-------
3.4 Sunmaiy of Results
Tbc ranks assigned to each of the ten representative plants using the pollution prevention
metrics examined in the present study are shown in Table 3.7. The important points regarding
each approach and how each compares with other alternatives have been stated in previous
sections.. A few points are evident in Table 3.7:
•	PPF and R/T ranks correspond closely. Both take throughput into account in
assessing the efficiency of the chemical usage leading to the reported releases. In
two cases, plants A and B, PPF assigns better scores than R/T does. The
comparatively favorable PPF rank occurs because those two plants handle only
acute toxins, while many competing plants handle chemicals that are both acutely
toxic and VOG This difference reflects PPFs capability of dealing with the multi-
dimensional risks associated with many chemicals.
•	TRJ release based rants and TRJ ratio based ranks are neither correlated with
each other nor with PPF or R/T ranks. The two TRJ metrics are valuable for
assessing progress over time by aggregate sets of Facilities, e.g.. the trend in TRI
releases over time at New Jersey chemical facilities. However, not taking into
account the level of chemical usage at each facility limits the effectiveness of TRJ
release based inter-facility comparisons. Likewise, not taking level of chemical
usage into account means that TRI ratio comparisons are subject to base-year
limitations. This limitation results in Eras that have achieved substantial pollution
prevention advances prior to the base year not only receiving no credit for those
achievements, but also being penalized in the sense that their efforts to achieve
further improvements are likely to be more costly than efforts at firms that did
little prior to the base year.
3-28

-------


Table 3.7


COMPARATIVE POLLUTION PREVENTION RANKS OF SELECTED PLANTS
1991
Plant ID
PPF Rank
TRI ReJ Rank
TRI Ratio
Rank
Rel-to-Thpt
Rank
A
28
44
26
45 1
B
32
43
37
47 1
C
1
5
17
1
D
32
61
57
29
E
36
63
9
34
F
36
42
62
33
G
5
11
37
5
H
4
25
58
4
I
51
62
24
49
J
3
27
18
3
3-29

-------
4.0
PERSPECTIVE PROVIDED BY OTHER DATA
A limited dumber of publicly available sources of information provide perspective on the
pollution prevention and other environmental activities of the New Jersey chemical firms examined
in this study. Public data sources accessed with reasonable effort include: records of litigation
proceedings (dockets); published rankings, such as those of The Council on Economic Priorities
(CEP) and Fortune magazine; and local newspaper articles or feature stories. The difficulty in using
these types of sources is that the information tends to be sparse, sporadic, and focused on the larger
firms.
Two additional indicators of each firm's environmental performance can be obtained from
the CRTK survey data. The first indicator is determined by assessing the "apparent quality" of the
materials accounting data submitted. As described in Chapter 2, the chemical records retained for
analysis in the present study are those that passed a two-percent materials-accounting-balance
screening. By reviewing all the records submitted by each facility, an approximate gage of the quality
of the firm's reports can be inferred by determining what fraction of their total reports passed the
2-percent screening test, and the extent of materials accounting balance error in the non-acceptable
records. The second indicator of a firm's environmental performance that can be obtained from the
CRTK survey submissions is the extent to which waste generation has been reduced through source
reduction.1 Because of its year-to-year nature, this information is subject to base-year dependence,
but it does indicate whether the firm has given any priority to pollution prevention.
State records documenting the performance of facilities under their discharge and/or RCRA
permits represent another potential indication of firms' environmental performance. These records
are not kept in a centralized location, however, and are not computerized. Accessing the records
requires going to each of the New Jersey regional offices and searching through file cabinets of
written reports. In addition, the files for the air, water, and solid waste programs are maintained in
separate files within the regional offices, and each of these files must be accessed in order to arrive
at a holistic picture of the firm's performance. This level of effort is justified when a detailed
evaluation of a firm is undertaken, but involves too great an effort for the screening analysis
'Source reduction data collected via the CRTK survey include the share of reduction attributable
to each source reduction mode, e.g., materials substitution, including the quantity and identity of
chemical replacement; reformulation or redesign of product; process or procedure modifications;
equipment or technology modifications; improved operations; discontinuance of operations; export
of use; or miscellaneous. In addition, the facility is asked whether there are plans to reduce the use
of the sub6tance or its generation as waste in the next two to five years and, if so, the quantity of
reduction projected to be achieved.
4-1

-------
undertaken in the present study.
The supplemental information assembled for the 81 New Jersey facilities is summarized in
Table 4.1, listed in order of total firm employment Facility identification number is listed in the first
column; total firm employment is given in the second column; docket information in the third column
identifies how many environmental litigation proceedings have been brought against that specific
facility; environmental performance ratings established by CEP and Fortune magazine are shown in
the fourth and fifth columns; and the last four columns show how many of the chemical records
submitted by each facility designated that source reduction efforts had been accomplished in 1990
and/or 1991.
Sections 4.1 through 4.4 discuss the extent of correspondence between the environmental
performance at the firms, suggested by the supplemental data, vis-a-vis the rankings determined via
PPF analysis.
4.1 Litigation (or Docket) Information
The Right-to-Know Computer Network (RTK Net) based in Washington, D.C is a joint
effort of OMB Watch and Unison Institute. RTK Net contains information on U.S. facilities relating
to their operations, size, environmental performance, and litigation proceedings. We accessed RTK
Net to gather as much information as possible on the 81 New Jersey chemical facilities of interest.
We focused on the docket file, which reports litigation brought against a facility for environmental
reasons.
In Table 4.2, the PPF rankings for the years 1989 through 1991 are shown for those New
Jersey chemical plants for which docket information is reported.3 As in Table 4.1, the plants are
listed by firm size. The following findings emerge from the data in Tables 4.1 and 4.2:
•	Most of the dockets pertain to very large facilities, employing more
that 10,000 workers. Litigation was reported for 17 of the 33 (52
percent) very large facilities among the 79 New Jersey facilities for
which total employment could be identified.
•	Litigation proceedings are reported in RTK Net for only 9 of the 46
(20 percent) small, medium, and large facilities (employing less than
10,000 workers).
facilities 30, 35,53, 59,60, and 62, for which docket proceedings are identified in Table 4.1, are
not included in Table 4.2. These six facilities reported zero waste generation for two or more of the
three years from 1989 to 1991, and were excluded from PPF analysis.
4-2

-------
1 .-x-y- x»#y^8?f?.
#\\7&m0#k

^.Doawt
cs>„*
AViyiw
fortun#


1W¥
4MH

44ggr.w,
33
UNK




2

2
47
UNK




3

7
50
UNK
1



1

3
26
8



1
1
1

21
17



2

2
4
60
18




1

1
79
18




2

10
44
20




1

3
56
25




5
2

59
25




3

3
58
26




2

3
45
28




4

3
28
30




3

7
23
35




6
1
9
49
35




2

1
22
45



1
1
1
1
71
59
1



3

4
34
60




1

5
42
60



3
1
1
6
40
62




3

9
62
65




7

8
48
75




5

4
68
80




5
1
11
41
92




8

9
70
100




6

9
24
120




1

7
38
125




1
2

57
150
2



2
2
5
1
160
1



5

9
52
160
2


1
7

a
51
200



2
1
2
3
75
200



1


6
67
260




5

13
35
300




3

4
37
300




4

3
46
325
1



6
1
16
20
470




6

6
65
700




2

2
32
725




1

1
25
1,000



1
4
2
4
29
1,311




¦ - 2

e

-------

T«bh* 44
- -^3 S5&v,,;
'V	"'W'VW*'» 4-
_ 5? ?saJ.
-------
Table 42
Plants for which RTK NET has Docket Information
PLANT ID#
Plant Size
(based on
PPF RANKING BASED ON WASTES j

total # of
employees)
1989
(n=70 plants)
1990
(n»64 plants)
1991 |
(n=65 plants) |
50
unknown
42
NA
18 |
71
small
31
28
32 |
57
small
28
26
58 |
1
medium
NA
13
18
52
medium
61
61
66
46
medium
22
42
47
7
large
42
18
49
39
large
38
40
46
81
very large
31
55
54
77
very large
54
NA
11
78
very large
34
32
7
64
very large
42
17
30 1
14
very large
31
32
52 I
54
very large
3
10
15
5
very large
49
45
63
11
very large
6
NA
20
19
very large
8
52
17
17
very large
22
28
30
66
very large
60
47
62
—„			1
4-5
-------
•	PPF did not rank any of the 19 plants listed in Table 4.2 among the
top 20-percent of facilities for more than one of the three years from
1989 to 1991.
•	Twelve of the 19 plants in Table 4.2 moved downward in PPF rank
over the three-year period, three plants maintained a relatively steady
rank, and only three plants improved in rank.
•	All seven plants involved in more than one docket proceeding have
lower PPF ranks in 1989 than in 1991.
These findings indicate a concentration of litigation proceedings among the very large firms,
and an absence of litigation-involved firms among the highest PPF ranks. Regarding the former
result, a skewed distribution is not surprising given the greater attention given to very large firms in
the design and administration of environmental regulations.
The lack of litigation-involved firms among PPF-ranked leaders and the downward trend in
performance over time by these firms, particularly those involved in multiple dockets, is somewhat
surprising, however. Most of the dockets pertain to facility performance in the early and mid 1980s,
while the PPF ranks are based on performance in the late 1980s and early 1990s. It would have
seemed likely, especially among the very large firms, that litigation would spur the firm to greater
emphasis on pollution prevention and control- The results shown in Table 4.1 suggest that this is not
the case. Involvement in litigation seems to be an indication that 8 firm is reactive, rather than
proactive, in environmental outlook, and can be expected to lag in environmental performance until
a major shift in corporate perspective occurs.
When considering the preceding results, it is important to bear in mind that: (a) PPF is
intended as a screening analysis, (b) the results presented here are subject to the inadequacies of
reporting under the TRI and CRTK surveys, and (c) the specific characteristics of the firms involved
must be reviewed before conclusions can be drawn. Nevertheless, the results do suggest that PPF-
type screening represents a useful addition to the available array of analytic methods for use in
comparatively assessing facility environmental performance.
4J Published Rankings of Firm Environmental Performance.
Two published sources of firm rankings according to environmental performance were
reviewed: (1) The Council on Economic Priorities (CEP) ranked firms within four categories - top
performers, mixed performers, nothing outstanding either positively or negatively, and poor
performers; and (2) Fortune magazine evaluated the environmental performance of 130 of the largest
U.S. manufacturers and identified the ten leading firms, ten firms that have significantly improved
their performance over the last five years, and ten laggards.
The CEP rankings for 1991 and the 1993 Fortune magazine rankings are shown in Table 4.1
for those New Jersey chemical firms receiving ranks. Only six of the very largest firms were ranked -
- three firms operating four facilities were ranked by CEP, and three firms operating five facilities
were ranked by Fortune magazine. Two of the CEP-ranked firms received "mixed performer" ranks,
4-6
-------
and the third received a "nothing outstanding" rank. One of the Fortune magazine ranked firms was
judged one of the ten "most improved" performers, and the other two were judged environmental
"laggards."
Table 43
Comparison of CEP, Fortune Magazine, and PPF Rankings
PLANT ID#
Total
Employ-
ment
CEP
Rank
Fortune
Magazine
Rank
... .
PPF Score Based on Wastes

1989
1990
1991
64
12,000
M

55
23
33
43
16,200

I
69
64
16 |
61
30,000
M*

35
33
27 1
16
39,281

P
19
56
37 1
17
39,281

P
53
33
33
9
94,000
M

31
48
29
31
94,000
M

23
17
15
12
143,961

P
18
23
33
13
143,961

P
19
25
29 |
Table 4.3 compares the PPF ranks for these firms with the ranks assigned by CEP and
Fortune magazine. The comparison indicates the following:
•	None of the firms are rated "top performers" by CEP or Fortune magazine. This is
consistent with the PPF rankings, which rate none of these firms within the top 20
percent of New Jersey chemical facilities.
•	Two of the three facilities owned by firms rated "mixed performers" by CEP exhibit
erratic PPF rankings, moving from a 55 to 23 to 33 ranking over the 1989 to 1991
period in one case, and from 31 to 48 to 29 in the other case. The third facility
improved gradually in PPF rank, from 23 to 17 to 15, over the three year period.
•	The firm judged as one of the ten "most improved" in environmental performance by
Fortune magazine significantly improved in PPF rank, moving from a 69 to 64 to 16
ranking from 1989 to 1991.
4-7
-------
• Three of the four facilities owned by the two firms judged poor performers by
Fortune magazine declined steadily in PPF rankings from 1989 to 1991. The fourth
facility improved from a PPF rank of 55 in 1989 (poor) to a rank of 33 (intermediate)
in 1990 and retained that rank in 1991.
In summary, the PPF rankings have been found to correspond reasonably well with the limited
amount of published data that is available to rank firms on the basis of environmental performance.
44 Quality of Materials Accounting Data
As discussed in Chapter 2, the quality of data collected via the CRTK and TRI surveys is of
uncertain quality. Since the accuracy of PPF and other analyses depend directly on the quality of the
data, it is important lo select the more reliable records for analysis, and over the longer term to move
toward increasingly belter data quality. In the present study, a materials accounting balance threshold
of two-percent was used to screen out records judged to be of unacceptable quality. In addition,
records in which zero waste generation was reported for throughput quantities exceeding the survey
reporting threshold were also judged unacceptable. Such a report indicates thai 100 percent of
entering chemical was successfully transformed to useful product, Le., a no-loss condition has been
maintained over the full year.
Having excluded records because of unacceptable materials accounting balance and/or zero
waste generation, the 81 chemical facilities were assessed by PPF on the basis of those records judged
acceptable. An insight into the probable reliability of Lhe data submitted by each of these facilities
can be obtained by examining the overall quality of all records submitted by each facility. For
example, data from a facility that submitted five records, each of which passed the two-percent
threshold and finite waste generation tests, is judged mare reliable than data from a facility that
submitted 7 records, only three of which passed the screening test likewise, if the 4 non-acceptable
records from the latter facility had an average materials accounting error of four-percent, the records
from this facility would be judged of higher reliability than the records from another facility having
3 acceptable records and 4 non-acceptable records averaging a ten-percent materials balance error.
Table 4.4 presents the data quality assessment results compared to the PPF rankings for the
81 New Jersey facilities. Each facility is designated as having submitted data that are of excellent,
good, poor, or mixed quality on the following basis: (1) facilities in which all records satisfy the
materials balance and finite waste thresholds are judged to have submitted data of excellent quality;
(2) facilities for which 70 percent or more of the records satisfy the two screening criteria and the
non-acceptable records have a materials accounting error of less than 5 percent are judged to have
submitted data of good quality; facilities for which less that 30 percent of the records are acceptable
or the average materials accounting balance error of the non-acceptable records is 10-percent or more
are judged to have submitted records of poor data quality; and (4) all other facilities are judged to
have submitted data of mated data quality.
4-8
-------
COMPARISON OF PPF RATING ANDDATA QUALITY RATING
PPF RATINGBASEDMff
on

Excellent
Good
Mixed
Poor
7 0 2 3 1
6 0 2 4
14 3 3 3
6 0 15
TOTAL
33 3 8 15
* PPF Ratings are based on the percentile PPF rank, as follows:
Excellent (Top 20%), Good (21%-40%), Mixed (41%-80%), and Poor (Bottom 20%)
** Data Quality Ratings are based on # of acceptable reports (meeting the
< = 2.0% MBE and total waste > 0 requirements), as follows:
Excellent (100% of reports acceptable); Good (70% or more of reports acceptable
and MBE% of unacceptable reports < 5.0%); Poor (less than 30% of reports
acceptable OR the MBE% of unacceptable reports greater than 10%);
Mixed (all other facilities)
Tfae following observations can be made based on the data quality results shown in Table 4.4:
•	Over half of the 59 facilities evaluated submitted data of excellent quality, as defined
by the criteria above, in their 1990 CRTK survey responses.
•	Data quality was either excellent or poor among the facilities evaluated. Fifty-six
percent submitted excellent quality data, and 25 percent submitted poor quality data.
Only 19 percent submitted reports judged to be of either good or mixed quality.
•	Little correlation was found between the PPF ratings of facilities and the quality of
data submitted by those facilities.
•	One potential use of the result is to recompute the PPF analysis placing greater
uncertainty on data from those facilities submitting poorer quality data. The largest
modification of PPF results would occur if one of the poor-data-quality facilities
functions as a PPF frontier facility.
4-9
-------
4.4 Implementation of Source Reduction Initiatives
The source reduction information for each plant given in Table 4.1 is compared in Table 4 .5
with PPF rankings for those plants in each of the three years. To illustrate, plant 8 reported that no
source reduction was undertaken witfa regard to the four chemicals they reported in 1990, nor with
regard to the 14 chemicals reported in 1991. Plant 8 was ranked 20 by PPF in 1989 and 1990, then
fell off somewhat to 32 in 1991. Plant 13, listed at the bottom of the first page of Table 4.5,
implemented source reduction for the one chemical it reported in 1990, but exercised no source
reduction for the three chemicals it reported in 1991. Plant 13 had a PPF rank of 17 in 1989, 18 in
1990, then fell off to 27 in 1991. The following observations can be made with regard to the data
in Table 4.5:
•	There seems to be little correlation between a facility's reporting source reduction
activity in the CRTK survey and that facility being highly ranked by PPF.
•	It seems likely that a consistent pattern of source reduction over a period of years will
result in higher PPF rankings. The lack of such correlation in the two years examined
may stem from the base year effect already discussed with regard to TR1 ratios.
4-10
-------
-\ ^ <5^-;":., J& Sf^ :: > ..-^' rt.;:-Jv v 3,
a&acma^twi Pfy	r
>;&oarc* Rvctacl^on"
S!
8
36
63
73
69
6
7
39
80
27
10
81
76
77
76
64
74
14
2
3
43
54
4
5
18
53
55
15
72
61
11
15
16
17
30
66
9
31
12
13
1,850
2,000
3,000
3,500
4,000
4.180
4,180
5,400
6,000
6,919
9,500
10,200
10,500
10,500
10,500
12,000
13,931
14,415
16,000
16,000
16,200
17,000
17,324
17,324
19 000
19,366
19,366
27,000
27,000
30,000
37,700
37,756
39,261
39,261
55,000
86,000
94,000
94,000
143,961
143,961
3
1
1
8
1
3
3
10
4
2
1
7
1
10
8
8
17
2
6
4
2
5
2
1
2
3
8
5
11
1
7
11
6
1
25
1
imy
3
1
19
«»1N
14
13
3
8
1
1
19
10
5
7
6
9
7
1
6
15
10
7
10
3
9
6
2
2
4
2
2
14
6
13
3
6
16
19
36
3
IfiW
=20
58
26
42
38
42
11
22
31
54
34
42
1
31
38
to
3
7
49
2
13
13
34
34
6
8
9
22
60
57
51
17
l«
*•>»»
20
9
7
18
40
32
11
30
55
32
17
4
32
21
24
10
1
45
5
30
26
B
50
52
56
28
47
49
15
21
13
32
25
9
8
49
46
48
14
20
54
11
7
30
1
52
22
25
13
15
1
63
37
45
9
59
20
17
56
30
62
23
12
51
27
Mo taa
Total Employ Total number ol employees in company (CRTK, 1991 and FINDS. 1994)
Source Reduction- Number of chemic&i* for which the company employs sourca reduction techniques
(Y = Yes), and the number (or which H doesn't (N = No) The years 1990 and 1991 are represented (CRTK. 1991)
-------



bxtuctioo.
>. - , " 
-------
5.0
POTENTIAL POLICY AND PRIVATE SECTOR APPLICATIONS
The analyses conducted in the two phases of the PPF project help bridge the gap between
the policy analysis functions of EPA and its regulatory responsibilities. As EPA attempts to develop
more collaborative arrangements with the regulated population, its' data and information management
needs will alter and grow. In the following sections, we discuss potential policy actions that might
help EPA address some of these data issues, and comment on how the work reported here and in
the earlier report, titled "Pollution Prevention Frontiers (PPF): An Approach to Measuring Pollution
Prevention Progress," may assist in this effort Specific issues addressed are:
•	how PPF supports comprehensive approaches to data collection and evaluation;
•	how EPA can use PPF for monitoring and evaluation purposes; and
•	how PPF can be used for diagnosis and managerial control within industry.
5.1 Policy Options for Data Collection and Evaluation
The demand for reducing the burden on industry of government regulations poses a dilemma
for EPA. The dilemma arises from the fact that a collaborative approach increases the flexibility in
the interactions between EPA and the organizations it oversees. This greater flexibility creates the
opportunity for increased diversity in the approaches undertaken to achieve pollution prevention and
control. In order to fairly monitor the resulting broader set of options, different types of information
are required to insure that the various regulatory and enforcement goals are met. As a result,
coincident with greater latitude in their interactions with EPA, the regulated organizations may be
required to report additional information in order to enable EPA to fairly evaluate the spectrum of
proposed environmental compliance options. Hence, with the increase in flexibility in ways of
achieving pollution prevention and control, there may also be a requisite increase in the amount and
variety of data collected.
Among the initial steps in moving toward such a more flexible system, EPA may need to re-
evaluate what additional types of data must be collected. As part of this evaluation, EPA may also
need to develop a more comprehensive information evaluation and management system in order to
support the evaluation of the more diverse approaches undertaken to satisfy environmental mandates.
The emerging call for increased reliance on risk-based mandates illustrates the nature of more
5-1
-------
complex regulatory analyses that appear inevitable as the relative benefits and costs of diverse types
of options must be comparatively evaluated.
Some of the criteria we implemented in determining the adequacy of the data and the
analytical tools for evaluation of pollution prevention and control efforts include: (1) public
availability; (2) independence from a base year; (3) utility for screening non-regulatory interactions,
(4) the ability to synergistically incorporate multidimensional risks.
•	Public availability of data: For the analytic approach to be useful in policy
applications, it is essential that the data employed be publicly available.
Discussion of policy issues is hampered when the data on which analyses are
based cannot be subjected to broad public scrutiny. PPF has been designed
to extract as much information as reasonable from the most extensive public
assembly of data characterizing pollution prevention performance at industrial
facilities. As discussed in footnote 7 of Chapter 1, more detailed evaluation
of firm performance is possible, but the information requirements rise sharply
when the needed data are not publicly available.
•	Independence from a base yean The ability of a firm or plant to reduce its
pollution levels depends upon the level of success it has already had with
pollution prevention efforts. Gearly, firms that have not made any effort will
seem to make big strides, particularly in percentage terms, when they first
begin to implement pollution prevention and control strategies. Thus, in
order to be fair, performance evaluations should take previous efforts into
consideration or, alternatively, conduct evaluations that are not based an
arbitrary starting date.
•	Utility for screening non-regulatory interactions: As the scope of EPA's
activities continues to expand beyond regulatory actions to include increased
efforts in information dissemination, technical assistance and other forms of
support, EPA will need to determine how to effectively couple these activities
with economic and other types of incentives. The information required for
such activity is qualitatively different from that used for regulation and
control. Furthermore, the analyses must become more sophisticated in order
to determine who should receive what form of assistance, i.e., incentives or
disincentives. Oftentimes, the data required for such evaluations are episodic,
i.e.f are not available on a continuous basis, hence the data systems required
are different from those traditionally used for monitoring purposes.
•	Multidimensional Risks: It has been know for some time that pollutants are
frequently not confined to a single hazard category, nor do they necessarily
pose the same risks when in the presence of other pollutants or products.
The data systems required to monitor these multidimensional risks and the
analytical tools required to evaluate such data must be capable of analyzing
the synergistic nature of combinations of pollutants and their corresponding
risks.
5-2
-------
Over the two phases of the PPF project, we have concluded that it is possible, albeit for a
limited number of chemicals used in certain segments of the chemical industry, to collect data and
analyze them so as to provide information that meets the above criteria, and thereby to evaluate the
comparative effectiveness of pollution prevention efforts at industrial firms. This determination is
based on an audit of the data publicly available on these chemicals, and the process of embedding
these data within a materials accounting framework in order to evaluate the quality of these data.
Information Audit: Policy implementation, particularly for an agency such as the EPA, relies
heavily upon the Agency's ability to acquire and analyze reliable data. Therefore, in the first part of
this study we invested heavily in conducting a thorough information audit We identified by source
and nature the publicly available data that might be used to monitor pollution prevention efforts.
The major sources of the data identified were the TRI survey and the New Jersey Worker and
Community Right to Know (CRTK) Act survey.
The purpose of this audit was to organize in one location all publicly available, internally
consistent data that pertain to the evaluation of the pollution prevention efforts of a significant
industrial sector -- New Jersey chemical manufacturers categorized in Standard Industrial Code 28.
This effort identified for consideration by policy makers the type of pollution-prevention-relevant data
that could be assembled with a reasonable degree of effort Having identified ail publicly available
data, the project undertook: (1) to evaluate the reliability of these data, and, more importantly, (2)
to determine the value of these data in supporting decision making processes within EPA and within
industry.
Evaluation of Data Reliability.	The framework, we used to evaluate these data is the materials
accounting (balance) approach. This approach assesses the reasonableness of reported data in terms
of the balance between the quantity of chemical input versus the quantity of chemical consumed or
output The National Academy of Science in reviewing the utility of materials accounting data notes
that the materials accounting approach "could be useful in comparing operations in a given industry
or in different industries that use the same listed chemicaL It could also be useful for assessing the
reasonableness of reported release estimates."1 As our study demonstrates, reconciling such data
given the current state in which they exist is, while useful, a laborious and painstaking process.
Given that the collection and transmission of much of the data used in this study is required
by various state or federal agencies, the policy issue facing EPA is whether it should invest in a
comprehensive information system to maintain these data in a readily accessible fashion. Crucial
information for this decision would be a comprehensive audit of the data that are currently collected.
Although limited to one aspect of the chemical industry, we have demonstrated the feasibility of such
an exercise. As the demand For reducing the burden of complying with various government rules and
regulation increases, EPA, having conducted such an audit will be well positioned to comment on the
type of data being currently collected, and eventually determine where data redundancies exist, where
reporting requirements might be changed in light of up-to-date information on potential hazards and
risks associated with the chemicals under consideration and how these data serve the needs of EPA
in the execution of its various policy formulation, implementation and evaluation functions.
¦National Academy of Sciences. Tracking Toxic Substances at Industrial Facilities: Engineering Mass Balance
versus Materials Accounting' National Academy Press, Washington DC, 1990, p. 3-4.
5-3
-------
We have also demonstrated that it is possible to build simple checks into the system that are
based upon materials accounting principles. These consistency checks would ensure a minimal level
of data reliability. In addition to helping evaluate data quality by determining whether the materials
accounting balance is achieved, the materials accounting approach offers the capability lo conduct
alternative analyses. Throughput data, which represents a central component of materials accounting
data, not only allows analyses that are independent of base year, but also provides valuable indicators
of data quality. For example, the reasonableness of reported waste-to-throughput ratios can be
assessed in terms of the activities being undertaken in the use of each chemical.
Materials accounting data also provide an initial basis for life cycle analyses of the reported
chemicals. To illustrate, between 1988 and 1991 New Jersey firms reported shipping off-site asor-in
products approximately 7 million pounds of 1,1,1-trichloroetbane. Over the same period, 2.3 million
pounds of 1,1,1-tnchloroethane were reported as released to the environment during manufacturing
activities at these facilities. The TRI data record only the quantity released during manufacturing.
However, during use of 1,1,1-trichloroethane in coatings and other applications, nearly all of the
chemical is released to the environment As a result, within a period of three to twelve months most
of the chemical shipped off-site within products will have been released to the environment. Thus,
a more accurate estimate of 1,1,1-trichloroethane releases to the environment associated with the
New Jersey facilities between 1988 and 1991 is 9.3 million pounds. From this life cycle perspective,
TRI reporting identifies only 25 percent (2.3/9.3) of the actual environmental load.
As the 1,1,1-trichloroethane example illustrates, a materials accounting framework provides
the perspective for evaluating environmental releases in other phases of the chemical's life cycle.
Collection of such data on a national basis would enable the characterization of chemical Qows from
origin to final consumption. Presently available data are insufficient to accomplish this task. A
general body of knowledge exists that identifies for major chemicals the general types of uses, and
in some cases approximate estimates of usage quantities, and TRI data codifies the reported
quantities of specified chemicals that are released during at manufacturing facilities. The materials
accounting framework would establish a comprehensive database identifying how TRI chemicals flow
into, between and out of manufacturing facilities, as well as the environmental releases that occur
during manufacturing operations. This type of approach is one of three models EPA is considering
for updating and expanding the Chemical Usage Inventory (CUI) program.
Evaluation of Data Value: To assess the value of materials accounting data, it is important: (1)
to clearly identify and define the criteria by which the pollution prevention efforts are to be judged,
and (2) to determine the types of data summaries and analyses that will be used to make these
judgments. It is clear from the analyses presented here that the types of measures that can be
constructed are severely limited by the availability of data. Hence, in order to fully exploit the
potential of PPF as an analytical tool, EPA, in conjunction with industry and other stakeholders,
would have to determine the criteria by which the pollution prevention efforts to are to be judged.
This study demonstrates the flexibility of the PPF approach in developing different measures of
performance that emphasize different aspects of pollution prevention performance. Once the
evaluation criteria have been established, PPF, in conjunction with other data analysis tools, can be
used to develop the measures necessary to determine whether the established performance criteria
are being met. Thus, PPF analysis offers another check of the value of the data being collected in
terms of its utility for analysis.
In this context, determining the nature and type of data that are necessary for EPA to fulfill
5-4
-------
its expanded policy analysis and performance role represents a central data management challenge
facing EPA. The Agency will have to rethink its data needs, preferably with inputs from industry and
other stakeholders, and design a database management system sufficient to utilize these data in
conduct of increasingly complex policy analyses.
5.2 Potential Applications of Pollution Prevention Frontiers
PPF indices can be developed and used by EPA and industry to serve a variety of diagnosis,
monitoring and control functions. Illustrative applications are discussed for the public and private
sectors in the following sections.
Potential Applications within EPA
PPF indices can be used by EPA for both policy formulation and evaluation.
Diagnosis: For policy formulation, PPF can be used as a diagnostic tool to explore the
implications of using different criteria and their associated measures. Given the flexibility of the PPF
indices, EPA could use this approach as a data exploration and reduction tool that yields simple
qualitative rankings. Analysis of the different measures obtained by using various criteria could help
determine the implications of implementing different policy options. These PPF indices, could then
also serve as points of departure for discussions, with the industry and other stakeholders, on
evaluation criteria and their implications in terms of data requirements and implicit incentives. PPF
is particularly suited for such discussions since the concepts underlying the technique are intuitively
straightforward in that the index is based on comparisons with the best observed practice.
Furthermore, the criteria used in these computations are explicitly stated, and the decision rule for
is that each plant be shown in its best possible light These measures an be, therefore, be perceived
as being fair in that they take into consideration the context within which the different plants operate.
Attention Focusing and Screening: Given the current quality of the data and the exploratory nature
of the indices, the PPF approach has not yet been validated for use for valuative and control
purposes. However, even at this formative stage it can be used to distinguish between leaders and
laggers, in terms of pollution prevention efforts, within the industry. This qualitative partitioning into
two groups can serve as a screening device for focusing attention on plants whose performance seems
to be exceptionally good or poor. Validation of the index as a viable measurement tool can then be
done through case studies of the plants that have been thus identified.
Evaluation and Monitoring: Once the PPF indices have been validated they can be used for
evaluating pollution prevention efforts. The evaluation can be done with respect to the performance
of other plants in the industry or against a pre-defined benchmark. Incentive structures could be
designed to effect improvements relative to other plants as well as for rewarding innovations that
improve the overall performance of the industry. Improvements over time, of the whole industry or
of individual plants can be studied by analyzing movements in the PPF frontier. Thus, the PPF
approach yields information not only on average improvements in the industry but also plant specific
information relative to the performance of the industry. Thus a variety of incentive structures can
be designed which reward or punish plants for their performance, relative to themselves, relative to
the best observed practice, relative to some industry average, or some combination of these relative
measures.
5-5
-------
Information Dissemination: PPF, unlike many other data reduction techniques yields plant specific
information which can be useful for developing databases where the best practices in the industry can
be documented and information can be shared, subject to confidentiality restrictions, on how
performance might be enhanced by emulating the best practice.
Additionally, the public release of plant specific rankings of pollution prevention activity can
serve as a powerful motivator for improving performance without having to resort to specific
incentives.2
Potential Applications within Industry
Much of the value of the PPF analysis for EPA can be translated fairly readily to individual
companies or plants. In particular, companies can maintain their own PPF indices for comparisons
among multiple plants or vis-a-vis plants of other companies, e.g., benchmark facilities. In this way,
PPF results can be used as a motivator for effective self regulation and compliance
For a variety of reasons, mainly pertaining to the confidentiality of the production activities,
detailed data of the type needed for monitoring and assessing managerial performance are seldom
publicly available. However, companies, can implement PPF indices for their own internal monitoring
and use PPF for managerial control.
Diagnosis: As industry attempts to design new technologies in response to emerging
environmental imperatives, PPF analysis can be used constructively at the process level. PPF can be
first used to establish performance benchmarks, and then to monitor process performance with
respect to these benchmarks. Using data not available in the public domain, companies could
construct indices that go well beyond the ones discussed in this report These indices would provide
insights regarding the relative performance of various processes and/or manufacturing activities, as
well as other managerial and economic initiatives.
Managerial Control-.	Measures of organizational effectiveness, as well as those of managerial
output and outcomes, can be developed using the computational techniques underlying PPF. The
precise nature of these measures would be designed to support specific organizational goals and
objectives. However, the basic logic of the PPF approach could implemented using data not
pertaining to pollution prevention activity.
Process Control-. The PPF indices constructed to measure pollution prevent effort can be used,
as is or in some modified form, to monitor process quality. Using concepts similar to those from
Statistical Quality Control, PPF indices can be implemented to provide a multidimensional view of
process performance. Process monitoring using pollution prevention as a criterion for superior
performance can be readily done using the PPF indices without any modification. However, other
aspects of process performance, including for instance output measures, downtime, and other
engineering criteria would have to be incorporated to obtain a complete indicator of process
performance.
^Evidence of the effectiveness of such public release of information can be seen, for instance, by the
influence of the J.D. Power and Associates' survey of consumer satisfaction in the car industry, or the
FAA's release of 'on-time" data for the airlines, and even by the effect of occasional rankings of
companies' environmental performance by magazines such as Fortune and Business Week.
5-6
-------
6.0
SUMMARY AND CONCLUSIONS
As stated in Section 1.1, the objectives of Phase 2 of the PPF project were:
•	To assess the nature and quality of materials accounting data collected in New
Jersey for the years 1989 through 1991, and identify any potential benefits
that might be gained by extending the Toxics Release Inventory (TRI) survey
to include additional data elements contained in the New Jersey data.
•	To compare the results obtained via PPF analysis of pollution prevention
performance at New Jersey chemical firms over the 1989 through 1991 period
vis-a-vis results obtained by alternative measurement approaches.
•	To compare the facility performance ratings obtained via PPF analysis with
ratings suggested by other types of data that offer insights into environmental
performance at the New Jersey facilities.
•	To identify potential policy applications of PPF analysis in the public and
private sectors.
The following sections summarize the findings achieved with regard to each objective during the
present study.
6.1 Usefulness of Materials Accounting Data
Materials accounting data have been shown to provide an extremely valuable, arguably
needed, perspective on the release and waste generation estimates reported in response to the annual
TRI surveys. Useful applications of this data include:
•	Materials accounting data provide a useful check on whether the TRI-
reported data are internally consistent with overall usage data for the
chemical at the reporting facility. While exact closure may not be
established, depending upon the method of estimating some of the
chemical usage variables, the quantity of chemical available for use
6-1
-------
during the year should approximately equal the quantity used, i.e., an
approximate materials accounting balance should be achieved. When
the data fail to satisfy this consistence check, further examination of
the data submission is in order.
•	Materials accounting data also provide a basis for life cycle analysis of
the chemicals of interest. As the 1,1,1 trichloroethane example in
Chapter 5 illustrates, the bulk of a chemical's releases may occur in
pre- or post-manufacturing phases of its life cycle. Materials
accounting data provide useful input into analyses to more precisely
pin down the sources of greatest environmental releases for the
chemicals of interest.
•	Collection of materials accounting data on a national basis would
facilitate establishment of a reasonably complete picture of the flows
of chemicals within our economy and the points at which most
significant environmental releases occur.1 Total annual production
quantities of most organic chemicals are already reported to the
Census. National materials accounting data would extend this
information by identifying where the bulk of this production occurred
and where the chemical was consumed in what types of applications.
This data could be used for life cycle analyses, as described above, and
gaps in the pattern of chemical flows could be readily identified and
studied.
•	The New Jersey survey also collects information on the nature and
intensity of the manufacturing activity using the reported chemicaL
For example, in 1990 5,000 pounds of the chemical may have been
consumed in the production of 200 refrigerators, and in 1991 the
chemicals usage may have grown to 6,000 pounds in the production of
220 refrigerators and 50 desks. This type of information can be used
in production normalized assessments of the effectiveness of pollution
prevention at facilities producing similar types of products.
6.2 PPF vis-a-vis Other Pollution Prevention Metrics
Comparison of PPF with pollution prevention assessments based on total TRI releases, year-
to-year ratios of TRI releases, and waste-to-throughput ratios have shown that each have certain
advantages. Use of throughput as a normalizing factor was found to provide an invaluable
perspective on the efficiency of chemical usage.
'Collection of facility-specific materials accounting data is one of three approaches EPA is
currently considering for collection of chemical use inventory (CUI) data.
6-2
-------
Comparisons of facility ranks based on total TRI releases or year-to-
year TRI ratios were found to bear little relationship to ranks based
on release-to-throughput ratios:
- Total TRI releases are primarily an indication of the size of
the facility. An illustration in Chapter 3 shows how a single
facility reporting the same release quantity of a chemical as
Gve smaller facility has nearly four times the throughput as
those five facilities, i.e., is nearly four times as efficient in
chemical usage even though reporting larger total releases.
Year-to-year TRI ratios, on the other hand, suffer from base-
year dependence. Facilities that have done little pollution
prevention prior to the base year may achieve significant year-
to-year improvements at lesser efforts than facilities that had
proactively conducted substantial pollution prevention prior to
the base year.
Waste-to-throughput ratios are not effective as a basis for assessing
improvements in dispersive chemical usages. These include
applications, such as volatile solvent use, where the quantity of
materials entering the application identically equals that leaving the
system, because the entering materials directly replace losses to the
environment In such cases, the waste-to-throughput ratio remains
equal to unity irrespective of changes in the efficiency of usage.
Relatively few dispersive chemical applications were reported at the
New Jersey chemical facilities analyzed, however. Thus, the benefits
of the throughput ratio in surmounting the base year barrier at the
New Jersey facilities substantially out weighed the difficulties in its use
in the few reported dispersive chemical applications.
If a situation were encountered in which a significant fraction of the
chemical uses did involve dispersive applications, the units of
production data being collected by the New Jersey survey might be
used (instead of throughput) to normalize waste generation.
PPF analysis represents an important extension of waste-to-
throughput ratios when the chemical population poses multiple risks,
such as carcinogens, mutagens, acute toxins, urban ozone precursors,
global warming contributors, and others. PPF was designed to easily
accommodate such multiple risk factors and, as discussed in Chapter
5, future policy evaluations will increasingly have to deal with such
complexities.
6-3
-------
63 PPF Ranks Compared to Other Facility Ratings
Publicly available data providing comparative rankings of tbe environmental performance of
industrial facilities and firms are limited. Information that was identified included: litigation
proceedings (or dockets); published rankings by tbe Council on Economic Priorities (CEP) and by
Fortune magazine; assessments of the data quality of reports submitted in CRTK survey; and
indications in the CRTK survey that source reduction had been undertaken. Although the
supplemental data was limited in quantity, the associated facility ratings corresponded reasonably well
with facility rankings generated by PPF. The following results were obtained:
•	Over half of the very large New Jersey chemical facilities (employing
more than 10,000 workers) have been involved in litigation, as
compared to litigation involvement by only one-fifth of all smaller
facilities. This skewed distribution is not surprising given the greater
attention given larger, technically sophisticated firms in regulatory
design and enforcement
•	None of the New Jersey facilities involved in litigation were ranked by
PPF among the top 20-percent performing facilities. Sixty percent of
the litigation-involved facilities moved down in PPF rank between
1989 and 1991, 20 percent remained steady in rank, and only 20
percent improved in PPF rank. None of the plants involved in more
than one litigation proceeding improved in PPF rank over the three
year period. As discussed in Chapter 4, these results suggest that:
(a)	PPF ranks are consistent with the litigation results,
assuming that participation in litigation represents a
failure in environmental performance; and
(b)	involvement in litigation seems to indicate that a
firm is reactive, rather than proactive, in
environmental outlook, and will continue to lag in
performance until a fundamental shift in perspective
occurs.
•	None of New Jersey firms are rated top performers by CEP or
Fortune magazine. PPF rankings are consistent with this result in that
none of the firms are ranked in the top 20 percent by PPF. The
movement of tbe firms in PPF rank over the 1989 to 1991 period was
consistent with the ratings of those firms by CEP and Fortune
magazine. For example, three of the four facilities owned by two
firms ranked poor performers by Fortune magazine declined steadily
in PPF rank over the three year period, and the fourth facility
improved from an initially poor rank, but then remained in the middle
rank of facilities.
6-4
-------
• Neither facility ratings based on the data quality of their survey
submissions nor the reported undertaking of source reduction by
facilities appeared to bis correlated with the PPF ranks for these
facilities. The former finding suggests that sensitivity analyses of PPF
rankings be conducted with each facility's performance weighted
according to the estimated quality of its data. The latter finding is
probably attributable to the base year effect, Le., some of the facilities
undertaking source reduction in the two years for which this
information has been collected may not have initiated much pollution
prevention in prior years.
6.4 Potential PPF Policy and Private Sector Applications
Economic considerations and the growing importance of pollution prevention are bringing
increasing pressure to bear on EPA to develop a more collaborative relationship with the regulated
population. As greater flexibility is introduced in the way environmental goals may be achieved,
EPA's need for more comprehensive data characterizing a wider range of environmental actions will
continue to expand. For example, increased consideration of the risk implications of alternative
regulatory approaches will require more refined delineation of the context within which releases will
occur. As discussed in Chapter 3, PPF has the capability to accommodate and analyze such a wide
variety of data. Potential applications in the public policy and private sector arenas include:
• PPF applications in policy analysis fall into four general categories:
- Diagnosis: As EPA explores different, more flexible
monitoring 8nd reporting arrangements, PPF provides
a multidimensional approach to assessing the value of
available data for valuative and decision making
purposes.
Attention Focusing and Screening: PPF provides a
qualitative partitioning of plants into leaders and
laggert. This partitioning can serve as a valuable first
step in identifying exceptional plants and focusing
attention on their exceptionally poor or good
performance.
Evaluation and Monitoring: Once the validity of the
PPF measures has been established, they can be
implemented in a management system for evaluating
and monitoring pollution prevention activity.
Information Dissemination: EPA can serve as a data
clearing bouse and use PPF to provide plant-specific
information, in contrast to the industry averages
generally available for valuative purposes.
6-5
-------
• Many of the beneGts of PPF analysis thai apply to the public sector
can be transferred directly to private sector applications. Three
potential applications are:
- Diagnosis: Industry can use PPF measures to
diagnose the implications, with respect to EPA's
criteria, of different pollution prevention options.
Compared to the analyses that can be done with
publicly available data, industry will be able lo conduct
mush richer analyses and obtain valuable insights into
the relative merits of different options.
Managerial and Process Control: As a generic
approach for constructing multidimensional indices,
the computational models underlying PPF can be
readily modified to develop indices of managerial and
process performance and to link these measures to
other measures of pollution prevention.
6-6
-------
APPENDIX A
OPERATIONALIZING MEASUREMENT WITHIN THE
MATERIALS ACCOUNTING FRAMEWORK
-------
APPENDIX A
Operationaliring Measurement Within the Materials Accounting Framework
The purpose of this appendix is to describe the mathematical basis for the construction of
Pollution Prevention Frontiers and the ensuing indices for comparing the pollution prevention
performance of industrial facilities. Other uses of DEA for environmental applications are briefly
considered.
A fundamental goal of pollution prevention is to minimize the amount of pollutant or waste
(Qwx) generated per unit of product output (P). The objective for developing an analytical
framework for comparing the performance of a plant in terms of its success at reducing pollutants
per unit of production is threefold: (a) for focusing attention, (b) for diagnosis, and (c) for feedback.
Finally, if and when the validity and reliability of the performance indicators have been established,
the information could be used for control. Implicit in the analytical framework is a metric to measure
the level of performance of each plant. This appendix discusses the nature of the PPF metric
employed to determine the level of performance at each plant
The value of the proposed metric depends crucially on the criteria established for evaluation
and the ability to reliably operationalize the performance criteria into valid measures (Desai, 1992).
The Materials Accounting approach discussed in Section 1 of Chapter 2 of the report offers a
framework within which the measurement of pollution prevention efforts can be operationalized.
Chemical processes, in general, use and produce chemicals according to well defined formulae.
Therefore, it is theoretically possible to identify for each process the minimal amount of pollutants
that will be generated by that process. However, for a variety of reasons, these theoretical minima
are rarely attained. Oftentimes, the problem is further compounded by the fact that a particular
chemical can be produced a number of ways and different plants use different technologies for
reducing pollution. Hence, the motivation for PPF is the development of a common index on which
various plants, producing a variety of chemicals using different chemical processes, can be evaluated
in terms of the amount of pollutant generated per unit of production activity.
The problem of evaluating or ranking plants when they produce a single product and a single
pollutant is trivial. The plant that generates the minimum amount of pollutant per unit product is
obviously the best The performance of other plants can be evaluated relative to the performance
of this plant The required index is simply an ordering of this ratio of pollutant to product, the
smallest implying the best performance and the largest indicating the worst The problem of
constructing an index, however, becomes non-trivial when multiple products are produced resulting
in multiple pollutants. The extension of the simple ratio analysis to multiple dimensions can be
effected in a number of ways as documented, for instance, in the literature on economic indices
[Caves, Christensen and Diewert, 1982; Malmquist 1953; Tornquist, 1936]. The PPF index is the
result of implementing a mathematical-programming-based index-construction technique in the
context of pollution prevention. This index-construction technique, called Data Envelopment
Analysis (DEA), has its basis in operations research and the economic theory of productioa
While this index has its origins in the economic theory of production, its mathematical
formulation is independent of that theory. In the context of comparing the pollution prevention
efforts of plants, the index to be developed corresponds to the following verbal statement:
A-l
-------
Find a multivariate ratio which (1) characterizes each plant in terms of its activity levels
and pollutants and (2) provides an ordering, from best to worst, of plants with similar
activity and pollutants.
Considering, for the purposes of illustration, the level of activity at a plant to be measured by the
amount of product generated, then, a multivariate ratio that meets the above description can be
expressed, for each of n plants, as:
V
E.
k " "
1 4 j SJI	(1)
Where:	P^ = Amount of product h being produced at plant j
I9 = Amount of pollutant q being produced at plant j and
ukJ and wv are weights assigned to the products and the pollutants.
Thus, we have in (A.1) the ratio of a virtual pollutant to a virtual product, where the virtual pollutant
is the weighted linear combination of pollutants and the virtual product is the weighted linear
combination of products. Given this characterization of the plants' activity, the computational issue
to be addressed is (a) how should the weights u and w be obtained and, subsequently, (b) how should
the ordering of these plants be achieved.
In creating DEA, Charnes, Cooper and Rhodes (1978) proposed a mathematical programming
formulation of this problem that simultaneously solves for the weights, and provides a relative
ordering of these plants. The weights are to be selected such that each plant obtains a score that
places it in the best possible light Ideally, each plant would want to minimize (1) or, conversely,
maximize the inverse of (1). Charnes, Cooper and Rhodes (CCR) formulated the problem in terms
of maximizing the inverse of (1). Since the weights are to be "objectively" assigned so as to
maximize this ratio, the size of the ratio has no upper limit In order to bound the maximum value
that any plant could obtain, CCR proposed an upper bound of 1. CCR operationalized the above
verbal statement of the problem as follows:
Find weights uh. and w^such that the ratio of virtual products to virtual pollutants
(ie., inverse of (1)) is maximized subject to the constraint that no ratio exceeds
unity.
A-2
-------
The mathematical expression for this statement is:
Maximize
(2)
such that
Vi-U.-vi
(3)
and Uy t 0 V j.
(4)
where ^ is value of the ratio associated with plant;. Note tbat this value is bounded above at 1 and
is always positive. This fractional program was, as mentioned above, developed in the context of
production theory. In that context, the numerator denotes a virtual output, being the combination
of actual outputs, and the denominator denotes a virtual input, being the combination of actual
inputs.
Charnes and Cooper (1962) developed a transformation that yields a linear equivalent for the
fractional program (2)-(4), thereby considerably simplifying the computation of ej The fractional
program yields an infinite number of solutions. For example, if (u, w) is an optimal solution, then
(au,aw) is also optimal for «t>0. Transforming the variables (u, w) into (ji, v), CCR offered the
following equivalent linear programming formulation, which is representative of one of the potential
solutions of (2)-(4):
max z-^Pj
H.v
(5)
s.t. vT Ij - 1
(6)
A-3
-------
^.P - v7 i 0	(7)
li'.v* t 0.	(8)
The corresponding dual of (5)-(8) is as follows, where X is a vector of weights and 6 is the measure
of performance.
min 6,
6,A
(9)
s.t. PXtPj	(10)
eij - ix t o,	(ii)
0 free, X i 0.	(12)
Over the decade-and-a-half since Charnes, Cooper, and Rhodes' original work, a number of
variations of these linear programs have been developed and implemented in a variety of contexts.
For instance, the ratio (2) can be maximized either by maximizing the numerator for a given value
of the denominator or the denominator can be minimiTpH for a given value of the numerator. Each
of these options yields a different linear program. Similarly, different assumptions about the weights
yield different sets of constraints and restrictions on the values of the weights, thereby modifying (5)-
(8) and (9)-(12). These different assumptions affect not only the membership of the set of plants that
define the frontier, but also how differences in the size of the plants and their activity levels affect
their performance scores. For a more detailed discussion of these models and the underlying
assumptions, see Seiford and Thrall (1990); additionally, see Seiford (1990) for a list of approximately
four hundred papers pertaining to the theory and applications of DEA.
One group of investigators has used DEA to examine the economic implications of
"undesirable outputs," or wastes. Fare, Grosskopf and Pasurka (1986) modeled "the effects of
environmental controls when these are viewed as restricting disposal of outputs." They considered
the cases of strong (free) disposability of outputs, in which no economic penalty is incurred, and weak
(restricted) disposability, in which undesired outputs are subject to environmental regulations. Their
study of U.S. steam electric plants "captured the opportunity or indirect cost of those regulations as
the difference in output when disposablity is free (i.e., unregulated) and when disposability is
restricted." They estimated the lack of disposability 'cost' to average 16 million kilowatt-hours in lost
potential output for each of the 100 electric utilities they examined. A subsequent study by Fare,
Grosskopf, Lovell, and Pasurka (1989) considered two modifications to the standard Farrell approach
to efficiency measurement: relaxing the restriction on production technology "to allow for the fact
A-4
-------
that undesirable outputs may be freely disposable." and modifying the efficiency measures to allow
for "asymmetric treatment of desirable and undesirable outputs." The resulting hyperbolic efficiency
measure was found in a study of 30 U.S. pulp and paper mills to be "very sensitive to whether or not
undesirable outputs were included," and it was determined that "departures of measured productivity
assuming strong disposabiiity from those assuming weak disposabiiity can be converted into measures
of output and revenue loss due to any lack of strong disposabiiity of undesirable outputs."
In summary, the analysis of pollution prevention can be modeled within in a number of
theoretical contexts. We have formulated it within the materials accounting framework suggested in
the National Academy of Science (1990) study. Even within this framework, a number of
interpretations can be developed by varying the criteria used to define pollution prevention activity.
For instance, as we have demonstrated, pollution prevention efforts in a plant can be evaluated with
respect to throughput of a particular chemical, or with respect to the amount of desirable products
produced, or more generally, in terms of the effects of the pollution on public health and general well
being. PPF analysis is intended as a screening device to help focus attention on plants, which under
a given model, appear to be leaders or laggers in terms of pollution prevention efforts. Development
of precise recommendations on bow the leaders may be emulated or the laggers helped to improve
their performance requires further scrutiny and study.
A-5
-------
References
Caves, D.W., L.R. Christensen and W.E. Diewert. "The Economic Theory of Index Numbers and
the Measurement of Input Output and Productivity," Econometrica, 50, pp. 1393-1414, 1982.
Charnes, A. and W.W. Cooper. "Programming with Linear Fractional Functional," Naval
Research Logistics Quarterly, 9, pp. 181-185, 1961
Charnes, A., W.W. Cooper and E. Rhodes. "Measuring the Efficiency of Decision Making
Units," European Journal of Operational Research, 2, pp. 429-444, 1978.
Desai, A. "Data Envelopment Analysis: A Clarification," Evaluation and Research in Education,
6, pp. 39-42, 1991
Fare, R., S. Grosskopf and C. Pasurka. "Effects on Relative Efficiency in Electric Power
Generation Due to Environmental Controls," Resources and Energy, 8, pp. 167-184, 1986.
Fare, R., S. Grosskopf, C.A.K Lovell and C. Pasurka. "Multilateral Productivity Comparisons
when Some Outputs are Undesirable: A Nonparametric Approach," Review of Economics and
Statistics, 71, pp. 90-98, 1989.
Malmquist, S. "Index Numbers and Indifference Surfaces," Trabajos de Estatistica, 4, pp. 209-241
1953.
National Academy of Science. "Tracking Toxic Substances at Industrial Facilities: Engineering
Mass Balance versus Materials Accounting," Commission to Evaluate Mass Balance Information
for Facilities Handling Toxic Substances, Board on Environmental Studies and Toxicology,
Commission on Geo&ciences, Environment and Resources, National Academy Press, 1990.
Seiford, LM. and R.M. Thrall. "Recent Developments in DEA," Journal of Econometrics, 46,
pp. 7-38, 1990.
Seiford, L.M. "A Bibliography of Data Envelopment Analysis (1978-1990), Version 5.0, Technical
Report, Department of Industrial Engineering, University of Massachusetts, Amherst, MA, 1990.
Tornquist, L "The Bank of Finland's Consumption Price Index," Bank of Finland Monthly
Bulletin, 10, pp. 1-8, 1936.
A-6
-------
APPENDIX B
Features or the PPF Approach*
Cognitive complexity - PPF reduces multiple variables to a single performance measure
thereby making inter-facility comparisons relatively simple.
Best observed practice - Estimates of performance are based on comparisons with the best
pattern of pollution prevention found in practice rather than comparisons with a
theoretical construct
Comparison - Performance is a relative concept based on the comparison of similar plants.
In particular, for each plant, its score on the performance index is a measure of the
amount of pollution for the plant relative to the pollution output that other plants
produce while maintaining at least the same level activity.
Fairness of the comparisons - The choice of weights assigned to the various activities and
pollutants considered in the construction of the performance index is such that it shows
each plant in its best light. Hence, there is a perceived fairness in the weighting of the
variables. However, as discussed below, the choice of variables is critical to the definition
of good performance
Multiple activities, multiple pollutants - The method uses information on the plant's
activities and the levels of pollution generated. It allows for the simultaneous
incorporation of a multiplicity of measures of plant activity and pollution generation in the
creation of the performance index
Model choice and variable choice - While PPF does not pose any restrictions on the choice
of variables, it should be noted that the choice of any subset of variables implies a specific
model and choice of criteria. The performance score that a plant receives depends
critically on the criteria used for determining performance. Therefore, it is advisable that
a number of representations of pollution prevention performance be explored in order to
capture the multiple facets of superior performance. Different specifications are
particularly important since the frontier, which is defined by actual plants, it is not always
robust with respect to extreme variations in the data.
Values - The choice of a model determines which performance criteria are valued. Hence
data reduction and variable selection is very important when direct comparisons are
involved and there is a potential for setting up dysfunctional incentives.
Controls - By making a distinction between plant or other characteristics which are outside
the control of the plant managers and those that are within their control, it is possible to
This section is based on the authors' personal experiences, Epstein and Henderson (1989) and
Schinnar (1980).
-------
identify what portion of the plant's performance is directly attributable to the manager's
performance versus being due to other factors outside the control of the managers.
Comparison sets - The PPF approach partitions the data into comparison sets wherein
plants with similar mixes of activity and pollutants are evaluated relative to each other and
with respect to a common frontier. Plants which belong to a common comparison set are
all compared to a common reference set of plants that define the best practice frontier.
These plants on the frontier that exhibit the best observed practice can therefore be
selected as candidates for greater scrutiny and more detailed case studies which might shed
light on the factors that contribute to their superior performance. Similarly, attention can
be focused on plants that appear to lag in their performance and more detailed
comparisons of leaders and laggers could potentially yield insights into how to improve
pollution prevention efforts.
Unit and level of analysis - In PPF analyses, the observation which is the unit of analysis is
also the level at which results are obtained. Unlike statistical techniques where the
analysis is based on the means and variance of the data and the results reflect the
aggregate features of the data, PPF yields information pertinent to each observation as
well as results on aggregate information.
Longiiudinal analyses - PPF is well suited for monitoring performance over time.
Comparisons can be conducted relative to a plant's own previous performance or with
respect to other plants included in the dataset.
Computational complexity - PPF scores are obtained by solving linear programs, one for
each plant being evaluated. Optimizers for solving linear programs are readily available,
however, specialized software developed to take advantage of some special features of the
problem can result in computational efficiencies.
References
Epstein M.K. and J.C. Henderson. "Data Envelopment Analysis for Managerial Control
and Diagnosis, Decision Sciences, 20, pp. 90-119,1989.
Schinnar, A.P. "Measuring Productive Efficiency of Public Service Provision,"
CDSE/HTJD project report, Fels Discussion Paper, No. 143, University of Pennsylvania,
July, 1980.
-------
APPENDIX C
TABULATED RESULTS FOR ALL ANALYSES
-------
APPENDIX C
TABULATED RESULTS FOR ALL ANALYSES
-------

70-
Release-Based PPF Ranks vs
Waste-Based PPF Ranks
1991
60-
TJ~
81
-M-
jfi

«
jc
c
(0
cc
U_
CL
CL
50-
67
57
42
80
47
-216-
66
40-
78
45
2
22
15
66
TO
®
op so-
il)
»
«
CD
© 20-
36 1
3 M
TB 53~
70
16
12
16
19
_55_
61
52
10-

77
,f*7.
54
"TT
74
J.

_5S
-r~
60
i
10
T~
20
—I-
30
—r~
40
50
70
Waste-Based PPF Ranks
-------
Release-Based PPF Ranks vs
Waste-Based PPF Ranks
70-
1990
60-
4#+"
V)
JC
c
«
-------
Release-Based PPF Ranks vs
Waste-Based PPF Ranks
70-
1989
60-
67
"51—*®"

73
49
W
C
(0
CC
Q.
Q_
T3
©
V)
CO
m
I
OJ
(A
CO
0)
cr
60-
17
38
40-
16
45	21 M
72
41
30-
I 71 61 9
57 76
-W-
~JT
52
66
-66-
20-
10
19
56
~w

14

10-
16
15
13
a27
3*°
12
"55"
7^1
.li-
ft
-r-
10
"~r~
20
-r~
30
-T"
50
40
Waste-Based PPF Ranks
60
70
-------
70
SO
50
(0
* 40-
co
0c
u.
0- 30-
Q.
20
10-
PPF Ranks and Waste Ranks vs
Waste-to-Throughput Ranks
1991
49"«

2i7
2*V
-rfa-
66
61
16
65
68
15

¥-
79
67
-4S-
7M1 + e
-64+?	—
333
*1
-w—»e—

38
1®1	ft
54 >	*1
* ^
50
Sfi V *
40
41_1
10
-r-
40
20	30
Waste-to-Throughput Ranks
T~
50
60
¦70
Waste Ranks
60
50
CO
-40 C
<0
QC
•30 IS
0)
~-
(/)
I
20
¦10
70
-------
70-
60-
50-
c/>
* 40-
co
CC
Q- 30-
20-
10-
29*
-54-
PPF Ranks and Waste Ranks vs
Waste-to-Throughput Ranks
1990
19
2?


Jfr.
28


79
fl^65
67	
802684
-550-

1B7
717
33
_L2_
38
70

it
LSfi_
18


34
V
—r~
10
I
20
—l—
30
-r~
40
-r~
50
-I-
60
Waste-to-Throughput Ranks
70
Waste Ranks
60
-SO
40
CO
JC
c
(0
ac
¦30 (0
20
10
70
-------
PPF Ranks and Waste Ranks vs
Waste-to-Throughput Ranks
1989
70-
60-


50-
i ate 33
5 79
(/)
J*
C
<0
CC
40-

2BB21
58281
Q. 30-
20-

512
4807
-4&U-
3SL
10-

3
/



-r~
10
—T"
20
-r-
30
—r-
40
—T"
50
—r-
60
Waste-to-Throughput Ranks
70
60
30
70
Waste Ranks
-50
40
V)
c
<0
DC
®
CO
(C
20
10
-------
/

70-
Waste Ranks vs PPF Ranks
1991
so-
le
-55-
-fiz_
12
06
52
50-
46
30
65
61
-W-
81
28

tn
C 40-
(0
cr
®
V)
(0 30-
38
17
54
51
42
80
-68-
27
28
45
47
10
21
56
24
57

74
40
20-
20
13
79
31
15
_5J2_
-64-
77
41
63
71
10-
-W-
72
18
22
-aa-
43
78
69
11
_zs_
70
-r-
10
-r-
20
—J—
30
-r~
40
-r-
50
I
60
PPF Ranks
70
-------
70-
Waste Ranks vs PPF Ranks
1990
60-
JJL
-a-
50- -
so
C 40-
(0
0c
a)
CO 30-
12
55
67
66
-li-
es
41
-ei-
io
5 r
17
51
42
49
14
-a*-
JL
64
34
10
21
3
38
80
41-
74
20-
27
20
45
47
78
24
23
-2fl_
40
57
31
71 22
10-
~*TT
16
4 37
76
69
70
33
-T-
20
—r
30
-1—
50
-J-
60
10
40
PPF Ranks
70
-------
Waste Ranks vs PPF Ranks
1989
70
60
50
-------
70-
1991 Waste Ranks vs
Waste-to-Throughput Ranks
60-
12
16
52
66
-55-
_SZ_
50-
40
65
39

58
49
61
-M-
10
C 40-
54
17
80
38
61
42
-ee-
IC
cc
w
(0 30-
27
29
45
47
10
56
24
57
21
-C3-
74
40
20-
20
13
79
31
15
-64-
41
50
10-
71
63
77
19
72
22
-37-
-69-
34
78
43
1«9
7B
70
-r~
60
10
20	30	40	50
Waste-to-Throughput Ranks
70
-------
70
1990 Waste Ranks vs
Waste-to-Th rough put Ranks
60

50-
12
66
5	67 9
55	5
46	65 41
	53—	61	
40
10	49
42
17	51
14	28
	39	25	
54
30
34
2	19	21
80	3
38	78
	66	—	
74 29
20
27
45 47
24 23
79 24
-22_
40
13
67®
31 71 22
10-
_64	15_
	672~
4 37	fta	33
,7. #<>0
18 1
	 	1	1	1	1	1	r
10	20	30	40	50	60	70
Waste-to-Throughput Ranks
-------
1989 Waste Ranks vs
Waste-to-Throughput Ranks
70-
12
60-

ie
55
51
77
65
50-
JZ_
28
41
-49-
B0
0)
C 40-
(0
GC
54
14

-58-
38
27
81
C/5	
Cfl 30-
20
45
23
79
29
67
20-
340	68 5842
13
47 64
73
_81_
10-
74-
-87-
-6-57	71-796	B1 5q
22
33 44
19
-18.
7*1
15
70
TT
60
10
20	30	40	50
Waste-to-Throughput Ranks
70
-------
APPENDIX D
GRAPHICAL RESULTS FOR RELEASE-BASED ANALYSES
-------

Excellent
Good
Mixed
Poor
5 0 4 4
2	0 2 8
3	5 7 7
2 3 4 9
TOTAL
12 8 17 28

pp.;®
	....J
-sVc ,«• i • vr*r-« *'>?«;«, J
Excellent
Good
Mixed
Poor
18 0 10 11
15	0 5 17
31 "to 11 14
16	4 6 19
TOTAL
80 14 32 61
* PPF Ratings are based on the percentile PPF rank, as follows:
Excellent (Top 20%), Good (21%-40%), Mixed (41%-80%), and Poor (Bottom 20%)
** Data Quality Ratings are based on # of acceptable reports (meeting the
<= 2.0% MBE and total waste > 0 requirements), as follows:
Excellent (100% of reports acceptable); Good (70% or more of reports acceptable
and MBE% of unacceptable reports < 5.0%); Poor (less than 30% of reports
acceptable OR the MBE% of unacceptable reports greater than 10%);
Mixed (all other facilities)
-------
PISM$i




awflili
fBBSBra
33SSMII
PlPf


Excellent
6
0
4
4


Good
7
0
1
5


Mixed
14
2
1
4


Poor
8
1
1
5
TOTAL
35
3
7
18
ILK:?:.1 ivit;V - :S;	=?: h:S;V:	5:"^ = T-1--;.:-.
}*(*' -	i'.Vf." l1 - 'f:iSn i:[* *c.V *y	 0 requirements), as follows:
Excellent (100% of reports acceptable); Good (70% or more of reports acceptable
and MBE% of unacceptable reports < 5.0%); Poor (less than 30% of reports
acceptable OR the MBE% of unacceptable reports greater than 10%);
Mixed (all other facilities)
-------

K >^4
N	4£#*v	'¦*<«¦"• $$$•*«*. ..*• ' ^	<*< ¦yr&iAr"--\" •>?"•>	^ii.,|u,vW.v
• ^..>-iuQ5isa^aaaig...>rx' asa^L-.^. , s _
Jibkdw-k^ -•••

"' '! J" »¦.~
' ~lc,^ vY;
•?/:*•<•¦•: * vv'iwiu'^..
, i * jft


mMrAhtwMrnrli

36
2,000
16
7
9
67.6
44%
P

63
3,000
3
3
0

100%
E
25
73
3^00
5
3
2

80*
M

68
4,000


0

100%
E
9
6
4,160
1

0

100%
E
6
7
4,160
20
13
7
35.3
65%
P
49
39
6,400
2

1
2.3
50%
M
46
60
6,000
10
7
3
22.0
70%
P
48
27
6,919
5
4
1
99.9
60%
P
14
10
9,500
7
7
0

100%
E
20
61
10^00
6
4
2
49.9
67%
P
54
76
10,500
15
6
7
9.1
53%
M

76
10,500
2

1
4.4
50%
M
7
77
10,500
6
4
4
33.1
50%
P
11
64
12,000
9
6
3
35.2
67%
P
30
74
13,931
1

0

100%
E
1
14
14,415
16
15
1
6.8
94%
M
52
2
16,000
13
12
1
56.6
92%
P
22
3
16,000
9
3
6
29.6
33%
P
25
43
16,200
9
4
5
58.0
44%
P
13
54
17,000
3
2
1

67%
M
15
4
17,324
9
9
0

100%
E
1
5
17,524
6

4
11.2
33%
P
63
16
19,000
4
4
0

100%
E

53
19,366
2
1
1
3.0
50%
M

55
19^66
6
5
1
37.9
83%
P
37
72
27,000
2

1
2.1
50%
M
9
15
27,000
5
4
1
96.7
60%
P
45
61
30,000
7
1
6
15.6
14%
P
59
11
37,700
19
9
10
17.5
47%
P
20
19
37,756
6
6
2
66.8
75%
P
17
17
36,261
3
2
1

67%
M
30
16
39,261
14
10
4
42.4
71%
P
56
30
55,000
6
6
2
8.7
75%
I
M

66
66,000
21
11
10
27.7
52%
P
62
31
94,000
1

0

100%
E
12
9
94,000
19
15
4
49.7
79%
P
23
13
143,961
3
3
0

100%
E
27
12
143,961
54
29
25
20.5
54%
P
51 1
-------
OtLX"'* "v >" ^'t?* v	. ,UJ<8
47
UNK
7
8
1
2.0
88%
a
80
n
UMK
2
1
1
U
50*
M
27
90
UNK
9
2
1

87%
M
18
26
•


1

0%
P

21
17
4
8
1
94
89%
a
80
78
18
8
8
2

78*
a
42
•0
18
1

1

0%
p

44
20
1

1

0*
p

54
28
2
2
0

100%
e
39
M
28
2

2

0*
p

M
28
9
2

24
87*
M
64
46
28
4
9


78*
a
98
at
30
7
7
0

100*
e
87
44
M

1
0

100*
E
88
23
94
7
8


88*
Q
81
23
48
2
1
1
118 X
80*
P
41
71
8*
4

2
74
80*
M
92
42
80
7
8
2
94
71*
Q
40
94
•0
8
2
9

40*
M
8
40
82
t
9
8
94.1
99*
P
3
42
88
2

2

0*
P

44
78
4
4
0

100*
e

"
"
12
10
2
9.7
89*
a
44
41
93
•
8
4
94.2
86*
p
18
70
100
8
e
2
904
78*
p
23
24
120
8
4
2
2J
87*i
M
83
M
128
2
1
1
9J
80%
M
27
87
180
7
4
9
884
87%
P
88
1
180
a
8
0

100%
E
18
82
180
8
8
9
44
m
M
88
81
200
8
9
2
4.1
80%
M
32
78
200
8
4
2

87%
M
6
87
280
19
8
8
18 J
98%
P
43
37
900
9
2
1
174
87%
P
4
as
900
4
1
9
28.2
26%
P

44
928
17
18
2
•94
aa%
P
47
20
470
8
4
2
44
87%
M
38
44
700
2
1

10.7
80%
P
88
92
728


1
1004
0%
P

28
1,000
8
8

9.7
89%
a

2>
1.911
8
8
0

100%
E
32
8
1,880
14
12
2
4.2
88%
a
32
-------
M
2300
3
8
1

m
O

U
3,000
1

0

100*
E

73
MOO
2
1
1
73J
BOW
P

H
4300
1

1

0*
P
3
•
4,iao
1

0

100%
e
7
7
4,180
12
4
3
473
m
p
18
33
8,400
1

0

100%
E
40
ao
3,000
4
3
1

7m
Q
32
27
•31*
3
1
2
18.4
m
P
11
10
3,800
3
8
1
113
33%
P
30
•1
10,200
4
3
1
73
78*
M
88
73
10300
1

0

10Cm
E
32
73
10300
3
4
2
373
37*
P

77
10300
3
3
8
200573
33*
P

34
12300
4
4
0

100%
E
17
74
13331
1
1
0

100*
E
4
14
14,410
3
3
1
2.1
33%
a
32
2
13300
4
4
0

100%
E
21
•
13300
3
2
3
33.1
40%
P
24
41
13300
13
13
3
88.1
33%
P

S4
17300
2
2
0

100%
E
10
5
17334
4
4
0

100%
E
48
*
17324
8
8
0

100%
E
1
1*
13300
8
3
2
3.7
30%
M
8
U
13333
2

0

100%
E
30
53
13333
7
7
0

100%
E

72
27300
1
1
0

100%
E
8
19
27300
8
2
3
3.4
40%
M
2#
•1
30300
3
1
2
33.1
33%
P
80
11
37.700
11
3
3
83.1
27%
P

1*
37,733
7
7
0

100%
E
82
17
33,231
2

0

100%
E
23
13
33331
11
7
4
10.7
34%
P
83
30
H300
3
4
1

0O%
O

M
33300
3
3
1

33%
a
47
31
34300


0

100%
E
18
3
34300
3
4
3
4.9
37%
M
43
13
143331
4
2
2
2J
80%
M
13
13
143331
23
14
11
90.2
83%
P
21
-------
-------



usmwiMa!!
ffeU
R£PO«T*


SllMI

M
2,000
•
s
1

83*
Q

a
3,000
1
1
0

100*
E

73
3300
2
1
1
7IJ
80*
P
80
M
4300
1

1

0*
P

•
4,180
1
1
0

100%
e
2*
7
4,180
12
4
•
ATS
S3*
p
42
M
6300
1
1
0

100*
E
30
*o
»JXX
4
*
1

78*
a
42
27
Mit
3
1
2
18 A
33*
p
11
10
•300
•
8
1
1IJ
m
p
22
•1
10,200
4
3
1
73
75*
M
31
78
10300

1
0

100*
e
34
77
10300
•
3
8
200*7.0
38*
p
84
7*
10300
•
4
2
373
87*
p

*4
12300
4
4
0

100*
E
42
74
133*1
1
1
0

100*
E
1
14
143H
•
•
1
2.1
8**
0
31
2
1*300
4
4
0

100%
E
»
a
1*300
¦
2
3
3*.1
40*
P
10
41
18300
i*
1*
3
SS.1
*3*
P

84
17300

2
0

100*
E
3
4
17 J24
8
8
0

100*
E
7
S
17324
4
4
0

100*
E
4*
1*
1*300
•
3
2
*.7
*0*
M
2
53
1*38*
7
7
0

100*
E

U
1*3**
2
2
0

100*
E
13
72
27300
1
1
0

100*
E
34
1*
27300
8
2
3
*3
40*
M
13
•1
30300
3
1
2
M.I
33*
P
34
11
37,700
11
3
•
88.1
37*
P
•
1*
37,7*4
7
7
0

100*
E
0
17
3*3*1
2
2
0

100*
E
22
1*
3*3*1
11
7
4
10.7
*4*
P
•
30
88300
•
4
1

80*
a

**
**300
•
•
1

W*
a
*0
31
*4300
1
1
0

100*
E

•
*4300
•
4
2
43
*7*
II
87
13
1433*1
4
2
2
23
80*
M
17
12
1433*1
28
13
12
303
82*
P
81
-------
»
UNK
1
pttatfTV;
wspwre:
i
Hffi
0
'WWfP'
100%
E
.¦SSSfeiW**
81
47
UNK
3
3
c

100%
E
43
SO
UNK
3
3
0

100%
E
43
28
•
1

1

0%
P

11
17
2
1
1
14.7
50%
P
«
7*
18
2
2
a

100%
e
4»
SO
1*
1

1

0%
p

44
30
1
1
0

100%
E
55
54
38
1
1
8

100%
e
25
n
38
1

1

0%
p

a
at
1
1
0

100%

94
48
38
3
3
0

100%

>0
28
30
2
2
0

100%
E
51
4*
H
1
1
0

100%

58
23
38
4
4
0

100%
E
58
n
46
1
1
0

100%

38
71
68
3
3
0

100%
E
31
42
•0
4
4
0

100%
E
28
*4
80
1
1
0

100%

15
40
82
3
1
a
48J3
33%
P
18
82
¦8
1

i

0%
P

44
78
2
2


100%
E

W
80
S
3
2
4J0
80%
U
28
41
•2
f

3
ru
40%
P
82
70
100
8
3
3
41J
80%
P
19
M
120
1
1
0

100%
E

M
128
1
1
0

100%
E
42
57
150
3
3


100%
E
28
1
180
3
1
2
183J
m
P

53
180
8
8
1
96 M
83%
P
81
78
300
1
1
0

100%
E
5
St
aoo
3

1
XIA
•7%
P
42
87
380
8
4
2
U
87%
II
12
ST
300
3
3
0

100%
E
17
M
300
a
1
1

80%
W

48
338
8
«
0

100%
E
22
20
470
3
3
1
41*4
87%
P
21
M
700
2
2
0

100%
E
S3
U
725
1
1


100%
E

28
1,000
4
3
1
294
7S%
P

28
1.311
2
1
1

50%
H
4
e
1#60
4
4
0

100%
E
2Q
-------
-------

RESULTS AND RANKS OF THE THREE APPROACHES




BASED ON WASTES






1991




CHTK WASTES

WST/THPT RATIO
PP* ANALYSIS
®LANT 10
WASTES
HANK
RATIO RANK
SCORE
RANK
1
3
1
0 00000
10
1 1000%
16
2
3000
42
0 00200
10
07000%
22
3
3 939
41
0 00372
41
0 3000%
25
4
0
4
000000
1
100 0000
1
s
270000
02
0 02190
96
0 0067%
S3
6
01
19
000011
11
16 4000%
6
7
70 700
90
0 03119
90
0 0231%
40
0
000
20
000180
39
0 1000%
12
0
9 030
47
0 00312
16
0 0000%
23
to
1.360
14
000021
10
04000%
20
11
10
0
0 00019
12
06000%
20
12
07)412
03
0 00*73
40
0 0162%
91
13
307
29
000042
22
03000%
27
14
7 011
90
0 00779
47
0 0160%
92
18
137
34
0,00140
30
0 0303%
49
19
16.033333
00
0 10300
02
00120%
90
17
0 000
40
0.00013
29
02000%
10
10
42
11
0 00000
0
2.1000%
17

001
29
0 00104
32
0 07*4%
16
at
1 412
36
0 01909
94
0 0070%
00
22
00
14
0 00101
14
00979%
41
23
000
30
0 01023
99
0 00*7%
01
24
930
11
000004
40
0 0130%
91
29






20






27
2.113
10
000034
20
94000%
U
n
10 032
91
0.00071
91
0 0127%
97
20
1.110
30
0 00000
27
0 1000%
12
30






31
100
22
000027
10
79000%
12
12






S)
20
10
040041
21
0.3009%
27
14
19
8
0 00003
0
71 0000%
0
a






30






17
20
10
0.00002
4
064000%
4
10
0.570
40
0007*1
40
04000%
27
»
0.000
91
0.00370
42
04327%
40
40
290
a
0.00008
2
004000%
3
41
71
17
040000
34
34000%
10
42
940
40
0.02013
90
00009%
40
4]
19
0
ootffn
10
04000%
13
44






46
1.470
30
0 00127
20
00000%
10
40
46429
97
040477
43
00290%
47
47
1.770
37
0.00030
46
04102%
90
40






41
11 101
94
044400
01
0 0027%
00
00
100
21
040120
10
14000%
10
91
4JH
44
040111
20
0 1000%
12
92
2402427
04
0 14000
03
0 0011%
00
93






94
4 102
43
040003
0
34009%
19
90
100 900
00
040100
31
0 0709%
17
90
1.000
12
0 00104
33
0 0747%
30
97
1JM
SI
041071
92
04114%
90
90
344*7
m
042SS
57
04om
•4
90






00






01
ajto
90
0 29001
00
04009%
90
02






03
00
10
0 00172
40
04000%
29
•4
129
20
0 00063
2)
0 2009%
30
00
01579
90
0 19200
04
0 0130%
96
00
3 077432
09
0 20740
06
00002%
62
07
2U009
01
004440
00
00449%
43
00
1.409
40
0 00200
37
0 0411%
44
00
10
0
040019
13
134000%
0
70
7
3
040022
17
06000%
23
71
110
10
000004
20
0 1000%
32
72
29
12
040019
11
13 0000%
0
71






74
301
29
0 00002
2
100 0000
1
79
1
1
000003
9
77 0009%
9
70






77
07
10
0 00017
19
11 0009%
11
70
0
4
0
7
024000%
7
70
970
27
001300
93
00012%
42
00
9937
40
0 00611
44
0 0239%
40
ai
0 420
92
0 00030
90
0 0111%
94
-------

RESULTS AND RANKS OF THE THREE APPROACHES




BASED ON WASTES






1990




CAT* WASTES

WST/THPT RATIO
PP* ANALYSIS
plant io
WASTES
RANK
RATIO RANK
SCORE
RANK
i
3
2
000009
6
4 9000%
13
2
3 391
37
0 00291
34
0 7000%
21
3
3 535
34
0 00399
U
05000%
24
4
20
9
0 00000
1
100 0000
1
5
139 174
M
0 00049
42
0 0329%
45
S
39
10
0 00010
10
17 9000%
7
7
910
20
000029
19
1 0000%
19
a
407
19
0 00079
24
09000%
20
0
249 961
59
009734
99
0 0271%
49
10
14 997
47
0 00197
29
0 2000%
30
12
39.194
54
000000
19
07000%
21
13
450
19
000029
17
1 0000%
19
14
9 314
42
0 00292
33
0 1000%
32
15
90
11
000092
22
04000%
29
19
30.279.799
91
019314
90
00000%
56
17
1.700
43
0 00091
29
03000%
29
it
1
1
000004
9
43 4000%
5
19
4,049
39
ooiaoo
44
00227%
52
2D
901
21
0 00346
31
0 1000%
32
21
8902
39
006017
54
00049%
90
23
120
15
000200
30
0 1000%
32
23
1 500
29
0 02143
50
00129%
50
24
1.497
25
0.01543
47
0 0179%
53
29






20






27
1.997
29
0 00021
14
9 0000%
11
a
9.093
41
0 00097
43
00279%
49
a
970
22
0 WW?
7
54000%
12
30






31
100
14
0 00060
21
37000%
15
32






>3
30
9
0.00044
30
04000%
23
34
3413
33
0.00007
19
44000%
13
SB






39






37
20
9
04)0002
3
974000%
3
39
3 400
32
0.00093
39
0.5000%
24
39
IJOO
40
0.00979
37
00727%
40
40
97
13
0.00002
2
1004000
1
41
30.700
53
0.23094
91
00093%
57
42
12J10
46
0.07907
39
04230%
51
a






44






46
1 749
27
0.00397
39
04969%
39
49
21.910
51
0.00430
40
00032%
42
47
1404
29
0 00090
41
00401%
a
49






49
13.100
49
044740
53
00069%
59
90






SI
104Z7
44
040062
39
04791%
30
S3
1410.792
90
0 10902
99
0403V%
91
U






54
4jn
30
000002
5
92000%
10
90
90 190
50
0 00140
27
0.3000%
30
M






57
2f0
17
0 00071
23
04000%
29
50
19,477
40
041000
49
04171%
54
99






•0






•1
90.379
50
047300
97
00200%
90
92






a






*4
90
11
00009
15
1 1000%
17
96
04.000
52
049154
56
00297%
49
99
449 792
50
0 03440
52
0 0292%
47
97
191J07
57
0 02992
91
00706%
41
99
2.92
30
040349
32
0 1000%
32
99
10
5
000012
12
149000%
9
70
7
4
000013
13
24000%
19
71
120
15
040070
29
04000%
26
72
25
9
000011
11
194000%
9
73






74
790
23
0 00002
3
99 4000%
4
79
9
3
000007
9
294000%
9
79






77






79
2 979
31
001303
49
0 1000%
32
79
1 296
24
001232
45
00943%
44
90
3 943
39
0 00193
29
0 1000%
32
61
20 030
49
001992
«4
0 0145%
55
-------

RESULTS AND RANKS OF THE THREE APPROACHES




BASED ON WASTES






IMS




CflTK WASTES
WST/THPT RATIO
PPF ANALYSIS M
^LANT 10
WASTES
RANK
RATIO RANK
SCOPE
rank
1
2
3 790
40
0 00274
30
0 4000%
30
3
500
10
0 00017
10
7 1000%
10
4
1 290
39
0 00011
7
10 9000%
7
s
63.074
90
0 00637
40
0 2000%
40
e
900
10
000106
a
1 1000%
a
7
4900
42
0 00300
44
0 3000%
42
s
371
0
0 00007
20
1 9000%
20
9
126 160
00
0O»»
97
00411%
37
10
0 6Z7
40
0 00009
22
1 2000%
22
11
9
3
0 00010
0
11 0000%
9
12
092403
63
000991
90
0 1000%
91
13
960
22
0 00047
10
2.0000%
17
14
4620
44
0 00100
31
0 0000%
31
18
20
4
ooooa
13
5 1000%
13
ie
32.432
97
040010
0
73000%
0
17
10400
91
0 00102
a
1 2000%
22
10
1
1
0 OOOQS
2
41 7000%
2
19
291
7
0 00010
9
7 9000%
6
20
1 286
36
040143
a
1 4000%
21
21
900
10
oooia
41
04000%
30
22
ao
6
0 00334
40
04000%
30
23
1 900
37
0 023*2
90
0 0630%
90
24






a






2D






27
1 062
36
040010
11
04000%
11
2i
10,000
92
000900
91
0 1000%
91
2B
647
24
0 00004
4
a 4000%
4
30






31






32






n
900
10
041110
93
0 1000%
91
34
1.000
a
00003*
19
90000%
19
36






30






37
900
10
000040
17
24000%
17
30
3400
41
040419
40
04000%
42
30
9421
46
040301
39
0 4000%
30
40
1000
a
040027
10
44000%
10
41
11 441
93
046007
02
0 0200%
02
42
1400
a
040130
a
06000%
a
43






44
900
10
0-030*4
96
04661%
90
40
1 140
33
0 00147
30
04000%
30
44
0460
47
000060
21
1 3000%
a
47
1.000
a
000402
46
04000%
42
41






40
0400
90
046341
99
0 0071%
90
90
400
9
040360
43
04000%
42
91
13464
64
040400
47
04000%
42
92
700477
02
0 00601
01
00211%
01
93






94
4406
43
0 00003
3
37 6000%
3
96
14 600
96
0 00031
13
9 1000%
13
90
1 000
a
000112
a
1 1000%
a
57
900
10
oooia
27
04000%
a
90
2477
40
04C22D
34
04000%
34
90






0D






61
961
21
040346
37
04000%
34
02






03






64
1000
a
0 00479
40
03000%
42
06
42400
96
0 19170
63
00079%
63
00
440466
01
003200
00
0 0964%
60
07
960
a
000021
12
5 7000%
12
00
1000
a
0 00100
24
1 1000%
a
00






70
94
9
0 00060
19
2.2000%
10
71
900
10
0 00206
33
0 6000%
31
72
900
10
0 00233
a
04000%
34
73
1 000
a
0 02941
90
00406%
96
74
900
10
000001
1
1004000
1
73
9
2
0 00006
9
22.1000%
3
70






77
17 200
96
0010a
94
0 0740%
94
76
900
10
0 00242
30
0 9000%
34
79
1 190
34
001103
92
02000%
40
90
6447
40
0 00361
42
03000%
42
91 1
2 000
36
000200
32
0 0000%
31
-------

RESULTS AND HANKS OF THE THREE APPROACHES




BASED ON RELEASES





1991




TAJ AEl£AS€S
REL/THPT ratio
PPf ANALYSIS (In %]
=\ANT 10
RELEASE

RATIO
RANK
SCORE
Rank
1
3
2
0 000061
11
I 5000%
16
2
3 900
44
0002000
49
0 3000%
29
1
3936
43
0 003716
47
a 2000%
12
4
9
5
0000001
1
100 0000
i
9
290 029
94
0020030
00
0 0001%
03
0
61
IT
0 000109
13
7 5000%
9
7
400
29
0000200
19
0 7000%
a
0
000
31
0 001660
42
0 1000%
30
9
9030
49
0 003121
40
03000%
a
to
1 390
36
0000212
20
0 0000%
a
11
10
7
0000149
19
0 0000%
a
12
96 794
01
0000660
2D
0 2000%
32
1}
307
20
0 000423
29
0 3000%
a
14
90 096
00
0 019986
«
0 0267%
92
15
194
24
0.002004
43
0 0913%
47
10
190 210
«
0 001001
34
0 1000%
30
17
6006
92
0000021
31
02000%
32
18






19
42
ift
0 000060
10
2.1000%
17
20
001
30
0.001030
30
0 0764%
43
21
1 412
37
0 019000
90
0 0070%
01
29
00
19
0.001007
41
0 0079%
40
23
000
32
0.010220
90
0 0007%
02
24
030
33
0000090
96
0 0130%
90
S






20






rt
2.113
41
0 000930
24
2.4000%
10
2D
10 032
90
0.000710
9«
00127%
00
2B
3.110
42
0 £00000
39
0 1000%
30
30






31
190
23
0000207
22
3 1000%
14
32






33
3D
11
0 000400
a
04000%
a
34
19
9
0000020
7
29 1000%
7
30






30
31
14
OAOtlO
14
04000%
22
37
20
11
0400022
9
30 4009%
9
30
0.370
91
0-007929
93
01000%
30
39
em
99
0403737
40
04927%
90
40
229
a
0400039
4
404000%
4
41
71
10
0000003
30
1 4000%
16
42
94a
47
OjQ2>1Z7
02
04201%
99
43
19
9
0000333
23
24000%
19
44






46
1.470
30
0
*
1
30
0 0983%
42
40
29.200
99
o^osm
- 44
0 0400%
40
47
1 007
39
00000*7
91
0 0201%
99
40






m
11 101
57
0444006
04
00027%
06
90
1
t
0400000
2
1004000
1
31
4 294
40
0401100
36
0 1000%
30
92
7J73
94
0400000
27
1 2000%
21
93






M
4 109
49
0400034
0
30000%
13
96
0.200
90
ooooou
12
1 3000%
»
90
1,000
39
0 001944
30
0 0747%
44
67
167
34
040M61
94
04140%
90
90
MJ07
90
043B2BO
01
04063%
64
90






00






•1
190
22
0401461
37
0 0000%
a
«2






03






04
IS
21
0 000620
a
02000%
32
06






00
02.734
02
000429
40
04200%
91
07
82406
09
0444404
03
0 0104%
90
00
1 900
40
0 001000
40
0 0727%
45
00
10
7
0 000194
10
9 3000%
10
70
7
4
0 000219
21
0 0000%
a
71
110
2D
0000030
32
0 1000%
30
72
29
13
0000194
17
9 3000%
10
73






74
301
27
0000020
3
4! 000(7%
3
79
3
2
0000020
0
31 0000%
6
70






77
07
«0
0 000173
10
47000%
12
70
9
9
0000032
0
294000%
0
70
970
a
0 013000
97
00612%
40
00
9 937
40
0 006130
90
0 0239%
94
01
7 177
93
0 007007
92
0 0174%
97
-------

RESULTS AND RANKS OF THE THREE APPROACHES



BASED ON RELEASES






1990




TRI releases
REL/THPT RATIO
PPP ANALYSIS flfi %)
P\_ANT ro
RELEASES
RANK
RATIO
RANK
SCORE
RANK
1
3
3
0000000
9
4 9000%
19
2
9 301
42
0 002000
a
0 7000%
a
3
3 330
37
0 003AA3
43
0 9000%
30
4
a
6
0 000003
1
1004000
i
s
90 010
37
0000009
40
0 0419%
90
6
M
12
0 000102
13
17 9000%
9
7
910
a
0000292
19
1 0000%
a
S
407
21
0 000704
30
0 9000%
27
9
19401
90
0 01*903
M
0 0029%
49
10
14397
90
0 001974
39
02000%
38
12
25449
94
0 000279
21
1 1000%
a
19
436
a
0 000297
a
1 0000%
24
14
04.&0
90
0000200
22
144000%
10
15
00
19
0.000917
27
04000%
31
19
19,100402
04
0401000
90
0 0197%
90
17
• TOO
49
0000000
92
0.9000%
93
10
1
1
0.000042
7
49 4000%
9
10
40
11
0.000074
11
17000%
17
ao
091
a*
0 00940!
39
0 1000%
37
21
9402
43
0.000109
90
0 0040%
90
22
120
10
0.001909
a
0 1000%
97
a
1400
a
0-021433
94
0 0120%
99
34
1 407
a
04)19027
91
0 017V%
S3
a






a






27
1407
a
0.000230
17
90000%
19
a
0 093
40
0400974
40
00270%
92
a
070
a
0.000040
0
54000%
14
90






31
100
10
0.000408
a
3 7000%
17
32






a
»
0
O.OOOMI
a
04000%
a
34
tin
a
0400900
a
44000%
19
X






M
1Q0J00
90
0 420019
01
04049%
01
37
a
0
0400019
3
07 0000%
3
U
27
10
0400091
9
904000%
9
a
IJOO
44
0409771
42
04727%
49
40
07
10
0400019
2
1004000
1
41
19.021
91
0190019
00
0 0190%
57
42
110
17
0400717
29
24000%
21
49






44






40
1.740
31
0402009
40
04000%
49
40
12310
40
0408910
a
0 1000%
J7
47
14S7
92
0400070
46
0 0410%
40
40






40
11100
40
0447409
57
04060%
90
90






91
1QJ27
47
040959
41
04701%
44
92
4 119
a
0400200
a
24000%
a
u






94
4.279
40
0400094
9
04000%
12
96
9.100
41
0400004
12
34000%
10
90






57
270
a
0400714
a
04000%
9t
90
10.477
u
0410009
u
04171%
99
90






od






01
1J9
30
0405470
44
04000%
a
02






«






04
00
19
0400259
10
1 1000%
a
<0






00
102.1a
90
0407022
47
00604%
47
«7
101407
02
ooaoa
90
04700%
49
00
2.920
94
0 003470
97
0 1000%
97
eo
10
9
04001 a
19
14 0000%
11
70
7
4
04001a
19
24000%
a
71
1®
10
0000707
31
04000%
a
72
a
0
0400111
14
104000%
9
73






74
700
a
0 000011
4
99 4000%
4
»
0
9
0 000072
10
34000%
7
70






77
100 000
91
04073*4
02
04030%
02
78
2470
a
04190a
90
0 1000%
37
70
1 200
27
0 012922
40
0 0949%
91
00
3 049
a
0 001994
>4
0 1000%
37
CI
20 030
99
ooiooa
93
0 0145%
90
-------
RESULTS ANO RANKS OF THE THREE APPROACHES
BASED ON RELEASES
1969
°LANT IC
TK RELEASES
RELEASE RANK
REL/THPT RATIO
RATIO RANK
ppp ANALYSIS Oh X)
SCOPE RANK
\
2
9 790
40
0 002730
37
04000%
a
3
MO
12
0 000160
11
7 1000%
11
4
1 500
39
0 000130
10
t 7000%
10
S
08 900
m
0 009100
47
0 2000*
49
0
1 500
39
0 003104
40
04000%
30
7
1,000
29
0 000070
21
1 4000%
22
a
300
11
0 000001
23
1 9000%
21
0
6 100
92
0 001003
32
0 0000%
32
10
10.4a
S4
0 001107
27
1 0000%
a
u
•
3
0 000104
a
11 0000%
a
12
41.003
94
0000431
10
20000%
1a
19
560
10
0000407
17
2.9000%
*7
14
1 wi
o
0000062
22
1 3000%
24
19
30
0
0000092
19
34000%
19
It
02*12
07
0000470
10
2.9000%
10
17
07.000
00
0000101
90
01000%
40
ie
212
0
0.000100
40
03000%
49
it
1400
30
0 0010a
34
1 2000%
a
2D
1.2n
34
0 001430
a
1 4000%
22
71
790
20
0004071
40
0 2000%
45
23
900
12
0000100
91
02000%
49
a
1 900
30
0023422
63
00030%
03
2*
2.000
44
0 0X3407
09
00300%
a
29
279
10
0 004371
44
0.3000%
43
2B






Z7
1.002
40
0 000170
12
04000%
12
a
tOJOO
90
0.000000
90
0 1000%
90
a
647
*
0400041
9
94000%
9
30
12
4
0000040
0
24000%
0
31
200
0
0.001121
a
1 1000%
a
32






M
900
12
0.011100
90
01000%
90
34
1 000
29
0.000342
13
94000%
13
3B






M






37
2UOOO
41
0.00 ua
31
07000%
a
31
a
9
0.000033
3
394000%
3
3i
9471
40
0003144
a
04000%
a
40
1400
2B
0000270
14
44000%
14
41
1^0
32
0400464
90
02000%
40
42
IM
a
0401303
a
0JOOO%
a
4]
10400
90
Q2B0M4
00
04047%
00
44
900
t2
ooaoot
02
00501%
02
40
9J10
90
0407440
94
0 2000%
40
41
1.7M
40
0000000
a
2.1000%
a
47
1J90
>2
0000031
40
0 2000%
40
41






40
MOO
93
0 03dt1«
64
04371%
04
00
3
1
00000a
2
414000%
2
SI
19441
01
0000000
40
0-2000%
40
u
37.237
OS
0402003
a
0 4000%
a
91






94
9JO0
47
0400030
4
314000%
4
M
1.100
91
0400131
0
0 2000%
0
91
1,000
a
040110
a
1 1000%
a
07
1,000
a
O4O207O
a
04000%
a
60
19411
00
0413001
00
04000%
00
00






•0






01
700
22
040041
41
04000%
a
09













•4
IJOO
a
0 007132
93
02000%
49
00
900
12
0401700
a
07000%
X
00
11,400
oa
0002303
a
04000%
a
07
100.030
00
0 030400
00
04304%
00
oo
120
49
04034a
42
0 4000%
a
00






TO
34
7
0 OOOBBB
10
2JDOO%
10
71
790
»
0 003074
a
04000%
a
72
791
a
0003407
43
04000%
43
n
2.000
41
0000031
07
04204%
07
74
900
tt
0000012
1
1004000
1
79
9
2
0400004
7
22.1000%
7
71






77
17 200
90
OO102Z7
01
0 0740%
01
78
900
12
0002410
34
04000%
a
71
1 100
31
0011027
97
02000%
40
00
10 413
90
0000000
92
0 2000%
40
61
1)000
97
0 012000
N
0 0020%
a
-------


TBI BATtO AND % CHANGE IN TOTAL RELEASES


*CLEA$£S


Tflr furro
% CNANC£


rflr **rrc
SCfeUriOC

PmWTJD
n*00
A* 00
R* 01
PUWT ID#
ao/ao
U "OBO

PLANT C»
ei«c
90 TOOl
Rank
1
0
1
3
l
NA
NA
NA
1
100 00%
S90COV
37
2
1 790
9 301
3 HO
2
W 23%
-0 7902%
41
I
73 00%
an»r%
9
3
900
3 US
3 930
3
707 00%
007 0000%
03
i
100 00%
00000%
V
4
1 900
20
•
4
1 33%
40 0007%
12
4
45 00%
90 0000%
17
9
6BM

390 oa
9
Hi 40%
41 4070%
90
9
290 00%
190 9004%
90
9
1 W
90
61
B
ian
•00 0007%
19
O
103 30%
31000%
49
7
t 000
910
400
7
il 00%
-40 0000%
20
7
01 37%
4 0279%
32
a
300
407
MO
a
10711%
J 1063%
44
9
107 00%
07 0702%
92
f
0 140
18 061
9 030
»
1990 79%
<490 7467%
07
0
4 *4%
00 9417%
10
10
i0 < at
(4 907
1 350
10
1)0-06%
30 0700%
40
10
031%
00 0000%
11
i \
»
&
10
11
0 00%
•100 0000%
r
11
NA
NA
NA
12
43*00
25440
M764
12
02.13%
47 0747%
33
13
21943%
119 0322%
37
U
MO
400
307
13
70 40%
21 9617%
30
13
00 00%
104407%
a
14
74 200
MM
90 000
14
4-230 77%
4130 7211%
oa
14
M 70%
-40 2004%
22
13
30
00
104
19
20047%
100 0000%
94
11
323.33%
223 um
00
K
02432
10,190 JOS
106 210
10
106*1 13%
10401 1341%
00
IS
1 00%
•40 0004%
e
17
47.000
9.700
oooe
17
11 19%
•404000%
21
17
00 00%
414072%
s
14
212
1
0
10
0 47%
40493%
10
10
000%
100 0000%
i
it
1400
40
42
10
307%
-00 090%
19
10
00 71%
-14 2007%
31
20
1 200
oot
001
20
11 12%
-404700%
30

100 00%
00000%
37
21
790
M63
1.411
21
11300%
•13 0000%
04
21
»01%
704029%
13
22
900
120
00
21
at oo%
-704000%
a
22
4000%
40 0000%
10
a
1400
1400
000
23
10040%
00000%
42
a
9343%
-40 0007%
21
*
UBS
1,407
830
M
sejn
-444102%
31
21
9642%
•44 1420%
23
3
270
0
0
a
000%
• 100 0000%
1
a
NA
NA
NA
as
0
0
0
20
HA
MA
KA
a
NA
NA
NA
37
i«a
1 007
2.113
27
no in
IB 1332%
40
27
10044%
0 3412%
40
as
toooo
SOU
10032
a
dim
17 0079%
30
a
11044%
104300%
47
9
K7
on
3.110
20
7010%
-304473%
30
a
404 10%
3041701%
02
30
19
0
0
X
000%
-1004000%
1
30
NA
NA
NA
31
a
too
100
31
40 70%
412106%
a
31
10040%
00 0000%
93
33
30
0402^90
0
32
31*741*017%
111740004007%
70
32
000%
1004000%
1
33
900
20
20
33
400%
•00 0000%
10
33
100 90%
00000%
IT
34
1000
9JB13
19
31
301-30%
auooo%
00
34
043%
-004720%
0
30
0
0
0
30
NA
NA
NA
a
NA
NA
NA
30
tl
101 MO
11
30
NA
NA
NA
a
000%
-004710%
3
IT
2000
20
20
37
140%
404000%
11
37
10040%
00000%
37
30
a
27
1570
31
00.10%
-04000%
40
30
24)3343%
242394*33%
64
30
9071
(JOG
a ,000
3t
140J0%
404007%
92
30
9040%
40140%
30
40
1.000
07
29
¦40
0 70%
404000%
10
40
ai4o%
1314660%
96
*1
iJOO
(Mfl
71
41
101340%
14104000%
00
41
040%
-004040%
7
41
1000
Itfl
Ma
42
1»40%
404009%
to
42
404041%
47404001%
00
43
lOJBi
46 464
19
43
43041%
330.2300%
01
43
0 03%
404072%
4
44
900
Q
0
44
a oe%
•1004000%
1
44
NA
NA
HA
44
4 J it
1 741
1 470
40
30.01%
-004007%
z?
46
04 71%
19 2021%
a
44
J 790
tU1l
&400
40
mm
223 7000%
90
40
20140%
1014010%
90
47
1.290
1 tsa
1*7
47
H0 10%
40 1000%
91
47
0240%
-7 1196%
34
41
900
0
4,031
40
MA
NA
NA
40
NA
NA
NA
44
OJOO
13.100
11 111
40
i»m
374047%
tl
40
00 43%
14 9729%
30
90
3
5
1
90
«on
• fOO 4000%
1
90
NA
NA
NA
»l
14J41
10427
4 2U
91
n«%
-42J10>%
33
91
30 27%
-40 7120%
19
U
17.297
4 113
7J73
92
1149%
404040%
20
92
10340%
034110%
64
U
172.100
0.1 to
H37
1)
MA
NA
NA
92
MA
NA
NA
M
3JOO
4.273
4.102
94
0040%
•104140%
37
94
0040%
-4 00*9%
30
90
• 100
5.100
0.200
98
0407%
•394290%
34
96
11040%
10 4006%
40
90
1.000
0
1.000
90
040%
-1004000%
1
90
MA
NA
NA
57
1000
z?o
097
57
vw%
•734007%
a
57
904 44%
34 4444%
01
51
14*11
10,477
34 JOT
90
10UA
3-9408%
42
90
12747%
3747T*
40
at
0
0
0
9i
HA
NA
NA
90
KA
NA
NA
m
0
0
0
•0
MA
NA
MA
00
MA
NA
NA
01
790
\JSM
190
CI
3U39%
1034337%
97
01
044%
40 1030%
12
aa
0
0
0
a
NA
NA
NA
02
NA
NA
NA
*»
47.000
0
0
03
040%
• 1004000%
1
03
NA
NA
NA
«4
1400
00
1»
M
400%
404000%
10
04
20643%
100433)%
90
«
BOO
0
0
»
0 00%
• 100 0000%
1
06
NA
NA
NA
40
31.400
102.1 a
92.734
oo
32B 10%
229.1962%
90
00
01 42%
164030%
24
«r
too JOO
i»JJ07
292406
97
140 74%
*0 7301%
93
07
19020%
90 2027%
91

3390
in
1MB
CO
00 77%
-10 2300%
30
00
7440%
a 1420%
37
m
0
to
to
AO
NA
NA
NA
00
100 00%
00000%
37
70
W
7
7
70
3D 40%
-TO 4110%
a
70
10040%
00000%
37
T1
790
120
110
>1
10 00%
-044000%
22
71
0147%
44330%
U
72
711
a
a
72
1J3*%
-004711%
14
72
10000%
00000%
37
T3
LOGO
0
0
73
000%
-too 0000%
1
73
NA
NA
NA
74
900
700
301
74
197 20%
07 2000%
90
74
44 47%
414207%
10
TS
9
«
3
ra
13040%
204000%
47
79
9000%
40 0000%
10
TO
000
7
1470
70
HA
NA
NA
70
NA
NA
MA
T7
17J30
100400
*7
77
SQ4J0*
0)4 2009%
00
77
004%
-004063%
9
TS
430
2xm
0
70
979 29%
479 2000%
02
70
0131%
40 0171%
0
79
1 190
i zoo
570
70
111 74%
117301%
40
70
4440%
404100%
10
00
H413
3 043
9437
10
22.30%
774042%
34
•0
19140%
914001%
90
01
13 000
20 030
7 127
ftl
194 00%
94 0700%
94
•1
39 W%
44 4164%
14
-------
R*port*d Throughput tor tha 61 Plant*
1989-1991
»LANT 10
voc
!M 1
ACUTE
TOTAL
vex
m
ACUTE
TOTAL
voc
	mr-
ACUTE
TOTAL
1
96 4&
96 423
M 423
U 496
W 8M
U 606
M UO
yo uo
34 630
2
0
2.192.096
2.102.006
0
1 900400
1 909 309
0
1 410 774
1 410 774
3
0
2997 039
2.907 039
0
910 400
910 400
0
960 574
960474
4
asoooso
10 BOO 790
10 000 700
7 199 471
7 290 471
7 200 471
7 193 797
7 329 797
7 329 747
3
13 192.346
9 97 493
13 192449
14 999477
0 907 140
14 909 an
12912401
7 790 241
1291291
a
0
474 140
474 140
0
977 3*9
377 349
0
991 290
SOt 90
7
491 734
1 137,714
1 137 734
339 207
1 044 463
1 944 403
940 117
2 290 696
2 29469
6
78400
3094*3
303 043
104,900
414 400
319 000
72. MO
299 196
39 217
0
0
4 3904)00
4400 000
0
3 00461
3 906 961
0
i 904 000
1 904 09
to
3 1949
ft 713,540
9 713449
2977401
4 717407
6 717 307
1 29 993
9 424 180
9 434 100
11
99470
90 970
99470
41 219
41 219
41 219
90 740
99 740
9 749
12
99499.409
09499 400
90 734 700
99 003409
934949
n on 947
19*1 779
100 041 770
19 212.479
13
1 24249
1 341000
1442400
1409 409
I4n.406
1 59 49
970.400
970 490
970 49
14
2JMJSO
2.402462
339497479
3 497 799
3.990430
320419440
049 430
1 09 790
319 99 027
13
99.291
9491
99 291
97 277
97.277
97 277
90427
90427
9427
10
124 121 AS
197411 2W
197411 S4
119 400 333
199 On 237
in On 237
11291 an
191 014 970
191.914 979
17
10 921JUT
10 921407
10.991407
io on 401
to 39401
1009401
10400.700
10 900 700
1049.79
ie
0
34.700
34,700
a
23 721
23.721
0
9440
949
19
1,307 £29
1407 229
1.367 229
on 4oi
on40i
9nm
79 on
79 on
720 09
20
901 2B9
973490
901409
an 711
29.207
m 711
40049
37349
40049
21
169.749
193 740
133 749
115401
119404
113404
90400
90 XO
949
22
77.091
77.091
77 001
ao in
n m
90 133
33.200
33 200
33 29
23
949
99400
99400
90407
90497
99407
43 097
43 997
43497
24
90 100
90 190
90 100
96 157
n 197
» 197
91464
9<4M
91404
29
0
9349
03420
0
47407
47407
0
22494
22404
20
0
son
3 903
0
9.721
$721
0
9292
9 202
27
0
9904 090
9464409
0
9.040.407
9440.407
0
9 97 1tt
9.207 19
2ft
1 007 444
1 109,429
1.109.439
997 4M
919407
919407
on 014
1 033 2U
1,09 214
a
20,421409
20.421490
20,421490
14,00049
U4n400
144n400
3 140 191
3 149 191
3,140 101
30
9

299432
0
31432
91432
0
122.79
122.79
11
0
112403
192402
0
S1J0
aot4n
0
979000
079400
U
0
100J14
100.114
0
9402,400
04K4B0
0
102.714
102.714
u
44404
44 004
44 994
49420
40420
49420
40 on
49 000
4049
M
0
4.111410
4.131419
0
9412.773
0412.773
0
93449
93449
30
0
39429
9429
0
97412
07412
0
91400
0149
M
0
0
0
114409
291400
91400
13240
236442
29442
17
0
1409490
1403400
0
I4n420
I4n420
0
9949
00049
9
0
090470
909470
0
904.400
904,409
0
99 79
99,79
9
1,90)712
1401732
1403.732
2.19402
2.19402
2.19402
2.19402
119402
2120402
40
0
3.706419
3.700419
0
3400436
3409436
0
1144949
1144949
41
0
199474
103474
0
1*419
139419
0
119,701
119,701
42
0
722403
722402
0
193>40O
193.400
0
19249
10249
41
0
41466
41409
0
4i4n
41439
0
464»
4049
44
0
34400
21 JDO
0
03497
03407
0
47400
4749
48
174 ItS
791400
701400
334.441
900 jn
9004n
2M402
1 19.212
1 19.211
44
940749
9497 000
9407 000
4470433
4470433
4470433
9034419
9 034 419
9.034 419
47
949,701
249.791
949 791
279406
279490
279 on
270414
279414
279414
40
0
0
0
0
0
0
0
0
0
41
an. to*
29,104
an. 104
271411
0*421
279431
90.774
240 774
29,774
SO
0
100400
103400
0
129400
19400
0
122279
122jn
91
3.119.404
3.119.4*4
3.119.404
3.104404
3.104404
3.104404
3470.79
3970 79
3470.79
32
9.90.093
13470*19
13470419
94n400
19409400
19.4949
9474400
I3.773.4n
19 77349
U
0
0
940249
0
0
S7.U1.700
0
0
99400
94
14149 49
141,107430
141 107430
1294»400
127.101440
127 101440
121.79640
12149400
1214949
sa
$2410400
42410400
92410400
92400400

92400 000
99410400
99410.000
941049
9i
900.409
900.490
900.400
729407
739457
79497
O004n
909496
0949
97
949
mm
309.121
27419
377400
577409
9.404
11949
11949
56
1469.19
1.300.129
1406.13
1.210401
1 210491
1 210401
10949
1 On 990
109460
99
0
4SJD0
49400
0
94400
04400
0
131.032
131432
40
71M0
7*440
73440
73,400
73,400
73.4n
044n
94400
0449
•1
0
29404
224,404
0
270404
Z79404
0
103.410
103410
92
199JT
1 177.29
1,204444
190.7*4
1.07X731
1 107412
194.174
914 99
93439
n
0
49400
49400
0
72400
72400
0
29410
2411
94
»to jm
21040
210433
239400
230400
2M4n
29.19
29.19
29.19
ae
0
200400
200400
0
300400
300400
0
40249
40249
99
1249149
134304a
13 039449
12.491049
I3 000 4n
13409,409
11.093419
14 09.407
14433 407
97
0
2.709 134
2.700,134
0
9 222.117
0 222.117
0
5 977 247
9 077 247
99
949300
049400
•49400
1 090 496
1.000 496
14n.4n
1,193464
1 19464
1 19464
9
0
97,221
97 221
0
90404
90404
0
9.110
9.110
TO
91 221
01.221
9<,221
91443
91.043
91443
31470
31470
31470
71
244.000
2M400
244 000
132.400
192.400
192.400
131490
131.290
13149
72
0
214 732
214.732
0
229. in
229.19
0
io2.ni
102431
73
0
94400
34400
0
71,030
7149
0
37 49
37.49
74
0
41 034420
41 OtJZ
0
41 440 7n
41 440 709
0
1949 240
194949
79
0
92400
92400
0
93400
93 000
0
11949
11949
79
0
0
900400
0
0
33449
0
0
19472
77
9
1400494
1 009404
0
310400
319409
0
97494
397404
79
0
200496
209 090
0
2D0 407
200 407
0
aim
20109
79
104 299
72432
104 299
104 an
7X032
104 an
42400
10472
4249
90
1 440.000
2340 400
2449 400
1 491 274
1,9n 104
i on 104
924,000
1 079 29
1 079 29
9t
199 464
1 002.420
1 002 420
207 931
1 090 490
1 069 400
202,700
1 009 442
t 00ft 442
-------
R*pon*o WntM lor th» Bi Plint*
1989-1991
°UV4T 10
VOC
1900
ACUTE
TOTAL
VOC/ACUTE
VOC
ACUTE
TOTAL
VOC/ACUTE
VOC
'Ml
ACUTE
TOTAL
VOC/ACUTE
1



3
3
3
3
1
3
2
0
5 750
5 750
0
5 391
S 391
0
1 HO
3 930
3
0
500
500
0
3939
3 909
0
] US
3935
4
790
1 250
1 290
10
20
20
5
»
0
5
aiP74
aooa
63474
174
'21066
iU 174
270 000
270 000
270 000
e
c
500
500
0
59
59
0
91
61
7
1JOO
BOO
4500
26
910
510
4 979
70 700
70 700
t
44
307
371
71
330
407
177
503
900
»
0
12B 100
129 100
0
240 661
244 691
0
5 930
5 630
10
300
9.027
9 027
5 830
14497
14 997
902
1.399
1 399
It
0
9
S



10
10
10
12
IttNt
9024*3
991403
30 194
36 164
30,194
673412
073412
677 466
13
900
SAO
500
496
456
458
197
167
367
14
J 444
4020
4 920
7290
9414
9 314
4792
7 913
57 019
15
2D
20
20
90
90
90
337
337
337
19
27419
32.432
32.432
1 43*30
30,279,798
30 279 799
346 667
16 831433
IB 833431
17
10400
1QJ00
10 soo
9 700
9 700
9700
9 608
6609
6608
19
0
1
1
0
1
1



18
231
391
291
9040
6.049
6 040
42
42
42
20
1 293
703
1 293
991
646
661
961
440
961
21
500
500
900
9962
9462
6962
1 413
1 *2
14ta
33
290
ao
390
120
120
120
60
00
00
23
1500
1400
I 900
1900
1 900
t900
600
500
600
24
29



1 467
1467
1 467
630
930
630
29
27
0
1 962
taa2
0
1467
1497
0
2.111
2.111
a
10 on
10400
10406
9474
9.043
9443
9 749
10 ou
10 032
a
647
047
647
•70
970
970
3,110
3.110
1110
10





100



j?



0
100

0
?60
160
u






a
208466
208489
31
SCO
900
500
ao
2D
2D
20
20
a
34
0
1000
1 OOO
0
1913
3413
0
16
19
39









31






29
191480
191480
37
0
soo
500
0
20
20
9
a
a
36
0
3400
1400
0
3.400
3 400
0
6 570
•470
39
V421
5 421
4,421
6400
9400
6400
6,000
6 000
6400
40
0
000
1400
0
97
97
0
22
28
41
0
11 441
11.441
0
30.760
30 700
0
71
71
42
a
1.000
t 000
0
12J10
1Z210
0
9429
942
O






0
16
19
44
0
500
900






46
i on
1.14ft
1 t4i
1414
1,748
1 748
1449
1479
1.479
44
9480
9.400
9400
21419
214*9
21419
4949
46426
4649
47
1 ooo
1 000
1 000
1904
1404
1 904
1 776
1 776
1 776
4®









40
9400
0 300
9400
13 100
13.100
11 100
11.191
11 191
11 101
SO
0
400
400



0
196
196
91
13484
13464
1)464
10427
104Z7
10427
4.283
4,299
4.2B3
S3
769480
781.477
789.477
020.404
1 910,762
1410.792
973448
2.302427
2402.627
S3









54
4 407
4900
4408
4 200
4,273
4.273
4 032
4 102
4 102
sa
M no
14400
14.900
80.190
BO 190
90 190
100 2D0
108400
108 200
se
1400
1400
1,000



1000
1 000
1 000
57
mo
900
900
zro
279
27©
f *4
1 244
1244
»
urn
2J7T
2477
19.477
19 477
19 477
24467
24487
24467
99









90









91
0
981
961
0
20471
2D 471
0
29419
29410
n









o






0
96
96
W
1 000
1000
1 000
60
60
60
129
129
129
06
0
42300
42400
0
34 000
24400
0
91479
61479
00
SIBH
440MB
449490
18 211
449 792
449,762
62.299
3077432
1077432
97
0
500
969
0
191407
191407
0
32.096
292496
U
1 000
000
1 000
249
2,929
2.629
3 466
3 406
3,466
m



0
to
10
0
10
10
70
34
34
>4
7
7
7
7
7
7
71
500
900
900
t20
120
130
110
110
110
7a
0
900
900
0
a
a
0
a
a
73
0
000
1 000


766



74
0
BOO
MO
0
796

0
361
161
79
0
9
9
0
9
6
0
3
1
70









77
0
IT 200
17 200



0
67
67
70
0
900
900
0
2479
2476
0
9
9
79
190
700
1 190
1 2B8
636
1 266
976
102
976
90
6 447
9447
9 047
a in
3643
1643
9 937
9917
9 937
91
500
2000
?000
19 903
20 030
20 030
6964
9 426
9 49
-------





Reported R»Im«m for the 61 Plants








1909-1991










1000


1001





total


TOTAL


TOTAL
tD

voc
ACUTE
/OC/ACUTE
VOC
ACUTE VOCrfcCUfg
VOC
ACUTE
VOC- ACUTE
\




)
3
3
3
3
3
2
/
0
3 750
3 750
0
3.301
3 301
0
3 960
3000
3

0
300
900
0
3539
3539
0
333
3 539
4
~
1 000
SOO
1 500
10
30
20
3
0
0
3

60 MO
87 290
U300
00*19
96*10
96*10
290 620
290 020
290 03
0

0
900
1 900
0
90
H
0
01
81
7

MO
000
1 000
256
510
310
221
460
466
8

00
314
300
71
330
407
177
303
000
9

0
8 100
B 1M
0
120 031
126 851
0
S030
3630
10

MO
10420
10 420
)«M
14 907
14 007
902
360
1 390
11

0
0
0



10
10
10
12

41 003
«1 003
41 603
23 730
23 736
23 646
93.166
53 160
36 704
13

MO
900
900
496
436

307
367
307
14

021
2.223
2223
4 067
0 060
94 240
2402
4 903
M 006
18

30
30
M
00
00
80
104
104
104
10

04 710
92432
02*32
727 460
10 106*02
10 196*02
106*02
100 210
106 216
17

87 000
87 000
07 000
0 700
0 700
0 700
8 006
8606
0606
10

0
212
212
0
1
1



16

1*06
1906
1.906
40
40
40
42
42
42
20

1 2B3
TO
i 203
001
640
001
061
*44
001
21

no
790
7M
8*62
0 092
0002
1 412
1412
1412
23

900
300
300
120
120
120
60
60
00
23

l MO
1900
1300
1 900
1 900
1*00
000
600
600
24

2.000
2006
2.606
1,467
1.467
1 467
630
630
630
29

0
770
270






20










37

0
062
1 602
0
1*07
1067
0
2.113
2.113
21

10*79
10*06
10*06
0*74
0*43
0.043
0 748
10*32
10Q32
29

«7
847
647
070
670
070
3.110
110
3.110
30

0
18
12






31

0
396
200
0
too
100
0
160
too
32

0
30
JO
0
0*62.290
9*tt230



u

000
900
900
20
20
20
20
a
a
M

0
000
1 000
0
3*13
3*13
0
19
19
16










30




Q
106*66
106*66
22
31
31
37

0
2.000
2X00
0
2D
ao
0
a
a
30

0
20
26
0
27
27
0
8*70
0670
30

5*71
9*71
9*71
8*00
0*00
0*00
0 000
0J000
0 000
40
0
1000
1*00
0
67
07
0
220
230
41

0
190
1 290
8
io*ai
10*21
0
71
71
42

0
1,000
1 000
0
110
110
0
3*a
9*a
43

0
10M
10 066
0
46*64
49*64
0
15
19
44

0
900
900






49

3oa
9411
9*11
1 014
1 746
1,746
1346
1.470
1 470
40

1.700
3790
3.790
12*18
12*16
12*16
a*oo
29,200
a*oo
47

1 20
290
1 ao
i*zr
1*27
1*27
1 007
1,607
1607
m

»J00
0*00
0*00
13.100
13,100
13,100
11 101
11 101
11,101
90

0
3
3



0
1
1
51

18*41
18*41
16*41
10*27
10*27
10*27
4 263
4.2B3
4263
52

90
37.237
37,237
404
4.113
4,113
24A
7*73
7*73
U










94

3061
3J0I
9*06
4.200
4 Z73
4,273
4032
4.102
4,102
50

8 100
0100
6100
9 160
9,100
9 tOO
0200
0200
0 200
M

1 000
000
1.000



1 000
1.000
1 000
57

790
1*00
1*00
270
270
270
067
067
067
M

11*11
10J11
16*11
10.477
10.477
10 477
24*67
24*67
M*67
90










60










11

0
790
790
0
i*a
1*29
0
190
1M
sa










S3

0
47*00
47 000






64

1.000
1*00
1*00
60
60
00
ia
129
ia
oe

0
MO
900






•a
~
31 106
31.406
31.406
27 063
102.128
102.126
62.070
62.734
62,724
07

0
106 *30
106*30
0
101*07
181*07
0
252.000
a 2.000
sa

j.ao
3 290
3.290
2. OS
2.029
2.629
1066
1*60
\ 960
Si




0
10
10
0
10
10
70

34
34
34
7
7
7
7
7
7
71

790
790
790
120
120
120
110
110
110
72

0
791
791
0
20
a
0
a
a
73
S
0
2.000
2.000






74
0
300
300
0
7B0
766
0
361
361
79

0
9
9
0
8
0
0
3
3
70










77

0
17 200
17 200
0
100 000
100 600
0
67
67
70

0
300
900
0
2*78
2*76
0
0
0
Tt

1 190
TOO
1 150
1 2BB
631
1 260
976
102
576
00

4 447
10 413
10 413
3 177
3043
36a
9 937
9937
9 937
01

11 MO
13 000
13 000
to 909
20 030
20 030
8 293
7 127
7 127
-------
APPENDIX E
GRAPHICAL RESULTS FOR WASTE-BASED ANALYSES
-------
Distribution of 1991 PPF Rankings
SIC 265
Distribution of 1991 PPF Rankings
SIC 286
Distribution of 1991 PPF Rankings
SIC 289
4
Q_
1 CW20'J80%0%0o/€0W0,380990%00%
Percentile of PPF Rankings
Distribution of 1991 PPF Rankings
TOTAL
101
2
E
3
Z
I
ITI
1 ii
I : I :|
• \ v • • •• • I
;l J .1 I
IllfeiJ
10%2O%3O%4O«a0«aO%7O%BO
-------
Distribution of 1991 PPF Rankings
SIC 281
c
ra
Q.
o
L_
a)
.a
E
3
Z
10?fiO?8O?iO?6O?fiO'VO?6O,9OO%0O%
Percentile of PPF Rankings
Distribution of 1991 PPF Rankings
SIC 282
10'KO'KO'MO'SOTO'KOWO'KO'Um
Percentile of PPF Rankings
Distribution of 1991 PPF Rankings
SIC 283

10%20%30%40%30%60%70%SO%SmS 00%
Percentile of PPF Rankings
Distribution of 1991 PPF Rankings
SIC 284
4
10%20%30%40%90%60%70%90%80'n 00%
Percentile of PPF Rankings
-------
Distribution of 1991 PPF Ranking*
SMALL PLANTS
to* 20% »* «0* 90* 80* TO*
Paranoia of PPF nuanq*
~litribution ol 1991 PPF Ranking*
MEDIUM PLANTS
Pninli at PPF Ranlangi
Distribution of 1991 PPF Ranking*
ALL PLANTS
io* jo* jo* 40* so* ao* ro* so* ao* 100*
Pareantta of PPF Ranknga
-------
70-
60-
50-
W
J* 40-
C
(C
£E
U.
CL 30-
20-
10-
1991 PPF Ranks and TRI Release Ranks vs
Refease-to-Throughput Ranks
Very Large Plants (> 10,000 employees)
B1
66
15
1ft
64217
13
61
29
11
-55-
19
+
54
78
>7
-fa-
st3
74
—I—
10
—r~
20
30
-r~
40
-r~
50
Release-to-Throughput Ranks
I
60
14
70
70
TBI Ranks
60
¦50
40
m
c
ta
X
o
tn
&
0)
30 ®
GC
20
10
-------
70-
60-
50-
-------
70-
60-
50-
W
C
(0
oc
U.
0.
Q_
40
30
20-
10
1991 PPF Ranks and TRI Release Ranks vs
Release-to-Throughput Ranks
Medium Plants (151-1,000 employees)
-t-




+
67



20
46

51
52

1










3*®
~
-1	
-I	r
	r
	1—
	1		
10	20	30	40	50
Release-to-Throughput Ranks
60
70
70
TRI Ranks
60
-50
40
30
W
C
(0
cc
a>
CO
(0
a>
a>
CC
20
10
-------
70
SO
SO'
J2 40
S
oc
U_
Q. 30
Q.
20
10-
1991 PPF Ranks and TRI Release Ranks vs
Release-to-Throughput Ranks
Small Plants (1-150 employees)
10	20	30	40	50
Release-to-Throughput Ranks
60
58 49
tt02?3
47 +
42
*


,»»2
45
79 +


71

+
70


m '
~
34
40
	1	

	1	1	
	1	 -T	
70
70
TRI Rank
-60
50
Ui
JSC
C
(0
40 OC
d)
v>
<0
_Q)
30 »
20
10
-------
1991 PPF Ranks and TRI Release Ranks vs
Release-to-Throughput Ranks
SIC 289
70
00
50
(0
t 40
(0
a:
LL
OL 30
Q.
20
10
40
23
zO
42



15
28®
7® +




*•


33




+


Vf
10	20	30	40	50
Release-to-Throughput Ranks
60
70
30
70
¦60
TRI Releases
•50
-40
CO
c
(0
oc
0)
(0
(0
0)
©
cc
20
10
-------
70
60
50-
V)
t 40
(0
cc
LL
Q_ 30
CL
20
10-
1991 PPF Ranks and TRI Release Ranks vs
Release-to-Throughput Ranks
SIC 286



¦¥
¦¥


f8"




+
81
80
67




46
-t-




49






16






64217




36
7
!>




+
+
>7
43




78 8
+
rx




»
I
	r
	1	

	1	
i
10	20	30	40	50
Release-to-Throughput Ranks
60
70
70
60
+
TRI Releases
50
(/)
c
(0
40 DC
(V
(/)
(0
30 ©
£E
20
10
-------
1991 PPF Ranks and TRI Release Ranks vs
Release-to-Throughput Ranks
SIC 285
70-
60-
JLL
57
SO-
OT
* 40-
(0
DC
U_
a. 30-
Q_
20-
10
10-
37
—r~
10
70
20	30	40	50
Release-to-Throughput Ranks
60
70
60
TRI Releases
50
40
30
(0
c
(0
CE
(D
V)
(0
0)
a)
cc
20
10
-------
70-
80-
50-
40-
t/i
c
(0
CC
LL
£L 30-
Q_
20-
10-
1991 PPF Ranks and TRI Release Ranks vs
Release-to-Throughput Ranks
SIC 284
TT
47
22
729
38
-55-
31
-T"
10
-r-
20
30
-1-
40
-T-
SO
-r-
60
•70
-40
30
70
Release-to-Throughput Ranks
60
TRI Releases
so
en
J*
c
(0
c:
CD
V)
«
0)
-------
36-
34-
32-
30-
<0
C 28-
(0
-------
1991 PPF Ranks and TRI Release Ranks vs
Release-to-Throughput Ranks
SIC 282
60-
50-
V»
68
40-
V>
JC
c
CO
CC 30-
LL
Q_
Q.
70
10
20-
54
10-
75
10
20
30
-r-
40
-r-
50
-r-
60
Release-to-Throughput Ranks
70
70
-60 TRI Releases
-50
-40
30
20
W
-X
c
(0
CC
©
0)
0>
oc
-10
-------
1991 PPF Ranks and TRI Release Ranks vs
Release-to-Throughput Ranks
SIC 281
30-
25-
20-
<0
¦X
C
<0
DC
IL
CL
Q.
15-
10-
13
27
34
40
74
T"
5
-r—
10
~T~
15
—r-
20
-r-
25
30
-r~
35
40
—r~
45
-50
50
Release-to-Throughput Ranks
-45
-40
TRI Releases
-35
-30
25 «
w
c
to
cc
o
(/)
CO
o
V
CC
» EE
-15
-10
-------
-------
70-
60-
50-
W
J* 40-
C
<0
QC
Li.
Q_ 30-
£L
20-
PPF Ranks and TRI Release Ranks vs
Release-to-Throughput Ranks
1991
81
ir
2I
r
•49

vf
42
46 •*
14
78
~ 45
72981
•4217
38
70
333
61
29
3i1 10
10-

	55	
1* «¦ +
" * >7
fl 0W
5»
4*
it*
T-
10
-r~
20
-T-
30
40	50
Release-to-Throughput Ranks
60
70
TRI Ranks
-80
¦50
tn
iC
c
<0
-40 CC
©
W
(0
Q)
30 0)
CC
20
10
70
-------
70-
PPF Ranks and TRI Release Ranks vs
ReJease-to-Throughput Ranks
1990
-70
60-
50-
40-
w
c
(0
cc
u.
Q_ 30-
20-
10-

23
ei
24^-

3#
1
-------
PPF Ranks and TRI Release Ranks vs
Release-to-Throughput Ranks
1989
70-
60-
-56-
4*

116 33
S0-
*0-
V)
JC
c
<0
tr
u.
0. 30-
+ 2«7S1flX2M5
725
52ZT«9§Bf8
/6S7
id3
-68T-


20-
20
10-

¥
¦iF*1
if*
l*
i*"
10
20
—r-
30
-I-
40
—r—
50
-r-
SO
Release-to-Throughput Ranks
70
TCI Ranks
.00
50

30 ®
cc
if
-20
10
70
-------
TRI Ratio Ranks vs PPF Ranks
1991/1990

38 42
» 42 ST

3. " , "
no67

S5 00 58
27 M

37 19 1 70 33 3 20 3#
71 «
19

45 "
" M
" « 24
14 23

„ re a
51 79
81 21

78 41
78 77
43 38
	1	1	1	1	 T I
0	10	20	30	40	50	60	70
PPF Ranks (1991)
-------

yi
TRI Ratio Ranks vs PPF Ranks
1990/1989
70
14
3
78
16
9 41 77
21 "
74
15 61
46
66
307 81
47.
75 „
8
2
10
ft
49
79
58 23
w
54 28 1 3
» «

68
28
51 24
31 7
70
mtk
57
T*
KU
23
45
40
. " 19
* 37 --
33


IV
SO
o>
oo
0> 50
O
O)
0>
c. 40
W
c
(0
(E 30
(8
E 2.
10
10
20
i
30
40
90
60
PPF Ranks (1990)
70
-------
70-
TRI Release Ranks vs PPF Ranks
1991
60-
16
67
JL2_
66
-M-
60-
62
46
39
2B

-65-
17
38
61
V)
c
(0
cc 40-
®
w
co
JD
® 30-
54
9
2
27
51
29
4*°
-66-
10
45
47
21
56
*h
23
-se-
er:
l-
20-
74
40
13
31
61
79
15
TT
64
-W-
10-
77
37
72
19
22
36
.33.
"3T
4
,5fi_
76
69
13"
11
70
75
-r~
30
—r-
40
—J—
50
—T~
60
-1—
10
—r~
20
PPF Ranks
70
-------
TRI Release Ranks vs PPF Ranks
1990

16
67 0 77

« , M
14 5
" M8,4,
4A * •

2a 49
17 51
39 28
, 38 21
3i 58 2

" « , 60
!7M »
Aft 47

23
7."
" » 20
13 8

42 71 22
.0 3'
a 64 15
in 19

00 7j
4 37 M 33
7. " . 70
18 1
	u	1	1	1	1	1	1	
H	 	1	I	1	I	1	1
0	10	20	30	40	50	60	70
PPF Ranks
-------
TRI Release Ranks vs PPF Ranks
1989

1« 67
s 17
12 S2
«® 51
ai »<%

"Jfr
80 81
28 43
10 . «
55 «3

*
14
j[T 37 73

«7 1fl
4 20 6 64 23

.. 340 7 56 42 57
ZD 72
61 72

13 71 "
74 3 „ 65 78 22 33 44

31 18
50
-**	1	1	1	1	1	1
0	10	20	30	40	50	60	70
PPF Ranks
-------
70-
TRI Ratio Ranks vs
Release-to-Throughput Ranks
1991/1990
60-
O
O) mn
O) 50-
Oi
O)
40-
w
c 30-
(0
E 20-
cc
I-
40
56
29
38
42
-45-
_5Z_
d2
31
52
40
-66-
58
_fiZ_
27
28
37 1
54
	LL_
692 70 33
7
20
^J9
71
47
13
45
-49-
68
17
66
24
74
75
22
_22_
14
51
10-

61
81
79
21
34
78
41
"T5~
77
43
10
20	30	40	50
Release-to-Throughput Ranks (1991)
I
60
70
-------
70-
TRI Ratio Ranks
Release-to Throughput Ranks
1990/1989
60-
14
Tto
41
-64-
76
21
O)
O) 50-
74
46
15
61
66
81
39
47,
67
O)
ZZ, 40 ~
<0
c
to
oc 30-
75
-96—
54
10
27
49
79
58
23
29
55
68
28
13
12
-ee-
51
24
(0
5 20-
31
57
70

2k.
6S*
45
40
42
10-
37
19
^2
64
33

10
20	30	40	50
Release-to-Throughput Ranks (1990)
60
70

-------