INTEGRATING ENVIRONMENTAL ACTIVITIES INTO A GENERAL EQUILIBRIUM FRAMEWORK:

ENVIRONMENTAL POLICY ANALYSIS 457 et al. 1993 identify polluting production processes and abatement technologies of each process. 7 Although there are diverse CGE models being applied to envi- ronmental policy analysis, environmental CGE modeling is still in its early stages. This is due to two reasons. The first is a relatively incomplete and unsophisticated specification of models of eco- nomic and environmental interactions. In many models, environ- mental externalities are weakly defined under rigid assumptions or exogenously determined, which undermines the effectiveness and comprehensiveness of an environmental CGE model. The second reason is a lack of a well-defined environmental data frame- work which can provide a solid basis for numerical specification of an environmental CGE model. In addition, most of the existing environmental CGE models are still stylized and ad hoc in nature. Even though a few models are applied to environmental issues in the real world, these models have mostly been built for developed countries. Few models aim at environmental issues of developing countries.

3. INTEGRATING ENVIRONMENTAL ACTIVITIES INTO A GENERAL EQUILIBRIUM FRAMEWORK:

AN ENVIRONMENTAL CGE MODEL This section summarizes the technical specification of an envi- ronmental CGE model developed by the authors. The model aims at simulating environmental policies and environmental protec- tion programs, such as pollution taxes and environmental stan- dards, in developing countries. 3A. Economy-Environment Interactions: Diagrams As mentioned earlier, the classical CGE approach assumes profit-maximizing producers and utility-maximizing consumers. 7 Due to space limitation, we cannot cover environmental CGE models thoroughly. Please see Xie 1995 for a detailed discussion. Other recent attempts to improve environ- mental CGE modeling are mentioned briefly as follows. Whalley and Wigle use a multi- country CGE model to examine international trade and global environmental questions. Espinosa and Smith 1994 report a CGE model of the world for measuring the environmen- tal consequences of international trade policy. Their model uses existing non-market valuation estimates within a specification for consumer preferences and have feedback effects on market demands. Noticing the importance of a lack of property right in environ- mental degradation of many developing countries, Devarajan 1993 suggests possible ways of incorporating the misuse of property right into environmental CGE models for developing countries. Persson 1994 formulates the situation of undefined and well-defined property rights in the CGE model of deforestation in Costa Rica. 458 J. Xie and S. Saltzman Figure 1. The pollution-production interactions. Consumers’ and producers’ optimal behaviors are affected, in one way or another, by the effects of pollution emissions on production and consumption and the implementation of pollution control policies. On the production side of the CGE model, each producer deter- mines his optimal output level by minimizing the costs of his inputs and maximizing the profits of his outputs. When pollution occurs in production processes and certain pollution emissions taxation, are required, the profit maximization problem of the producer is subject to changes. The producer will adjust his output level based on new costs and new production functions containing pollution effects. Figure 1 illustrates the impacts of pollution on production. First, Figure 1 shows that the producer’s total cost includes not only the costs of factor inputs but also pollution related costs due to environmental protection requirements. There are two types of pollution control costs specified in the figure. One is the pollution emission taxes and another the costs of removing pollution in order to comply with environmental standards. Here, no changes in production technology under pollution control requirements are assumed. Second, pollution, in many cases, affects productivity directly. Pollution emissions degrade environmental quality, which then affects the quality and quantity of production factors: fixed capital, labor, and land. The degradation of fixed capital and labor causes productivity to decrease. How about the consumption side of the model? Without pollu- tion effects, household demands for two goods, C 1 and C 2 , in the CGE model, can be derived from the utility maximization problem of the representative household. ENVIRONMENTAL POLICY ANALYSIS 459 Figure 2. The pollution-consumption interactions. However, when pollution occurs, it influences the household’s decisions on consumption. The interactions among pollution, pol- lution control and consumption are described in Figure 2. Figure 2 shows two types of possible impacts of pollution on the consumption side of an economy: the decrease of utility and the changes in household disposable income. The household suf- fers from environmental degradation caused by consumption and production pollution emissions. Therefore, the more pollution the household “consumes,” the less utility it has. Two possible effects of pollution control activities on household disposable income are also identified in Figure 2. The first is the payment of the household for disposing of its waste. The payments for trash tax and motor vehicle waste gas emission tax are two common types of this kind. The second is a nominal increase of household through polluters’ compensating the environmental damage to the household. For example, a farmer may get compensation from a factory that discharges waste water into his farmland. To capture the influence of pollution on consumption, the util- ity-maximizing problem of household consumption needs to be altered in the following three ways. First, the utility function needs to reflect the pollution effects. To model the negative impacts of pollution emissions on utility, the utility function needs to include environmental quality or pollution emissions. Second, household expenditure on its waste disposal should be subtracted from its income. Third, the compensation for environmental damage to the household, if any, should be added to household income. 460 J. Xie and S. Saltzman Figure 3. A pollution abatement sector. To eliminate pollution as a result of satisfying pollution emission requirements, pollution abatement is required in many economies. Figure 3 depicts a representative pollution abatement sector. Different from a production sector, the output of a pollution abatement sector A g is the value of the pollution cleanup. Pollu- tion cleanup can be viewed as special goods which are purchased at a certain price by polluters in order to reduce their pollution emission levels. The optimal output level of a pollution abatement sector can be determined in the way same as that of a production sector. The price of a pollution abatement service is assumed to be determined implicitly by the market. Such an assumption is likely true in a situation where pollution is treated in profit-seeking waste treatment plants, such as a private-owned waste water treat- ment plants. However, this assumption is unrealistic when pollu- tion abatement facilities or sectors are affiliated with an individual factory, because the in-built pollution abatement sectors or facili- ties will not likely pursue a maximum profit strategy on their own. When this is the case, we can fix the price of a pollution abatement service at the level of the average cost of the pollution abatement activity. 3B. General Features of the Model After discussing the representation of the environment– economy interactions in diagrams, the technical specification of the multi-sector environmental CGE model developed by the authors are presented as follows. The economic part of the model is similar to other applications of CGE models to developing ENVIRONMENTAL POLICY ANALYSIS 461 countries. 8 Specifically, it is an adapted version of the Cameroon models developed by Condon et al. 1986 and Devarajan et al. 1991. 9 However, the model presented here is unique in the way it integrates various pollution control activities with economic activities in a CGE framework. The environmental part of this model includes mainly: 1 pollution abatement activities and pol- lution abatement costs or payments of production sectors; 2 pollution taxes, such as production pollution emission taxes and household waste disposal taxes; 3 pollution control subsidies; 4 environmental compensation; 5 separately accounted envi- ronmental investment; and, 6 various pollution indicators includ- ing pollution cleanup ratios and the levels of pollution abatement and emissions. The model can be characterized as an integrated economic and environmental model in the line of the CGE ap- proach. The model assumes an economy having n production sectors and a representative household group. There are m types of pollutants generated from production and consumption processes and m pollution abatement sectors, each of which treats only one unique type of pollution. Production and pollution abatement sectors need intermediate inputs and two primary factors: capital and labor. Government and trades with the rest of the world are included in the model. The model is static with aggregate factor supplies exogenously determined. Following closely the notation convention used in Devarajan, Lewis and Robinson’s model, the model presented here adopts 1 upper case letters for endogenous variables, 2 Greek letters or lower case letters for parameters, and 3 upper case with bar or lower case for exogenous or control variables. The indices for sets used in the model are: ip and jp 5 1, 2, . . . n, representing production sectors; g and ia 5 1,2, . . . m, representing pollutants and pollution abatement sectors; and, i and j 5 1,2, . . . , n, n 1 1, . . . , n 1 m, containing both production and pollution abatement sectors. They appear as lower case subscripts. The model has over 50 equations or equation groups. They are divided into eight equation blocks. They are presented in Table 8 These kinds of CGE models for developing countries were pioneered by, among others, Adelman and Robinson 1978, Taylor 1979, and Dervis et al. 1982. 9 For convenience, the notation adopted in the economic equations of the model is close to that used in Devarajan et al.’s 1994 model. The layout and description of the equation blocks are also influenced by their work. 462 J. Xie and S. Saltzman 1. A list of variables and parameters with units used in the case study for China is also enclosed in the table. A brief description of the model by equation blocks is provided below. 10 3B-1. Price Block. There are a total of ten different groups of average prices in the model. They are composite good prices P, domestic good prices PD, capital input prices PK, the domestic process of imports and exports PM and PE, the prices of intermediate inputs PN, value-added prices PVA, the world prices of imports and exports PWM and PWE, and output prices PX. The relationships among these prices are sketched in Figure 4, and the equations defining prices in the model are presented in block 1 of Table 1. All of these equations in the price block, except for Equation 5, are standard in the literature of CGE models. Equation 5 shows the composition of production cost. The left-hand side of the equation is the total cash inflow of a sector, i.e., the product sale income plus the receipt of government subsidy. It is decomposed on the right-hand side into value-added prices PVA, indirect taxes tc, spending on intermediate inputs based on the fixed input–output IO coefficients a jp ,i , pollution emissions taxes PETAX g ,i and pollution abatement costs PACOST ia,i . PE- TAX g ,i and PACOST ia ,i are affected by pollution intensities, pollu- tion cleanup rates and prices. They are defined later in Equations 36 and 37 in pollution equation block. The trade-related prices in the non-tradable sectors are set to zero, so are the prices for trade-related variables in the pollution abatement sectors because no pollution abatement services are assumed to be tradable. Finally, notice that the prices of pollution abatement services PX i and P i iPia are defined in this block in the same way as product prices. 3B-2. Output and Factor Inputs. For simplicity, production is modeled by a Cobb–Douglas production function of two primary factors capital and labor in Equation 8. 11 The model assumes that a firm maximizes its profits by hiring capital and labor until 10 Again, space limitations preclude a more detailed discussion of the full model, see Xie 1995 in detail. 11 Other possible production specifications usually used in CGE models are single stage or nested CES production functions with the elasticity of substitution estimated separately. ENVIRONMENTAL POLICY ANALYSIS 463 Table 1: An Environmental Computable General-Equilibrium Model The model contains n production sectors, of which s sectors export products and t sectors import goods; m types of pollution and m corresponding pollution-abatement sectors; two primary factors: capital K and labor L, and intermediate inputs; one household group h; government G; and, the rest of the world ROW. The following environmental components are incorporated into the model costs or payments of production sectors for pollution cleaning services, pollution abatement activities, pollution emission taxes of production sectors, household waste disposal taxes, environmental compensation by production sectors to household, and separately accounted environmental investment. Sets ip, jp 5 1,2, . . . n. production sectors g, ia 5 1,2, . . . m. pollutants or pollution abatement sectors i, j 5 1,2, . . . n,n 1 1 . . . n 1 m. production and pollution abatement sectors ie 5 1,2, . . . t. sectors with exports im 5 1,2, . . . s. sectors with imports ic 5 1,2, . . . k. sectors providing goods for final consumption Equations Prices: 4n 1 m 1 t 1 s 1 1 equations 1 PE i 5 1 1 te i PWE i R i P ie domestic price of export goods 2 PM i 5 1 1 tm i PWM i R i P im domestic price of import goods 3 PX i 5 PE i E i 1 PD i XXD i XD i average price of output 4 P i 5 PM i M i 1 PD i XXD i X i price of composite goods 5 PX i XD i 1 SUB i 5 PVA i XD i 1 PX i XD i tc i 1 XD i S jp a jp,i P jp 1 S g PETAX g ,i 1 S g PACOST g,i activity cost composition 6 PK i 5 S j P j b j,i price for capital input 7 PINDEX 5 GDPVARGDP price index Output and factor and trade: 3n 1 m equations 8 XD i 5 ad i K l- a i L a i Cobb-Douglas production function 99 wldist i WL 5 a PVA i XD i L i wage and labor demand 9 ″ wkdist i WK 5 1-aPVA i XD i K i capital return and capital demand Trade: 2n 1 m 1 t 1 s equations 10 X i 5 ac i {d i M 2 pci i 1 1 2 d i XXD 2 pci i } 2 1pci Armington function 11 M i 5 XXD i {PD i PM i d i 1 2 d i } 11 1 pci i P im import demand 12 XD i 5 at i {g i E pti i 1 1 2 c i XXD pti i } 1pti CET function 13 E i 5 XXD i {PE i PD i 1 2 g i g i } 2 11 2 pti i P ie export supply Income, Tax and Saving: 15 equations 14 Y 5 YH 1 YC total income 15 YH 5 S i wldist i WL L i household labor income Continued 464 J. Xie and S. Saltzman Table 1: Continued 16 YC 5 S i wkdist i WK K i company’s capital return 17 YCTAX 5 YC-DEPR t yc corporate revenue tax 18 YHTAX 5 YC t yh personal income tax 19 INDTAX 5 S i PX i XD i tc i indirect tax 20 TARIFF 5 S im PM im M im tm im tariff 21 GR 5 YCTAX 1 YHTAX 1 TARIFF 1 INDTAX 1 ETAX 1 DTAX 1 DDEBT 1 FDEBT 1 DEFT government revenue 22 ESUB 5 S ie PE ie E ie te ie export subsidy 23 CSUB 5 S ip SUB ip subsidy to production 24 SG 5 GR-HSUB-CSUB-DSUB-ESUB-S i P i GD i government saving 25 SH 5 mps YH1-t yh 1 REMIT 1 DCMP-DTAX-DDEBT household saving 26 SC 5 YC-DEPR-YCTAX company saving 27 DEPR 5 S i d i PK i K i depreciation 28 SAVING 5 SH 1 SC 1 SG 1 DEPR-DEFT total saving Expenditures: 4n 1 m 1 2n equations 29 P i CD i 5 bh i YH1-t yh 1 REMIT-DDEBT 1 DCMP-DTAX1-mps 1 HSUB household consumption 30 GD i 5 bg i GC government consumption 31a INT i 5 S j a i,j XD j i P ip intermediate demand for goods 31b INT i 5 S j PACOST i,j P i i P ia intermediate demand for pollution cleanup 32 DST i 5 dstr i XD i inventory demand 33 PK ip DK ip 5 bK ip INVEST-S i P i DST i -EINV-BSPLUS nominal investment by sector of destination 34 ID ip 5 S jp b ip,jp DK jp investment demand by sector of origin Pollution: 7m 1 n 1 2mn 1 m 1 4 equations 35 PK ia DKE ia 5 bk ia EINV environ. investment by sector of destination 36 IDE ip 5 S ia b ip,ia DKE ia environ. investment demand by sector of origin 37 PETAX g,i 5 tpe g d g,i XD i 1 2 CL g impl g,i production pollution emission tax 38 PACOST g,i 5 PA g d g,i XD i CL g adj g,i pollution abatement cost 39 PA g 5 X0 g TDA0 g P g pollutant abatement price conversion 40 TDA g 5 X g TDA0 g X0 g total pollution abated 41 DA g 5 TDA g -GD g TDA0 g X0 g production pollution abated 42 CL g 5 DA g S i d g,i XD i cleanup rate for production pollution 43 DG g 5 S i d g,i XD i 1 S i dc g,i CD i 1 GD i total pollution generated 44 DE g 5 DG g -TDA g total pollution emitted 45 ETAX 5 S i S g PETAX g,i production pollution emission tax 46 DTAX 5 S g tpd g S ip dc g,ip CD ip consumption waste disposal tax 47 DSUB 5 S ia SUB ia subsidy to pollution abatement 48 DCMP 5 S g w g DE g environmental compensation Market Clearing: n 1 m 1 4 equations 49 X i 5 CD i 1 INT i 1 GD i 1 ID i 1 IDE i 1 DST i commodity equilibrium 509 S i L i 5 LS1-Runemp labor market equilibrium 50 ″ S i K i 5 KS capital market equilibrium 51 S im PM im M im 1 BSPLUS 5 S ie PE ie E ie 1 REMIT 1 FDEBT balance of payment Continued ENVIRONMENTAL POLICY ANALYSIS 465 Table 1: Continued 52 INVEST 5 SAVING saving-investment Social Welfare: 3 equation 53 U 5 P i CD bhi i P R TDA b pi g utility 54 GDPVA 5 S i PVA i XD i 1 INDTAX 1 ETAX 1 TARIFF-ESUB-CSUB- DSUB nominal GDP 55 RGDP 5 S i CD i 1 INT i 1 GD i 1 ID i 1 IDE i 1 DST i 1 S ie E ie -S im M im 1-tm im real GNP Endogenous Variables: BSPLUS surplus in balance of payment curr Y 100 million CD i private consumption of product i 990 Y 100 million CL g cleanup rate of pollutant g produced from production sector i unit less CSUB total subsidy to production sectors curr Y 100 million DA g total amount of production pollutant g being abated 100 million tons DCMP total environmental compensation curr Y 100 million DE g total amount of pollutant g emitted 100 million tons DEPR financial depreciation curr Y 100 million DG g total amount of pollutant g being generated 100 million tons DKE ia environmental investment by sector of destination 990 Y 100 million DK ip investment by sector of destination 990 Y 100 million DST i inventory demand for product i 990 Y 100 million DSUB total subsidy to pollution abatment activities curr Y 100 million DTAX total household waste disposal tax curr Y 100 million E ie export of products in sector i to the rest of the world 990 Y 100 million ESUB total export subsidies curr Y 100 million ETAX total production pollution emission tax curr Y 100 million GD i government consumption of product i 990 Y 100 million GDPVA nominal GDP curr Y 100 million GR government revenue curr Y 100 million IDE ip environmental investment demand by sector of origin 990 Y 100 million ID ip production investment demand by sector of origin 990 Y 100 million INDTAX indirerct tax curr Y 100 million INT i intermediate demand by sector of origin 990 Y 100 million INVEST total investment curr Y 100 million K i capital demand by production sector i 990 Y 100 million L i labor demand by production sector i million workers M im import from the rest of the world to the sector i 990 Y 100 million PACOST g,i pollution abatement costs curr Y 100 million PA g price of pollutant g abated curr yuanton PD i domestic prices unity PE ie domestic price of exports unity PETAX g,i pollution emission taxes curr Y 100 million P i price of composite good i unity PINDEX GDP deflator unity Continued 466 J. Xie and S. Saltzman Table 1: Continued PK i price of a unit of capital installed in secotr i unity PM im domestic price of imports unity PVA i net or value added price unity PX i average output price by sector i unity RGDP real GDP 990 Y 100 million Runemp unemployment rate unit less SAVING total saving curr Y 100 million SC company saving curr Y 100 million SG government saving curr Y 100 million SH household saving curr Y 100 million TARIFF tariff curr Y 100 million TDA g total amount of pollutant g being abated 100 million tons U utility unit less WK average capital return rate unit less XD i domestic output by sector 990 Y 100 million X i composite goods supply 990 Y 100 million XXD i local sales of locally produced goods 990 Y 100 million Y total income curr Y 100 million YC company revenue curr Y 100 million YCTAX transfer of company revenue to government curr Y 100 million YH personal income curr Y 100 million YHTAX personal income tax curr Y 100 million Exogenous or Policy Control Variables: DDEBT domestic debt curr Y 100 million DEFT government deficit curr Y 100 million EINV total investment on pollution abatment activities curr Y 100 million FDEBT government borrowing from abroad curr Y 100 million GC aggregate government spending 990 Y 100 million REMIT net remittances from abroad curr Y 100 million HSUB subsidy to household curr Y 100 million PWE ie world price of export product i unity PWM im world price of import product i unity KS aggregate capital supply 990 Y 100 million LS aggregate labor supply million workers R foreign exchange rate 990 yuan per 990 U.S. dollar SUB i government subsidies to sectors i curr Y 100 million WL economy-wide average wage rate curr Y 100 millionmillion workers mps household saving rate unit less tc i indirect tax rate on sector i unit less te ie export subsidy rate on sector i unit less tm im tariff rate on sector i unit less tpd g tax rate of household waste disposal curr yuanton tpe g tax rate of production pollution emission curr yuanton t yc of corporate revenue transferred to government unit less t yh personal income tax rate unit less w g environ. compensation of a unit of pollutant emitted curr yuanton Continued ENVIRONMENTAL POLICY ANALYSIS 467 Table 1: Continued Parameters: rc i Armington function exponent unit less rt i CET function exponent unit less d i Armington function share parameter unit less g i CET function share parameter unit less h i export demand elasticity unit less a i labor share parameter in a Cobb-Douglas production function unit less ac i Armington function shift parameter unit less ad i Cobb-Douglas production function shift parameter unit less a ij input-output technical coefficient 990 yuan990 yuan at i CET function shift parameter unit less b ij capital coefficient derived from capital composition matrix unit less, S i b ij 5 1 d g,i production pollution intensity ton990 yuan dc g,i consumption pollution intensity ton990 yuan d i capital depreciation rate for sector i unit less dstr i share of sectoral production for inventory demand unit less impl g,i adjustment factor for the implementation of production emission tax unit less adj g,i adjustment factor for pollution abatement payment in sector i unit less bg i expenditure share of government spendings unit less, S i b gi 5 1 bh i expenditure share of household spendings unit less, S i bh i 5 1 bk ia share of environmental investment by sector of destination unit less, S ia bk ia 5 1 bk ip share of investment by sector of destination unit less, S ip bk ip 5 1 bp g exponent of total pollution abatement in utility function unit less wldist i sector-specific parameter for wage of labor in sector i unit less wkdist i sector-specific parameter for capital return rate in sector i unit less X0 g pollution abatement output in the base year 990 Y 100 million TDA0 g total level of pollution abatement in the base year 100 million tons Note: Units are used for the case study of China; and Y stands for yuan, basic unit of Chinese money. One U.S. dollar was equal to 3.78 yuan in 1990. the factor price equals the marginal product revenue. The demands for capital and labor are determined by Equations 99 and 9 ″ . 3B-3. Trade Block. Table 1 lists the equations related to trade. Domestic goods supplied to the domestic market and the goods for trades are assumed to be differentiated and imperfect substitut- able in the model. To handle the imperfect substitution, CES- type functions are often employed in CGE models. Equation 10 aggregates imports and domestic sales into a composite good using a CES function. Similarly, sectoral output XD is defined in Equa- tion 12 as a constant elasticity of transformation CET function combining exports and domestic sales of output. 468 J. Xie and S. Saltzman Figure 4. The price system. The import demand M shown in Equation 11 is derived from minimizing the costs of using composite goods given in Equation 4 subject to the CES function of Equation 10. Equation 11 indicates that, an increase in the price of imports relative to the price of domestic sales will lead to a decrease in import demands. Equation 13 defines export supply. As for non-tradable sectors as well as pollution abatement sec- tors, the corresponding trade terms are set to zero, and the sectoral outputs and the sales of the composite goods equal domestically produced goods. 3B-4. Income, Tax, and Saving Block. Income, taxes, and sav- ings of households, enterprises and government are defined by Equations 14 to 28 in Table 1. These equations map the value added to incomes, to taxes, and to savings. 3B-5. Demand Block. The demand for commodities can be divided into household consumption demand, government con- sumption demand, intermediate inputs, investment demand and inventory. They are depicted by Equations 29 to 34 in Table 1. 3B-6. Pollution Block. Although the prices and quantities of pollution abatement services have already been determined in the price and output blocks blocks 1 and 2, the variables related to pollution taxes and pollution emissions still need to be calculated. The pollution block in Table 1 lists the equations defining the variables related to pollution and pollution control activities. ENVIRONMENTAL POLICY ANALYSIS 469 Among these pollution equations, Equations 37 and 38 define the pollution emission taxes PETAX g ,i and pollution-abatement costs PACOST g ,i , respectively. Equation 37 indicates that PE- TAX g ,i is a function of sectoral outputs XD i , pollution emission tax rates tpe g , pollution intensities d g ,i , and pollution cleanup rates CL g . While this equation is calibrated to base year observa- tions, the initial data rarely fit into this equation. This is usually due to the difficulty in collecting pollution taxes and measurement errors. There is frequently a discrepancy between the planned pollution emission tax and the actual tax collection. The discrep- ancy is often large in developing countries. For instance, the total amount of pollution emission taxes collected in China is only about half of the amount of taxes that the Chinese government should collect based on pollution discharge fees standards. To reflect the implementation difficulty, we need to introduce an adjustment factor impl g ,i into the equation. The unit-less adjust- ment factors can be estimated by calibrating Equation 37 to base year data. The sectorally specific factor can also take into account the differentiation of pollution cleanup rates across sectors, which is otherwise ignored by using the economy-wide average cleanup rate CL g in this model. Equation 38 indicates that pollution-abatement costs PA- COST g ,i by sectors and by pollutants are associated with sectoral outputs XD i , pollution intensities d g ,i , pollution cleaning rates CL g , and the prices of pollution-abatement services PA g . The abatement cost of a production sector is obtained from the amount of pollutants abated, i.e., d g ,i XD i CL g , times the price of pollution cleanup PA g . The units of both sides of the equation are dollar. 3B-7. Market Clearing and Model Closure. Under equilib- rium requirements, the demand and supply for goods must be equal to each other as is the demand and supply for factors. The market clearing conditions are given in market clearing block in Table 1. Because the CGE model is based on the general equilibrium condition, the equilibrium equations are not all independent. Ac- cording to Walras’ Law, one of equilibrium equations can be dropped without any effect on the simulation results of the model. The saving–investment equation is actually dropped from the model due to this reason. 3B-8. Social Welfare and GDP. The model adopts a Cobb– Douglas utility function to measure social welfare U since the 470 J. Xie and S. Saltzman economy has only one representative household. To reflect the effects of pollution cleaning on social welfare, social welfare func- tion shown in Equation 54 includes the level of pollution abate- ment TDA g for g 5 1, 2, . . . , m. Pollution-abatement services are assumed to be public goods which are purchased by production sectors under environmental laws or regulations andor the government. Households have no demand for pollution cleaning services. 12 Finally, nominal GDP GDPVA and real GDP RGDP are defined in Equations 54 and 55. They are used in Equation 7 to determine the price index.

4. NUMERICAL SPECIFICATION OF THE MODEL