Is There a Trade-off Between Trade Liberalization and Pollution Abatement?

Is There a Trade-off Between Trade Liberalization and Pollution Abatement? A Computable General Equilibrium Assessment Applied to Costa Rica Se´bastien Dessus, OECD Development Centre, Paris, France Maurizio Bussolo, Warwick University, Coventry, United Kingdom Recent literature elucidates important linkages between trade and environment, emphasizing on the possible reconciliation between trade liberalization and emission abatement policies. Facing the lack of robustness of qualitative results shown in theoretical work, this paper attempts a quantitative assessment of the interdependencies of these policies. A recursive dynamic CGE model for Costa Rica shows that environmental taxes marginally reduce growth yet allow a sharp decrease of emissions, and that outward orientation alone promotes growth, but induces a risk of specialization in dirty activities. Free trade combined with appropriate effluent taxes enhances factor reallocation towards competitive industries, and hence growth, while significantly abating emissions.  1998 Society for Policy Modeling. Published by Elsevier Science Inc.

1. INTRODUCTION Trade and environment linkages are under increasing scrutiny, and a vast literature has emerged on the subject (see Dean, 1992; Cropper and Oates, 1992; Beghin, Roland-Holst, and van der Mensbrugghe, 1994, for surveys). Important linkages have been identified, and major conclusions have been drawn in terms of Address correspondence to Maurizio Bussolo, Fedesarrollo, Calle 78 No 9-91, Santa Fe de Bogota, Columbia. Authors are indebted to C. Perroni, D. Roland-Holst, D. van der Mensbrugghe, and J. Whalley for thoughtful comments. Received December 1995; final draft accepted July 1996. Journal of Policy Modeling 20(1):11–31 (1998)  1998 Society for Policy Modeling Published by Elsevier Science Inc.

0161-8938/98/$19.00 PII S0161-8938(96)00092-0

12

S. Dessus and M. Bussolo

economic policy. Pressures to use trade instruments for environmental ends have been debunked as a blunt and inefficient approach to environmental policy (Anderson and Blackhurst, 1992). Even in a second-best world, the optimal policy to abate emissions would be a targeted uniform tax per unit of pollution because this would directly discourage the emissions of pollutants (Perroni and Wigle, 1994; Bragga, 1992). However, many questions still remain. Most countries are now engaged in a trade liberalization process, while being aware of environmental problems, and the coherence between environmental and trade policies is a major concern. Will the realization of comparative advantages induce a risk of specialization in dirty activities? Will the implementation of domestic environmental taxes affect international competitiveness? Facing the lack of robustness of qualitative results shown in theoretical work concerning these issues (Copeland and Taylor, 1995), most recent studies have been focused on measuring quantitatively the interdependencies of environmental and commercial policies. Empirical research tends to confirm that developing economies specialize in dirty industries (Hettige, Lucas, and Wheeler, 1992; Low and Yeats, 1992; Birdsall and Wheeler, 1992). However, studies do not find strong evidence that OECD countries’ stricter environmental regulations per se have influenced competitiveness (OECD, 1993; Tobey, 1990). This could suggest that developing economies have a real comparative advantage in dirty productions, and hence a trade-off between trade liberalization and environmental preservation could occur. These questions have been frequently addressed by using Computable General Equilibrium (CGE) models (Lee and RolandHolst, 1994; Perroni and Wigle, 1994). Their main advantage lies in the possibility of combining detailed and consistent real-world databases with a theoretically sound framework. Following this line of investigation, this paper aims at offering a quantitative analysis of the linkages between economic activity and the environment in Costa Rica, and specifically at evaluating the joint impact of environmental and commercial policies. This country appears to be a particularly relevant case in the context of these interrogations (Persson and Munasinghe, 1995). Its openness rate averaging 80 percent makes its economy very sensitive to the commercial policy reforms it has engaged in. Its exports are concentrated in the primary sectors, which require extensive use of polluting chemical intermediates. Abatement policies could impose additional costs and hence could modify its exports competitiveness. By contrast,

A CGE ASSESSMENT APPLIED TO COSTA RICA

13

a more open trade policy could imply an increase in polluting productions. Three main aspects of the CGE model presented in this paper account for its specificity with respect to previous analysis. First, it embodies a high level of disaggregation for pollutants, products, sectors, and types of households. This model can be used to simulate abatement policies targeted to specific air emissions, measuring, at the same time, the effect on related water and soil pollutants. Trade policy reform and the related resource reallocation do not have uniform outcomes across sectors. The expansion or contraction of specific activities have differentiated environmental consequences. The product disaggregation of the model allows the highlighting of certain environmental outcomes of trade policy. Moreover, income distribution issues arising from environmental and commercial policies, and the question of the redistribution of environmental taxes receipts, are briefly discussed and can be further investigated due to the detailed classification scheme of households. Second, this model explicitly includes dynamic features. Its simulations run to year 2010, allowing the introduction of exogenous factors such as productivity shifts and demographic changes that affect capital accumulation and growth trajectory. Comparing the trends of outputs and emissions derived from different scenarios reveals the dynamic interdependencies of environmental and commercial policies. Third, most economywide studies on growth and environment linkages rely on effluent intensities associated with output, and do not allow for substitution between nonpolluting and polluting factors (e.g., Lee and RolandHolst, 1994). Abating pollution is then achieved principally by reducing output in pollution-intensive sectors, with a significant cost in terms of growth. By contrast, in our model, pollution emissions are linked to polluting input use, rather than output. Technical adjustment by substituting nonpolluting factors to polluting factors may therefore be assessed. We first consider pollution-abatement policies alone, holding trade-policy parameters constant, and investigate their effects on growth, sectoral allocation, and trade. We observe that the cost in terms of growth of abating emissions is marginal, and that targeting one type of emission de facto reduces all the others pollutants examined. A second scenario considers trade liberalization alone. Trade distortions are removed progressively over time. Here we capture the pro-growth effect of trade liberalization as well as its environmental implications. In this case, we find that

14

S. Dessus and M. Bussolo

the risk of specialization in dirty activities is very high for Costa Rica, even in the absence of stricter international environmental regulation. In a third and last scenario, we combine environmental and trade policies and show how they mitigate each other in terms of negative effects (trade-induced pollution and anti-growth effects of environmental policy). This last combined scenario shows how free trade and environment protection can coexist without growth–environment trade-off, by setting targeted effluent taxes. 2. THE MODEL The model used in this paper originates directly from a prototype model (Beghin et al., 1996) built for the OECD Development Centre research program on sustainable development, environment, resource use, trade, and technology. It is calibrated on the data contained in a Social Accounting Matrix (SAM) estimated for the year 1991 (Bussolo and Roland-Holst, 1994). This SAM includes 10 household categories (5 urban and 5 rural), 40 sectors, 16 labor types (with rural/urban disaggregation), and 13 different polluting emissions. A detailed list of the model dimensions is presented in Appendix 2. The model is dynamic and solved recursively for the years 1992, 1995, 2000, 2005, and 2010. The following subsections briefly describe the main characteristics of the model. Production The Constant Elasticity of Substitution production function is a nested structure taking into account the optimizing behavior in the choice of production factors. It assumes constant returns to scale. Output results from two composite goods: non-energy intermediates and energy plus value added. The intermediate aggregate is obtained by combining all products in fixed proportions (Leontief structure). The valued-added and energy components are decomposed in two parts: aggregate labor and capital & energy. Labor is a composite of its 16 categories. The capital-energy bundle is further disaggregated into its basic components. By distinguishing between new and old capital, this model allows one to distinguish the allocation of capital existing at the beginning of the period, or already installed, from that resulting from contemporary investment (putty/semi-putty production function). Finally, the energy aggregate includes two types of energy that are substitutes: oil and electricity. Appendix 1 depicts the nested decision process

A CGE ASSESSMENT APPLIED TO COSTA RICA

15

in the choice of production factors. Substitution elasticities reflect adjustment possibilities in the demand for production factors originated from variations in their relative price. Consider in particular these values1: 0.00 between intermediates and value added with old capital plus energy; 0.50 between intermediates and valueadded aggregate incorporating new capital plus energy; 0.12 between aggregate labor and old capital-energy bundle; 1.00 between aggregate labor and new capital-energy bundle; 0.40 among different types of labor; 0.00 between old capital and energy; 0.80 between new capital and energy; 0.25 among different sources of energy associated with old capital; 2.00 among those associated with new capital. Income Distribution and Absorption Labor income is allocated to households according to a fixed coefficient distribution matrix derived from the original SAM. Likewise capital revenues are distributed among households, corporations, and rest of the world. Corporations save the after-tax residual of that revenue. Private consumption demand is obtained through maximization of household-specific utility function following the Extended Linear Expenditure System (ELES).2 Household utility is a function of consumption of different goods and saving. Elasticities are different for each household and product, and they vary in the range 0.20 (for basic products consumed by the household with highest income) to 1.30 (for services). Government and investment demands are disaggregated in sectoral demands once their total value is determined according to fixed coefficient functions. International Trade The model assumes imperfect substitution among goods originating from different geographical areas.3 Imports demand results from a CES aggregation function of domestic and imported goods. Export supply is symmetrically modeled as a Constant Elasticity of Transformation (CET) function. Producers decide to allocate their output to domestic or foreign markets responding to relative 1 These elasticities are derived from the available relevant literature. See for instance Burniaux, Nicoletti, and Oliveira-Martins (1992). 2 A useful reference for the ELES approach is found in Lluch (1973). 3 Armington (1969).

16

S. Dessus and M. Bussolo

prices. Elasticities between domestic and foreign products are of comparable magnitude for imports demand and exports supply. Their values are 3.00 for agricultural goods, 2.00 for manufactured goods, and 1.50 for services. The small-country assumption holds, Costa Rica being unable to change world prices; thus, its imports and exports prices are exogenous. Capital transfers are exogenous as well. The balance-of-payments equilibrium therefore determines the final value for the current account. Model Closure and Dynamics The equilibrium condition on the balance of payments is combined with other closure conditions so that the model can be solved for each period. First consider government budget. Its surplus/deficit is exogenous and the household income tax schedule shifts in order to achieve the predetermined net government position. Second, investment must equal savings, with these originating from households, government, and rest of the world. The dynamic structure of the model results from that last condition of equilibrium between savings and investment. A change in the savings volume influences capital accumulation in the following period. Exogenously determined growth rates are assumed for various other factors that affect the growth path of the economy, such as population and labor-supply growth rates, labor and capital-productivity growth rates, and energy-efficiency factor growth rate. Agents are assumed to be myopic and to base their decisions on static expectations about prices and quantities. The model dynamics is thus recursive, displaying a sequence of static equilibria. Emissions Emissions are determined by intermediate or final4 consumption of polluting products. In addition, certain industries display an autonomous emission component linked directly to their output levels. This is done so as to include some polluting production processes that would not be accounted for by considering only the vectors of their intermediate’s consumption. It is assumed that labor and capital do not pollute. Emissions coefficients associated 4 Final consumption, in this context, is restricted to households, government, and investment demand.

A CGE ASSESSMENT APPLIED TO COSTA RICA

17

with each type of consumption and production are derived from a previous study on the determinants of polluting intensity for the United States and adapted to the Costa Rican case (Dessus, Roland-Holst, van der Mensbrugghe, 1994)5. A change in sectoral output, or in consumption vectors, either in levels or composition, therefore affects emission volumes. Formally, the total value for a given polluting emission takes the following form: E5

oi oj ajCi,j 1 oi biXPi 1 oj ajXAj

(1)

where i is the sector index; j, the consumed product index; C, intermediate consumption; XP, output; XA, final consumption; aj, the emission volume associated to one unit consumption of product j; and bi, the emission volume associated with one unit production of sector i. Thus, the first two elements of the righthand-side expression represent production-generated emissions; the third one, consumption-generated emissions. In the following analysis, the two sources of pollution (production versus consumption) are considered separately, since the means to abating them differ. Actually, emissions from production can be reduced in three ways (Grossman and Krueger, 1992; Copeland and Taylor, 1994): either through a lower aggregate output (the scale effect), or through a change in the commodity composition (more or less of dirty or clean goods, the composition effect), or through the adoption of cleaner technologies (rebalancing the input mix in favor of less polluting factors, the technological effect). Abatement of consumption-originated emissions could not be achieved through the former effect, since consumption is only composed of final products. There are 13 types of polluting substances. Their volume is independently determined and measured in metric tons. Toxic emissions in air (TOXAIR), water (TOXWAT), and soil (TOXSOL) depend primarily on consumption of chemicals (especially fertilizers for water pollution), oil-derived products, and mineral products. Bio-accumulative emissions differ from the previous ones because of their long-term effects on bioorganisms, due to their high lead (or other heavy metal) concentration. Again, these 5 This study offers econometric estimates of the sectoral determinants of 13 types of emissions. Its calculations are based on the 1987 IPPS (Industrial Pollution Projection System) database developed at the World Bank for the United States (Hettige et al., 1994).

18

S. Dessus and M. Bussolo

are distinguished according to the medium where they are released: either into the air (BIOAIR), or water (BIOWAT), or soil (BIOSOL). These emissions result from the use of mineral and metal products, generally found in construction-related sectors. There are five types of toxic substances released in the air: sulfur dioxide (SO2), nitrogen dioxide (NO2), carbon monoxide (CO), volatile organic compounds (VOC), and suspended particulates (PART). Their levels depend primarily on fuels consumption (oil-and coal-derived products). Finally, two additional categories of water-polluting substances are considered: suspended solids (SS) and those measured by their biochemical oxygen demand (BOD). These emissions are related to the consumption of mineral products. As stated above, the households’ utility functions do not include any term directly related to environmental quality. In other words, pollution levels do not explicitly affect households’ utility. Despite the theoretical validity of such relationship, empirical applications would require estimates for utility values that households assign to environmental qualities. Unfortunately, the statistical information on which these estimates can be based is still too limited (Perroni and Wigle, 1994). An emission abatement policy will still have utility effects, in the traditional sense, through its effects on consumption and savings. Likewise, environmental degradation does not affect production factors productivity. Productivity gains resulting from a greener environment (in particular for cultivated land in Costa Rica6) are not measured in this model. Hence, the potential gains from environmental protection policies are very likely to be underestimated. Policy Instruments The model includes a variety of important instruments of economic policy: direct and indirect taxes on production, consumption, and revenues; tariffs and other taxes; and subsidies on international transactions. Each of these tax/subsidy is differentiated by sector, product, household, production factor, consumption type, or income source. A uniform tax on each unit of polluting 6 The related issue of optimal use of natural resources has been analyzed for Costa Rica in Persson and Munasinghe (1995). In particular, within a CGE framework, they consider the linkages between agriculture land demand and deforestation in two scenarios: with and without well-defined property rights on forest resources.

A CGE ASSESSMENT APPLIED TO COSTA RICA

19

emission is also introduced and paid by the agents producing that emission. The tax level can be endogenously determined if levels of emission (abatement) are to be targeted, otherwise it can be exogenously fixed. In this latter case, emission levels become endogenous. 3. THE REFERENCE AND ALTERNATIVE SCENARIOS We consider three basic scenarios in reference to an 18-year (1992–2010) base trend. The latter determines the reference growth and pollution trajectory in absence of environmental, trade, or any other policy reform and is considered as a benchmark scenario. The definition of a plausible evolution for the Costa Rican economy is based on several simplifying hypotheses. The following simulations should therefore not be considered as a forecast exercise, for CGE models are not adequate forecasting tools. In fact, the definition of a growth path, even when sustained with exogenous assumptions, only serves the purpose of establishing a scenario with no policy interventions. Impacts of environmental and trade policies are then evaluated against this reference scenario by measuring the variations in the economic and environmental aggregates. Fixing values for exogenous variables within a realistic confidence interval seems to have no major consequences: the relative variations of the different economic aggregates measured during the evaluation of alternative policies with respect to the reference scenario seem uninfluenced by those a priori choices. In order to define the reference path we make the following assumptions. The GDP growth rate up to 2010 is exogenously determined so that the capital productivity growth rate can be estimated.7 We assumed a 4.8 percent yearly average growth rate for GDP, corresponding to the historical growth rate of Costa Rica for 1955–90.8 Demographic trends are derived from a Costa 7 In the reference scenario the GDP growth rate is fixed and the capital productivity growth rate is endogenously determined. In the alternative policies simulations, the previously estimated capital productivity rate is exogenous and GDP growth rate becomes endogenous. 8 This yearly average growth rate results from the estimation of a linear trend of the logarithm of GDP at constant (1987 base year) prices (in brackets are shown t-Student values):

y 5 24.6 1 0.048 t R2 5 0.97 (602) (31.7)

20

S. Dessus and M. Bussolo

Rican government forecast study (DGEC, 1988). Population growth rate decreases from 2.4 percent in 1992 to 1.6 percent in 2010. Rural and urban populations grow by 2.0 and 2.8 percent, respectively, in 1992, and 1.3 and 1.8 percent in 2010. Labor force, 15- to 64-year-old people, represents 59.8 percent of total population in 1992 and becomes 65.8 percent in 2010. It is assumed that the labor supply growth rate coincides with that of labor force, that is, its rate goes from 2.8 percent in 1992 to 2.0 in 2010 (hence the participation rate remains constant). Finally, some additional hypotheses allowed the estimation of variations in composition of Costa Rican labor force in terms of independent and salaried, rural and urban, and skill degree categories. The yearly average growth rate of productivity for each type of labor was fixed at 0.75 percent for the whole time period.9 A further hypothesis concerns monetary transfers among agents and public expenditures. These are supposed to be growing at the same rate as GDP. The government budget deficit decreases during the simulation period so that it reaches its balance in 2010. The last growth hypothesis assumes that energy efficiency factor increases yearly by one percent. Apart from the latter, no other modification affects the current technology. However, this can become less polluting because of factor substitution due to changes in tax structure, production, and consumption. The first scenario considers a progressive reduction of each type of emission (i.e., it consists of 13 different experiments). A target in terms of emission abatement is fixed as follows: emissions levels are reduced with respect to the reference scenario by 2 percent in 1995, 8 percent in 2000, 17 percent in 2005, and 25 percent in the end of the period. The instrument used to reach this target is a uniform tax per unit of emission, the level of which is endogenously determined by the model. The second scenario simulates a policy of unilateral trade liberalization through a progressive reduction in import tariffs and export subsidies. These trade distortions are decreased ad valorem by 5 percent with respect to the benchmark case in 1995, by 27.5 percent in 2000, by 58.75 percent in 2005, and completely abolished in 2010, when domestic and international prices equalize. World prices remain unchanged. We 9 The aggregate labor productivity growth rate, which takes into account workers’ shifts towards more productive occupations, is measured as a yearly growth rate in average wage of 1.5 percent between 1991 and 2010.

A CGE ASSESSMENT APPLIED TO COSTA RICA

21

Table 1: Emission Elasticities: Separated Environmental and Commercial Policies With respect to production

With respect to consumption

Tox

Bio

Nox

Air

Wat

Tox

Bio

Nox

Air

Wat

Benchmark

0.85

1.04

0.75

0.81

1.01

0.96

1.13

0.93

0.94

0.97

TOXAIR TOXWAT TOXSOL BIOAIR BIOWAT BIOSOL SO2 NO2 CO VOC PART BOD SS

0.40 0.45 0.49 0.69 0.82 0.70 0.63 0.63 0.61 0.42 0.63 0.72 0.75

0.72 0.80 0.71 0.59 0.93 0.63 0.92 0.93 0.88 0.77 0.93 0.73 0.78

0.05 0.08 0.16 0.60 0.72 0.61 0.35 0.35 0.32 0.10 0.35 0.63 0.65

0.35 0.39 0.46 0.70 0.80 0.71 0.59 0.59 0.57 0.36 0.59 0.73 0.74

0.60 0.69 0.56 0.48 0.87 0.49 0.85 0.85 0.79 0.64 0.85 0.58 0.65

0.73 0.75 0.77 0.88 0.98 0.88 0.86 0.87 0.85 0.71 0.86 0.89 0.90

1.07 1.11 1.03 0.87 1.11 0.93 1.13 1.13 1.11 1.10 1.13 1.00 1.02

0.58 0.59 0.62 0.88 0.95 0.87 0.72 0.72 0.71 0.60 0.72 0.87 0.88

0.68 0.70 0.75 0.91 0.96 0.90 0.83 0.83 0.82 0.65 0.83 0.90 0.91

0.78 0.89 0.71 0.53 0.96 0.53 0.95 0.96 0.91 0.83 0.95 0.61 0.68

LIB

0.99

1.16

0.86

0.96

1.13

0.96

1.16

0.92

0.93

0.98

Notes: Tox: toxic pollutants; Bio: bio-accumulative metals; Nox: sulphur, nitrogen, and carbon oxides; Air: other air pollutants; Wat: other water pollutants. Elasticities are measured as the ratio between the yearly average growth rates of polluting emissions and production or consumption ones, during the 1991–2010 period.

therefore do not assume any change in international environmental regulations that could affect the competitiveness of Costa Rica. The last set of simulations combines the two types of reforms. In order to make our results more legible, emissions are aggregated in five groups, although the simulations have been carried out with the fully disaggregated model: toxic pollutants (TOXAIR, TOXWAT, TOXSOL), bio-accumulative metals (BIOAIR, BIOWAT, BIOSOL), oxide emissions (SO2, NO2, CO), other air pollutants (VOC, PART), and other water pollutants (BOD, SS). These aggregations are consistent in physical terms and do not hide relative variations of opposite sign. In fact, emissions show a high correlation degree within each group. 4. RESULTS Aggregate simulations results are presented in Table 1 for individual abatement taxes and for trade liberalization. This table

22

S. Dessus and M. Bussolo

displays, for each simulation, pollution volumes elasticities for the above-defined five emission groups with respect to production and consumption. Rows correspond to different scenarios: the first is the benchmark (in italic), the following 13 are emission-specific abatement policies, and the last represents trade liberalization (LIB). If, for instance, a uniform tax is levied on water pollutants (TOXWAT), so that 25-percent abatement with respect to the benchmark is the target for the year 2010, average yearly growth rates for bio-accumulative emissions from production in the period 1991–2010 would be equal to 0.80 times the average growth rate for production, or, considering consumption originated emissions, 1.11 times the corresponding growth rate for final demand. Environmental Policy The different progressive abatement policies examined have quite low costs in terms of output and allow one to separate to a significant degree long-term evolution in production from its pollution consequences. Moreover, it seems that a specific abatement policy not only reduces its targeted emissions, but also other pollutants. Invariably for each simulation and across pollutant groups, emissions elasticities with respect to production are smaller than in the benchmark. Clearly, substitutions effects among different types of emissions are not induced in the production processes. Aggregate reduction in emission volumes is primarily the result of decreased production-generated emissions. This is due to a shift of production towards less polluting activities as well as to the implementation of cleaner technologies. A detailed analysis decomposing the various reduction effects would show a significant lower output for those sectors producing highly polluting manufacturing goods,10 up to 20 percent with respect to the reference scenario in the year 2010 (composition effect). It would also illustrate that emissions abatement in the other industries is obtained through diminished pollution intensities (technology effect), as the result of substituting polluting intermediates with more labor and capital. This also partially explains the low cost in terms of value 10 Highly polluting manufacturing consists of pulp and paper, chemicals, oil refinery, iron and steel, and construction materials industries.

A CGE ASSESSMENT APPLIED TO COSTA RICA

23

added.11 The average yearly GDP growth rate in the simulations ranges between 4.78 and 4.83 percent, very close to the benchmark rate of 4.81 percent. It should also be noticed that these policies do not significantly affect Costa Rican external competitiveness, for its aggregate exports are at most (in the case of an air bioaccumulative emission abatement policy, BIOWAT) decreased by 4 percent in real term, with respect to the benchmark in 2010. These policies seem to be much less effective regarding the emission volumes generated through final consumption. This can be explained considering two related facts. First, households’ reactions to emissions taxes are sluggish. As already observed in OECD countries, this is primarily due to the households’ lower, with respect to enterprises, replacement rate of appliances and other durable goods (housing, vehicles), which are the main source of pollutants (fuels, chemicals). Second, most of the tax burden is on enterprises, these being the main polluters through their production activities. Trade Liberalization The trade liberalization scenario displays very different results from the previous abatement cases. Consider the last row in Table 1. The complete elimination of trade barriers by the year 2010 results in a stronger specialization in more polluting activities compared to the benchmark. This is only due to the structure of Costa Rican comparative advantages. Highly polluting manufacturing sectors and export agriculture record significant growth rates (6.4% and 6.2% against 4.8% and 2.5% in the reference case). Both composition and technology effects contribute to increased pollution elasticities, with the first being the larger component. In other words, Costa Rican specialization in polluting activities is mainly explained by a shift in the output composition towards more pollution-intensive products and, to a lesser extent, by the use of dirtier technologies across all industries. This composition effect reflects the full exploitation of Costa Rican comparative advantage in polluting sectors. The scale effect also plays 11 Notice that usually savings are higher when abatement policies are implemented, even with no anticipation of future emission taxes by the agents. In fact, revenues from these taxes are redistributed to households as a function of their income tax rates. Thus, the correspondence of high income tax rates and high savings within the same household group and the revenues redistribution scheme result in faster capital accumulation.

24

S. Dessus and M. Bussolo

an important role, raising aggregate emission levels up to 15–20 percent more than the reference scenario by the year 2010 (15% for oxides, 20% for bio-accumulative emissions, 19% for other types of emissions). Final consumption and output increase by 12.2 and 9.4 percent, respectively, compared to the benchmark. Trade liberalization results in a higher growth rate of GDP (5.2% against 4.8% in the reference case) and in expanded volumes of trade (an additional 31.5% for exports and 29.8% for imports with respect to the benchmark levels in 2010). As for abatement policy, trade liberalization has minor effects on final demand composition. Elasticities of emissions originating from consumption with respect to consumption do not differ significantly from the benchmark scenario. Combined Reforms The contrast between positive environmental results of pollution abatement policies and economic gains of trade liberalization stimulates the study of a scenario where both emission taxes and tariff elimination are combined. This might also be a more realistic case than that of isolated policies. Since the mid-1980s, Costa Rica has been attempting to integrate its economy to international markets. This is reflected in a significant increase in the export share in GDP (39% in 1992 against 27% in 1982) and in the capital account liberalization in 1992. The recent participation of the Costa Rican government to the Miami summit of American countries in December 1994 confirms its plans of regional integration. The signature of the Uruguay Round agreement is another proof. Furthermore, a policy of trade liberalization with no associated measures of pollution control seems quite implausible for two reasons. On the one hand, given the current Costa Rican comparative advantage structure, tariff elimination will favor the most polluting sectors. This tendency could be exacerbated if Costa Rican trade partners, among which OECD countries play a major role,12 should tighten their environmental policies. In these countries, production costs of locally polluting goods will rise, promoting cheap imports from Costa Rica. On the other hand, antipollution pressures could originate from neighbor countries, hence 12 Note that Costa Rican exports to OECD countries represent 62 percent of their total value in 1991, and imports coming from this group of countries are 63 percent, for that same year.

A CGE ASSESSMENT APPLIED TO COSTA RICA

25

Table 2: Emission Elasticities: Coordinated Environmental and Commercial Policies With respect to production Nox

Air

Wat

With respect to consumption

Tox

Bio

Tox

Bio

Nox

Air

Wat

Benchmark

0.85

1.04

0.75

0.81

1.01

0.96

1.13

0.93

0.94

0.97

TOXAIR 1 LIB TOXWAT 1 LIB TOXSOL 1 LIB BIOAIR 1 LIB BIOWAT 1 LIB BIOSOL 1 LIB SO2 1 LIB NO2 1 LIB CO 1 LIB VOC 1 LIB PART 1 LIB BOD 1 LIB SS 1 LIB

0.36 0.43 0.50 0.79 0.95 0.80 0.72 0.72 0.70 0.40 0.73 0.83 0.86

0.74 0.85 0.72 0.55 0.99 0.58 1.00 1.00 0.95 0.81 1.01 0.76 0.81

20.17 20.12 20.01 0.66 0.82 0.66 0.33 0.32 0.29 20.08 0.33 0.69 0.72

0.30 0.36 0.47 0.82 0.94 0.83 0.68 0.68 0.66 0.32 0.68 0.85 0.87

0.56 0.68 0.52 0.43 0.91 0.41 0.91 0.91 0.84 0.63 0.91 0.54 0.63

0.58 0.63 0.66 0.85 0.99 0.84 0.83 0.83 0.81 0.58 0.83 0.86 0.87

1.07 1.12 1.01 0.77 1.14 0.86 1.16 1.16 1.13 1.10 1.16 0.99 1.01

0.39 0.42 0.45 0.84 0.94 0.82 0.63 0.63 0.61 0.43 0.63 0.83 0.84

0.52 0.55 0.64 0.90 0.96 0.88 0.79 0.79 0.78 0.49 0.79 0.88 0.88

0.64 0.81 0.55 0.34 0.96 0.31 0.95 0.96 0.88 0.73 0.96 0.45 0.53

inducing the Costa Rican government to adopt higher environmental quality standards. These demands could even arise locally, once a more developed society discounts disutility from pollution at a higher rate (Copeland and Taylor, 1993). In addition, significant increases in environmental damages could negatively affect tourism, the second source of foreign exchange.13 All these factors contribute to make a combination of trade liberalization and abatement policy the most plausible scenario for Costa Rica in future years. Table 2 shows pollution elasticities for 13 experiments with coordinated environmental and trade policies. It is worth noting that both policies are now jointly implemented and that their time paths and rates are exactly the same as in the previous simulations. 13 Although tourism is not modeled as an explicit sector, it is possible to argue that a policy of mere trade liberalization could result in its contraction. In the simulations, such a policy would increase agricultural production for exportation (mainly bananas) by a factor of two with respect of the benchmark in 2010. This requires a more intensive and/ or extensive use of the land factor. In both cases, wild areas or simply noncultivated land would either be damaged (Lutz and Daly, 1990) or reduced with negative consequences for tourism.

26

S. Dessus and M. Bussolo

The coordination of environmental and commercial policies produces positive results: it significantly reduces emissions and yet allows income and consumption to grow. It also permits one to separate consumption from emissions to a larger extent than with a simple abatement policy (see Table 1). Emission elasticities with respect to consumption are almost invariably lower. In other words, the share of the demand for polluting products under a coordinated policies regime decreases more rapidly than under a pure abatement policy regime. This is due to a larger increase in households’ incomes that shifts consumption away from polluting manufactured goods and towards services. Output of polluting products increases according to Costa Rican comparative advantage. In the coordinated policies case, the composition effect is less important, but it is compensated by more substantial factor substitutions that reduce the emission volumes per unit of output. In fact, this technology effect is reflected in the substitution of polluting intermediates with more labor and capital. Larger savings, stemming from increased households’ incomes, reduce capital rental rates and allow enterprises to make the new investments necessary to reduce emissions without excessively rising output costs. Economic gains from coordinated environmental and commercial policies are quite substantial. Combined reforms lead to a GDP growth rate ranging between 5.2 and 5.3 percent for 11 of the 13 pollutants considered (and 5.0% and 4.9% for the two remaining items, namely bio-accumulative pollution in air and in water). The integration of the Costa Rican economy to the international markets and the reallocation of factors in the cleaner and more efficient sectors generate important gains that are cumulative. In fact, higher incomes generate larger savings that are used to finance more investment, so that from one period to another productive capacity is increased in a virtuous circle. The reduction of domestic prices through their alignment to world prices stimulates households’ consumption that registers higher values across all emissions taxes. 5. CONCLUSION Our empirical results support several findings. The further integration of Costa Rica to the world economy induces an important risk of environmental degradation, if it is not accompanied by voluntary environmental reforms. Actually, the new patterns of

A CGE ASSESSMENT APPLIED TO COSTA RICA

27

resource allocation and output composition under free trade are more pollution-intensive. Furthermore, a strong scale effect arises. Economic growth resulting of trade liberalization increases total pollution. Environmental reforms, through targeted fiscal policies linked to the utilization of polluting goods (in either intermediate or final use) could achieve significant pollution abatement without hampering economic growth or international competitiveness. The former results confirm those of previous studies (Beghin, RolandHolst, and van der Mensbrugghe, 1995; EEC, 1994; Bovenberg and Goulder, 1993). When coordinated with trade liberalization, these reforms are facilitated by economic gains that could reduce opposition and rent-seeking activities. Moreover, tariff removal enhances the adjustment capacity of the Costa Rican economy to domestic environmental regulations. Therefore, if trade and environmental policies appear to be independent, as defended by Perroni and Wigle (1994), their coordination is the best solution to benefit from the pro-growth effect of economic integration while mitigating environmental degradation. Finally, we need to emphasize some fundamental qualifiers concerning our conclusions, which would tend to reinforce the desirability and potential attractiveness of pollution control. Our quantitative approach overstates the negative consequences of controlling emissions. First, we only evaluate in this paper the economic cost implications of abating pollution, without taking into account the plausible benefits of reducing pollution in terms of growth, productivity, welfare, and health. Second, the model does not consider the introduction of cleaner technologies (investment in emission abatement equipment goods and/or “cleaner” capital), although those might be competitive with existing technologies if pollution were appropriately priced and might be easier to achieve as Costa Rica further integrates into the world market. Free trade and greater integration in world markets represent the way to maximize growth opportunities; when combined with appropriate environmental policies, they minimize the growth–environmental trade-offs faced by Costa Rica.

APPENDIX 1: PRODUCTION NESTING

Figure A1. (1) Each nest represents a different CES bundle. Substitution elasticities separated by a comma indicate respectively, the CES substitution elasticity for old capital and for new capital. The elasticity may take the value zero. Because of the putty/semi-putty specification, the nesting is replicated for each type of capital, that is, old and new. The values of the substitution elasticity will generally differ depending on the capital vintage, with typically lower elasticities for old capital. (2) Intermediate demand, both energy and non-energy, is further decomposed by region of origin according to the Armington specification. However, the Armington function is specified at the border and is not industry-specific. Substitution elasticities separated by a dash indicate the range of elasticity values. (3) The labor aggregate is composed of 16 different types of labor. For clarity only two are shown.

A CGE ASSESSMENT APPLIED TO COSTA RICA

29

APPENDIX 2: MODEL DIMENSIONS The model includes 40 sectors of production, and the corresponding 40 commodities: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Banana Coffee Sugar cane Cocoa Cereal Cotton Tobacco Livestock Forestry & fishing Other agriculture Meat and dairy Edible oils Grain milling Bakeries Refined sugar Other food Beverage Processed tobacco Textile Leather and footwear

21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40

Wood furniture Paper & printing Chemical Refined petroleum Tire & tube Rubber & plastic Ceramic & glass Building material Metal Electrical goods Transport equipment Other manufacturing Construction Finance Commerce & restaurant & hotel Transport & communication Other services Electricity Real estate Government services

The model includes 16 types of labor: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

Urban salaried professionals and managers Urban salaried white-collar workers Urban salaried industrial workers Urban salaried unskilled workers Urban independent professionals and managers Urban independent white-collar workers Urban independent industrial workers Urban independent unskilled workers Rural salaried professionals and managers Rural salaried white-collar workers Rural salaried industrial workers Rural salaried unskilled workers Rural independent professionals and managers Rural independent white-collar workers Rural independent industrial workers Rural independent unskilled workers

The model includes 10 households, depending on head of household’s professional activity: 1 2 3 4

Urban Urban Urban Urban

professionals and managers head of household white-collar workers, commerce, salesmen head of household industrial workers head of household unskilled workers head of household

30

S. Dessus and M. Bussolo 5 6 7 8 9 10

Urban non-active head of household Rural professionals and managers head of household Rural white-collar workers, commerce, salesmen head of household Rural industrial workers head of household Rural unskilled workers head of household Rural non-active head of household

REFERENCES Anderson, K., and Blackhurst, R. (1992) The Greening of World Trade Issues. Ann Arbor: University of Michigan Press. Armington, P. (1969) A Theory of Demand for Products Distinguished by Place of Production, IMF Staff Papers 16:159–178. Beghin, J., Dessus, S., Roland-Holst, D., and van der Mensbrugghe, D. (1996) General Equilibrium Modelling of Trade and the Environment. OECD Development Centre Technical Papers, forthcoming. Beghin, J., Roland-Holst, D., and van der Mensbrugghe, D. (1994) A Survey of the Trade and Environment Nexus: Global Dimensions, OECD Economic Studies 23:167–192. Beghin, J., Roland-Holst, D., and van der Mensbrugghe, D. (1995) Trade Liberalization and the Environment in the Pacific Basin: Co-ordinated Approaches to Mexican Trade and Environment Policy, American Journal of Agricultural Economics 77:778–785. Birdsall, N., and Wheeler, D. (1992) Trade Policy and Industrial Pollution in Latin America: Where Are the Pollution Havens? In International Trade and the Environment (P. Low, Ed.). World Bank Discussion Paper No. 159. Washington: The World Bank. Bovenberg, A.L., and Goulder, L.H. (1993) Integrating Environmental and Distortionary Taxes: General Equilibrium Analysis. Presented at the conference on “Market Approaches to Environmental Protection,” CEPR/IIS, Stanford University, December 3–4. Braga, C.A.P. (1992) Tropical Forests and Trade Policy: The Case of Indonesia and Brazil. In International Trade and the Environment (P. Low, Ed.). World Bank Discussion Paper No. 159. Washington: The World Bank. Burniaux, J.-M., Nicoletti, G., and Oliveira-Martins, J. (1992) GREEN: A Global Model for Quantifying the Costs of Policies to Curb CO2 Emissions. OECD Economic Studies 19:49–92. Bussolo, M., and Roland-Holst D. (1994) The 1991 Social Accounting Matrix for Costa Rica. Mimeo, OECD Development Centre, Paris. Copeland, B.R., and Taylor, S.M. (1993) Trade and Transboundary Pollution. University of British Columbia Discussion Paper, No 93-42. Copeland, B.R., and Taylor, S.M. (1994) North–South Trade and the Environment. Quarterly Journal of Economics 109:755–787. Copeland, B.R., and Taylor, S.M. (1995) Trade and the Environment: A Partial Synthesis. American Journal of Agricultural Economics 77:765–771. Cropper, M.L., and Oates, W.E. (1992) Environmental Economics: A Survey. Journal of Economic Literature 30:675–740. Dean, J.M. (1992) Trade and the Environment: A Survey of the Literature. In International Trade and the Environment (P. Low, Ed.). World Bank Discussion Paper No. 159. Washington: The World Bank.

A CGE ASSESSMENT APPLIED TO COSTA RICA

31

Dessus, S., Roland-Holst, D., and van der Mensbrugghe, D. (1994) Input-Based Pollution Estimates For Environmental Assessment in Developing Countries. OECD Development Centre Technical Papers No. 101:Paris. DGEC, Direccion General de Estadistica y Censos (1988) Estimaciones y proyecciones de poblacion 1950–2025. Ministerio de Planificacion Nacional y Politica Economica, San Jose, Costa Rica. EEC (1994) The Potential Benefits of Integration of Environmental and Economic Policies: An Incentive-Based Approach to Policy Integration. Directorate-General for Environment, Nuclear Safety and Civil Protection (DGXI), Brussels. Grossman, G.E., and Krueger, A.B. (1992) Environmental Impacts of a NAFTA. CEPR Discussion Paper Series, No. 644, April. Hettige, H., Lucas, R.E.B., and Wheeler, D. (1992) The Toxic Intensity of Industrial Production: Global Patterns, Trends and Trade Policy. American Economic Review 82: Hettige, H., Martin, P., Singh, M., and Wheeler, D. (1994) The Industrial Pollution Projection System. Mimeo, World Bank. Lee, H., and Roland-Holst, D. International Trade and the Transfer of Environment Cost and Benefits. In Applied Trade Policy Modelling (J. Francois and K. Reinert, Eds.). Cambridge: Cambridge University Press, forthcoming. Lluch, C. (1973) The Extended Linear Expenditure System, European Economic Review 4: Low, P., and Yeats, A. (1992) Do Dirty Industry Migrate? In International Trade and the Environment (P. Low, Ed.). World Bank Discussion Paper No. 159. Washington: The World Bank. Lutz, E., and Daly, H.E. (1990) Incentives, Regulations and Sustainable Land Use in Costa Rica. Working Paper No. 34, World Bank, Environment Department, Washington. Perroni, C., and Wigle, R.M. (1994) International Trade and Environmental Quality: How Important Are the Linkages? Canadian Journal of Economics 27:551–567. OECD (1993) Environmental Policies and Industrial Competitiveness. Paris: OECD Documents. Persson, A., and Munasinghe, M. (1995) Natural Resources Management and Economywide Policies in Costa Rica: A Computable General Equilibrium (CGE) Modelling Approach. The World Bank Economic Review 2:259–285 Tobey, J.A. (1990) The Effects of Domestic Environmental Policies on Patterns of World Trade: an Empirical Test, Kyklos 43:191–210.