll constant values were converted to 1982-1984 dollars for the
analysis to be consistent with the consumer price index (CPI), which has
1982-1984 as a base.
A-ll
-------
period t be p^ and qnc, respectively for n = 1, ..., N and
t = 0, 1, ...,T. Then the Laspeyres price index of the N
commodities for period t (relative to the base period 0) is
defined as
N I N
V* t 0 v^ 0 0
PL 5 2L, Pn ^n I JL Pn ^n • (A-l)
N=l I N=l
Real domestic prices and quantities of nonexterior
coatings were used to construct the price index for the
exterior coatings equations and real domestic prices, and
quantities of noninterior coatings were used to construct the
index for the interior coatings equations. Each index is
computed for the years 1981 through 1991, with 1981 serving as
the base year.
A.3 EVALUATION OF DATA QUALITY
The Current Industrial Report series is generally
considered a reliable source for quantities and values of
products shipped. Monthly and annual data were estimated from
a sample designed to measure activities of the entire paints
and allied products industries. Each annual report provides
data for 2 years, and the most recent figures were used for
the coatings analyses. In addition to the four representative
coatings products, the architectural coatings Census category
includes two other products: architectural lacquers and
architectural coatings, not elsewhere classified. These
categories were not included in the estimates because of
A-12
-------
insufficient data. However in 1991, these two product
categories combined represented only 1.3 percent of the total
value of shipments for the architectural coatings market.4
Statistics reported in the Current Industrial Reports at the
seven-digit SIC product level are based on Annual Surveys of
Manufactures and represent about 95 percent of total shipments
;in the paint industry (SIC 2851)-43 To produce estimates for
the entire industry, the Census Bureau inflates the quantity
and value figures reported in the annual survey by a factor
based on data reported by all establishments in the 1987
Census of Manufactures.13 The inflation factors for
architectural coating product categories are as follows: 1987
through 1991, 1.00; 1982 through 1986, 1.004; and for 1981,
1.04.c Quantity and value figures for the four product
categories used in the demand estimation are inflated using
these factors. Prior to 1981, data were not collected at the
more specific seven-digit SIC level. Using the longer time
series would provide more data points but would also preclude
analysis of the individual product categories, and
representativeness would be lost.
The export data used are the best publicly available;
however, combining export and domestic data to estimate
domestic consumption poses some problems. The classification
systems used to gather both types of data are different, and
the corresponding product categories used cannot always be
''The inflation factor for 1981 is based on 1977 Census relationships
and for 1982 through 1986 on 1982 Census relationships.
cThe 1991 quantities and values used in the model (values to impute
price) also include products in the special purpose and miscellaneous
allied paint products categories. The special purpose inflator for 1991 is
1.06, and the miscellaneous inflator in 1991 is 1.18.
A-13
-------
compared. For example, data from the U.S. Department of
Commerce publication U.S. Imports for Consumption and General
Imports, TSUSA Commodity by Country of Origin were not used
because the imported commodity classifications had no
comparable domestic output classification. Exclusion of
imports from the estimate of domestic consumption does not
pose a problem because in 1991 the value of imports for
architectural, OEM, and special purpose coatings (SIC 28511,
28512, 28513) combined represented less than 0.9 percent of
the total domestic value of shipments.44 Data from U.S.
Exports Schedule B Commodity by Country were available for
1981 through 1991, and the export categories correspond well
with the four domestic product categories except for 1989
through 1991.45~52 In 1989, the export codes and categories
changed and are no longer compatible with the domestic
categories. In addition, quantities are reported in kilograms
rather than gallons, as they were in previous years. For
these reasons, export data were not used to adjust domestic
consumption after 1988. The GNP data typically represent
income for the entire nation including income generated from
American businesses located overseas. The current price data
for the paint products and raw materials should be considered
reliable, though their accuracy may be affected by the
exclusion of imports for the coatings products and of exports
and imports for the raw materials prices. CPIs for all urban
consumers with a base of 1982 through 1984 were used in
calculating real prices.
The raw material prices used are representative of the
entire U.S. and export market for these products, rather than
just the U.S. supply to the paints and coatings industry. The
A-14
-------
alkyd resins were used in this estimation to represent an
input found only in solventborne coatings and acrylic resins
to represent an input found only in waterborne coatings.
However, some acrylic resin materials are used in some
solventborne coatings and alkyd resins are used as modifiers
in waterbornes. Exports and imports were not considered when
computing raw material supply prices because foreign trade
data were not available for alkyd and acrylic resins. In
1991, exports of titanium dioxide represented 17.9 percent of
the total domestic value shipped and imports were 10.9
percent.53
A.4 REFERENCES
1. U.S. Department of Commerce. Current Industrial Reports:
Paint and Allied Products, 1991. Washington, DC: U.S.
Government Printing Office. 1992.
2. Industry Insights. Architectural and Industrial
Maintenance Surface Coatings VOC Emissions Inventory-
Survey . Prepared for the National Paint and Coatings
Association in cooperation with the AIM Regulatory
Negotiation Industry Caucus. Final draft report. 1993.
3. Ref. 1.
4. Ref. 2.
5. Pindyck, Robert S., and Daniel L. Rubinfield.
Econometric Models and Economic Forecasts. 2nd Ed. New
York, McGraw-Hill, Inc. 1981.
6. U.S. Department of Commerce. Current Industrial Reports:
Paints and Allied Products, 1982. Washington, DC: U.S.
Government Printing Office. 1983.
A-15
-------
7. U.S. Department of Commerce. Current Industrial Reports
Paints and Allied Products, 1983. Washington, DC: U.S.
Government Printing Office. 1984.
8. U.S. Department of Commerce. Current Industrial Reports
Paints and Allied Products, 1984. Washington, DC: U.S.
Government Printing Office. 1985.
9. U.S. Department of Commerce. Current Industrial Reports
Paints and Allied Products, 1985. Washington, DC: U.S.
Government Printing Office. 1986.
10. U.S. Department of Commerce. Current Industrial Reports:
Paints and Allied Products, 1986. Washington, DC: U.S.
Government Printing Office. 1987.
11. U.S. Department of Commerce. Current Industrial Reports;
Paints and Allied Products, 1987. Washington, DC: U.S.
Government Printing Office. 1988.
12. U.S. Department of Commerce. Current Industrial Reports:
Paints and Allied Products, 1988. Washington, DC: U.S.
Government Printing Office. 1989.
13. U.S. Department of Commerce. Current Industrial Reports:
Paints and Allied Products, 1989. Washington, DC: U.S.
Government Printing Office. 1990.
14. U.S. Department of Commerce. Current Industrial Reports:
Paints and Allied Products, 1990. Washington, DC: U.S.
Government Printing Office. 1991.
15. U.S. Department of Commerce. U.S. Exports Schedule B
Commodity by Country. FT446/Annual 1981. 1982.
16. U.S. Department of Commerce. U.S. Exports Schedule B
Commodity by Country. FT446/Annual 1982. 1983.
17. U.S. Department of Commerce. U.S. Exports Schedule B
Commodity by Country. FT446/Annual 1983. 1984.
18. U.S. Department of Commerce. U.S. Exports Schedule B
Commodity by Country. FT446/Annual 1984. 1985.
A-16
-------
19. U.S. Department of Commerce. U.S. Exports Schedule E
Commodity by Country. FT446/Annual 1985. 1986.
20. U.S. Department of Commerce. U.S. Exports Schedule B
Commodity by Country. FT446/Annual 1986. 1987.
21. U.S. Department of Commerce. U.S. Exports Schedule E
Commodity by Country. FT446/Annual 1987. 1988.
22. U.S. Department of Commerce. U.S. Exports Schedule E
Commodity by Country. FT446/Annual 1988. 1989.
23. U.S. Department of Labor. Handbook of Labor Statistics.
Bulletin 2340. 1989.
24. U.S. Department of Commerce. Survey of Current Business.
v. 70, no. 7. July 1990.
25. U.S. Department of Commerce. Survey of Current Business.
v. 72, no. 7. July 1992.
26. U.S. Department of Commerce. Survey of Current Business.
v. 72, no. 12. December 1992.
27. Ref. 24.
28. Ref. 25.
29. Ref. 26.
30. U.S. Department of Commerce. Current Construction
Reports, Housing Completions, November 1992. Washington,
DC: U.S. Government Printing Office. 1993.
31. U.S. International Trade Commission. Synthetic Organic
Chemicals, U.S. Production and Sales, 1981. USITC
Publication 2470. 1982.
32. U.S. International Trade Commission. Synthetic Organic
Chemicals, U.S. Production and Sales, 1982. USITC
Publication 2470. 1983.
A-17
-------
33. U.S. International Trade Commission. Synthetic Organic
Chemicals, U.S. Production and Sales, 1983. USITC
Publication 2470. 1984.
34. U.S. International Trade Commission. Synthetic Organic
Chemicals, U.S. Production and Sales, 1984. USITC
Publication 2470. 1985.
35. U.S. International Trade Commission. Synthetic Organic
Chemicals, U.S. Production and Sales, 1985. USITC
Publication 2470. 1986.
36. U.S. International Trade Commission. Synthetic Organic
Chemicals, U.S. Production and Sales, 1986. USITC
Publication 2470. 1987.
37. U.S. International Trade Commission. Synthetic Organic
Chemicals, U.S. Production and Sales, 1987. USITC
Publication 2470. 1988.
38. U.S. International Trade Commission. Synthetic Organic
Chemicals, U.S. Production and Sales, 1988. USITC
Publication 2470. 1989.
39. U.S. International Trade Commission. Synthetic Organic
Chemicals, U.S. Production and Sales, 1989. USITC
Publication 2470. 1990.
40. U.S. International Trade Commission. Synthetic Organic
Chemicals, U.S. Production and Sales, 1990. USITC
Publication 2470. 1991.
41. U.S. Department of Commerce. Current Industrial Reports
Inorganic Chemicals, 1991. Washington, DC: U.S.
Government Printing Office. 1992.
42. Ref. 1.
43. Ref. 1.
44. Ref. 1.
45. Ref. 15
A-18
-------
46. Ref. 16
47. Ref. 17
48. Ref. 18
49. Ref. 19
50. Ref. 20
51. Ref. 21
52. Ref. 22
53. Ref. 41
A-19
-------
APPENDIX B
METHODOLOGY FOR COMPUTING MARKET AND
WELFARE ADJUSTMENTS
-------
B.I METHODOLOGY FOR COMPUTING SUPPLY EFFECTS
For the purposes of modeling the regulatory effects in
each market, producers are separated into four categories,
based on their response to the regulation:
• producers of products slated for withdrawal,
• producers of products on which exceedance fees are
paid,
• producers of products slated for reformulation, and
• producers unconstrained by the regulation.
The baseline (preregulatory) quantities from these groups
are denoted as follows: Qx, QF, QR, and Qu for groups 1, 2, 3,
and 4, respectively. Total baseline market output equals the
sum of the four components:
Q = QX + QF + QR + QU. (B-l)
Figure B-l depicts the aggregation of these subgroups
into a market supply function. The regulation causes a shift
in the aggregate supply function depicted in Figure B-l as a
result of two phenomena: an inward supply shift due to
eliminating Group 1 through product withdrawals (e.g., the
shift from S° to S1) , and an upward supply shift due to
imposing per-unit fees on the products from Group 2 (the shift
from S1 to S1') . There is no supply shift emanating from Group
3 because the reformulation is assumed not to affect marginal
production costs, and there is no shift from Group 4 because
the unconstrained producers experience no regulation-induced
B-l
-------
S/Q
P1
P0
Q'
Figure B-l. Single market effects of VOC content regulation.
change in their cost structure. So the full regulation-
related shift is from S° to S1', which leads to a new market
equilibrium. At the new equilibrium, price rises to P' and
quantity falls to Q'.a
B.2 DEMAND EFFECTS
Figure B-l depicts a partial equilibrium view of the
short-run effect of imposing content limits in one market
graphical analysis demonstrates that the post-regulatory market
effects are uncertain if we were to consider the possibility that the
reformulation process changes the marginal cost of producing the coating as
a result of changes in material or labor costs, for example. This
empirical issue can be resolved given sufficient data on the effect of VOC
content on production costs for all affected products. Unfortunately,
these data were not available for this study, and we cannot conduct the
appropriate empirical analysis to draw such conclusions.
B-2
-------
One must also consider the role of substitute products in
determining the equilibrium adjustments, which suggests a
multimarket perspective. Figure B-2 depicts the markets for
two products (A and B) that are demand substitutes. The price
of product B factors into product A's demand function and vice
versa:
DA = DA(PA, PB) (B-2)
DB = DB(PB, PA). (B-3)
Given that A and B are substitutes implies
6DA / 6PB > 0 (B-4)
6DB / 6PA > 0 . (B-5)
Suppose the supply of A is affected by the content limits
in the manner described above, but that the supply of B is
unaffected. This initiates a supply shift in market A from SA°
to SAR. Holding the initial demand function constant, this
shift would generate an equilibrium quantity of QA" and price
of PA". However, the associated price increase in market A
induces an outward shift in the demand for product B, which
raises the price of product B. Likewise, the increase in B's
price leads to an outward shift in the demand for product A,
which raises its price and so on. This interaction continues
until post-regulatory equilibrium is established at (PAR, QAR) .
(pBR' O •
B.3 COMPUTING CHANGES IN EQUILIBRIUM PRICES AND QUANTITIES
The change in equilibrium prices and quantities for the
products affected by the content limits and their substitutes
can be numerically computed by adjusting the equations in the
multimarket supply and demand system to reflect the imposition
B-3
-------
$/Q
? (Pg)
QA/t
S/Q
Product A
B
B
Dg (Pj)
QB QB
Product B
Figure B-2. Multiple market effects of VOC regulations
B-4
-------
of these limits. For each market, i, the equilibrium change
in quantity supplied of each product affected by the
regulations equals the sum of the supply changes from each of
the producer subgroups :
. (B-6)
The change (from baseline) in quantity supplied by the
withdrawal sector is simply the negative of the quantity
originally supplied by that group:
= - QiX. (B-7)
The change in quantity supplied from the fee-paying
sector is specified as follows:
AQiF = eJ(QiF/Pi) (APr FJ (B-8)
where e^ is the supply elasticity of the fee producers in
market i, APi is the change in equilibrium market price, and
other terms are as previously defined (without the
subscripts) . APi- F^ is the change in "net price" for the
fee-paying producers (i.e., the change in unit process less
the unit fee) .
The changes in quantity supplied from the reformulating
group and unconstrained groups, respectively, are
(B-9)
iU = elu(Qiu/P)APi. (B-10)
These producers respond to the increase in price with no
counteracting effect on costs. Given the higher price in the
post-regulatory equilibrium, output will increase from these
two groups of producers.
The aggregate change in equilibrium supply quantity can
now be restated by combining the preceding five equations:
B-5
-------
VFi> + eiR(QiR/Pi)APi
+ eui(Qui/P)iAPi . (B-ll)
The change in market demand for each product is given by
AQ,D = E,; (Q. / PJAP, + E,.j(Q, / PS)AP, (B-12)
1 J.A 1 11 1J 1 J J
where Eii is the own-price demand elasticity for product, i and
EAj is the associated cross-price demand elasticity between
products i and j. Consumer demand theory supports the
assertion that own-price elasticities are negative and that
cross-price elasticities of substitutes are positive. To
attain equilibrium, the change in quantity demanded must equal
the change in quantity supplied in both markets:
AQ^ = AC^3. (B-13)
This provides a system of M*3 equations in M*3 unknowns,
where M equals the number of markets affected by the
regulation. This can be reduced to an M*2 equation system,
simply by substituting AQiD=AQis=AQi. This system can be
solved simultaneously to compute the change in equilibrium
price and change in equilibrium quantity for each market. To
do this, baseline market data, model parameters (supply and
demand elasticities), and an empirical characterization of the
various supply shocks alluded to above are needed.
B.4 COMPUTING WELFARE EFFECTS
Changes in the market equilibrium cause changes in
resource allocation, which, when quantified, provide measures
of how the welfare costs of the regulation are distributed
across groups affected by the regulation. The groups focused
upon here are AIM coatings producers and consumers, because
the changes in prices and quantities directly affect their
B-6
-------
welfare. Since fee payments are considered, the government
sector is also included in the welfare analysis because they
collect the fee revenues. This study does not measure the
welfare benefits of reductions in VOC emissions, a value
against which these costs may be measured to determine the net
value to society of the proposed regulatory structure.
B. 3.1 Effects on AIM Coatings Producers
The profits earned at the new equilibrium to the profits
earned at the old equilibrium can be compared as a measure of
effects of the regulation on the individual producer. Forgone
baseline profits (7t°) provide a measure of the loss to
producers that choose to exit rather than reformulate:
A* = *R* - n° = -ic°. (B-14)
For the remaining producers, the change in profits is affected
by several factors, including the incurrence of the fixed
reformulation cost and any associated changes in price,
quantity, and marginal cost.
The remaining firms' costs may be affected through either
the reformulation cost or the fee payment. The effect of the
content limit on producers is generally not uniform and thus
raises some distributional considerations. As indicated
above, shifts in the aggregate supply function will cause the
market price to rise. For some producers, the benefits of the
price increase may outweigh the net costs of compliance. This
is certainly the case for producers of coatings with VOC
content below the regulatory standards, because they incur no
reformulation costs but would gain from the rise in market
price sparked by the compliance costs and/or product
withdrawals incurred by their competitors. Alternatively,
fixed reformulation costs may be substantial for some
producers, outweighing the positive price effect. The profit
effect will be negative for those producers. Other producers
B-7
-------
may fall in the midrange, where the price benefits and cost
effects essentially offset each other.
Changes in producer welfare are generally reported as
changes in producer surplus. The aggregate change in producer
surplus for the withdrawn-product producers equals the sum of
forgone profits from all withdrawn products in market i:
N*
APS,* = -£ n.. . (B-15)
j-i
The j subscript indicates forgone profits from the j ' th
product in market i. N/1 is the number of withdrawn products
in market i . The change in producer surplus from the
reformulating sector can be approximated as follows:
Q/ + 0.5*AQiR*APi- (R^N) . (B-16)
APi is the change in equilibrium price, AQ/ is the change in
equilibrium quantity from the reformulating producers, QiR is
the initial quantity of the reformulating producers, Rac is the
annualized reformulation costs, and NiR is the number of
products needing reformulation.
The change in producer surplus for the fee-paying
producers is
APS/ = (APi-FJ* (Q/ + AQiF)-0.5*AQiF*(APi-Fi) . (B-17)
The first term reflects the net revenue effects of the price
rise less the fee payment and the second term reflects changes
in deadweight loss .
Finally, the change in producer surplus for unconstrained
producers is
APS/ = AP^Q/ + 0.5*AQiu*APi (B-18)
with the Qiu reflecting the quantity supplied by these pro-
B-8
-------
ducers . Total (net) producer surplus effects is simply the
sum of the terms above:
ApSi = APSiX + APS/ + PS/ + APS/. (B-19)
B.3.2 Effects on AIM Coatings Consumers
Changes in consumer welfare are measured by the change in
consumer surplus, which quantifies losses due to a combination
of the higher price and reduced consumption quantity. This
change can be approximated as follows :
i = -APi*(Qi + AQA) + 0.5*APi*AQi. (B-20)
B.3.3 Effects on the Government Sector
The transfer of fees from the fee-paying producers to the
recipient of those fees must be considered. For the purposes
of the welfare analysis, the government is identified as the
"recipient" of the fees.
AGSi = Fi*(QiF + AQ/) . (B-21)
Ultimately, the government may choose to redistribute
those fees back to affected producers or consumers or back to
other members of society via the Treasury; however, for
purposes of quantifying these distributional flows, they are
assigned as gains to the government sector.
B.3.4 Net Welfare Effects
The net welfare effects are computed by taking the sum of
producer, consumer, and government surplus:
Avrcv = APSi + AcSi + AGSi. (B-22)
This calculation nets out any transfers from one group to
another within society (e.g., transfers from consumers to
producers through higher prices and transfers of fee revenues
from producers to the government) because these transfers do
B-9
-------
not affect the total sum of resource costs, just how they are
distributed within society. AWFi provides an estimate of the
net social costs of the regulation.
B-10
-------
APPENDIX C.
VOLATILE ORGANIC COMPOUND CONTENT LEVELS AND EMISSION
REDUCTIONS FOR SELECT ARCHITECTURAL COATINGS
FROM THE 1990 SURVEY DATA'.
-------
r
Units are grams of VOC per liter
o
o
to
§'
.
5
c"
Q.
I
O.
ft
X
n
1
1
c
1
",-.
?
5
c.
o
ft
X
§"
c
3
5
5
(TO
S
6
tt
n
C-
_
v:
5-
a
c
f^
c
ft
cr
All emission reductions represent
3
a.
o
|"
i«
c
<
5
S"
jg
a
S
B
Unless otherwise specified, units i
the manufacturer.
™
5"
*z
|
o
<
R
"S
ft
C
-*>
E
5
S"
B
5
ft
T3
O
0
1
O
c
•3
O.
V.
"
1
r.
o"
to
B
C.
C.
ft
o.
0
a
y
80
S
trt
CD
ft
i
X
§'
i
?
5
2
3
00
rt
c
3
3
a
i.
«•
03
£
i
to
O
i
?r
1'
£
a
n
n
n
3.
?
3
v:
c
c.
ft
•8
o
a
n
o.
c
3
*•<
C.
to
C
en
»
a*
W
^
Er
B"
rt
to
c
3-
8
ft
c
3
ft
S
ft"
<
R
ft
a.
o
c
cr
tr
?- Crt
^ c.
:- B
; Sealers (other than Lacquer Sandi
TVvt.ir.. r.^tlo^
3
21
n
£•
^ ^
^
S o
£ §
5 0
3 —
i j
^^2DOOOt~"n<<'"'n
c «» ^ o ^j£ ~3 »?«""
^ -3 ^J'Ts ^S Pg.-3 "Z Q
!itl||||f 5!ll
^ifaKg^n^r y = m
"• £. 5 "
c i.
0^^1^**UJ*.0^^0*.O^^^J
ocSooooooooooo
-s-----s--«--
— |O UJ U> U)
O LA LA OO ^O
J^ N) \f\ "LA "so yj 1*J ~C> "sO ^_ io V. "to
O
*"" ^ *o "~
zr is -Z. TO v> "Z.
£_ 2. o C to C
= 3 5 x 5 s
=. S s o T f
Coalings - Interior
Opaque
oatings
Coatings - Exterior
> and Undercoaters
al Maintenance Coatings
^ NJ tO
ii. Ui UJ tO U> U»
o o o o o o
UJ — UJ -~J K> O^
4* -t»- 4^ LA LA f>
V) LA Vj Lu "it*, "ifi.
UJ -J •— \C O LA
O ^s* 00 oo u> to
"oo u> 1y\ »- oo ~oo
-ti *O -f^ O^
-O OO Ui . — UJ
-o u* o ij to —
O* to u* ^^ -^j *.
W5
IT
5
is-
Clear and Semltransparent
LA
LA
LA
O
LA
U>
£
S
O
OO
ta—
OO
-o
H
^,
o
2
ft
rr
I
?
5
tn
u>
o
0,
to
0
8
to
LA
^
0
c
5
re
o
tt
ft
re
C
HE >
^ • (r
iti
<
D
f
r
tc
€
s> C <^
1 s. a
Ui
n 5
^ a ^
j c. 5
r- C ^
"cr§ 3
O
A
m
? 1 D
o < £
D « «
—' D
ft
cr
0
m F
D m
og
d o
0 >
2 -2.
•n "" 0
o o g
H S ?
m m c
- 3 z
o H o
w g o
c n 5
» £ H
< 3 m
m tn Z
D c fn
> i P
I [
8>
> 2
H C
Is
w
O
2
^
-a
m
D
5?
-------
APPENDIX D:
Sensitivity Analysis of National Cost Estimate
-------
APPENDIX D:
Sensitivity Analysis of National Cost Estimate
In the Economic Impact Analysis, the total social cost is extrapolated from detailed
information received from a survey of 116 companies in the industry. The method of
extrapolation assumed the volume of the products not included in the survey and the companies
that produce them have similar characteristics as those in the survey. Because this does not
account for several of the companies in the industry that produce products with lower annual
volume relative to those of the survey, the analysis provides a lower bound of the actual
compliance costs that will result from the regulation. An alternative method of accounting for
nonsurveyed products and their associated reformulation costs is presented below to provide and
upper bound of total social cost of the regulation. The actual cost imposed on society is likely to
be between the upper and lower bounds presented by the two methodologies.
Current Method:
National reformulation costs are currently estimated in the Economic Impact Analysis by
computing reformulation cost for the survey populations and multiplying this number by the
ratio of national market volume in each of the 13 product categories to the volume of surveyed
products.
National Reformulation Cost = Reformulation Cost of Survey Products * (Market Volume by
category/Survey Volume by category)
This is completed for each AIM coating market segment and summed across all market segments
to get the industry total.
For this method to be an accurate measurement of national reformulation costs, the ratio
of market volume to survey volume must equal the ratio of the number of a market products to
the number of survey products. This is because reformulation costs are incurred on a per
product, rather than a per unit, basis. Therefore, the average product size of nonsurveyed
products must be the same as the average product size of surveyed products. With little
information on the nonsurveyed products, the analysis implicitly assumed that the surveyed
products are representative of all products that are produced in the industry and thereby used this
information to calculate the national estimate of social cost.
While the survey comprises about three-quarters of the total AIM coatings market
volume, the surveyed companies (116 in number) account for less than one-quarter of the more
than 500 companies reported to produce AIM coatings. As a result, the companies not
responding to the survey are likely to be small AIM coating producers, which would increase the
total number of nonsurveyed products subject to reformulation.
-------
Based on survey responses from small companies, the data indicate that average product
size for small company products is much smaller (about one-sixth the size) of the average size of
all products in the survey. If there exists a combination of the factors of: (a) the nonsurveyed
companies being small companies, and (b) small company products being lower in volume, then
the ratio of market volume to survey volume understates the ratio of the number of market
products to survey products. In other words, the survey potentially omits a relatively large
number of small products. In addition, it is expected that some surveyed companies combined
products to some extent for reporting purposes. The extent to which this occurred is not known,
but would tend to understate the total number of products. As a result, the methodology used in
the El A best represents a lower bound of national reformulation costs by not adequately
accounting for the number of nonsurveyed products. To determine an upper bound of costs, the
methodology can be modified to assume that all of the nonsurveyed products are small volume.
Since the EPA has recently received information that indicates that not all of the companies
omitted from the survey can be considered small since some large companies were also omitted,
this alternative measure of national cost would produce an upper bound of the estimate of social
cost.
Alternative Method
For each of the 13 defined market segments in the AIM coatings industry there is data on
total market volume derived from the Census of Manufacturers data for the baseline year (1991)
and the total volume of surveyed products for that category. From the data, the total volume that
is omitted from the survey is computed as follows:
Omitted volume = Market volume - Survey volume (by category)
If the average size of omitted products is known, the number of omitted products can be
estimated as follows.
Omitted products = Omitted volume/Average volume of an omitted product.
If the proportion of omitted products needing reformulation is known, then the number of omitted
product reformulations can be computed as:
Omitted product reformulations — Omitted products * Proportion of omitted products needing
reformulation,
and the corresponding reformulation costs are then:
Cost of Omitted Product Reformulations = Omitted product reformulations * Reformulation
cost per product.
-------
The national reformulation costs can then be computed as'
National Reformulation Costs = Survey Product Reformulation Cost + Omitted Product
Reformulation Cost.
Since there is not specific data on nonsurveyed products, the average product volume is not
available to directly determine the number of omitted products and thus the number of omitted
product reformulations needed. However, the information from the survey can be used to impute
the values for the nonsurveyed products. If the analysis assumes that: (1) the average size of the
omitted products in each market segment equals the average size of small company products
reported for that market segment in the survey, and (2) the omitted product reformulation rate is
each market segment equal the reformulation rate for small company products reported for that
market segment in the survey data1, then the estimation of national costs will increase due to an
increase in the number of product reformulations assumed to occur in the nonsurveyed
population
Table 1 presents the total number of reformulations resulting from this methodology and
the total "static" reformulations costs for the entire AIM coatings industry, which refers to the
cost resulting from the imposition of the reformulation costs on all products that exceed the limits
set by the regulation without consideration of the potential for products to be withdrawn from the
market. In Table 1, the ratio of estimated total AIM reformulations to total survey reformulations
is approximately 2:1, which is in contrast to the estimate derived from the original methodology,
based on total market volume to survey volume of 1.28.1.
The social welfare cost estimate is also presented in Table 1, which captures the potential
responses of the affected firms by allowing affected producers to either reformulate or withdraw a
product from the market. In this way, producers can minimize the impacts of the regulation. If it
is less costly to the firm to withdraw the product than it is to reformulate, the presumption is that
they will do so
Summary:
In summary, modifying the cost aggregation methodology to bound the costs for
nonsurveyed AIM products raises the social cost estimate for the regulation to about $38.9
million, which is 55% higher than the lower bound estimate provided by the original analysis of
the Economic Impact Analysis. These estimates derived largely from modifying the methodology
to explicitly estimate the costs incurred by the nonsurveyed AIM coatings producer, under the
assumption that the omitted products are products by small companies and possess characteristics
similar the small company products reported in the survey. Because reformulation costs are
incurred on a per product basis, rather than a per unit (volume) basis, the process of aggregating
1 In addition, there is also a potential for a greater number of the nonsurveyed
products to exceed given VOC limits since small volume products are often used in specialty
(niche) markets This bias would increase the estimate of national cost, however, since not all of
the nonsurveyed products are small volume products, or in specialty markets, it is likely that this
bias is captured within th euper and lower bound estimates provided in these analyses.
-------
the cost estimates from the survey population to the national population needs to be based on the
ratio of total products in the market to total products in the survey The aggregation as
previously done used the ratio of market quantity for each category to survey quantity as the
aggregation factor. Because the ratio of products exceeds the ratio of product quantities, the
previous cost estimate serves as a lower bound and the new estimate provides an upper bound of
national costs
Table 1. National Reformulation Costs under the
Alternative Methodology
No. Reform-
ulations for
Surveyed
Products
1,111
No. Reform-
ulations for
Nonsurveyed
Products
1,234
Total
Reformulations
2,345
Total Static
Reformulation
Cost
$41.6 million
Total Social
Cost
$38.9 million
-------