wPs alO.k POLICY RESEARCH WORKING PAPER 2002 Accounting for Toxicity It is important for environmental regulators to Risks in Pollution Control weight pollutants for their relative toxicity risk when Does It Matter? developing priorities for pollution control efforts at the industrial or regional level. Susmita Dasgupta But at high levels of Benoit Laplante aggregation, the choice of Craig Meisner indicator need not be the subject of immense debate. The World Bank Development Research Group Environment and Infrastructure November 1998 POLICY RESEARCH WORKING PAPER 2002 Summary findings The accounting and public release of information about from those indicators. They apply those factors to the industrial toxic pollution emissions is meeting increasing 3,426 industrial municipalities of Brazil and explore Rio criticism in that these listings typically do not account for de Janeiro and Sao Paulo in detail. the different toxicity risks associated with different After ranking states and municipalities for their pollutants. A firm emitting a large amount of a relatively pollution intensity, results indicate that at the state level, harmless substance is ranked as a heavier polluter than a risk-weighted rankings remain largely the same across the firm emitting a small quantity of a potent substance. 10 sets of toxicity risk factors used in this paper. By and Such "unweighted" rankings of firms, it is argued, may large the result also holds true at the municipal level. lead to a misallocation of resources and a wrong Although at the state level the unweighted ranking is prioritization of efforts in pollution control. This is a relatively similar to the risk-weighted ranking, at the particular problem in developing countries, where municipal level significant differences were found sources for pollution control are typically scarce. between the risk-weighted and unweighted rankings. To account for varying toxicity risk, a number of These findings suggest that it is important for organizations have developed thresholds or exposure environmental regulators to weight pollutants for their limits for various pollutants. But many toxicity risk relative toxicity risk when developing priorities for factors and methods are currently available, and different pollution control efforts at the industrial or regional risk indicators yield different results and hence priorities. level. But at high levels of aggregation, the choice of So Dasgupta, Laplante, and Meisner review seven risk indicator need not be the subject of immense debate. methods and construct 10 sets of toxicity risk factors This paper - a product of Environment and Infrastructure, Development Research Group - is part of a larger study on industrial pollution. Copies of this paper are available free from the World Bank, 1818 H Street NW, Washington, DC 20433. Please contact Yasmin D'Souza, room MC2-622, telephone 202-473-1449, fax 202-522-3230, Internet address ydsouza@worldbank.org. Susmita Dasgupta may be contacted at sdasguptaCaworldbank.org. November 1998. (36 pages) The Policy Research Working Paper Series disserninates the findings of work in progress to encourage the excange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polisbed. The papers carry the names of the authors and shoutld be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the view of the World Bank, its Executive Directors, or the countries they represent. Produced by the Policy Research Dissemination Center ACCOUNTING FOR TOXICITY RISKS IN POLLUTION CONTROL: DOES IT MATTER? By Susmita Dasgupta Benoit Laplante and Craig Meisner The World Bank Development Research Group Environment & Infrastructure For correspondence, communicate with: Susmita Dasgupta, Development Research Group, The World Bank, 1818 H Street, N.W., Washington, D.C. 20433; or at sdasgupta@worldbank.org. Acknowledgments We would like to thank the researchers at the Instituto Brasileiro de Geografia e Estatisticsa (IBGE) in Brazil who supplied us with the data necessary to conduct this analysis. Our most sincere thanks to Manjula Singh and Mala Hettige for their work with the Industrial Pollution Projection System. Finally we thank Mr. David Wheeler of the Development Research Group of the World Bank for his comments. Usual disclaimers apply. . . Executive Summary The accounting and public release of information pertaining to industrial toxic pollution emissions is meeting increasing criticism in that these listings typically do not account for the different toxicity risks associated with different pollutants: A firm emitting a large quantity of a relatively harmless substance would rank as a larger polluter than another firm emitting a small quantity of a very potent substance. It is argued that such "unweighted" rankings of firms may lead to a misallocation of resources and wrong prioritization of effort devoted to pollution control. This may be of particular importance for developing countries where resources devoted to pollution control are typically scarce. In an attempt to account for the relative differences in chemical toxicity, a number of organizations have developed thresholds or exposure limits for various pollutants to account for their various toxicity risk. Given the large number of toxicity risk factors and methodologies currently available, a crucial issue pertains to the possibility that different risk indicators may yield different results, thus leading to different sets of priorities. In this paper, we review seven risk methodologies currently available, and construct 10 different sets of toxicity risk factors from these indicators. We then apply these factors to the 3426 industrialized municipals of Brazil; we further explore in more detail, the case of Rio de Janeiro and Sao Paulo. Upon ranking states and municipals for their pollution intensity, results indicate that at the state level, risk weighted rankings remain largely the same across the 10 different sets of toxicity risk factors used in this paper. This result by and large also holds true at the municipal level. Moreover, at the state level, the unweighted ranking is relatively similar to the risk weighted ranking. However, at the municipal level, significant differences were found between the risk weighted and unweighted rankings. These findings suggest that it is of importance for environmental regulators to engage into weighting pollutants for their relative toxicity risk when prioritizing pollution control effort either at the industrial or regional level. The findings however suggest that at high levels of aggregation, the choice of a particular indicator should not be a matter of immense debate. . . I. Introduction An increasing number of environmental regulators in developed and developing countries have embarked on programs to account for the release of toxic pollution by industrial firms. Some of these programs also involve the public release of the information thus collected from industrial polluters. An example of such a program is the Toxics Release Inventory (TRI) published annually by the United States Environmental Protection Agency. These listings are meeting an increasing amount of criticism in that they typically do not account for the different toxicity risk associated with different pollutants: A firm emitting a large quantity of a relatively harmless substance would rank higher than another firm emitting a small quantity of a very potent substance. It is argued that such "unweighted" rankings of firms may lead to a misallocation of resources and wrong prioritization of effort devoted to pollution control. To the extent that the information is publicly available, it may also lead to a false sense of security, or alarm. Indeed, results from previous applications of risk-weighting factors have been significantly different than priority rankings based solely on volume-based techniques which do not account for the heterogeneous risk of pollutants.' Acknowledging these differences in chemical toxicity, a number of organizations have developed indices, based on alternative methodologies, to weigh pollutants to account for their relative toxicity risk. Given the large number of indicators currently Among others, see Hettige and Wheeler (1996), Horvath et al. (1995), Laplante and Smits (1998) and Swanson et al. (1997). 1 available, a recurring concern and crucial issue for environmental regulators pertains to the possibility that different methodologies may yield different results, and indicate different sets of priorities. This may be of particular importance for environmental regulators of developing countries where resources devoted to pollution control are typically scarce. In this paper, we review seven indicators currently available and from these construct ten different sets of toxicity risk factors. We then apply these toxicity risk factors to the 3426 industrialized municipals of Brazil,2 and discuss in more detail the estimates obtained for Sao Paulo and Rio de Janeiro (henceforth Rio). Our findings suggest that it is of importance for environmental regulators to engage into weighing pollutants for their relative risk factors when prioritizing pollution control effort. The findings however suggest that the choice of a particular risk indicator should not be a matter of immense debate as the relative impact of different risk indicators appears to be small for the prioritization of pollution control effort at high levels of industrial aggregation. In the next section, we describe the risk exposure indicators chosen for analysis in this paper, and for each index indicate how the relative risk factors were constructed. In Section III, we apply these factors to Brazil and focus more specifically on Rio and Sao 2 There were a total of 4974 municipals as of August 31, 1995; 3426 of these presented some degree of industrialization (IBGE, Directoria de Geociencias, Departamento de Estudos Territoriais, 1995). 2 Paulo. We first provide a brief description of the Brazilian industrial and environmental regulatory contexts. We then estimate toxic pollution emissions for each industry and region of Brazil with the help of the Industrial Pollution Projection System. Finally, we apply the various sets of toxicity risk factors to these pollution estimates. We briefly conclude in Section IV. II. Risk weighting methodologies and risk factors (i) Risk weighting methodologies In both developing and developed countries, risk assessment has recently become an integral component of the formulation of pollution regulation. In a context of limited resources, the identification and prioritization of intervention (defined in terms of industrial sectors or geographical areas) based on an assessment of risk is imperative for the reduction of toxic-related health problems. In the United States for example, various governmental and non-governmental organizations, such as the Environmental Protection Agency (EPA) and the American Conference of Governmental Industrial Hygienists (ACGIH) regularly publish comprehensive lists of hazardous chemicals which may serve as guidelines for explicitly incorporating toxic risk in the prioritization of pollution control effort. However, each of these organizations follows its own method of classifying chemical hazards or risk, where the choice of an indicator is for the most part dictated by regulatory requirements. For the purpose of the current comparative analysis, our interest lies not only in comparing risk-weighted and unweighted rankings of pollution intensive areas, but also to compare the risk weighted rankings under alternative measures of risk (such as short term vs long term exposure). 3 In Table 1, we list seven widely recognized toxicity indices along with the organization using these indices. The classification enforceable / non-enforceable indicates whether or not the index is enforceable by law. Table 1 Toxicity indices l Source I Index name | Classification American Conference of Governmental Threshold Limit Values (TLV) Medium-term Industrial Hygienists (ACGIH) Non-enforceable U.S. Department of Labor, Occupational Safety Permissible Exposure Limits Medium-term and Health Administration (OSHA) PEL) Enforceable National Institute for Occupational Safety and Recommended Exposure Limits Medium-term Health (NIOSH) (REL) Enforceable Deutsche Forschungsgemeinschaft Maximum Concentration Values Medium-term (DFG, Federal Republic of Germany) in the Workplace (MAK) Enforceable Santa Clara Center for Occupational Safety and Health-Based Exposure Limits Long-term Health (SCCOSH) (HBEL) Non-enforceable U.S. Department of Energy, Subcommittee on Temporary Emergency Exposure Short-term Consequence Assessment and Protective Limits (TEEL) Enforceable* Action (SCAPA) U.S. Enviromnental Protection Agency, Sector Toxics Release Inventory Long-term Facility Indexing Project (SFIP) Indicators toxicity weights (TRI Non-enforceable toxicity weights) TEEL are enforceable only on U.S. Department of Energy sites. In Table 2, we provide a brief definition of each index. Note that TLV, PEL, REL, and MAK all have the same definition. However, while PEL, REL and MAK have all been largely adopted from TLV, the number of chemicals covered by each index is different. Moreover, PEL, REL and MAK are enforceable by law whereas TLV act only as recommendations. TLV, PEL, REL and MAK have been broadly classified as medium term since exposure is defined as a time-weighted average per working day and / or working week with the number of years of exposure not being explicitly referred to (i.e. versus chronic or lifetime exposure). 4 Two other indicators, HBEL and TEEL, are of interest as they define polar extremes of relative toxicity: HBEL are maximum lifetime daily exposure concentrations, while TEEL can be interpreted as maximum short term exposure concentrations. The last indicator, the U.S. EPA TRI toxicity weighting system, is a new initiative by the EPA to address public criticism that the TRI inventory has been purely based on volumes of emissions.3 This very recent development of risk incorporation into the TRI accounting of chemical loads is a valuable opportunity for the purpose of the current exercise. (ii) Riskfactors In order to incorporate risk into the analysis, we must aggregate the releases of various industrial pollutants within a specific region. In order to do so, we must initially convert each chemical into equivalent weights for each of the ten sets of toxicity indicators presented in Table 2. These weights were calculated by normalizing each chemical with respect to a reference chemical, chosen to be sulfuric acid: (1) wii = (Reference value for sulfuric acid provided by index j) (Reference value of chemical i provided by index j) where wij is the sulfuric acid equivalent risk factor associated with chemical i upon using toxicity index j (j = I to 10).4 3 The TRI is understood as being one of the largest sources of environmental information available to the public (Hamilton, 1995; Konar and Cohen, 1997). 4 For example, suppose there are three pollutants A, B, and C, with pollutant A representing sulfuric acid. Each of the 10 indices provides a threshold or exposure limit for pollutants A, B and C. Let the limits provided by index j be RAj, RBj, and Rcj respectively. Then, wAj would be RAj/RAj or 1, wj would be RAj/RBj and wcj would be RAj/RCj. 5 Let Qix be the estimated (unweighted) pollution load of chemical i in region x. Then, (2) Qi, - wij gives the estimated risk weighted releases of pollutant i (in sulfuric acid equivalent) in region x upon using index j. Aggregating over all pollutants yields: n (3) Pj = LQi. wi i=l where Px1 is the estimated total risk weighted releases of pollution in region x upon using index j, and n is the number of chemicals covered by index j. III. Estimates of risk weighted pollution load in Brazil (i) The Brazilian context The industrial sector in Brazil accounts presently for nearly 35% of GDP, and represents a significant share of the total working population, across the five major regions of Brazil (Table 3). However, the relative importance of industrial activity across Brazilian states varies considerably and in some regions represents a very significant share of total employment. As can be seen in Table 4, manufacturing employment as a percentage of the total working population is quite high in suchi states as Sao Paulo and Rio de Janeiro. These highly industrialized states (with the exception of the Distrito Federal) also enjoy a higher GDP per head while less industrialized states lag far behind (Table 5). Within the manufacturing sector, some the largest employers are in the wearing apparel, motor vehicle, metal and textile industrial sectors (Table 6). 6 Table 2 Description of index Index | Description 1997 ACGIH Threshold Limit Values Time-weighted average (TWA) exposure concentration that cannot be exceeded for a conventional 8-hour workday and a (TLV) 40-hour workweek. 1993-97 OSHA Permissible Exposure TWA exposure concentration that cannot be exceeded for a conventional 8-hour workday and a 40-hour workweek. Limits (PEL) 1994 NIOSH Risk Exposure Limits TWA exposure concentration that cannot be exceeded for a conventional 8-hour workday and a 40-hour workweek. (REL) 1996 DFG (MAK) TWA exposure concentration that cannot be exceeded for a conventional 8-hour workday and a 40-hour workweek. 1995 SCCOSH Health-Based Exposure Maximum lifetime daily exposure concentration for a conventional 8-hour workday, 240-days/year, for 40 years. Limits (HBEL - Non-Cancer) 1998 U.S. DOE SCAPA Temporary Derived mostly from TLV-STEL and PEL-STEL. A 15-minute time-weighted average exposure concentration that Emergency Exposure Limits (TEEL-O) should not be exceeded at any time during the workday. 1998 U.S. DOE SCAPA Temporary Derived mostly from TLV-C and PEL-C. A time-weighted average concentration that is not to be exceeded during any Emergency Exposure Limits (TEEL-1) part of the working exposure. 1998 U.S. DOE SCAPA Temporary Measure of toxicity (TCLo and TDLo) estimated from human or human-equivalent toxicity data from Sax's Dangerous Emergency Exposure Limits (TEEL-2) Properties of Industrial Materials (1996/97) with a maximum of 500 mg/m3 for particulate materials, if no hierarchy- based values could be estimated using methodology outlined in NIOSH (1994). 1998 U.S. DOE SCAPA Temporary Measure of lethality (LCLo, LDLo, and LD50) estimated from human or human-equivalent toxicity data from Sax's Emergency Exposure Limits (TEEL-3) Dangerous Properties of Industrial Materials (1996/97) with a maximum of 500 mg/m3 for particulate materials, if no hierarchy-based values could be estimated using methodology outlined in NIOSH (1994). 1998 U.S. EPA SFIP TRI Relative Risk- A chronic human health proportional weighting system utilizing Reference Dose values (RfD) for cancer and non-cancer Based Chronic Human Health Indicator effects, along with weight-of-evidence measures for carcinogens. RfDs are derived from a combination of NOAEL, Toxicity Weights (TRI Toxicity LOAEL, uncertainty factors in intraspecies variability, interspecies extrapolation and extrapolation from subchronic to Weights) chronic data. Note: STEL - Short-Term Exposure Limit; C - Ceiling limit; TCLo - Toxic Concentration Low; TDLo - Toxic Dose Low; LCLo - Lethal Concentration Low; LDLo - Lethal Dose Low; LD50 - Lethal Dose Fifty; NOAEL - No Observable Adverse Effect Level; LOAEL - Lowest Observable Adverse Effect Level See Appendix A for more details on TCLo, TDLo, LCLo, LDLo and LD50. 7 Table 3: Industrial employment per region, 1993 Region Total Working Industrial % Employed in Population Employment (1) Industry Sudeste t2 28 700 970 7305 969 25.46 Sul (1) 11 560 445 2 535 344 21.93 Norte (4) 2 555 088 490 426 19.19 Centro-oeste (5) 4 601 976 704 640 15.31 Nordeste (6) 18 968 726 2 724 173 14.36 (1) - Industry includes the construction sector. (2) - Sudeste includes: Minas Gerais, Espirito Santo, Rio de Janeiro, Sao Paulo. (3) - Sul includes: Parana, Santa Catarina, Rio Grande do Sul. (4) - Norte includes: Rond6nia, Acre, Amazonas, Roraima, Para, Amapa, Tocantins. (5) - Centro-oeste includes: Mato Grosso do Sul, Mato Grosso, Goias, Distrito Federal. (6) - Nordeste includes: Maranhao, Piaui, CearA, Rio Grande do Norte, Parafba, Pemambuco, Alagoas, Sergipe, Bahia. Table 4: Level of manufacturing employment, 1994 State [ Manufacturing Population of % of Working Population | Employment (1) | Working Age (2) in Manufacturing Santa Catarina 313 259 2 757 602 11.36 Sao Paulo 2 082 706 19 789 464 10.52 Rio Grande do Sul 494 381 5 528 990 8.94 Parana 273 241 4 784 951 5.71 Minas Gerais 421 575 9 000 056 4.68 Rio de Janeiro 364 493 7 783 014 4.68 Pemambuco 136 808 3 756 185 3.64 Alagoas 46 871 1 327 357 3.53 Espirito Santo 53 506 1 543 563 3.47 Amazonas 34 404. 1 158 283 2.97 Ceara 93 484 3 330 191 2.81 Mato Grosso do Sul 24 346 1 057 585 2.30 Goias 53 221 2 423 371 2.20 RioGrandedoNorte 28 175 1 320 106 2.13 Sergipe 16 414 814 387 2.02 Para 47 986 2 749 896 1.75 Paraiba 25 075 1 607 131 1.56 Mato Grosso 16 757 1 322 892 1.27 Rondonia 9 351 760 053 1.23 Bahia 73 754 6 271 747 1.18 Distrito Federal 11 673 1 028 217 1.14 Piaui 12 685 1 302 481 0.97 Amapa 1 290 161 986 0.80 Acre 1 670 220 835 0.76 Maranhao 16 649 2 419 597 0.69 Tocantins 2 174 505 368 0.43 Roraima 386 156 300 0.25 Total 4 656 334 84 881608 5.49 (1) - Source: Instituto Brasileiro de Geografia e Estatisticsa (IBGE), 1994. (2) - Source: IBGE, Diretoria de Pesquisas, Departamento de Populacao e Indicadores Sociais, Censos Demograficos de 1980 e 1991. Calculated as the total population between ages 17 to 60. 8 Table 5 Gross domestic product per capita, 1994 Highest ($) Lowest ($) Distrito Federal 7 080 Paraiba 1 108 Sao Paulo 4 666 Maranhao 1 055 Rio de Janeiro 4 386 Sergipe 958 Parand 3 674 Tocantins 901 Rio Grande do Sul 3 670 Piaui 835 Source: Instituto de Pesquisa Econ6mica Aplicada (IPEA). Table 6 Top 10 manufacturing employers Brazil | Rio de Janeiro Sao Paulo Industry Percentage Industry Percentage Industry Percentage share* share share Wearing apparel 7.75 Wearing apparel 12.07 Motor vehicles 10.57 Motor vehicles 6.51 Iron & steel 7.20 Wearing apparel 6.69 Footwear 5.01 Printing & publishing 7.04 Fabricated metal 6.02 _____________________ products l Fabricated metal 4.44 Plastic products, N.E.C. 4.87 Plastic products, N.E.C. 4.84 products l Sugar factories & 4.32 Drugs & medicines 4.26 Spinning, weaving & 4.01 refineries finishing textiles Spinning, weaving & 4.14 Bakery products 4.07 Electrical apparatus & 3.93 finishing textiles supplies, N.E.C. Iron & steel 3.86 Fabricated metal 4.03 Printing & publishing 3.59 products Plastic products, N.E.C. 3.81 Shipbuilding & 3.91 Sugar factories & 3.12 repairing _ refineries Sawmills, planing & 3.40 Spinning, weaving & 3.55 Footwear 3.02 other wood mills finishing textiles Printing & publishing 3.39 Nonmetallic mineral 2.59 Iron & steel 2.76 products, N.E.C. * - Percentage share of all manufacturing employment. An unfortunate consequence of industrialization is the increasing emission of toxics to the environment, and subsequently the requirement for prioritization, regulation and control. Brazil's experience with environmental regulation has been both a success 9 story, due to the significant experience of some state agencies, and a failure as a consequence of recent fiscal constraints and a lack of political support. Environmental legislation in Brazil dates back to 1973 and was modeled mostly after the American experience, relying heavily on standards and licenses. The objectives of environmental policy are defined in terms of minimum ambient environmental standards which the Federal Government has established for air and water. Brazil's pollution control policy is centralized around a licensing system that requires a valid environmental license for every potentially polluting activity. States have implemented their own licensing systems based on the national framework. Since 1974, most States have created Environmental Protection Agencies (OEMAS - orgdos estaduais de meio ambiente) which are in charge of licensing, monitoring and enforcing environmental regulations. States have implemented different systems of fines for environmental violations; fines are normally a function of the estimated level of damages resulting from the violation. However, as in most cases, the effective implementation of those fines has proved challenging. Municipalities are playing an increasingly important role in pollution management and are currently responsible for zoning, water, sanitation, solid waste and drainage services. In addition, larger municipalities are assuming licensing functions for activities which have the potential of being significant sources of local pollution. However, the 10 administrative capacity to implement, monitor and enforce the terms of the licenses is typically limited. Pollution control in the states of Rio and Sao Paulo is the responsibility of FEEMA (Fundacdo Estadual de Engenharia do Meio Ambiente) and CETESB (Conpanhia de Tecnologica de Saneamento Ambiental) respectively. These two agencies have often been acknowledged as leading environmental agencies in the developing world. They are the largest state environmental agencies in Brazil with staff of 2 200 and 900, and supported budgets of approximately US$ 90 and US$ 25 million for the year 1997 respectively (World Bank, 1998). Most other state environmental agencies have much smaller staff and budgets, and have suffered a serious decline in recent years due to fiscal constraints. The larger agencies have also experienced a serious decline in their effectiveness. Environmental management in the state of Rio has deteriorated as a result of the fiscal crisis and lack of political support under previous administrations. FEEMA is paralyzed by a lack of accountability, an excessive number of poorly paid and unmotivated staff, and serious budget rigidities. It is argued that the numerous bureaucratic environmental requirements, and a serious lack of reliable environmental information and planning prevent the State environmental agencies from adequately performing its core functions. Given the poor economic and political environment within which state environmental agencies must operate, there is a strong need for the application of tested 11 methodologies in order to prioritize pollution control effort. This is especially the case for state environmental agencies which are currently preparing comprehensive restructuring and modernization plans aimed at improving their effectiveness. These plans involve an important decentralization of roles and responsibilities, while still retaining a supervisory role. To these ends, the application of tested methodologies to estimate pollution load on a regional basis can provide regulators with crucial information pertaining to areas of high pollution intensity. This is especially the case on matters of toxic emissions for which information, on a plant and / or regional level, has never been thoroughly collected in Brazil. (ii) Estimating pollution load in Brazil As indicated earlier, in order to estimate the total releases of (sulfuric equivalent) chemicals in Brazil, we must first obtain (unweighted) estimates of emissions of chemicals (Qix in equation (2)). Despite the existing legislative and institutional apparatus, it is generally recognized that Brazilian environmental authorities (like most environmental authorities of developing countries) lack the necessary information on plant-level emissions to set priorities, strategies, and action plans. This is especially the case with toxic chemicals whose releases are typically not monitored. As a response to this insufficiency of information, Hettige et al. (1995) have developed the Industrial Pollution Projection System (IPPS) to exploit the fact that industrial pollution is heavily affected by the scale of industrial activity and its sectoral 12 composition. IPPS operates through sector estimates of pollution intensity (pollution per unit of activity). The system combines data from industrial activity (such as production and employment) with data on pollution emissions to calculate pollution intensities, i.e. the level of pollution emissions per unit of industrial activity (Pollution intensity = Pollution emissions / Measure of industrial activity). The model can be refined to include only chemical intensities by using information solely on chemical releases along with industrial activity (Chemical intensity = Chemical emissions / Measure of industrial activity). As illustrated in Figure I, chemical intensities have initially been calculated with data available in the United States from the U.S. Manufacturing Census and the U.S. Environmental Protection Agency (EPA). The Census maintains a database known as the Longitudinal Research Database (LRD) which contains information from the Census of Manufactures (CM) and the Annual Survey of Manufactures (ASM). While the CM contains information on all manufacturing establishments in the United States, the ASM seeks further and more detailed information on a subset of those companies. Once an establishment has been selected to be part of the ASM, information is collected from the chosen company once a year, for a period of 5 years. The LRD thus contains detailed information on approximately 200,000 plants. The EPA maintains a number of databases 13 on pollution emissions including the Toxics Release Inventory (TRI), which was used to calculate chemical intensities for each industry.5 An immediate difficulty with the calculation of intensities is the measure of industrial activity. While physical volume of output would be the ideal unit of measurement, industries and even establishments within a given industry use different units to report the volume of their production, thus not allowing for comparison across industries. The value of output and plant-level employment (also contained in the LRD) however do offer such common units of measurement. Combining the LRD's database with the EPA TRI database, it was possible to calculate 329 chemical intensities for each industrial sector (at the 4 digit International Standard Industrial Classification (ISIC) level), using both the value of output and plant-level employment.6 5 At the time, the TRI contained information on annual emissions for more than 300 toxic chemicals to the environment. Manufacturing establishments that (1) employed 10 full-time employees or more and (2) produced, imported or processed 25,000 pounds or more of any listed chemical had to report the nature and quantity of the chemical produced, imported, or processed. rI 1987, approximately 20,000 enterprises reported their releases of such chemicals. A listing of the 329 selected chemicals is provided in Appendix B. 6 It should be understood that since different industries will emit a different number and composition of chemicals, each industry will have a different number of estimated chemical intensities. These intensities may be obtained from the web site http://www.worldbank.org/nipr 14 Figure 1 Industrial Pollution Projection System Pollution Intensity US Manufacturing Census US EPA EMISSIONS (200,000 plants) (20,000 plants) ~~mi MJ1Data Source: Policy Research Department, The World Bank For Brazil, we were able to obtain 1994 employment figures for every industrial sector, for all 3426 industrialized municipals. Multiplying these figures by the associated employment-based chemical intensity obtained from IPPS resulted in the total chemical load per industrial sector per municipal (or state).' Summing across all industries in a state or municipal yielded the estimated releases of chemicals in the state or municipality. In what follows, these estimates, unweighted for toxicity risk, are referred to as the "Qvolume" ranking (or Qi, in equation 2) since it is solely based on the total volume of releases, with no account of relative risk. These estimates can finally be weighted for 7 While absolute estimates of pollution emissions differ upon using employment-based pollution intensities and value-based pollution intensities, Hettige et al. (1 995) and Laplante and Smits (1 998) have shown that the ranking of industrial sectors (from largest polluters to smallest) remains the same. Moreover, Wheeler et al, (1 998) in a comparative analysis of 12 countries, including Brazil, have found the ratio (Pollution emission / Employment) to be relatively constant across countries for the same industrial sectors. 15 their relative toxicity as shown previously (multiplied by wij ). Results are presented for each of the 10 indices shown in Table 2. (iii) Pollution load in Brazil We first begin at the aggregate level, looking at pollution emissions at the state level for the entire country. In Table 7, observe that Sao Paulo ranks -first, perhaps unsurprisingly, in terms of pollution emissions, both weighted and unweighted. More importantly, observe that the risk weighted rankings of the states are relatively similar across indices. HBEL and TRI offer slightly different rankings which may be explained by the fact that these are the long-term exposure indices in our analysis. The volume ranking, while providing a different ranking than the risk weighted rankings also performs relatively well, especially for those states ranked as the largest and those ranked as the smallest producers of toxic emissions. Hence, despite a few notable exceptions, state level rankings remain, for the most part, consistent across the risk indicators and with the volume indicator. 16 Table 7 Ranking of Brazilian states, indexed on TLV State | TLV PEL REL [MAK [ HBEL |TEEL-O] TEEL-1 TEEL-2 TEEL-3 TRI Volume Sao Paulo 1 1 1 1 1 1 1 1 1 1 1 Parana 2 2 5 2 2 4 2 2 2 5 4 Minas Gerais 3 4 2 4 6 2 4 4 3 4 2 Santa Catarina 4 3 6 3 3 6 3 3 4 6 3 Rio Grande do Sul 5 6 4 5 5 5 5 5 6 3 6 Para 6 5 7 6 4 7 7 6 7 8 10 Rio de Janeiro 7 7 3 7 7 3 6 7 5 2 5 Ceara 8 10 9 8 14 8 8 8 8 10 7 Bahia 9 8 8 9 10 9 9 9 9 7 8 Mato Grosso do Sul 10 9 13 10 8 12 11 10 11 11 17 Pernambuco 11 12 10 11 15 10 10 11 10 9 9 Rond6nia 12 13 16 13 9 15 14 13 15 18 22 Espirito Santo 13 11 11 12 12 11 13 12 12 13 12 Amazonas 14 14 14 15 11 14 16 14 16 15 18 Maranhao 15 15 17 17 13 17 17 15 17 14 21 Rio Grande do Norte 16 17 15 14 18 16 12 16 13 17 11 Goias 17 16 12 16 17 13 15 17 14 12 13 Mato Grosso 18 18 19 19 16 19 19 18 19 20 20 Alagoas 19 19 18 18 21 18 18 19 18 19 16 Paraiba 20 20 20 20 24 20 20 20 20 22 14 Piaui 21 22 21 21 25 21 21 21 21 16 19 Distrito Federal 22 21 22 23 20 22 23 23 23 21 23 Sergipe 23 23 23 22 23 23 22 22 22 24 15 Acre 24 24 24 24 19 24 24 24 24 25 24 Tocantins 25 25 25 25 22 25 25 25 25 26 25 Amapa 26 26 26 26 26 26 26 26 26 23 26 Roraima 27 27 27 27 27 27 27 27 27 27 27 * - TLVs have been in use since 1946 and are the basis of numerous other indices (such as PEL, REL and MAK). In addition, TLVs are continuously revised to account for recent scientific research. 17 (iv) Pollution load in Sdo Paulo & Rio de Janeiro If we examine our estimates at a dis-aggregate (municipal) level for Sao Paulo and Rio, observe in Tables 8 and 9 that the rankings of municipals remain more or less identical across the risk-weighed indices, with the exceptions of HBEL and TRI for which the ranking of municipals vary the most with the other risk-weighted rankings. This is especially the case for Sao Paulo. Most striking however is the markedly different ranking obtained when risk is unweighted (volume ranking). Table 8 Top 20 municipalities for Rio de Janeiro, indexed on TLV (1) Municipal TLV PEL REL MAK HBEL TEEL-0 TEEL-1 TEEL-2 TEEL-3 TRI Volume code Ij= 330455 1 1 1 1 1 1 1 1 1 1 1 330630 2 2 2 2 2 2 2 2 3 2 11 330170 3 3 3 3 3 5 3 3 2 3 2 330350 4 4 4 4 6 7 4 4 4 4 8 330340 5 6 5 7 11 9 8 6 7 8 4 330040 6 5 6 6 4 6 5 5 8 9 14 330330 7 10 8 5 5 3 6 8 9 6 6 330490 8 7 7 8 7 8 7 7 5 7 9 330010 9 31 14 9 8 4 11 20 24 15 17 330390 10 9 10 10 9 11 9 9 10 10 3 330025 11 12 II 11 16 14 10 10 6 5 16 330510 12 13 9 13 19 15 16 15 14 12 15 330030 13 11 13 12 10 10 14 11 15 13 22 330200 14 18 17 15 17 12 17 12 20 16 26 330580 15 15 15 17 15 20 19 16 19 20 10 330100 16 14 18 14 14 16 12 14 11 11 19 330240 17 19 20 19 20 17 21 18 22 19 30 330250 18 17 19 16 18 19 13 17 16 23 13 330080 19 21 31 22 12 29 24 23 29 27 39 330190 20 16 16 24 23 21 22 19 13 21 25 (1) - Total number of industrialized municipals is 75. 18 Table 9 Top 20 municipalities for Sao Paulo, indexed on TLV Municipal TLV PEL REL MAK HBEL TEEL-0 TEEL-1 TEEL-2 TEEL-3 | TRI Volume code 1_ l 355030 1 1 1 1 1 1 1 1 1 l 1 354870 2 2 2 2 2 2 2 2 2 3 3 354840 3 4 3 4 40 9 7 4 8 11 23 351880 4 3 4 3 4 4 3 3 3 2 4 354780 5 5 5 5 10 5 5 5 4 4 14 351380 6 6- 6 6 6 4 6 6 6 5 6 350950 7 7 9 7 12 7 4 7 7 8 5 351350 8 9 12 8 8 6 8 8 5 9 26 l 353800 9 13 7 18 31 15 26 15 25 31 66 354520 10 8 30 9 4 19 9 9 14 22 29 352590 11 11 16 10 7 13 10 10 11 14 10 355220 12 10 11 13 13 11 15 11 16 16 8 353070 13 17 14 11 18 24 11 16 17 45 19 352310 14 15 8 23 53 22 28 25 24 32 33 353870 15 12 13 15 11 12 16 13 19 19 21 350570 16 16 10 14 19 16 12 14 12 7 16 355250 17 14 38 12 5 27 13 12 15 27 11 354880 18 18 17 16 14 8 21 20 22 10 34 351300 19 23 15 29 41 31 31 28 29 28 50 353060 20 22 22 22 20 20 23 23 20 17 30 (1) - Total number of industrialized municipals is 548. These results are confirmed upon calculating the Spearman rank correlation coefficient between each index. This coefficient, noted rs, is calculated as follows: n 6E(Ri -Rj)2 (4) rs = 1- i=1 1 n(n2-1 where Rj and Rj is the rank of the municipal under methodology i and j, and n is the number of pairs of ranks (n = 75 for Rio and 548 for Sao Paulo). Observe in Tables 10 and 11 that the rank correlation coefficients are significantly lower for the volume-based index, and higher across the risk-weighted indices with the exception of the long-term exposure HBEL and TRI indices. 19 Table 10 Rank correlation coefficients for Rio de Janeiro TLV PEL REL MAK HBEL TEEL-0 TEEL-1 TEEL-2 TEEL-3 TRI VOL TLV 1.0000 PEL 0.9586 1.0000 REL 0.9695 0.9839 1.0000 MAK 0.9916 0.9630 0.9706 1.0000 HBEL 0.9759 0.9687 0.9619 0.9691 1.0000 TEEL-0 0.9800 0.9630 0.9809 0.9870 0.9617 1.0000 TEEL-1 0.9822 0.9619 0.9688 0.9940 0.9557 0.9878 1.0000 TEEL-2 0.9864 0.9818 0.9830 0.9879 0.9747 0.9884 0.9871 1.0000 TEEL-3 0.9744 0.9725 0.9757 0.9838 0.9492 0.9815 0.9880 0.9881 1.0000 TRI 0.9613 0.9451 0.9609 0.9648 0.9390 0.9627 0.9621 0.9565 0.9602 1.0000 VOL 0.9368 0.8991 0.9209 0.9512 0.8944 0.9452 0.9514 0.9376 0.9570 0.9239 1.0000 Table 11 Rank correlation coefficients for SAo Paulo TLV PEL REL MAK HBEL TEEL-0 TEEL-1 TEEL-2 TEEL-3 TRI VOL TLV 1.0000 PEL 0.9704 1.0000 REL 0.9684 0.9818 1.0000 MAK 0.9941 0.9751 0.9704 1.0000 HBEL 0.9560 0.9671 0.9351 0.9543 1.0000 TEEL-0 0.9791 0.9795 0.9812 0.9864 0.9524 1.0000 TEEL-1 0.9849 0.9729 0.9687 0.9966 0.9459 0.9858 1.0000 TEEL-2 0.9849 0.9904 0.9765 0.9921 0.9654 0.9873 0.9911 1.0000 TEEL-3 0.9693 0.9775 0.9712 0.9834 0.9352 0.9802 0.9874 0.9897 1.0000 TRI 0.9596 0.9512 0.9585 0.9606 0.9427 0.9661 0.9570 0.9566 0.9510 1.0000 VOL 0.9373 0.9172 0.9349 0.9529 0.8827 0.9518 0.9594 0.9429 0.9540 0.9300 1.0000 20 Figures 2 and 3 compare the TLV and volume rankings of all municipals of Rio (Figure 2) and Sao Paulo (Figure 3). In these figures, a darker shading is to be interpreted as more pollution intensive.8 Note in both figures that where the volume ranking would indicate a number of municipals as high priority (dark shade), the TLV index ranks a large number of these same municipal quite low, implying that the pollution loads are relatively non-toxic. These results indicate first that accounting for risk does make a significant difference in tenns of identifying the areas (or industrial sectors) that should be deserving attention. They also indicate that in both Sao Paulo and Rio, a significant reduction of emissions of toxic chemicals could be obtained by allocating monitoring and control resources in a relatively small number or municipals. Another noticeable result across the risk-weighted indices is between the short term (TEEL) and long term indices (HBEL and TRI). Note the lower correlation coefficients in Tables 10 and 11. In comparing the HBEL ranking with the short term lethal exposure index TEEL-3 in Figures 4 and 5, we observe a number of areas which have a larger potential to be lethal in the long term (i.e. Municipal code 330010; ranked 8" by HBEL, 24"' by TEEL-3). Thus at greater levels of dis-aggregation, it appears that the outlook of risk (short or long term) becomes increasingly significant. 8 Graphs were constructed using ArcView 3.0 Geographical Information System. Estimates were divided into 7 shaded categories to highlight changes in relative ranking. For comparative purposes, the "volume" legend in Figures 2 and 3 were adjusted to match that of the TLV scale. 21 Figure 2: Volume vs. TLV risk-weighted ranking of municipals for the Rio de Janeiro region Volume 0.005 - 27.821 r 27.821 - 91.451 3 91.451 - 146.206 - 146 206 - 337.532 337.532 - 572.416 572A16 -1057.959 _ 1057.959 - 59664.334 4~~~~~ TLV W0.005 -27.821 Municipal 330610 EJ27.821 -91.451 m91.451 -146.206 146.206 -337.532 337.532 -572.416 n572416 - 1057.959 1057.959 - 59664.334 W E S Figure 3: Volume vs. TLV risk-weighted ranking of municipals for the Sao Paulo region 20.00 1- 57.19 57.19 -208.91 i 208.91 - 448.986 448.986 - 1068.414 1068414 - 1917.471 1917.471 -3706.26 _ 3706.26 - 215857.414 vTW+V Figure 4: Acute (TEEL-3) vs. chronic-weighted (HBEL) ranking of municipals for the Rio de Janeiro region HBEL 0.001 - 6.13 6.13 - 19.364 19.364-40.827 40.827 - 80.14 8014 -126.993 _ 126.993 - 327.719 327.719 -5942.927 Municipal , TIEEL3 301 Il0.001 - 6.13 33 10-- 6.13 - 19.364X.:J PM 19364 -40.827 40.827 - 80.14 80.14 - 126.993 126.993 -327719 327719 -5942.927 N > W~~~~~~~~w E r Figure 5: Acute (TEEL-3) vs. chronic-weighted (HBEL) ranking of municipals for the Sao Paulo region HBELI W 0001 -22.387 352090 | 22.387 - 66.873 66.873 - 130.484 130.484 - 235.272 235.272 - 366.99 366.99 - 1127.663 307 1127.663 - 17721,629 Municipal 352230 T:L:<; W0001 -22.387 EZI 22.387 -66.873 M 66.73 - 130.484N 130.484 - 235.272 235.272 - 366.99 36699 - 11277663 1127,663 - 17721.629 IV. Conclusion The accounting and public release of information pertaining to industrial toxic pollution emissions is meeting increasing criticism in that these listings typically do not account for the different toxicity risks associated with different pollutants: A firm emitting a large quantity of a relatively harmless substance would rank as a larger polluter than another firm emitting a small quantity of a very potent substance. It is argued that such "unweighted" rankings of firms may lead to a misallocation of resources and wrong prioritization of effort devoted to pollution control. This may be of particular importance for developing countries where resources devoted to pollution control are typically scarce. In an attempt to account for the relative differences in chemical toxicity, a number of organizations have developed thresholds or exposure limits for various pollutants to account for their various toxicity risk. Given the large number of toxicity risk factors and methodologies currently available, a crucial issue pertains to the possibility that different risk indicators may yield different results, thus leading to different sets of priorities. In this paper, we have reviewed and applied to Brazil seven risk methodologies currently available, and constructed 10 different sets of toxicity risk factors from these indicators. Upon ranking states and municipals for their pollution intensity, results indicate that at the state level, risk weighted rankings remain largely the same across the 10 different sets of toxicity risk factors used in this paper. This result by and large also holds true at the 26 municipal level with the exception of the long-term exposure indices (HBEL and TRI) which offer different rankings of pollution intensive municipals. Moreover, at the state level, the unweighted ranking is relatively similar to the risk weighted ranking. However, at the municipal level, significant differences were found between the risk weighted and unweighted rankings. These findings suggest that it is of importance for environmental regulators to engage into weighting pollutants for their relative toxicity risk when prioritizing pollution control effort either at the industrial or regional level. This exercise appears to be of greater importance as one seeks to determine prioritization of pollution control effort at a greater level of dis-aggregation. 27 References American Conference of Governmental Industrial Hygienists (1997), Guide to Occupational Exposure Values - 1997, Cincinnati, ACGIH. Hamilton, J.T. (1995), "Pollution as News: Media and Stock Market Reactions to the Toxic Release Inventory Data", Journal of Environmental Economics and Management, 28: 98-113. Health Based Exposure Limits Committee (1995), Chemical Exposure Guidelines, Santa Clara Center for Occupational Safety and Health (SCCOSH) at http://turva.me.tut.fi/cgi-bin/wilma/chesa.840022811 .html (obtained November 1997). Hettige, M., D. Wheeler, (1996), An Environmental Performance Analysis System for Industrial Plants in Mexico, mimeo, Development Research Group, The World Bank, Washington, D.C. Hettige, M., P. Martin, M. Singh and D. Wheeler (1995), The Industrial Pollution Projection System, Working Paper Series, No. 1431, Development Research Group, The World Bank, Washington, D.C. Hettige, H., Mani, M. and D. Wheeler (1998), Industrial Pollution in Economic Development (Kuznets Revisited), Working Paper Series, No. 1876, Development Research Group, The World Bank, Washington, D.C. Horvath, A., C. T. Hendrickson, L. B. Lave, F. C. McMichael and T-S. Wu (1995), "Toxic Emissions Indices for Green Design and Inventory", Environmental Science & Technology, 29, 2: 86A-90A. Instituto Brasileiro de Geografia e Estatisticsa (IBGE), Anuario Estatistico do Brasil, 1995. IBGE, Directoria de Geociencias, Departamento de Estudos Territoriais, 1995 at http://www.ibge.org/english/brasil/other6.html (obtained June 1, 1998). Instituto de Pesquisa Econ6mica Aplicada (IPEA), Boletim Conjuntural, 1994. Konar, S. and M. A. Cohen (1997), "Information as Regulation: The Effect of Community Right-to-Know Laws on Toxic Emissions", Journal of Environmental Economics and Management, 32, 1: 109-124. Laplante, B. and K. Smits (1998), Industrial Pollution in Latvia, mimeo, Development Research Group, The World Bank, Washington, D.C. 28 Lewis, R. J., Sr. (Ed.) (1997), Sax's Dangerous Properties of Industrial Materials, 9h Ed. Van Nostrand Reinhold, New York, August. NIOSH Pocket Guide to Chemical Hazards (1994), U.S. Department of Health and Human Services, Public Health Service, Centers for Disease Control, June. Swanson, M., G. A. Davis, L. E. Kincaid, T. W. Schultz, J. E. Bartmess, S.L. Jones and E. L. George (1997), "A Screening Method for Ranking and Scoring Chemicals by Potential Human Health and Environmental Impacts", Environmental Toxicology and Chemistry, 16, 2: 372-383. "Temporary Emergency Exposure Limits" (1998), U.S. Department of Energy, Subcommittee on Consequence Assessment and Protective Action (SCAPA), Rev. 1 3,WSMS-SAE-98-000 1 at http://dewey.tis.eh.doe.gov/web/chem safety/doe reg.html (obtained February, 1998). "Toxic Release Inventory Indicators Toxicity Weights" (1998), U.S. Environmental Protection Agency, Office of Pollution Prevention and Toxics, Sector Facility Indexing Project, Washington, D.C. at http://es.epa.gov/oeca/sfi/sabrev.html (obtained January, 1998). World Bank (1996), Brazil, Managing Environmental Pollution in the State of Rio de Janeiro, Vol. I & II, Report No. 15488-BR. World Bank (1998), Brazil, Managing Environmental Pollution Problems: The Brown EnvironmentalAgenda, Vol. I: Policy Report, No. 16635-BR. 29 Appendix A Further details of risk indicators TCLo - Toxic Concentration Low - the lowest concentration of a substance in air to which humans or animals have been exposed for any given period of time that has produced any toxic effect in humans or produced a carcinogenic, neoplastigenic, or teratogenic effect in animals or humans. TDLo - Toxic Dose Low - the lowest dose of a substance introduced by any route, other than inhalation, over any given period of time and reported to produce any toxic effect in humans or to produce carcinogenic, neoplastigenic, or teratogenic effects in animals or humans. LCLo - Lethal Concentration Low - the lowest concentration of a substance in air, other than LC50, which has been reported to have caused death in humans or animals. The reported concentrations may be entered for periods of exposure which are less than 24 hours (acute) or greater than 24 hours (subacute and chronic). LDLo - Lethal Dose Low - the lowest dose (other than LD50) of a substance introduced by any route, other than inhalation, over any given period of time in one or more divided portions and reported to have caused death in humans or animals. LD50 - Lethal Dose Fifty - a calculated dose of a substance which is expected to cause the death of 50% of an entire defined experimental animal population. It is determined from the exposure to the substance by any route other than inhalation of a significant number from that population. 30 Appendix B Chemical substances in analysis CAS Substance 71556 1, ,1 -TRICHLOROETHANE (METHYL CHLOROFORM) 79345 1,1,2,2-TETRACHLOROETHANE 76131 1, 1,2-TRICHLORO- 1,2,2-TRIFLUOROETHANE (FREON 11 33 79005 1,1,2-TRICHLOROETHANE 57147 1,1-DIMETHYL HYDRAZINE _ 120821 1,2,4-TRICHLOROBENZENE 95636 1,2,4-TRIMETHYLBENZENE (PSEUDOCUMENE) 106887 1,2-BUTYLENE OXIDE (1,2-EPOXYBUTANE) 96128 1,2-DIBROMO-3-CHLOROPROPANE (DBCP) 106934 1,2-DIBROMOETHANE (EDB) (ETHYLENE DIBROMIDE) 107062 1,2-DICHLOROETHANE (ETHYLENE DICHLORIDE) 540590 1,2-DICHLOROETHYLENE 78875 1,2-DICHLOROPROPANE (PROPYLENE DICHLORIDE) 122667 1,2-DIPHENYLHYDRAZINE (HYDRAZOBENZENE) 106990 1,3-BUTADIENE 541731 1,3-DICHLOROBENZENE (M-ISOMER) 542756 1,3-DICHLOROPROPYLENE 123911 1,4-DIOXANE (1,4-DIETHYLENE DIOXIDE) 82280 I -AMINO-2-METHYLANTHRAQUINONE 95954 2,4,5-TRICHLOROPHENOL 88062 2,4,6-TRICHLOROPHENOL 94757 2,4-D (DICHLOROPHENOXYACETIC ACID) 39156417 2,4-DIAMINO ANISOLE SULFATE 615054 2,4-DIAMINOSANISOLE 95807 2,4-DIAMINOTOLUENE 120832 2,4-DICHLOROPHENOL 105679 2,4-DIMETHYLPHENOL 51285 2,4-DINITROPHENOL 121142 2,4-DINITROTOLUENE 606202 2,6-DINITROTOLUENE 87627 2,6-XYLIDINE 53963 2-ACETOAMINOFLUORENE 117793 2-AMINOANTHRAQUINONE 532274 2-CHLOROACETOPHENONE (ALPHA) (PHENACYL CH-IORIDE) 110805 2-ETHOXYETHANOL (ETHYLENE GLYCOL MONOETHYL ETHER; CELLOSOLVE) 109864 2-METHOXYETHANOL (ETHYLENE GLYCOL MONOMETHYL ETHER. METHYL CELLOSOLVE) 88755 2-NITROPHENOL 79469 2-NITROPROPANE 90437 2-PHENYLPHENOL (SODIUM SALT) 91941 3,3'-DICHLOROBENZIDINE (AZO DYE) 119904 3,3'-DIMETHOXYBENZIDINE (AZO DYE; o-DIANISIDINE) 119937 3,3'-DIMETHYLBENZIDINE (AZO DYE; o-TOLIDINE) 101804 4,4'-DIAMINODIPHENYL ETHER (4,4'-OXYDIANILINE) 80057 4,4'-ISOPROPYLIDENEDIPHENOL (BISPHENOL A) 101144 4,4'-METHYLENE BIS(2-CHLOROANILINE) (MBOCA) 101611 4,4'-METHYLENE BIS(N,N-DIMETHYL) BENZELAMINE 101779 4,4'-METHYLENE DIANILINE (4,4'-DIAMINODIPHENYLMETHANE) 139651 4,4'-THIODIANILINE 534521 4,6-DINITRO-O-CRESOL 60093 4-AMINOAZOBENZENE 31 CAS Substance 92671 4-AMINODIPHENYL (P-isomer) 60117 4-DIMETHYLAMINOAZOBENZENE 92933 4-NITRODIPHENYL (P-isomer) 100027 4-NITROPHENOL 99592 5-NITRO-O-ANISIDINE 75070 ACETALDEHYDE 60355 ACETAMIDE 67641 ACETONE 75058 ACETONITRILE 107028 ACROLEIN 79061 ACRYLAMIDE 79107 ACRYLIC ACID 107131 ACRYLONITRILE (VINYL CYANIDE) 309002 ALDRIN (1,4,5,8-DIMETHANONAPHTHALENE) 107051 ALLYL CHLORIDE 7429905 ALUMINUM (FUME OR DUST) 1344281 ALUMINUM OXIDE (FIBROUS FORM) 97563 AMINOAZOTOLUENE, 0-ISOMER (C.I. SOLVENT YELLOW 3) 7664417 AMMONIA 6484522 AMMONIUM NITRATE (SOLUTION) 7783202 AMMONIUM SULFATE (SOLUTION) 62533 ANILINE 90040 ANISIDINE (0-ISOMER) 104949 ANISIDINE (P-ISOMER) 134292 ANISIDINE HYDROCHLORIDE (0-ISOMER) 120127 ANTHRACENE 7440360 ANTIMONY 7440382 ARSENIC 1332214 ASBESTOS (FRIABLE) 492808 AURAMINE (C.I. SOLVENT YELLOW 34) 7440393 BARIUM 98873 BENZAL CHLORIDE 55210 BENZAMIDE 71432 BENZENE 92875 BENZIDINE 98077 BENZOIC TRICHLORIDE (BENZYL TRICHLORIDE; TRICHLOROMETHYLBENZENE) 98884 BENZOYL CHLORIDE 94360 BENZOYL PEROXIDE 100447 BENZYL CHLORIDE 7440417 BERYLLIUM 92524 BIPHENYL (DIPHENYL) 108601 BIS(2-CHLORO-I -METHYLETHYL) ETHER (DICHLOROISOPROPYL ETHER) 111444 BIS(2-CHLOROETHYL) ETHER (DICHLOROETHYL ETHER; 2,2'-DICHLORODIETHYL ETHER) 103231 BIS(2-ETHYLHEXYL) ADIPATE 542881 BIS(CHLOROMETHYL) ETHER (DICHLOROMETHYL ETHER) (BCME) 75252 BROMOFORM (TRIBROMOMETHANE) 141322 BUTYL ACRYLATE (ACRYLIC ACID & N-BUYTL ESTER) 78922 BUTYL ALCOHOL (SEC-BUTANOL) 75650 BUTYL ALCOHOL (TERT-BUTANOL) 85687 BUTYL BENZYL PHTHALATE 123728 BUTYRALDEHYDE 2650182 C.I. ACID BLUE 9, DIAMMONIUM SALT 3844459 C.I. ACID BLUE 9, DISODIUM SALT 4680788 C.I. ACID GREEN 3 32 CAS Substance 569642 C.l. BASIC GREEN 4 989388 C.I. BASIC RED 1 1937377 C.I. DIRECT BLACK 38 2602462 C.I. DIRECT BLUE 6 16071866 C.I. DIRECT BROWN 95 2832408 C.I. DISPERSE YELLOW 3 81889 C.I. FOOD RED 15 3761533 C.I. FOOD RED 5 3118976 C.I. SOLVENT ORANGE 7 842079 C.I. SOLVENT YELLOW 14 128665 C.I. VAT YELLOW 4 7440439 CADMIUM 156627 CALCIUM CYANAMIDE 133062 CAPTAN 63252 CARBARYL (SEVIN) 75150 CARBON DISULFIDE 56235 CARBON TETRACHLORIDE (TETRACHLOROMETHANE) 463581 CARBONYL SULFIDE 120809 CATECHOL (PYROCATECHOL) 133904 CHLORAMBEN (3-AMINO-2,5-DICHLOROBENZOIC ACID) 57749 CHLORDANE 7782505 CHLORINE 10049044 CHLORINE DIOXIDE 79118 CHLOROACETIC ACID 108907 CHLOROBENZENE (CHLORINATED BENZENE) 510156 CHLOROBENZILATE (4,4'-DICHLORO-BENZILIC ACID ETHYL ESTER) 67663 CHLOROFORM 107302 CHLOROMETHYL METHYL ETHER (CMME) 126998 CHLOROPRENE (BETA-CHLOROPRENE; NEOPRENE) 1897456 CHLOROTHALONIL 7440473 CHROMIUM 7440484 COBALT 7440508 COPPER (FUME OR DUST) 120718 CRESIDINE (P-ISOMER) 1319773 CRESOL (ALL ISOMERS) 108394 CRESOL (M-ISOMER) 95487 CRESOL (0-ISOMER) 106445 CRESOL (P-ISOMER) 98828 CUMENE 80159 CUMENE HYDROPEROXIDE 135206 CUPFERRON 110827 CYCLOHEXANE 1163195 DECABROMODIPHENYL OXIDE 117817 DI (2-ETHYLHEXYL) OR (SEC-OCTYL) PHTHALATE (DEHP) 2303164 DIALLATE 25376458 DIAMINOTOLUENE (MIXED ISOMERS) 334883 DIAZOMETHANE 132649 DIBENZOFURAN 84742 DIBUTYL PHTHALATE 25321226 DICHLOROBENZENE (MIXED ISOMERS) 95501 DICHLOROBENZENE 1,2-(O-ISOMER) 106467 DICHLOROBENZENE 1,4-(P-ISOMER) 75274 DICHLOROBROMOMETHANE (BROMOCHLORO.) 75092 DICHLOROMETHANE (METHYLENE CHLORIDE) 33 CAS Substance 62737 DICHLORVOS 115322 DICOFOL 1464535 DIEPOXYBUTANE 111422 DIETHANOLAMINE 84662 DIETHYL PHTHALATE 64675 DIETHYL SULFATE 131113 DIMETHYL PHTHALATE 77781 DIMETHYL SULFATE 121697 DIMETHYLANILINE (N,N-DIMETHYLANILINE) 79447 DIMETHYLCARBAMOYL CHLORIDE 117840 DI-N-OCTYL PHTHALATE 106898 EPICHLOROHYDRIN (I -CHLORO-2,3-EPOXYPROPANE) 140885 ETHYL ACRYLATE (ACRYLIC ACID & ETHYL ESTER) 100414 ETHYL BENZENE 75003 ETHYL CHLORIDE (CHLOROETHANE) 541413 ETHYL CHLOROFORMATE 74851 ETHYLENE 107211 ETHYLENE GLYCOL 75218 ETHYLENE OXIDE 96457 ETHYLENE THIOUREA (2-IMIDAZOLIDINETHIONE) 151564 ETHYLENEIMINE 2164172 FLUOMETURON 50000 FORMALDEHYDE 76448 HEPTACHLOR 87683 HEXACHLORO- 1,3-BUTADIENE 118741 HEXACHLOROBENZENE 77474 HEXACHLOROCYCLOPENTADIENE 67721 HEXACHLOROETHANE 1335871 HEXACHLORONAPHTHALENE 680319 HEXAMETHYL PHOSPHORAMIDE 302012 HYDRAZINE 10034932 HYDRAZINE SULFATE 7647010 HYDROCHLORIC ACID (HYDROGEN CHLORIDE) 74908 HYDROGEN CYANIDE 7664393 HYDROGEN FLUORIDE (HYDROFLUORIC ACID) 123319 HYDROQUINONE (DIHYDROXYBENZENE) 78842 ISOBUTYRALDEHYDE 67630 ISOPROPYL ALCOHOL (MANUFACTURING, STRONG-ACID PROCESS ONLY, NO PROCESS) 7439921 LEAD 58899 LINDANE (HEXACHLOROCYCLOHEXANE-gamma) 108316 MALEIC ANHYDRIDE 12427382 MANEB 7439965 MANGANESE 108781 MELAMINE 7439976 MERCURY 67561 METHANOL (METHYL ALCOHOL) 72435 METHOXYCHLOR 96333 METHYL ACRYLATE 74839 METHYL BROMIDE (BROMOMETHANE) 74873 METHYL CHLORIDE 78933 METHYL ETHYL KETONE (MEK; 2-BUTANONE) 60344 METHYL HYDRAZINE 74884 METHYL IODIDE 108101 METHYL ISOBUTYL KETONE (HEXONE) 34 CAS Substance 624839 METHYL ISOCYANATE 80626 METHYL METHACRYLATE (METHACRYLIC ACID METHYL ESTER) 101688 METHYLENE BISPHENYL ISOCYANATE (DIPHENYLMETHANE-4,4'-DIISOCYANATE; MDI) 74953 METHYLENE BROMIDE 1634044 METHYL-TERT-BUTYL ETHER 90948 MICHLER'S KETONE 1313275 MOLYBDENUM TRIOXIDE 505602 MUSTARD GAS (2,2'-DICHLORODIETHYL SULFIDE) 91203 NAPHTHALENE 134327 NAPHTHYLAMINE (ALPHA or 2-NAPHTHYLAMINE) 91598 NAPHTHYLAMINE (BETA or 2-NAPHTHYLAMINE) 71363 N-BUTANOL (N-BUTYL ALCOHOL) 7440020 NICKEL 7697372 NITRIC ACID 139139 NITRILOTRIACETIC ACID 98953 NITROBENZENE 1836755 NITROFEN 51752 NITROGEN MUSTARD (N-METHYL-BIS(2-CHLOROETHYL)AMINE) 55630 NITROGLYCERIN (NG) 156105 NITROSODIPHENYLAMINE (P-ISOMER) 55185 N-NITROSODIETHYLAMINE (NDEA) 62759 N-NITROSODIMETHYLAMINE (N,N-DIMETHYLNITROSOAMINE) 924163 N-NITROSODI-N-BUTYLAMINE (DBN) 621647 N-NITROSODI-N-PROPYLAMINE (NDPA) 86306 N-NITROSODIPHENYLAMINE 4549400 N-NITROSOMETHYLVINYLAMINE 59892 N-NITROSOMORPHOLINE (NMOR) 759739 N-NITROSO-N-ETHYLUREA 684935 N-NITROSO-N-METHYLUREA 16543558 N-NITROSONORNICOTINE 100754 N-NITROSOPIPERIDINE (NPIP) 2234131 OCTACHLORONAPHTHALENE 20816120 OSMIUM TETROXIDE 56382 PARATHION 87865 PENTACHLOROPHENOL 79210 PERACETIC ACID 108952 PHENOL 106503 PHENYLENEDIAMINE (P-ISOMER) 75445 PHOSGENE (CARBONYL CHLORIDE) 7664382 PHOSPHORIC ACID 7723140 PHOSPHORUS (YELLOW OR WHITE) 85449 PHTHALIC ANHYDRIDE 88891 PICRIC ACID (2,4,6-TRINITROPHENOL) 1336363 POLYCHLORINATED BIPHENYLS (CHLORODIPHENYLS, 54% CHLORINE) 1120714 PROPANE SULTONE, 1,3- 57578 PROPIOLACTONE (BETA-PROPIOLACTONE) 123386 PROPIONALDEHYDE 114261 PROPOXUR (BAYGON) 115071 PROPYLENE 75569 PROPYLENE OXIDE (1,2-EPOXYPROPANE) 75558 PROPYLENEIMINE (2-METHYLAZIRIDINE) 110861 PYRIDINE 91225 QUINOLINE 106514 QUINONE (P-BENZOQUINONE) 35 CAS Substance 82688 QUINTOZENE (PENTACHLORONITROBENZENE) 81072 SACCHARIN (MANUFACTURING ONLY, NO PROCESSOR REPORTING) 94597 SAFROLE 7782492 SELENIUM 7440224 SILVER 1310732 SODIUM HYDROXIDE (SOLUTION) 7757826 SODIUM SULFATE (SOLUTION) 100425 STYRENE (PHENYLETHYLENE; VINYL BENZENE) 96093 STYRENE OXIDE 7664939 SULFURIC ACID 100210 TEREPHTHALIC ACID 127184 TETRACHLOROETHYLENE (PERCHLOROETHYLENE) 961115 TETRACHLORVINPHOS (STIROFOS) 7440280 THALLIUM 62555 THIOACETAMIDE 62566 THIOUREA 1314201 THORIUM DIOXIDE 13463677 TITANIUM DIOXIDE 7550450 TITANIUM TETRACHLORIDE 108883 TOLUENE (TOLUOL) 584849 TOLUENE-2,4-DIISOCYANATE (TDI) 91087 TOLUENE-2,6-DIISOCYANATE 95534 TOLUIDINE (0-ISOMER) 636215 TOLUIDINE HYDROCHLORIDE (0-ISOMER) 8001352 TOXAPHENE (CHLORINATED CAMPHENE) 68768 TRIAZIQUONE 52686 TRICHLORFON 79016 TRICHLOROETHYLENE 1582098 TRIFLURALIN (2,6-DINITRO-N,N-DIPROPYL-4-(TRIFLUOROMETHYL) BENZENAMINE) 126727 TRIS(2,3-DIBROMOPROPYL) PHOSPHATE 51796 URETHANE (CARBAMIC ACID, ETHYL ESTER) 1314621 VANADIUM (PENTAOXIDE; FUME OR DUST) 108054 VINYL ACETATE 593602 VINYL BROMIDE (BROMOETHENE) 75014 VINYL CHLORIDE 75354 VINYLIDENE CHLORIDE (l,1-DICHLOROETHYLENE) 108383 XYLENE (M-ISOMER) 1330207 XYLENE (MIXED ISOMERS) 95476 XYLENE (0-ISOMER) 106423 XYLENE (P-ISOMER) 1314132 ZINC OXIDE (FUME OR DUST) 12122677 ZINEB 36 Policy Research Working Paper Series Contact Title Author Date for paper WPS1979 Banking on Crises: Expensive Gerard Caprio, Jr. September 1998 P. Sintim-Aboagye Lessons from Recent Financial 38526 Crises WPS1980 The Effect of Household Wealth Deon Filmer September 1998 S. Fallon on Educational Attainment: Lant Pritchett 38009 Demographic and Health Survey Evidence WPS1981 Evaluating Public Expenditures James E. Anderson September 1998 L. Tabada When Governments Must Rely Will Martin 36896 on Distortionary Taxation WPS1982 Analyzing Financial Sectors in Alan Roe September 1998 D. Cortijo Transition: With Special Reference Paul Siegelbaum 84005 to the Former Soviet Union Tim King WPS1983 Pension Reform in Small Developing Thomas Charles Glaessner September 1998 M. Navarro Countries Salvador Valdes-Prieto 84722 WPS1984 NAFTA, Capital Mobility, and Thomas Charles Glaessner September 1998 M. Navarro Mexico's Financial System Daniel Oks 84722 WPS1985 The Optimality of Being Efficient: Lawrence M. Ausubel September 1998 S. Vivas Designing Auctions Peter Cramton 82809 WPS1986 Pufting Auction Theory to Work: Paul Milgrom September 1998 S. Vivas The Simultaneous Ascending Auction 82809 WPS1987 Political Economy and Political Ariel Dinar September 1998 F. Toppin Risks of Institutional Reform in Trichur K. Balakrishnan 30450 the Water Sector Joseph 'Wambia WPS1988 The Informal Sector, Firm Dynamics, Alec R. Levenson September 1998 T. Gomez and Institutional Participation William F. Maloney 32127 WPS1989 Contingent Government Liabilities: Hana Polackova October 1998 A. Panton A Hidden Risk for Fiscal Stability 85433 WPS1 990 The East Asia Crisis and Corporate Michael Pomerleano October 1998 N. Dacanay Finances: The Untold Micro Story 34068 WPS1991 Reducing Air Pollution from Urban Mark Heil October 1998 R. Yazigi Passenger Transport: A Framework Sheoli Pargal 37176 for Policy Analysis WPS1 992 The Present Outlook for Trade John Croome October 1998 L. Tabada Negotiations in the World Trade 36896 Organization Policy Research Working Paper Series Contact Title Author Date for paper WPS1993 Financial Safety Nets and Incentive Philip L. Brock October 1998 K. Labrie Structures in Latin America 38256 WPS1994 Estimating Wealth Effects without Deon Filmer October 1998 S. Fallon Expenditure Data - or Tears: Lant Pritchett 38009 with an Application to Educational Enrollments in States of India WPS1995 What Macroeconomic Policies Mansoor Dailami October 1998 B. Nedrow Are Sound?" Nadeem ul Haque 31585 WPS1996 Namibia's Social Safety Net: Kalinidhi Subbarao October 1998 P. Lizarondo Issues and Options for Reform 87199 WPS1997 On Measuring Literacy Kaushik Basu October 1998 M. Mason James E. Foster 30809 WPS1 998 The Structure and Determinants of Sudharshan Canagarajah October 1998 A. Garscadden Inequality and Poverty Reduction Dipak Mazumdar 38400 in Ghana, 1988-92 Xiao Ye WPS1999 Heterogeneity among Mexico's Wendy V. Cunningham October 1998 T. Gomez Micro-Enterprises: An Application William F. Maloney 32127 of Factor and Cluster Analysis WPS2000 GATT Experience with Safeguards: J. Michael Finger Oct:ober 1998 L. Tabada Making Economic and Political 36896 Sense of the Possibilities that the GATT Allows to Restrict Imports WPS2001 Measuring the Dynamic Gains from Romain Wacziarg November 1998 S. Crow Trade 30763