\V/PS 1% 941 POLICY RESEARCH WORKING PAPER 1877 What Improves Strengthened enforcement raises the price of pollution Environmental Performance? and provides an incentive to reduce it. A cost-effective . . ~~~~~~~~~~~~~~~~~complemnent to stricter Evidence from Mexican Industry eorcement ts etrecte enforcement is effective environmental management Susmita Dasgupta and training programs within Hemamala Hettige plants. David Wheeler The World Bank Development Research Group January 1998 LICY RESEARCH WORKING PAPER 1877 Summary findin gs Using new survey evidence, Dasgupta, Hettige, and of environmental specialists, and assigning environmental Wheeler analyze the effects of regulation, plant-level tasks to general managers is more effective than using management policies, and plant and firm characteristics special environmental managers. on environmental performance in Mexican factories. Regulatory pressure works. Plants that have They focus especially on management policies: the experienced regulatory inspections and enforcement are degree of effort to improve environmental performance significantly cleaner than those that have not. and the type of management strategy adopted. Public scrutiny promotes stronger environmental They index effort with two variables: adoption of ISO policies. Publicly traded Mexican firms are significantly 14000-type procedures for pollution management and clearer than privately held firms. use of plant personnel for environmental inspection and Size matters. Large plants in multiplant firms are much control. Proxies for strategic orientation are two indices more likely to adopt policies that improve environmental of mainstreaming: assigning environmental respon- performance. sibilities to general managers instead of specialized OE,CD influences do not matter. It is generally environmental managers, and providing environmental assurned that plants linked to OECD economies show training for all plant employees, not just specialists. superior environmental performance, but they find no Detailed survey data let them test the performance evidence that OECD links - including multinational impact of such factors as ownership, scale, sector, trade ownership, trade, management training, or management and other business relationships, local regulatory experience - affect environmental performance. enforcement, local community pressure, management New technology is not significantly cleaner. They find education and experience, and workers' general no evidence that plants with newer equipment perform education. Their findings: better environmentally (once other factors are accounted Process is important. Plants that institute ISO 14000- for). type internal management procedures show superior Education promotes clean production. Plants with environmental performance. more highly educated workers show significantly better Mainstreaming works. Environmental training for all environmental management efforts and performance. plant personnel is more effective than developing a cadre This paper - a product of the Development Research Group - is part of a larger effort in the group to understand the determinants of environmental performance in developing countries. The study was funded by the Bank's Research Support Budget under the research project "The Economics of Industrial Pollution Control in Developing Countries" (RPO 680- 20). Copies of this paper are available free from the World Bank, 1818 H Street N`W, Washington, DC 20433. Please contact David Wheeler, room MC2-529, telephone 202-473-3401, fax 202-522-3230, Internet address dwheelerl @worldbank.org. January 1998. (24 pages) The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange 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 polished. The papers carry the namies of the authors and should 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 countrzes they represent. Produced by the Policy Research DLissemination Center What Improves Environmental Performance? Evidence from Mexican Industry Susmita Dasgupta* Hemamala Hettige David Wheeler Development Research Group World Bank * The authors are respectively Economist, Development Research Group, World Bank; Economist, Asian Development Bank; and Principal Economist, Development Research Group, World Bank. We gratefully acknowledge the contribution of our World Bank colleagues Violeta Rosenthal, Paul Martin and Michele de Nevers, who organized, planned, and helped to implement the survey which provided the database for this paper. Our thanks to Paul Martin and Richard Wells for useful suggestions and comments. We would also like to acknowledge the work of our collaborators in the Mexican survey team, led by specialists from the Monterrey Institute of Technology. Financial support was provided by the World Bank's Country Department LA2 and the World Bank's Research Comniittee, under RPO 68-020. EXECUTIVE SUMMARY This paper uses new survey evidence to analyze the effects of regulation, plant-level management policies, and plant/firm characteristics on the environmental performance of Mexican factories. We focus particularly on management policies: the degree of effort to improve environmental performance, and the type of management strategy which is adopted. Effort is indexed by two variables: adoption of ISO 14000-type procedures for pollution management,1 and use of plant personnel for environmental inspection and control. Strategic orientation is proxied by two indices of 'mainstreaming:' assignment of environmental responsibilities to general managers, instead of specialized environmental managers; and general environmental training for plant employees, as opposed to training for specialists only. Our detailed survey data also enable us to test the performance impact of many other factors, including ownership, scale, sector, trade and other business relationships, local regulatory enforcement, local community pressure, management education and experience, and workers' general education. Recognizing that plant-level management policies and environmental performance are simultaneously determined, the paper uses two-stage least squares for econometric estimation. In the first-stage equations, the four management policy indices are regressed on exogenous measures of regulation and the characteristics of plants, firms and markets. The first-stage results are used to instrument the management indices in the second-stage regression, which measures their impact on actual performance, along with the effects of workers' education and the cost of pollution control. The implications of our results can be summarized as follows: (1) Process is important. Plants which institute ISO 14000-type internal management procedures exhibit superior enviromnental performance. (2) Mainstreaming works. Environmental training for all plant personnel is more effective than developing a cadre of environmental specialists; assigning environmental tasks to general managers is more effective than using special environmental managers. (3) Regulatory pressure works. Plants which have experienced regulatory inspections and enforcement are significantly cleaner than their counterparts. (4) Public scrutiny promotes stronger environmental policies. Publicly-traded Mexican firms are significantly cleaner than their privately-held counterparts (5) Size matters. Large plants in multi-plant firms are much more likely to adopt policies which improve environmental performance. (6) OECD influences don't matter. Analyses of pollution control in developing countries generally assume that plants linked to the OECD economies have superior environmental performance. However, we do not find a significant role for any OECD See Appendix 2 for a description of ISO 14000 environmental management standards. i linkage: multinational ownership, trade, management training, or management experience. (7) New technology isn't significantly cleaner. We find no evidence that plants with newer equipment have better environmental performance, once other factors are accounted for. (8) Education promotes clean production. Plants with more highly-educated workers have significantly greater environmental management effort and better performance. Our results have some interesting implications for the World Bank and other development assistance institutions. The significance of our regulatory variables underscores the importance of strengthened enforcement, which has been a traditional focus of technical assistance projects. Stricter enforcement raises theprice ofpollution and provides an important incentive for pollution reduction. However, our results also highlight the potential of programs which promote more effective environmental management and training within plants. Such programs can increase the elcasticity of industry's response to regulation, and may in some cases provide very cost-effective complements to stricter enforcement. ii 1. Introduction Researchers in economics and management typically differ in their treatment of management decisions. Assuming that firms optimize with known production sets (including management techniques), economists prefer to analyze responses to exogenous changes in markets, regulations and other variables. The conventional, if generally unspoken, view is that management practices are endogenous. Some management researchers, on the other hand, are comfortable with the assumption that managers have incomplete knowledge of production sets and techniques for optimization. This assumption validates empirical studies which relate management choices to differences in profitability. Although we are economists, we adopt the management research perspective in this paper because the assumption of incomplete information seems appropriate. For plant and firm managers in developing countries, improved environmental performance implies experimentation with new production sets. In Mexico, for example, serious environmental regulation is largely a phenomenon of the 1990's. Both regulators and businesses are still adjusting to the environmental era, and there is considerable uncertainty about the relative effectiveness of alternative approaches to regulatory policy and plant-level environmental management. Using recent survey evidence from Mexico, this paper analyzes the relationships linking management and environmental performance in a large sample of industrial facilities. We focus on two questions. First, why does plant-level environmental performance in developing countries vary so widely when regulation is weak? Conventional theory would suggest little or no pollution control effort when the 'price of pollution' is very low. However, many factories in developing countries have already adopted significant pollution control measures. Recent work on industrial pollution in Asia has suggested the importance of three factors:2 formal regulation (where it exists); informal regulation, or community pressure; and plant and firm characteristics which affect both the cost of abatement and the incentive to abate. In Asia, the characteristics which best explain variations in environmental performance are plant size (a plus, because of public visibility and scale economies in, abatement); government ownership (a distinct minus, because government facilities are less efficient and typically shielded from formal or informal regulatory pressure); and production sector (because of generic variations in process technologies and abatement costs). Contrary to the conventional view, this research has not found any evidence that multinational ownership has a significant effect on environmental performance. However, there is strong evidence that local community pressure (informal regulation) has an impact. While the Asian results provide some initial insights, data limitations have prevented a comprehensive analysis. Important but unanswered questions include: 2 See Pargal and Wheeler, 1996; Hettige, et. al., 1996; Hartman, et. al., 1996; Wang and Wheeler, 1996. 1 1. Are plants with newer equipment more likely to comiply with regulations, both because newer technologies are cleaner and because end-of-pipe abatement is less costly for plants which use newer processes? 2. Are plants which export heavily to the OECD affected by environmentalist concerns in their client countries? 3. Are plants cleaner if they have managers whose training and/or experience has been in OECD countries? 4. Are firms whose shares are publicly traded more sensitive to environmental issues, and more likely to seek improved environmental performance, than privately-held firms? 5. Do scale economnies in pollution control originate at the firm level, as well as the plant level? 6. Do plants with more human capital control pollution more effectively? Our second question is: How do manufacturing facilities achieve better environmental performance? We can address this question because the Mexican industry survey includes detailed questions about environmental management and training practices. In the paper, we investigate interplant differences in these practices and their impact on environmental performance. Our empirical assessment of internal policies recognizes an inevitable simultaneity problem: Positive correlations between superior environrnental performance and indices of environmental management do not imply causal roles for the latter. Firms will almost certainly change their environmental management practices when optimal pollution abatement levels are altered by exogenous changes in regulations or prices. We use two-stage least squares to address this simultaneity problem. In the first-stage equations, we regress internal management indices on exogenous determinants of pollution control activity. Our management indices captu:re both relative effort to improve performance and the type of strategy adopted by the plant. Effort is indexed by two variables: degree of adoption of ISO 14000-type procedures for pollution management; and expanded use of plant personnel for environmental inspection and control. Strategic orientation is captured by two measures of 'mainstreaming' environmental concerns: assignment of environmental responsibilities to general managers, rather than specialized environmental managers; and general environmental training for plant employees, as opposed to training focused on environmental specialists. Our detailed survey data enable us to test the effect of many exogenous factors on management effort and strategic orientation. These include plant and firm scale, ownership, human resource quality, trade relationships, OECD training and experience of plant managers, experience of formal regulation in previous years, and degree of informal regulation by local communities. The results of the first-stage equations are, of course, interesting in themselves. They also enable us to construct instrumented versions of the management indices for the second-stage equation. The dependent variable in this equation is a measure of compliance with Mexican environmental regulations. Righthand variables include the four instrumented policy indices, a measure of general employee 2 education, and proxies for pollution abatement costs (sector, scale and technology vintage). We believe that this two-stage analysis can provide useful information for both public and private managers in developing countries. For private managers, it can help reduce uncertainty about appropriate strategies for improving environmental performance. For public managers, it can help identify the most important sources of leverage to reduce excessive pollution. By identifying effective internal management options, it may also suggest new possibilities for cost-effective intervention. Economists have traditionally argued that regulators should use market-based instruments whenever possible, leaving industry to determine appropriate ways of adjusting to stricter regulation. We support this proposition, but our results suggest a possible extension: By increasing the elasticity of firms' response to new incentives, appropriate technical assistance for improved internal management may be a cost-effective adjunct to market-based regulation. The rest of the paper is organized as follows. Section 2 introduces Mexico's pollution control system, and Section 3 describes the survey which has been used for our analysis. The analytical approach and estimating equations are developed in Section 4; results are discussed in Section 5; and Section 6 explores the implications with a simulation exercise. Section 7 concludes the paper. 2. Industrial Pollution and Regulation in Mexico During the past forty years, Mexico has built one of the largest industrial economies in the developing world. Until recently, however, environmental regulation has received scant attention. The result for Mexico City has been air pollution which is among the worst in the world, and other cities are also suffering from excessive pollution (Hettige, et. al., 1996). Table 1 provides recent emissions estimates for Mexico's urban regions. At the national level, an institutional response to this problem began in the late 1980's and has accelerated during the past few years. In its first phase of development, the national regulatory system has had two prominent characteristics: an emphasis on command-and- control regulation, and a multi-institutional approach to monitoring and enforcement. Under command-and-control regulation, the environmental performance of polluters is evaluated according to compliance with numerous licenses and permits issued to each plant. 3 Table 1: Pollution from Industrial Sources in Urban Mexico, 1994 (tons/year) Pollution Sources TSPa SO2" Coe NOXd HCe Total Total 6,345 143,024 1,161,231 203,142 313,060 145,508 1,965,965 Urban Area == _ Tula-Vito-Apaxco 13 21,503 339,763 2,787 66,270 12,540 442,863 Manzanillo, Col. 11 18,881 207,045 2,292 52,296 414 280,928 Toluca-Lerma, Edo. MexL 67 13,704 203,170 2,006 42,779 1,274 262,933 Salamanca, Gto. 49 11,391 117,675 =3,565 21,930 4,073 158,634 Tijuana, B.C. 114 3,053 11,749 117,552 6,032 231 138,617 Coatzacoalcos-Minatitlan 78 9,944 5,818 25,053 35,125 58,479 134,419 Tampico-Altamira 14 5,876 66,323 824 8,876 30,660 112,559 ZMVM 4,623 6,358 26,051 8,693 31,520 33,099 105,721 Torre6n, Coah. 94 4,585 59,092 8,052 639 178 72,546 Monterrey, N.L. 85 9,724 22,360 2,164 8,375 115 42,738 La Paz, B.C.S. 25 3,038 31,128 =267 6,007 45 40,485 Merida, Yuc. 42 2,497 28,932 1,057 7,806 79 40,371 Cd. Juairez, Chih. 135 4,968 5,751 17,199 9,160 1,195 38,273 Guadalajara, Jal. 423 15,045 10,634 1,624 3,184 49 30,536 Other cities 572, 12,457 25,740 10,007 13,061 3,077 64,342 a b c d e Total Suspended Particulates; Sulfur Dioxide; Carbon Monoxide; Nitrogen Oxides; Hydrocarbons Source: SEMARNAP The second prominent characteristic of Mexican environmental management has been multi-institutional administration. Different aspects of industrial pollution control have been the responsibility of different agencies within the Secretaria del Medio Ambiente, Recursos Naturalesy Pesca (SEMARNAP). The principal units are the Instituto Nacional de Ecologia (INE), responsible for monitoring air and toxic emissions; Comisi6n Nacional del Agua (CNA), with similar responsibility for water emissions; and Procuraduria Federal de Protecci6n al Ambiente (PROFE PA), which is charged with factory inspections and enforcement of all pollution regulations. During the 1990's, PROFEPA has expanded its activities from a few inspections per year to several thousand. Mexican environmental policy continues to evolve rapidly, reflecting rising consciousness of pollution problems and Mexico's higher international profile (e.g., negotiations to enter NAFTA and the OECD). In April 1997, President Zedillo introduced a new approach to pollution control in Mexico, based in part on rapid development and implementation of the Sistema Integrado de Regulaci6n Directa y Gesti6n Ambiental de la Industria (SIRG). The SIRG is intended to promote Mexico's interest in bothL international competitiveness and strong environmental protection. It will integrate regulation of emissions to air, ground and water; stress cost-effective regulatory instruments; permit much broader public access to environmental information; and operate from an integrated information system which will employ all the data resources of SEMARNAP. 4 For the purposes of this paper, two aspects of Mexico's recent pollution control experience are most important. First, the system is quite new. Many Mexican factories have not yet been inspected, although PROFEPA's activities are widely known. Second, business managers in Mexico now acknowledge that they will have to respond to stricter environmental standards. Our survey of Mexican industry therefore provides a snapshot of an industrial system in transition. Faced with credible evidence of the government's intent to regulate more strictly, Mexican firms have begun experimenting with new approaches to environmental management and training. 3. The National Survey of Industrial Polluters The data used for this paper were produced by a large survey of Mexican manufacturers carried out in the fall of 1995. The survey focused on four sectors which are estimated to generate between 75% and 95% of Mexico's total industrial pollution: Food, chemicals, non-metallic minerals, and metals. Confidential, in-depth interviews were conducted at 236 facilities, chosen to represent Mexican factories in a set of categories defined by sector, size class, and location. Tables 2 and 3 provide information on the distribution of plants surveyed. Summary information on other variables can be found in Appendix 1. Table 2: Sectoral and Size-Class Distribution of Plants Surveyed Sizea/Sector Food Chemicals Non-Metallic Metals Minerals Large 21 18 12 20 Medium 22 21 12 18 Small 19 23 27 23 a Size classes are defined by employment ranges as follows: Small (16 -100 employees); Medium (100 - 250 ); Large (250 +) Table 3: Regional and Sectoral Distribution of Plants Surveyed Regiona/Sector Food Chemicals Non-Metallic Metals Minerals Medium City 21 20 18 21 Industry 19 20 9 21 Corridor b I III Large City 22 22 24 19 a An industrial area which extends between two urban regions The survey was designed by a World Bank team which included the authors. It was conducted by the Monterrey Institute of Technology (MT), with the explicit support of Mexico's National Environment Ministry (SEMARNAP) and the Mexican National Association of Industries. To minimize reporting bias, the survey was conducted only after agreement by all sponsors (the Government, the Bank and the Industries Association) that 5 the MIT team would not reveal the identity of the plants surveyed or the respondents within the plants. The survey provides very detailed information about environmental performance and its determinants, including plant, firm and market characteristics; sources of environmental information; the quality and costliness of relationships with regulators; the cost of pollution control; and the measures taken by plant management to improve performance. Our analysis relies solely on self-reported environmental performance, since we had no access to independently-audited data on pollution and regulatory compliance. Is self- assessed performance a credible measure? Useful evidence is provided by Table 4, which summarizes the respondents' overall assessment of their facilities' compliance with Mexican regulations. Table 4: Self-Assessed Environmental Performance Environmental Performance Number % of of Plants Total Excellent: Far more than necessary for compliance 23 10 Good: Almost always in compliance 84 38 Fair: Occasionally compliant 99 44 Poor: Never in compliance 10 4 Very Poor: Far below compliance; very damaging 8 4 With confidentiality reasonably well-assured, 52% of survey respondents replied that their plants were not in compliance with regulations. Only 1 0%/o rated their facilities as Excellent, and approximately the same number as Poor or Very Poor. We have no benchmarks, so we can only make an educated guess about the degree of upward bias in this self-assessment. Suggestive evidence is provided by recent research on Indonesia, a country with similar regulations but much weaker monitoring and enforcement. Independent auditing of a large sample of Indonesian plants has shown that approximately 64% are non-compliant (Afsah, et. al., 1996), with 3% in the 'extremely damaging' category. This evidence suggests that the degree of upward bias in the Mexican self- assessment may not be large. In any case, our analysis focuses on relative, not absolute, performance. Independent assessment of conditions in the surveyed plants by the MIT team indicated a high correlation between self-assessment and observable conditions. 6 4. Modeling Environmental Performance In the short run, plant managers typically respond to stricter regulation with some investment in end-of-pipe abatement equipment. They also begin a long-run process of experimentation and adaptation, in which products and processes are re-assessed in the search for new optimum arrangements. Formally, this response can be captured by extending the standard KLEM production frontier into a new dimension: 'use of environmental services,' or pollution (P). In the KLEMP model, stricter regulation imposes a non-zero pollution price (either implicitly, through enforcement of standards, or explicitly, through the collection of pollution charges or imposition of tradable permit systems). Faced with a new set of relative input prices, profit-maximizing firms reveal a new set of input demand equations, including one for environmental demand.3 The environmental demand function differs from the others in one important respect: The incremental cost of pollution to the firm is an expected value, not a set price. It is influenced by many variables, including stochastic enforcement actions, variable financial penalties, reputational damage, and the uncertain cost of extensive negotiations with regulators, local communities, and other environmental stakeholders. Plants and firms with different characteristics will assess enforcement probabilities and expected costs in different ways. Although much empirical research remains to be done, economic logic does suggest some hypotheses about the determinants of plant and firm responses. Determinants of Environmental Effort Scale: Improving environmental performance requires some fixed investmenit and the application of specialized skills. Large facilities should therefore find it less costly at the margin to undertake performance-improving measures. At a higher organizational level, the same argument should hold true for multi-plant firms: Specialized resources can be spread over more units, allowing for more ambitious environmental initiatives. Scale should also work through the 'visibility effect' (see Pargal and Wheeler, 1996): Larger polluters are more detectable by surrounding communities, and may well be under stronger pressure to abate. Thus, scale may increase the incentive to abate as well as the elasticity of response. Both factors should push factories toward more pollution control effort. Public Information: Recent research in North America, Latin America and Asia (Laplante, et. al., 1994, 1997) has shown that environmental performance affects stock prices in both developed and developing countries. Publicly-traded firms whose good performance is publicized receive premia in the market; bad performers are discounted. Since this provides an additional incentive to control pollution, we would expect publicly- traded firms to have stronger environmental management than their privately-held counterparts. 3 For an extended discussion see Pargal and Wheeler (1996). 7 OECD Linkages: It is plausible to suppose that industrial facilities will have stronger environmental management if they are linked to the OECD economies through ownership, trade, and professional relationships. To minimize internal transactions costs, multinationals may use OECD-standard environmental management practices in developing countries. The rapid growth of green labeling in North America and Western Europe may pressure developing-country exporters toward greater environmental effort. Local managers trained in OECD business or technical schools, or with OECD operations experience, may adopt practices which reflect the environmental standards of their host countries. Education: More highly-educated workers and managers should respond more efficiently to new incentives, so we would expect more environmental effort from factories with greater human resources. Regulation: Since 1990, PROFEPA has rapidly expanded its inspection and enforcement program. A significant proportion of Mexican factories havre now been inspected at least once. In addition, local authorities have inspection and enforcement programs in some urban areas. Recent research for the U.S. (Magat and Viscusi, 1990) and Canada (Laplante and Rilstone, 1996) has suggested that inspections have a significant impact on environmental management, even in cases where factories are in formal compliance with regulations. If the regulatory system is working in Mexico, we would expect to observe the same impact. Technology vintage: More recent vintages (particularly imported equipment) should reflect stricter environmental regulation by incorporating environment-saving innovations. In addition, end-of-pipe pollution abatement equipment is generally less costly to install when it is combined with more recent technologies. Since better environmental performance should cost less in newer plants, we might expect to observe more environmental effort in such facilities. Sector: Industry sectors vary significantly in pollution intensity (pollution per unit of output) because of differences in abatement costs (Dasgupta, et. al., 1996; Hartman, et. al., 1997). Faced with identical expected penalties for pollution, we would expect sectors with lower abatement costs to exhibit greater environmental effort. Management Options and Environmental Performance Since Mexico has only recently entered the environmental era, it does not seem realistic to assume that Mexican firms are fully-informed about the profitability of alternative strategies for pollution control. Under these conditions, we believe that new information can be generated by studying the impact of management on environmental performance. We use the survey information to define two indices of management effort in this context: adoption of ISO 14000-type environmental management procedures; and expanded use of personnel for environmental inspection and control. To study the effect of alternative strategies, we also define two indices of 'mainstreaming:' assignment of environmental 8 responsibilities to general managers, rather than specialized environmental managers; and general environmental training for plant employees, as opposed to training focused on environmental specialists. The rationale for one of our effort indices seems clear: Expanded use of personnel for environmental inspection and control reflects the standard logic of resource reallocation when relative prices change. However, the other indices require more explanation. ISO 14000 Sequence: Why should a 'workbook' approach to environmental management yield better performance? In general, workbook exercises will improve overall understanding and performance only if they induce appropriate learning. In the case of ISO 14000, the basic 'exercises' are defined as follows:4 1. An initial review by management to identify environmental issues of concern (e.g. excessive use of polluting inputs; the potential for a serious environmental accident); 2. Establishment of priorities for action, taking into account factors such as local environmental regulations and potential costs; 3. Establishment of an environmental policy statement, to be signed by the CEO; 4. Development of performance targets based on the policy statement (e.g. reduction of heavy metals emissions by 50% by some future date); 5. Implementation of an environmental management system as part of reaching the performance targets; 6. Implementation review; performance measurement. Following the ISO 14000 sequence will not, of course, guarantee any improvement in environmental performance. However, it seems likely that plants which complete these steps will be informed, organized, and motivated in ways which distinguish them from other facilities. As Table 5 shows, the factories in our sample vary widely in their adoption of such practices. 4 See Appendix 2 for a more detailed discussion. 9 Table 5: Adoption of ISO 14000 Procedures Adoption Number Percent Score of Plants S 25 111 47.0 25 < S < 50 45 19.1 50 < S < 75 38 16.1 75s