Industrial Pollution Prevention Trends in New Jersey
December 1996 - Michael Aucott - Debra Wachspress - Jeanne Herb

Trends

A. State and National Trends:

1. Releases and Transfers
Direct releases to environmental media, and indirect releases (transfers), both as reported on federal Form R, have shown marked declines since 1988 at both the national and the New Jersey level. At the national level, releases declined from a total of approximately 4.9 billion pounds in 1988 to 2.8 billion pounds in 1993. In New Jersey, similar declines are evident in both releases and transfers from 1987 to 1993. The greatest overall decline in New Jersey has been in transfers. Much of this decline in transfers is in quantities shipped off site for disposal or treatment. The rates of decline of both transfers and releases are similar; about 20% per year. These trends, which are not normalized for production, are apparent in the graph below.

2. Nonproduct Output
Nonproduct output (NPO) represents all quantities of listed chemicals which left production processes other than that which was shipped as, or in, product. Total NPO is essentially synonymous with total "production-related wastes" which are reported on the federal SARA 313 (Toxics Release Inventory) form R, Section 8. Only data for TRI facilities (i.e., those covered by SARA 313 reporting requirements) was used for NPO trends. Economic and employment data was based on all industrial facilities within the indicated SIC code. Chemicals covered in the review were those which were included on the 1993 SARA 313 list, minus those chemicals which were delisted in 1994 or for which the threshold determinations were substantially modified in 1994

From 1990 to 1993, NPO by all U.S. TRI facilities in SIC codes 26, 28, 30, 33, and 34 increased slightly, while both employment at all facilities in these SIC codes, and value added, corrected to constant 1983 dollars, remained essentially the same.

In contrast to the national trend, NPO by New Jersey TRI facilities in these 5 SIC codes declined during this period. Employment at all New Jersey industrial facilities in the 5 SIC codes also declined, although NPO appears to have declined at a greater rate. NPO was determined for all TRI chemicals, except those delisted (see above) from TRI Section 8 data, and was also calculated from the New Jersey Release and Pollution Prevention Report data (DEQ-114) for just those 167 chemicals EHS originally covered under the New Jersey Community Right-to-Know (RTK) reporting requirements.

NPO can also be measured with a larger group of chemicals than the EHS list from DEQ-114 data starting with 1991, when this report began collecting data for essentially the same list of chemicals as the TRI list. NPO trends shown with this method show a similar pattern of decline, as the graph below indicates.

Theoretically, the NPO for all chemicals, except those delisted, etc. in 1994,, as determined from the TRI Section 8 data and the NPO as reported on the DEQ-114 should be identical. They are not, however. NPO quantities as determined from Section 8 data are about 20 percent higher than NPO quantities as determined from DEQ-114 data. One reason may be that quantities of chemicals reported on the DEQ-114 as "destroyed on site" are counted as NPO; while "consumed" in a process are not. Perhaps quantities reported (correctly) on the DEQ-114 as consumed are reported in TRI Section 8 as "treated on site", inflating the TRI quantities. Other differences may exist as well. For instance, quantities recycled on site may be counted more than once, which would also inflate the Section 8 quantities. Ongoing quality assurance and quality control efforts by the Department are leading to improvements in facilities' quantification of their NPO and use quantities. In the future, the discrepancy between federally-reported NPO and state-reported NPO may decrease due to these efforts, and because of efforts by the federal EPA to clarify the definitions of these Section 8 quantities.

Although there is not close agreement in the total quantities reported on the TRI and DEQ-114 for all chemicals, and, as discussed above, the total quantities for the EHS (RTK) chemicals are smaller, a declining trend is evident in all cases.

Reasons for this apparent decline in NPO are clarified somewhat by the TRI Section 8 data. These data track seven different categories of NPO. These are releases, energy recovery on-site, energy recovery off-site, recycled on-site, recycled off-site, treated on-site, and treated off-site. Changes in each of these quantities from 1990 (projected backward from the 1991 data), through 1993 are shown below.

NPO cannot be tracked farther back than 1990 by any method, because the necessary data were not part of the reporting programs.

3. Use
Use is defined in the pollution prevention rules as the total quantity of a chemical brought on site and produced on site, including quantities that are recycled out of process and re-introduced to a process on site, adjusted for changes in inventory from one year to the next. Use is thus synonymous with "throughput." Use, or throughput, can also be determined from outputs. In an outputs-based calculation, use equals the sum of quantities consumed (i.e., chemically reacted within processes), shipped off site as or in products, and lost as NPO.

Federal data does not include use (i.e., throughput) quantities, so no comparisons between trends in use between New Jersey and the U.S. as a whole can be made. Use data has been collected in New Jersey since 1987. Prior to 1991, only information on the 167 original chemicals on the EHS list was reported. The 1987 New Jersey use data, like the 1987 federal TRI release data, appears to contain many inconsistencies, gaps, and reporting errors. These flaws are not surprising for the first year of a major new reporting system. Unfortunately, they preclude the use of 1987 data in this analysis.

However, a data set representing use of the EHS chemicals from 1988 through 1994, as reported on New Jersey's DEQ-114 and its precursor, the DEQ-100, is available. As discussed above, in 1994 several chemicals were delisted from the federal TRI reporting list, or their definitions were substantially changed. Reporting of these on the DEQ-114 was also terminated or changed significantly. For consistency of 1994 data with that of earlier years, these chemicals have been removed from this analysis. Shown below are total use quantities for all reporting facilities each year in SIC codes 26, 28, 30, 33, and 34. Chemicals included are those 1993 TRI chemicals that were also EHS chemicals, minus the delisted and other chemicals discussed above.

When use trends for all TRI chemicals are considered, only data from 1991 on is available. This trend is similar. See graph below.

A variety of uses of chemicals exist. Chemicals can be "used" because they are produced as a product, or because they are brought on site to be used as a formulation component in the manufacture of a product or an article. Sometimes options exist to replace a listed chemical in these types of feedstock, or raw material use, while still producing the same product. In other cases, no alternatives exist, or none are feasible. In these situations, reductions in use would not be expected. In some, cases, however, chemical use represents activities which result in most of the chemical being transformed to nonproduct output, such as when a solvent is brought on site and subsequently used for degreasing. In these cases, NPO reduction and use reduction would occur simultaneously.

Quantities consumed (i.e., chemically reacted as part of a production process) and quantities shipped off site as or in a product account for most of the total use quantity. That portion of chemical use which becomes NPO is relatively small. The graph below shows these proportions.

Data show that the portion of total use which is NPO has become smaller every year. This steadily shrinking portion is perhaps a better indicator of pollution prevention progress in New Jersey than the graphs displayed earlier which show declining NPO quantities without reference to total use in a given year. The percent measure, in effect, normalizes for production, to the degree that total use is related to production. The trend in NPO as percent of total use is shown in the graph below

It should be noted that total use quantities, and to a lesser extent total NPO quantities, are dominated by a relatively small number of facilities. For instance, according to a preliminary review of 1994 DEQ-114 data, the largest 20 facilities, in total quantity consumed, account for 95% of all the TRI chemicals consumed by covered facilities in New Jersey. 95% of the total quantity shipped was contributed by the 69 largest shippers, and 95% of the total NPO use was accounted for by 78 facilities. Statewide trends in use and NPO are thus strongly influenced by actions of the largest facilities.

4. Trends in Relation to 1987 Quantities
The Pollution Prevention Act states that a goal of the law is ".. a significant reduction over five years after the preparation of the pollution prevention plans required by this act, calculated on the basis of 1987 amounts, in the use of hazardous substances at industrial facilities, and a 50% reduction over five years after the preparation of the pollution prevention plans required by this act, calculated on the basis of 1987 amounts, in the generation of hazardous substances as nonproduct output;..." (N.J.S.A. 13:1D-36).

Determining 1987 quantities is problematic, as noted above, because there were no measures of NPO reported at that time. TRI data became available for the first time in 1987, but only for releases, including transfers. Use data, although it was reported in 1987 on the DEQ-100, is of questionable quality, and only covers the original 167 EHS chemicals. The DEQ-100, which became the DEQ-114, started collecting data which could be used to determine NPO in 1990, but only for those 167 EHS chemicals. It was not until 1991 that a full-fledged reporting program, including NPO and use data for all TRI chemicals, was in place in New Jersey.

Thus 1987 quantities, and their relationship to trends since then, can only be estimated. As discussed above, NPO in New Jersey has shown a consistent downward trend. The greatest rate of decline, approaching 30% per year, has been for EHS chemicals, not including those delisted, etc. in 1994. The smallest rate of decline of NPO has been about 11.5% per year, based on TRI Section 8 quantities. Projecting back to 1987 from the 1994 NPO quantity reveals that, even with this most conservative reduction rate estimate, NPO quantities in 1987 were probably higher by more than 50% than they were in 1994. During the period 1991 to 1994, a reduction of approximately 1/3 has taken place. See graph below.

The projection back to 1987, as discussed above, is not normalized for production. The average drop in NPO as percent of use, approaching 12% per year from 1990 through 1994, argues that at least a portion of the observed NPO reductions have been achieved through greater efficiency, (i.e., pollution prevention) although substantial contribution to these reductions from the shutting down of processes cannot be ruled out.

Use reductions are harder to estimate. Only the DEQ-100 and DEQ-114 provide use data. Prior to 1991, these data are limited to the original 167 EHS chemicals, as discussed earlier. The 1989 and 1988 data appear to contain some examples of double counting, which was corrected in later years through aggressive QA/QC by the NJDEP Right to Know Program. Use trends over the last few years do not show a clear trend, and it is likely that some of the use reductions seen from 1990 to 1992 were due to process shut-downs. However, as discussed above, the portion of use which became NPO appears to show a consistent drop, which argues that at least some of the use reduction has been due to true pollution prevention.

5. Earlier Reductions Reported in the Pollution Prevention Plan Summaries
Because it was clear that many companies had made substantial pollution prevention progress prior to the planning base year of 1993, an optional section was included in the Pollution Prevention Plan Summary Reports to allow facilities to record these gains. Only 56 out of 435 reporting facilities chose to use this optional section.

The optional section reveals some major achievements by many companies. The top 5 in use, NPO, and release reductions, respectively, appear in the tables below.

NOTE: The reductions above are listed as supplied by the facilities on their reports. NPO reductions will typically lead to use reductions of similar quantity, as discussed above. Use, NPO, and release quantities may not have been reported in all cases by a facility in the optional section of the Pollution Plan Summary Report.

B. Trends in Each of the Five Major SIC codes, 26, 28, 30, 33, and 34:
 Five major SIC codes are included in the first group of facilities covered by New Jersey's Pollution Prevention Act. Most of these facilities have been involved in pollution prevention planning since before the date their plans were due to be completed, July 1, 1994. Information on pollution prevention goals and other information submitted in the Pollution Prevention Plan Summaries are available for analysis for facilities in these five SIC codes, but not for other SIC codes. For these reasons, analyses of pollution prevention activities and trends of individual SIC codes are limited to these five groups.

1. The SIC Code Classification
The SIC code system classifies facilities according to their economic activity. For industrial facilities, the most common or important type of product of a facility dictates its classification. The 5 priority SIC codes considered herein are 26, 28, 30, 33, and 34. SIC 26 facilities manufacture paper and allied products. SIC 28 facilities manufacture chemicals and allied products. SIC 30 includes rubber and miscellaneous plastics products manufacturers. SIC 33 contains primary metals industries. SIC 34 is manufacturers of fabricated metal products, except machinery and transportation equipment.

2. NPO Trends, by SIC Code
Trends in NPO, since 1990 have been plotted for each of the five groups. In addition, because it contains many more facilities, SIC code 28 has been broken down into several of its three-digit subsets for some analyses. These subsets are 283, 284, 286, and a group consisting of 281,282, and 289 lumped together. These subsets were chosen because data on their economic performance is available from U.S. Bureau of the Census data. SIC 283 consists of manufacturers of medicinal chemicals, botanical products, and pharmaceuticals. SIC 284 contains producers of soaps, detergents, specialty cleaning preparations, surface active agents and similar compounds, and perfumes, cosmetics, and other toilet preparations. SIC 286 is manufacturers of gum and wood chemicals, cyclic organic crudes and intermediates, organic dyes and pigments, and other industrial organic chemicals. SIC 281 contains those who produce alkalies and chlorine, industrial gases, and inorganic chemicals. SIC 282 includes manufacturers of plastics materials, resins, synthetic rubber, and manmade fibers. SIC 289 includes manufacturers of adhesives and sealants, explosives, printing ink, carbon black, and chemicals and preparations not elsewhere classified. Also plotted are changes in employment over time, and changes in value added. With these latter two measurements, all facilities in the SIC code in New Jersey are considered. For NPO and other chemical-related information, however, only data from those facilities covered by the pollution prevention rules are considered.

Much of the decline of NPO by covered New Jersey facilities is accounted for by those in SIC 28. See graph below:

3. Apparent Inverse Correlation Between Value Added Growth and NPO Reduction
 SIC 28 appears to be among the healthiest economically in New Jersey, showing a long term increase in value added by manufacture for all industrial facilities in SIC 28. Although the long term trend in employment is down, NPO data for the years 1990 through 1993 shows an apparently steeper rate of decline.

In fact, in every New Jersey SIC code or subset looked at in this analysis, NPO shows a declining trend when value added shows an increasing trend, and NPO shows an increasing trend when value added shows a decreasing trend.

In other words, it appears that NPO has been declining in New Jersey primarily in those industrial sectors which appear to be economically healthy (i.e., the value added trend since 1982 is positive). Conversely, NPO has been flat or actually increasing in those sectors which appear to be having trouble economically (i.e., where value shows a declining trend since 1982).

This same inverse relationship between trends in value added and NPO appears to exist for total U.S. data, although NPO does not show the decline at the national level that is apparent in New Jersey.

The table below summarizes these data. It shows trends in value added, employment, and NPO generation, in percent per year. The table's economic data is for years 1982 through 1991, based on data from 1982, 1987, 1990, and 1991 from the U.S. Bureau of the Census. These economic data apply to all the facilities in the sector. NPO data, for NJ facilities, is based on EHS chemicals only (167 chemicals on original NJ list), not including those delisted or whose definitions were substantially modified in 1994. U.S. NPO data is based on TRI Section 8 data. NPO data is for covered facilities only.

The apparent inverse relationship shown in the table is displayed by plotting the value added vs. NPO reduction trends for both NJ and U.S. data in the graphs below.

4. Reduction Goals, vs. Use and NPO Quantities: Major Chemicals in each SIC group
The chemicals used and generated as NPO in the greatest quantity within each SIC group, and the aggregate reduction goal for each of these chemicals, by SIC group, is shown in the graphs below. Only the top ten chemicals, or those used in relatively large quantities, have been shown for each SIC group.

5. Pollution Prevention Methods vs. Goals per Unit of Product
Section D of the Pollution Prevention Plan Summaries contains use and reduction goals, per unit of product for each listed chemical, for each targeted process. The goals are expressed as percent reduction (per unit of product) over the 5-year planning period. Also included with each goal is a section describing the pollution prevention methods which the facility intends to use to accomplish the reductions. These methods are listed in Appendix 2 of the Pollution Prevention Plan Summaries. They can be grouped into five broad categories, corresponding to the pollution prevention methods listed in the Pollution Prevention Act. These are:

• Raw Material Substitution
• Product Reformulation
• Production Process Redesign or Modification
• In-Process Recycling
• Improved Operation and Maintenance of Production Process Equipment

In order to gauge the relative importance of these various methods in the pollution prevention plan goals, Section D data have been dis-aggregated as part of this analysis. That is, individual records were split up into several records, each of which contains a single unique process (for a specific facility), a single chemical, and a single pollution prevention method. Then, the average use and NPO reduction goal associated with each specific method was determined, and the number of times each method appeared (in all of the Sections D) was counted. Those methods associated with the highest goals were determined.

For use reduction goals, the specific (non-aggregated) method associated with the highest reduction goals was the substitution of different raw materials. For NPO reduction, this method was also associated with the highest reduction goals. However, the method was not selected as often as some other methods. To account for the net effect of a method, the average goal per method was multiplied by the number of times that method appeared. This latter procedure permits a relative determination of the overall effectiveness of each type of method. Improved operation and maintenance, though typically having low 5-year goals relative to raw material substitution on individual Section D reports, appears relatively frequently. In other words, raw material substitution, when it can be employed, appears to offer large reductions. But it cannot be used as often as improved operation and maintenance. This latter method typically produces modest results, but it can be used more often. Thus it earns a higher overall score. The results are shown in graphic form below.

6. The SIC Question
The Standard Industrial Classification (SIC) System was developed under the direction of the U.S. Office of Management and Budget and extended by the U.S. Department of Commerce, Bureau of the Census. It is a federal system that was designed to track economic and labor statistics. The system is commonly used in comparing trends in the U.S. economy. The SIC code system was most recently revised and updated by the federal government in 1987. In recent years, several environmental laws in New Jersey have relied on the SIC code system to determine industrial coverage and compliance with environmental regulations. The classification system operates so that the industrial coverage is progressively narrower with the successive addition of digits. For instance, SIC 28, Chemicals & Allied Products contains a number of 3-digit codes. One is SIC 281, which represents industrial inorganic chemicals. Within this 3-digit code are several 4-digit groups, such as 2816, which represents inorganic pigments. (The 4-digit groups are subdivided farther; the most detailed, 8-digit level, represents specific products.)

Throughout the analysis in this report, SIC codes are accepted at face value. Because the SIC system is detailed and widely used, it is generally assumed that it is a good indicator of the types of activities actually taking place at a facility. However, as part of its analysis, the NJDEP studied the SIC code designations to assess its reliability for environmental regulatory purposes. The NJDEP analysis indicates that there is often little correspondence between a facility's reported SIC code and the actual business activity, at least insofar as this activity involves listed chemicals.

Two types of information were used by NJDEP in its analysis. The first is the description of the business activity which is provided by each facility as part of its annual Release and Pollution Prevention (DEQ-114) report. Using only these descriptions, NJDEP staff guessed the SIC code for each facility. These "guessed" codes were then compared with the actual code reported by the facility to the Department of Labor. The match between the NJDEP-guessed codes and the actual reported codes was only accurate about 1/3 of the time at the 4-digit level. Even at the 2-digit level, the guess and the reported code only matched about 2/3 rds of the time. See graph below.

Another method was also used to investigate the applicability of the SIC code. In this, the SIC code of the product of each process, which is reported on each Section C of the Pollution Prevention Plan Summary, was compared with the reported SIC code of the facility. Again, the match was not especially close. At the 4-digit level, only 66% of the product SIC codes matched the facility's reported code.

These discrepancies between codes based on products manufactured and recent reported descriptions of facilities' business and the reported SIC codes point up a possible problem when reported SIC codes are used to categorize businesses. The problem may not be with the SIC code system itself, but rather with the way the codes are assigned and updated.

Many environmental regulations, including New Jersey's Industrial Site Recovery Act (ISRA), state and federal air pollution rules, and others, are applied based on facilities' SIC codes, on the presumption that these codes indicate the types of pollutants and environmental impacts that may be associated with the facility. The poor correspondence between SIC code and the actual industrial activities involving toxic chemicals revealed by NJDEP's analysis casts doubts on this presumption.

7. Other Issues
Though not directly related to trend analysis, other useful information can be gleaned from the Pollution Prevention Plan Summaries.

a. Compliance

Based on Pollution Prevention Plan Summaries received and written assertions from facilities that they soon planned to submit summaries, NJDEP earlier forecast a greater than 90% compliance rate. However, many of the promised Plan Summaries have failed to arrive. To date, 433 Summaries have been received from the 491 facilities which NJDEP believes are covered by the Pollution Prevention Act. (It is possible that some of these facilities have dropped below reporting threshold or are no longer covered for some reason.) To date, Summaries have not been received from 58 of these 491 facilities, for an 88% compliance rate.

b. Goals

Section B of the Pollution Prevention Plan Summaries show chemical-specific, facility-wide five-year reduction goals for use and NPO. In the data used for this analysis, there were 1859 Section B records. Of these, 938 had a goal for use or NPO greater than zero. 921, or approximately 50% of the records had a goal of zero for both use and NPO. These records represented reports of 435 facilities (several were facility-wide permit pilot facilities, which have a different reporting schedule). Of these 435, 316, or approximately 73%, have either a use or an NPO goal greater than zero for at least one listed chemical on site. Thus, about 27% of all reporting facilities have goals of zero for all listed chemicals on site.

c. Common Types of Processes; Associated Goals

In Section C of the Pollution Prevention Plan Summaries, facilities are required to identify their processes using specific process descriptors provided in the instructions to the reporting form. The dis-aggregation approach described earlier with Section D analyses was also used for Section C process description data. This permitted the count of the number of times each process descriptor appeared. The approach also made possible the association of an average goal, for all chemicals involved, for each specific process descriptor. Also available through Section C analyses is some other process-related information.

It is clear that most processes at covered facilities are batch processes; only about 11% of the nearly 2000 manufacturing processes reported are continuous. About 60% of all processes produce chemicals, and about 30% produce articles. Of the 60% of processes which produce chemicals, over 90% are batch processes. Of the 30% of processes which produce articles about 80% operate in a batch mode.

The process descriptor "CP3", which stands for "formulation/blending" appears in the database 447 times. The next most common descriptors, and the number of times they appear, are: UC1 (Mixing), 324; UC2 (Packaging), 308; CP4 (Chemical synthesis), 293; UA9 (Drying), 244; UB4 (Filtration), 235; and AA4 (Cleaning of Equipment), 208.

The process descriptor associated with the highest average use reduction goal (for all chemicals involved with that process) is HA5, "Steel drum", perhaps indicating measures associated with handling of drummed materials. TA1 (No Reaction) and CP1 (Purification) are also associated with high use reduction goals. TA1 is the descriptor associated with the highest average NPO reduction goal. HA5 and CP3 are also associated with high NPO reduction goals.