POLICY RESEARCH WORKING PAPER 1761 In C-nina. evirn-nnienta Bending the Rules In regulators play by [he rules, bul often bend them n w,vav, Discretionary Pollution Control that reflectimportant in China environmental annd scial concerns Regulators giv'e little or no slack to heavu Susmita Dasgupta dischargers Old fac[ories pay Mainul Huq more. state-cxvned facturie, David Wheeler pay' higrier rates, anJ big employers get a discujnt The World Bank Policy Research Department Environment, Infrastructure, and Agriculture Division May 1997 I POLICY RESEARCH WORKING PAPER 1761 Summary findings Industry compliance with pollution regulations is far regulation, the economics of compliance, and regulatory from universal, even in North America. In developing discretion. They find: countries, compliance rates are often quite low, * Cost-sensitive plants will try to adjust emissions to particularly where budgets for regulation are low or the point where the marginal levy equals the marginal inspectors are corrupt. cost of abatement. And strictness of enforcement varies. Regulators are * In practice, local regulators have considerable reluctant to impose stiff penalties on financially strapped discretion in judging both compliance and appropriate plants that are major employers, and in many developing penalties for noncompliance. China's regulators play by countries state-owned plants are treated more leniently the rules, but often bend them. Underreporting and than their private-sector counterparts. underassessment are common in China. But variable But research on determinants of compliance and regulation is systematic, not random, and seems to reflect enforcement is rare, even in industrial societies. important environmental and social concerns. Old Dasgupta, Huq, and Wheeler use new plant-level data factories pay more, state-owned factories pay higher for China to analyze variations in both compliance and rates, and big employers get a discount. And regulators enforcement, with a focus on regulation of water give little or no slack to heavy dischargers. pollution. They look at the mechanics of official This paper - a product of the Environment, Infrastructure, and Agriculture Division, Policy Research Departrment - is part of a larger effort in the department to understand the economics of industrial pollution control 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 the paper are available free from the World Bank, 1818 H Street NW, Washington, DC 20433. Please contact Evelyn de Castro, room N10-019, telephone 202-458-9121, fax 202-522-3230, Internet address edecastro@worldbank.org. May 1997. (24 pages) The Policy Research Working Paper Saies dissemmates the findings of work in progress to encourage the excange of ideas about developmnent 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 names of the autbors 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 viewo of the World Bank, its Executive Directors, or the countries they represent. Produced by the Policy Research Dissemination Center Bending the Rules: Discretionary Pollution Control in China by Susmita Dasgupta* Mainul Huq David Wheeler *Respectively Eenoromist. Consultant and Principal Economist in the Environment, Infrastructuretand Agriculturxe Division, Policy Research Department, World Bank. This paper has been produced in a collaborative program supported by China's National Environmental Protection Agency (NEPA), the Tianjin Environmental' Protection Bureau (TEPB) an'd the World Bank's Country Department EA2. Funding has been provided by the World Baank's Research Support Budget. We are grateful to our colleagues in NEPA and TEPB, without whose assistance this study would not have been possible. Thanks also to C.H. Zhang and Hua Wang for their help and comments. 1. Introduction Industry compliance with pollution regulations is far from universal, even in North America (Magat and Viscusi, 1990; Laplante and Rilstone, 1995; Dion, Lanoie and Laplante, 1996). In developing countries, compliance rates are often quite low (Hettige, Huq, Pargal and Wheeler, 1996). Since budget-constrained regulatory agencies cannot monitor all facilities, some non-compliance is attributable to optimizing behavior: Firms may choose to remain non-compliant if the incremental cost of moving to compliance is greater than the expected loss associated with discovery and payment of penalties. Where inspectors are scarce or the courts lenient, non-compliance may be quite common (Afsah, Laplante and Makarim, 1996). In addition, of course, corruption of inspectors may play a significant role in some countries. Strictness of enforcement can also vary substantially across plants. In the US, for example, numerous press accounts and case studies have identified political pressure as a source of variation in local enforcement of national regulations (Wheeler, 1991). Environmental regulators have proven quite reluctant to impose stiff penalties on financially- strapped plants which are major employers (Deily and Gray, 1991). In many developing countries, state-owIned factories seem to have been treated more leniently than their private- sector counterparts (CETESB, 1994; Pargal and Wheeler, 1996; Huq, Hartman and Wheeler, 1996). Although anecdotes are plentiful, systematic research on determinants of compliance and enforcement is rare even in industrial societies because the necessary information is seldom provided by regulatory agencies.' To our knowledge, no such studies have been done for developing countries. In this paper, we use new plant-level data provided by China's National Environmental Protection Agency (NEPA) and the Tianjin Environmental Protection Bureau (TEPB) for an analysis of variations in both compliance and enforcement. These data provide a unique opportunity for regulatory analysis in a developing country, because NEPA has operated and documented a country-wide emissions charge system for over ten years. We focus on regulation of water pollution because the appropriate data are more plentiful in the factory sample available to us. The paper is organized as follows. In Section 2 we review China's system for ,enforcing industrial pollution regulations. Section 3 develops models for the analysis of compliance and enforcement, while Section 4 introduces the data. We discuss the econometric results in Section 5 and provide a summary and conclusions in Section 6. 2. China's Pollution Levy In China's regulatory system, emissions which exceed official standards are not treated as legal violations. Rather, Article 18 of China's Environmental Protection Law specifies that "in cases where the discharge of pollutants exceeds the limit set by the state, a compensation fee shall be charged according to the quantities and concentration of the pollutants released." This compensation fee, or levy, has been implemented nationally since 1982. Almost all of China's counties and cities are now operating the levy system, and approximately 300,000 factories have been charged for their emissions (NEPA, 1994). 1 In many countries, such records are protected by law. Even in the U.S., the Environmental Protection Agency has only recently announced plans to publish records of inspections, violations and emissions at the plant level. 2 The water pollution levy is not a true Pigovian charge; it is assessed on emissions which exceed established discharge standards for pollutant concentration in waste water. Chinese discharge standards vary across pollutants, industrial sectors, and "water environment function areas" which distinguish receiving waters by the quality of intended use. They also vary by age of plant, with more lenient standards for facilities established prior to 1979. With NEPA's permission, local areas can impose stricter standards and higher levy rates if they think it appropriate. Pollutant-specific levies are calculated by multiplying three elements together: a unit fee; the volume of waste water discharged; and the ratio of effluent concentration to the standard concentration.2 For plants with multiple pollutants, the maximum concentration ratio is used for levy assessment. Unit fees escalate with discharge volume.3 3. Model Specification 3.1 The Economics of Compliance Non-compliant factories simply have to pay the levy, so pollution control is largely an economic issue for Chinese managers. In the case of a single pollutant, the total levy for the jth non-compliant plant- is given by the formula:4 2 The termsr 'intiuent-i.anaviaulliuen' -^rer to the waste stream before and after end-of-pipe abatement. 3For more discussion of the levy and its impact on pollution, see Florig and Spofford (1994) and Wang and Wheeler (1996). 4 Model variable definitions are summarized in Table 1. 3 (3. 1) Lj=j 'Wj Ils where Lj = Expected total levy payment pj = Levy rate = Effluent concentration =s = Concentration standard Wj = Waste water volume Recent econometric work on factory-level abatement costs in China (Dasgupta, Huq and Wheeler, 1996) suggests the following model for the case of a single pollutant: (3.2) Aj =y l - where O 0 and Aj Total abatement cost g0j = Influent concentration ,u; = Effluent concentration. Zj = A vector of plant characteristics which affect costs (sector, age, scale, ownership, etc.) At the plant level, y, is significantly less than unity (i.e., abatement is subject to scale economies). While total abatement cost rises less-than-proportionately with scale of waste water treatment, marginal abatement cost increases with percent reduction in pollutant concentration from influent to effluent. Total pollution-related cost is therefore given by: (3.3) Cj = Lj + A1j = p.j i±Wj + yowjl{[ { -I l}HZ jm Cost minimization implies choosing an effluent concentration pt such that (3.4) '= 0 j 4 Cost-minimizing managers in regulated private firms and township-village enterprises will reduce pollution to the point where the expected levy rate is equal to the marginal cost of abatement. Managers in state-owned enterprises (SOEs) with hard budget constraints should exhibit similar behavior; their response to pollution charges may be less elastic where soft budget constraints persist. For a plant j which adjusts so as to minimize pollution-related costs, optimal effluent concentration is given by the solution to (3.4): 1 y I-1 Y2 1 1 M pm (3.5 ~ =yoy)12+]W7Y2+1 Y2J+1 Y2+1p'Y2+1 r]7Jz2+1 (3.5) lj=(Y 0G 2) j WJ Foj 1sj PJ lm m=l A cost-minimizing plant will be compliant if .j*.< tij. By conventional reasoning, p * should never be less than [tj because the levy is zero for discharges whose effluent concentration is below the standard. However, a number of factors may lead some plants to perform better than the standard requires. These include the market value of environmental reputation (particularly for large enterprises) and pressure from local communities (Pargal and Wheeler, 1996; Afsah, Laplante and Wheeler, 1996; Wheeler and Afsah, 1996). In equation (3.5), we have the following expectations about the impact of plant-level variables on effluent concentration and compliance: 1. Discharge Volume (WVj): Since waste water treatment is subject to very significant scale economies, optimal effluent concentration will fall as discharge volume increases, and the probability of compliance will increase. 2. Influent Intensity (-ji0): Optimal effluent concentration increases less-than- proportionately with influent concentration (0 < y2/(y2+1) < 1); the probability of compliance will decrease as influent concentration rises. 5 3. Concentration Standard (ppjg): As the standard is tightened, more abatement will be warranted to avoid higher levies. The impact on optimal effluent concentration is inversely related to the abatement cost elasticity (Y2), since managers faced with rapidly- escalating abatement costs will be more willing to pay additional fines. By the same reasoning, the probability of compliance will decrease as 'js increases. 4. Pollution levy rate (pj): Increases in the levy will reduce optimal effluent concentration and increase the probability of compliance. 5. Plant Scale: Abatement, process modification and production are joint activities. In larger plants, the fixed costs associated with engineering skills and other relevant inputs can be distributed across a larger number of activities. This should lower the cost of pollution reduction, through its impact on process modification and end-of-pipe abatement. We therefore expect optimal effluent concentration to fall (and the probability of compliance to rise) with increasing plant scale. We use total employment as our proxy for scale. 6. Age: Pollution control has been increasingly embodied in new process technologies. In addition, installation of end-of-pipe abatement equipment during plant construction is cheaper than retrofitting. With a steady increase in regulatory pressure since 1980, we expect newer plants in China to exhibit better environmental performance. Probability of compliance should therefore be negatively affected by age of plant. 7. Ownership: Even after age and scale are accounted for, we expect public ownership to increase pollution intensity and reduce the probability of compliance. State-owned enterprises (SOEs) are likely to be less efficient, creating more waste residuals and 6 pollution than their private counterparts. Soft budget constraints for many SOEs should also reduce their managers' responsiveness to pollution charges. 3.2 The Political Economy of Regulation Relying on anecdotal evidence, critics of China's pollution control system have asserted that enforcement of the levy system is relatively arbitrary. (Qu, 1991). Personal ties between regulators and plant managers and other forms of favoritism are commonly cited. To our knowledge, however, the sources of variation in regulatory enforcement have not been analyzed systematically. As we noted in the introduction, variation in enforcement may reflect social welfare concerns as well as personal ties. In practice, regulators have considerable discretion in two dimensions of regulation: The formal identification of factories as non-compliant (and therefore subject to the levy); and the strictness of their enforcement response (measured by the effective levy rate which is applied to excess discharges). For analytical purposes, we specify two adjustment equations which relate plant characteristics to officially-recognized effluent concentration and the effective levy rate. M=1 (3.7) li J= a~ oWj-. Wn H Z m-. (3.7) pOj= POWT, z1 - Poi The variables in these-two adjustment equations include: 1. Discharge-Volume (W1): Although Chinese regulations focus on concentration standards, actual pollution damage is also a subject of concern. The levy system 7 recognizes this problem by applying higher rates to large dischargers. It is also likely that Chinese regulators pay closer attention to such plants. We might therefore expect discharge volume to have a. positive effect on the regulators' identification of non- compliance, as well as on the levy rate. 2. Plant Scale: Given the political importance of employment, Chinese regulators may well be more lenient toward facilities which are large employers. 3. Age: Plants constructed prior to 1979 face laxer regulatory standards than newer facilities. Even so, plant managers could be expected to invoke 'grandfathering' arguments when confronted by regulators with evidence of non-compliance. If age has any impact on regulator discretion, we would expect it to be toward lenience in both identification of non-compliance and assessment of the levy. 4. Ownership: In mixed economies, it has often proven difficult for government regulators to punish violations by state-owned enterprises. Political and bureaucratic factors seem to prevent effective supervision of one government agenqcy by another. If China's experience is similar, we would expect laxer enforcement for its SOEs. However, they might well be given extra scrutiny for non-compliance even if enforcement is more difficult. 3.3 Compliance Equation Specification In our treatment of compliance, we distinguish between actual plant-level emissions intensity (p) and officially-recognized intensity (u). Our compliance equation incorporates 8 two sets of factors: economic calculations and regulator discretion. Substituting (3.5) into (3.6) we obtain: 1___ |a _+Y__ Y2 1 am+ O_m (3.8) jLI=aO@O 22 )2 Y2+1 7-2+1 72+1 H Zj.m Information problems have forced us to simplify (3.8) for econometric estimation. Our data base does not include observations on the local standards (IAsj) faced by individual factories. In addition, we have to exclude the pollution levy-rate (pj) because we use a non- zero levy as our indicator of non-compliance.5 To control for these factors, we introduce sector dummy variables (Sin) in (3.10). They proxy the effects of the following composite term in (3.8): 1 (3.9) Psi 2 pj Following (3.9), we expect effluent concentration to be higher in sectors with high concentration standards relative to their unit levy rates. The impact of the standard/levy ratio on effluent concentration and compliance will vary inversely with abatement cost elasticity. After substitution of dummy variables, the effluent concentration equation becomes: I +7-1 IL) M~ ja+-'~ (1 °Y2+1 y ) 2+1 nM Zj,m n 2+I N-1 OnSn (3. 10) =.1 - 0(y 2)2 in=i in fn=i The associated log-log form is: 5Technical issues of probit estimnation are discussed later in this section. 9 (3.1 lo j ={log ao + log(zO72)}+{w+ logWi+ t-g2 M Em ~~~N-1 m-l + m + +l logZmj + XO,nSjn With composite parameters r, this becomes: M N-1 (3.12) log ij =,no +i lwIog W +n, log go + Y-lr7zmllogZj + Xi'Z S. J g i~~ M-tI n=1 Sn? in Regardless of its true compliance status, a factory is judged compliant if ui < ,s. Since we have no observations on jt,, the effluent standard for each plant, we cannot observe officially-recognized.compliance directly.. However, we know that all levy-paying plants are recognized as non-compliant by NEPA and the TEPB. We therefore use our data set to construct a binary dependent variable Cj, whose value is 1 for factories which pay no levy and 0 otherwise. Assuming that log 4a is normally distributed6, the. probability of compliance (log a s log jQ) can be calculated from the cumulative norrnal probability distribution. The parameters of (3.12) can be estimated by probit, with Cj = 1 when the factory is officially compliant and zero. otherwise. As we noted previously, our use of the pollution levy for identifying non-compliant factories excludes the use of factory-specific levy rates on the righthand side of the estimating equation.7 In (3.12), Zmj includes measures of plant age, scale (employment) and state ownership (a dummy variable whose value is one for SOEs). Our expectations about the signs of estimated paraneters in-the.compliance equation are summarized in Table 2. 6 Since plant-level pollution intensities are highly skewed, the log-log specification of (3.12) has the advantage of imposing quasi-nornality in the underlying error distribution. 7 Given the match between left- and righthand zeros, the probit estimator obtains a spurious 'perfect fit' if the levy is included as an explanatory variable. 10 3.4 Enforcement Equation When formula (3.1) is actually applied by Chinese regulators, the effective levy rate is a function of sector, discharge volume, and the adjusted unit levy rate (from (3.7)): N-1 '~%Sj,J'~N-1i (3.13)5j = pLj H.l et W 00po W,'° HI et } JZ,I In=l n=l t- The effective levy is given by: (3.14)L. = p. -W. =P Pwz m ' J is J O0OJp~s = "?-- JM For multiple pollutants, the levy is based on the pollutant with the maximum ratio of effluent concentration to the regulatory standard. We specify the associated estimating equation in log-log form: log L log PP0O +0 log{max }+ {P ++ + 1} logW1 (3.15)s N-1 M + 6 S. + E 7t Z. +£ . n n Jn m=1 m Jm j With composite parameters co this becomes: [ PJ] N-1 logLj =coo + log{max + coW log Wj _ 1nS jn (3.16) M + z. + Z . m=l m JM j where ei is a random error term and Zj includes measures of age, scale and ownership. Since the maximum concentration/standard ratio is part of the levy formula, its inclusion in (3.16) is particularly importunt. We use national sectoral standards as our proxies for values of k5. 11 No levy is paid by factories whose officially-recognized effluent concentrations are all equal to or below the relevant standards. Since this is true for some plants in our sample, the dependent variable is left-censored at zero but takes on a broad range of positive values. We therefore estimate the parameters of (3.16) using tobit. Table 3 summarizes our prior predictions on signs. 12 4. Data For this analysis, NEPA and the TEPB have provided us with 1993 data for 328 factories scattered across China's urban/industrial areas.8 The sample has broad sectoral coverage (Table 4). Not surprisingly, it has very heavy public-sector representation: 291 plants are state-owned enterprises (SOEs), 20 are collectives, and 9 are joint ventures or wholly-owned by foreign firms.9 The sample plants appear to be older and larger than average, but nonetheless exhibit wide variation in age, scale, pollution intensity and abatemeni cost. Years of operation vary from 3 to 93, with the median at 35. The majority of plants were established prior to 1979, and therefore face lower emissions standards then newer facilities. Employment varies from 139 to 37,000; the median plant has 1781 employees. Some plants report significant abatement activity and effluent concentrations below the designated emission standards, while others show no sign of abatement effort, despite extremely high levels of influent concentration. Within the sample, influent COD concentrations can be as high as 100,000 mg/l, while effluent concentrations are as high as 22,800 mg/I.10 Abatement costs also vary substantially: The highest cost incurred by a plant in the sample is RMB 68.75 million, but the median is only RMB 0.3 million. The incidence of levies suggests that the environmental performance of SOEs follows the pattern observed elsewhere in Asia (Pargal and Wheeler, 1996; Hartman, Huq and Wheeler, 1996): 73% of 8 We believe this to be an approximately random sample from NEPA's database of 3000 top polluters. Of course, these plants are not a random sample of Chinese industrial facilities. As a group, they are likely to have higher-than average pollution intensity. Although this may affect average compliance status, we have no reason to believe that it will bias our estimates of incremental relationships. 9In the sample of 328 plants, 320 are identifiable by ownership class. 10 These maximum values contrast with official COD concentration standards in the range of 100-200 mg/l. 13 SOEs pay levies, compared with 52% of non-SOEs. Both the incidence of levy payments and the average levy differ substantially across sectors (Table 4). 5. Econometric Results 5.1 Probit Results: Determinants of Compliance Table 5 presents the full set of probit results for the compliance equation (based on 3.12); variables are successively deleted from the full specification until significant factors remain. Total sample size is constrained by data availability, particularly for the influent intensities (inclusion of the latter reduces the estimation sample size from 276 to 107). The results confirm our prior expectations in cases where the predicted signs are unambiguous (see Table 2): SOEs are significantly less likely to be compliant than plants in other ownership categories; large plants (measured by employment) are significantly more likely to be in compliance. Where prior expectations were ambiguous, our results are mixed. Older plants are significantly less likely to be in compliance, suggesting that the economic impact of age on abatement cost outweighs the effect of laxer regulations and any inclination toward leniency on the part of regulators. On the other hand, the results for discharge volume suggest that large polluters face compliance-related scrutiny which outweighs the abatement cost advantage of scale. Plants with large waste water discharges are significantly less likely to be judged compliant. In the initial regression, we incorporate influent intensity measures for all three major pollutants in the data set (total suspended solids (TSS); chemical oxygen demand (COD); biological oxygen demand (BOD)). There are clearly collinearity problems; when only BOD intensity is included, its estimated parameter has the expected sign but a low significance 14 level. The large standard error makes the point estimate [y2/(Y2+I) -.25] consistent with a wide range of abatement cost elasticities (y2)4 Dummy variables are included for sectors which are heavily represented in the tegression subsample. Their collective insignificance suggests that sectoral standards and'levies vary together [i.e., j./p remains approximately constant in (3.9)]. 5.2 Tobit Results: Determinants of Enforcement Our tobit results for equation (3.16) are reported in Table 6. In this case, heteroskedasticity across observationls can be a source of serious estimation error. Maddala and Nelson (1975) have shown that uncorrected estimates are inconsistent. Although our log- log specification is a common expedient for avoiding heteroskedasticity, Fishe, et. al. (1979) have shown that the log transformation does not solve'the problem in tobit models. We have therefore estimated the final form of our tobit equation with and without a heteroskedasticity correction. The corrected equation is estimated by maximum likelihood,' under the assumption that error variance is a function of output. 'The results are strongly consistent with the existence of heteroskedasticity [t(a) = 8.928] arid confirm the significance of output as the control variable (t(Output) = 2269). However, we find that in this case the parameter estimates in the corrected equation are nearly, identical to those in the uncorrected equation. Deletion of insignificant variables again leaas us to arop the sector dummies from the final equation. Our results for the two variables in the levy formula,conform to prior expectations: The elasticity for the mnaximu'm' concentration/standard ratio is positive and not significantly different from one; the waste water discharge elasticity is positive and significantly greater than one. Among the plant characteristics, the effect of facility scale is also as we expected: Large employers are assessed at much lower rates than other plants. However, our results contain two major surprises: Older firms are assessed at significantly higher rates, and SOEs at far higher rates than other facilities. Although these results are contrary to our expectations, they are consistent with the estimates for the same variables in the compliance equation. In China, state-owned enterprises are apparently subject to more rigorous enforcement than collectives and factories with private shareholders. I This is an important reversal of previous findings for mixed Asian economies (Pargal and Wheeler, 1996; Huq, Hartman and Wheeler, 1996, Hettige, Huq, Pargal and Wheeler, 1996). 6. Summary and Conclusions' In this paper, we have investigated the determinants of plant-level compliance and enforcement in China's water pollution levy system. Our study incorporates three factors: 1. The mechanics of official regulation: A plant is judged non-compliant if its officially- reported waste water discharge contains at least one pollutant whose concentration is above the regulatory standard. For non-compliant plants, the official levy incorporates several factors: A unit levy rate which varies by sector and, in some cases, by locality; a standard (upward) adjustment of the unit rate as discharge volume increases; and multiplication of the adjustment rate by (a) the plant's maximum effluent concentration/standard ratio and (b) the volume of waste water discharge. 2. The economics of compliance: Cost-sensitive plants will attempt to adjust emissions to the point where the marginal levy is equal to the marginal cost of abatement. Representation of private facilities in the subsample is too sparse for meaningful separation of collective and private-sector effects. 16 3. Regulators' discretion: In practice, local regulators have considerable discretion in judging both compliance and appropriate penalties for non-compliance. Our results suggest that all three factors play significant roles in compliance and enforcement. The compliance results highlight the significance of economic factors. As expected, compliance probability is negatively related to state ownership and age, positively related to plant scale. However, our results also suggest that regulators' discretion has a strong effect on outcomes. Abatement economics imply higher rates of compliance for large dischargers, because China regulates effluent concentration and marginal costs are lower in large abatement facilities.12 However, waste stream volume has a large negative impact on reported compliance, suggesting that regulators give little or no slack to large dischargers. Our enforcement results indicate that assessment of the levy is typically consistent with theform dictated by regulatory statutes. The effective levy rate goes up sharply with discharge volume, as mandated, and the levy-elasticity of plant-specific maximum concentration/ standard ratios is not significantly different from one. However, the results also suggest that the substance of levy assessment reflects a large measure of regulator discretion: Old factories pay more, state-owned factories pay higher rates, and big employers get a discount. We conclude that China's regulators play by the rules; but frequently bend them. Compliant factories would be unlikely to accept non-compliant status, so the estimated impact of plant characJ ristics on reported compliance and -enforcement suggests that under- reporting and under-assessment are common in China. In this paper, we have found that 12 See Dasgupta, Huq and Wheeler (1996) for evidence on abatement costs. 17 variable regulation is systezpatic, not random, and that it seems to reflect important environmental 'and social concerms. !8 REFERENCES Afsah, S., B. Laplante and D. Wheeler, "Controlling Industrial Pollution: A New Paradigm," Policy Research Department Working Paper No. 1672, World Bank. Afsah, S., B. Laplante and N. Makarim, 1996, "Program-Based Pollution Control Management: The Indonesian PROKASIH Program," Policy Research Department Working Paper No. 1602, World Bank. CETESB, 1994, Acao de CETESB em Cubatao: Situacao em Junho de 1994, Sao Paulo, Brazil: Companhia de Technologia de Saneamento Ambiental (CETESB). Dasgupta, S., M. Huq, D. Wheeler and C. Zhang, 1996, "Water Pollution Abatement by Chinese Industry: Cost Estimates and Policy Implications," Policy Research Department Working Paper No. 1630, World Bank. Deily, M.E. and W.B. Gray, 1991, "Enforcement of Pollution Regulations in a Declining Industry," Journal of Environmental Economics and Management, 21, pp. 260-274. Dion, C., P. Lanoie and B. Laplante, 1996, "Monitoring of Enviromnental Standards: Do Local Conditions Matter?" Policy Research Department Working Paper No. 1701, World Bank. Fishe, R.P.H., G.S. Maddala, and R.P. Trost, 1979, "Estimation of a Heteroskedastic Tobit Model," Manuscript, University of Florida. Florig, K. and W. Spofford, 1994, "Economic Incentives in China's Environmental Policy," Washington: Resources for the Future, October (mimeo). Hettige M., M. Huq, S. Pargal and D. Wheeler, 1996, "Determinants of Pollution Abatement in Developing Countries: Evidence from South and Southeast Asia," World Development, December. Huq, M., R. Hartman and D.Wheeler, 1996, "Why Paper Mills Clean Up: Determinants of Pollution Abatement in Four Asian Countries," Policy Research Department Working Paper No. 1710, World Bank. Laplante, B. and P. Rilstone, 1995, "Enviromnental Inspections and Emissions of the Pulp and Paper Industry," Policy Research Department Working Paper No. 1447, World Bank. Maddala, G.S. and F. Nelson, 1975, "Switching Regression Models with Exogenous and Endogenous Switching," Proceedings of the American Statistical Association (Business and Economics Section), pp. 423-6. 19 Magat, W. and W.K. Viscusi, 1990, "Effectiveness of the EPA's Regulatory Enforcement: The Case of Industrial Effluent Standards," Journal of Law and Economics, 33, pp. 331-60. NEPA, 1994, The Pollution Levy System, (Beijing: China Environmental Science Press). Pargal, S. and D. Wheeler, 1996, "Informal Regulation in Developing Countries: Evidence from Indonesia," Journal of Political Economy, December. Qu, Geping, 1991, Environmental Management in China, (Beijing: UNEP and China Environmental Science Press). Wang, H. and D. Wheeler, 1996, "Pricing Industrial Pollution in China: An Econometric Analysis of the Levy System," Policy Research Department Working Paper No. 1644, World Bank. Wheeler, D., 1991, "The Economics of Industrial Pollution Control: An International Perspective," World Bank, Industry and Energy Department Working Paper No. 60, January. Wheeler, D.,and S. Afsah, 1996, "Going Public on Pollution: Indonesia's New Public Disclosure Program," East Asian Executive Reports. World Bank, 1992, "China Environmental Strategy Paper," Environment, Human Resources and Urban Development Operations Division, China and Mongolia Department, World Bank, Washington, DC. 20 Table 1: Variable Definitions pj = Levy rate Lj = Expected total levy payment pj = Effective levy rate Lj = Actual total levy payment Wj = Waste water volume Aj = Total abatement cost Cj = Total pollution-related cost p., = Effluent concentration p,j = Officially-recognized effluent concentration p, = Concentration standard p0j = Influent conccntration Zj = A vector of plant characteristics which affect costs (sector, age, scale, ownership, etc.) Table 2: Predicted Signs, Compliance Equation Variable Predicted Regulators' Enterprise Full Sign Discretion Costs Discharge Volume (Wj) - + Influent Concentration (poj) 0 Scale (Employment) (Z,) + + + Age (Z2) + SOE (Z3) - or 0 Sectors (Z4 ... ZM) Variable Variable Variable Table 3: Predicted Signs, Enforcement Equation Variable Predicted Partial Partial Full Sign ,Effects Effects Discharge Volume (Wj) . + (>) Py>o 9>0 Influent Concentrati6n -Ratio (gi'o) + ((=) Scale (Employment) (Z-) - Age (Z2) SOE (Z3) Sectors (Z4 ... ZM) Variable 21 Table 4: Levy Incidence Across Sample Industry Sectors Total Number of Proportion Paying Mean * Sector Plants Levy Levy (10,000 RMB yuan) Food 46 0.80 39.95 Beverages 42 0.69 21.05 Textile 63 0.68 16.94 Leather 25 0.56 10.04 Pulp and Paper 26 0.85 146.36 Power 22 0.73 54.14 Petroleum RefLn.ing 9 0.78 119.57 Chemicals 41 0.78 83.25 Pharmaceuticals 11 1.00 28.44 Plastic 3 1.00 60.71 Cement 14 0.36 11.86 Iron and Steel 7 0.71 7.49 Others 19 0.37 21.68 * Calculated on the basis of levy-paying plants only. 22 Table 5: Probit Estimates (Compliance Equation) Dependent Variable: Compliance Status (1 if Compliant) Variable Descriptions: LDISCHARGE (W): Log (amount of waste water discharged) LTSSINF (koT): Log (TSS concentration in the influent) LCODINF (jtoc): Log (COD concentration in the influent) LBODINF (kiOB): Log (BOD concentration in the influent) LAGE (Zi): Log (age of the plant) SOE (Z2): Dummy variable = 1 if the plant is state owned = 0 otherwise LEMP (Z3): Log (number of employees) Model 1 Model 2 Model 3 Model 4 Coefficient Z Coefficient Z Coefficient Z Coefficien Z ._________________ t Intercept 2.609 1.206 1.481 0.771 1.524 0.804 -0.598 -0.811 LDISCHARGE -0.419** -2.626 -0.427** -3.009 -0.428** -3.043 -0.242* : -3.780 LTSSINF -0.021 -0.387 -0.025 -0.475 LCODINF 0.011 0.199 0.011 0.196 LBODINF -0.258 -1.385 -0.247 -1.462 -0.237 -1.439 LAGE -0.597** -2.427 -0.611** -2.573 -0.611** -2.600 -0.196* -1.660 SOE -1.948** -3.340 -1.769** -3.320 -1.767** -3.387 -0.548** -2.184 LEMP 0.296 1.016 0.445* 1.677 0.432* 1.636 0.308** 2.505 Paper -0.227 -0.294 . Food 0.172 0.255 Textiles 0.023 0.050 Petroleum 1.552 1.401 Cement 0.589 0.455 No. of obs. 107 _ 107 _ 107 276 Chi sq 30.82 28.39 28.03 21.88 Probability 0.002 0.0002 0.00 0.00 * Significant at 10% ** Significant at 5% 23 Table 6: Tobit Estimates Dependent Variable: Log [Effective Levy] Variable Descriptions: LCONCSTD (max -): Log {max (concentration of pollutant i in the effluent/ 's concentration standard for i)} where i= BOD, COD and TSS. LDISCHARGE (W): Log (amount of waste water discharged) LTSSINF (hOT): Log (TSS concentration in the influent) LCODINF (Itoc): Log (COD concentration in the influent) LBODINF (kOB): Log (BOD concentration in the influent) LAGE (Z1): Log (age of the plant) SOE (Z2): Dummy variable = 1 if the plant is state owned = 0 otherwise LEMP (Z3): Log (number of employees) a: Test statistic for heteroskedasticity Output: Value of output in millions of RMB Yuan Model 1 Model 2 | Model 3 Coefficient t Coefficient t Coefficient z Intercept -2.474 -0.413 1.283 0.235 0.754 0.154 LCONCSTD 0.850* 1.717 0.785* 1.659 0.758 0.879 LDISCHARGE 2.621 * * 5.792 2.502** 6.212 2.352** 4.567 LAGE 1.565** 2.005 1.500** 1.987 1.289* 1.681 SOE 7.893** 3.825 8.050** 3.917 7.808** 3.859 LEMP -1.961** -2.264 -2.375** -3.012 -2.060** - _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 2 .3 8 0 Paper -1.123 -0.647 _ Food -0.841 -0.447 Textile 0.762 0.588 Petrol -4.740 -1.419 Cement -2.110 -0.444 a ___________ 4.895** 8.928 Output 0.0003** 2.269 No. of obs. 133 133 133 Chi sq 55.90 52.72 Probability 0.00 0.00 * Significant at 10% ** Significant at 5% 24 Policy Research Working Paper Series Contact Title Author Date for paper WPS1746 The Role of Long-Term Finance: Gerard Caprio, Jr. April 1997 P. Sintim-Aboagye Theory and Evidence Asli Demirg0g-Kunt 38526 WPS1747 Protection and Trade in Services: Bemard Hoekman April 1997 J. Ngaine A Survey Carlos A. Primo Braga 37947 WPS1748 Has Agricultural Trade Liberalization Merlinda D. Ingco April 1997 J. Ngaine Improved Welfare in the Least-Developed 37947 Countries? Yes WPS1 749 Applying Economic Analysis to Gary McMahon April 1997 C. Bernardo Technical Assistance Projects 37699 WPS1750 Regional Integration and Foreign Magnus Bl6mstrom April 1997 J. Ngaine Direct Investment: A Conceptual Ari Kokko 37947 Framework and Three Cases WPS1751 Using Tariff Indices to Evaluate Eric Bond April 1997 J. Ngaine Preferential Trading Arrangements: 37947 An Application to Chile WPS1752 Ghana's Labor Market (1987-92) Sudharshan Canagarajah April 1997 B. Casely-Hayford Saji Thomas 34672 WPS1753 Can Capital Markets Create Paul Lanoie April 1997 R. Yazigi Incentives for Pollution Control? Benoit Laplante 37176 WPS1754 Research on Land Markets in South Rashid Faruqee April 1997 C. Anbiah Asia: What Have We Learned? Kevin Carey 81275 WPS1755 Survey Responses from Women Mari Pangestu April 1997 J. Israel Workers in Indonesia's Textile, Medelina K. Hendytio 85117 Garment, and Footwear Industries WPS1756 World Crude Oil Resources: G. C. Watkins April 1997 J. Jacobson Evidence from Estimating Supply Shane Streifel 33710 Functions for 41 Countries WPS1757 Using Economic Policy to Improve Rashid Faruqee April 1997 C. Anbiah Environmental protection in Pakistan 81275 WPS1758 The Restructuring of Large Firms Simeon Djankov April 1997 F. Hatab in Slovakia Gernard Pohl 35835 WPS1759 Institutional Obstacles to Doing Aymo Brunetti April 1997 M. Geller Business: Region-by-Region Results Gregory Kisunko 31393 from a Worldwide Survey of the Beatrice Weder Private Sector Policy Research Working Paper Series Contact Title Author Date for paper WPS1760 Credibility of Rules and Economic Aymo Brunetti April 1997 M. Geller Growth: Evidence from a Worldwide Gregory Kisunko 31393 Survey of the Private Sector Beatric Weder WPS1761 Bending the Rules: Discretionary Susmita Dasgupta May 1997 E. de Castro Pollution Control in China Mainul Huq 89121 David Wheeler