Document(s) 10 of 15

Introduction

Green or environmental-based trade restrictions are, in principle, imposed by importing countries to deprive an exporting country of any competitive advantage based on environmental destruction.¹ In practice, green trade restrictions effectively amount to non-tariff barriers to trade which have the effect of limiting access to foreign markets by countries in which the gap between marginal private and marginal social costs has a significant bearing on international competitiveness. Some, but by no means all, green trade restrictions are GATT-compatible. In particular, those which seek to limit access based purely on process and production method externalities are inadmissible under GATT rules (Deloitte & Touche 1994: 23). While green trade restrictions have not yet had a significant impact on trade flows, they may be present in many forms such as eco-labelling, deposit refunds, environmental taxes and others.

Little is known about the precise relationship between growth, distribution and the environment, except that it is exceedingly complex and necessitates a structural analysis (Pezzey 1992, Karshenas 1993 and Taylor 1996). Few general analytical results are available, but it is seems clear that, firstly, macro and environmental impacts cannot be separately analysed since growth and environmental deterioration often go hand in hand and, secondly, if green trade restrictions retard export growth and, therefore, GDP per capita, environmental degradation may follow. Gibson (1996) distinguishes extensive from intensive environmental degradation. The former refers to damage associated with distributively neutral output growth, while the latter comes about as a result of deterioration in the distribution of income which leads to environmental “mining”. Hence, green trade restrictions may inhibit extensive but advance intensive environmental decay.

¹ All views expressed in this chapter are those of the authors and not the Development Bank of Southern Africa (DBSA). The authors are grateful to Craig Mackenzie and Andre Roux for their valuable criticism. Lance Taylor and Stephen Gelb collaborated on an earlier version of the model. Projections in this paper were originally made in July 1995.

This paper investigates numerically the relationship between green trade restrictions, macroeconomic performance and environmental quality in South Africa using a multisectoral, dynamic computable general equilibrium model with both real and financial sectors. Three simulations are considered which attempt to capture some of the complex interactions of growth and distribution with key environmental indicators. For each simulation, a quadratic extensive environmental loss function is evaluated that assigns given weights to the emissions of greenhouse gases and water contaminants. The first addresses the effects of responding to green trade restrictions by internalising the local and global externalities of the South African energy sector. A second examines the issue of restrictions on foreign imports of South African mining products, in particular coal. A final simulation assesses the effects of restrictions on non-primary or manufactured exports due to consumption externalities generated by the energy-intensive methods by which they are produced. It is seen that the first option is macroeconomically superior to simply bearing the brunt of the green trade restrictions assumed in their second and third simulations.

The paper is organised as follows: the first section summarises the approach with a discussion of the relevant features of the computable general equilibrium model, key parameter settings and the base run; the second section discusses the environmental block and provides some estimates of emission levels; the third section presents the simulation results both with respect to their macroeconomic and environmental impact; and the fourth section concludes.

The model and the base run

The model employed for the simulations is a nine-sector, two-class structuralist computable general equilibrium system with both real and financial sectors. See Gibson and Van Seventer (1996) for full details of the model specification. Its heritage is Keynesian in that an independent investment function is specified so that output adjusts to bring savings into balance with investment. The investment function depends upon an accelerator, the cost of capital and a “crowding in” term which captures the positive effect of public sector investment on private sector accumulation. For each period, aggregate demand determines the level of output. Across periods, however, supply-side factors are considered. Accumulated past investment determines the capital stock which, in turn, sets the level of capacity output according to productivity parameters. With capacity determined in the previous period, and output in the current period, the level of capacity utilisation is known.

Private consumption is determined by disposable income according to a linear expenditure system. Current government expenditure, including employment, rises with fixed growth rates over the 1994 levels while investment is a residual determined by an exogenous budget deficit to GDP ratio and an endogenous level of government savings. Exports are determined in two ways. As seen in Figure 1, primary exports, which include agriculture and mining, are determined as a residual once output,X, and domestic demand, Dd, are known. Supply is set at full capacity utilisation, u, defined as the ratio of output to capacity output, Xc. The world price, P*, converted at the endogenous exchange rate, e, is taken as an exogenous parameter.

FIGURE 1 Market structure and sectors in the computable general equilibrium model

The market structure for non-primary goods is shown in the second panel of Figure 1. Demand, D, is determined as the sum of consumption, C., investment, I, including private and public and government expenditure, G; plus net exports, E. The price in the non-primary markets is determined as a mark-up, t, on costs, B. Exports in the non-primary sector are tied to capacity growth in those sectors. There is some switching built into the model in that non-primary exports rise as a proportion of output with the real exchange rate.

Wages adjust to inflation incompletely with endogenous labour productivity that depends upon various measures of labour market tightness. Public sector wages match those of the private sector. The exchange and interest rates are managed by the South African Reserve Bank (SARB) and are described in the model with a reaction function based on capacity utilisation, inflation and the state of the reserves. Firms and households allocate their portfolios of wealth in proportion to relative rates of return on equity, bonds, capital flight and money. The government borrows from the central bank and foreigners and the rest of the debt is placed with the public investment commissioner and the domestic financial sector. The latter create loans to themselves sufficient to absorb the residual public sector debt. The money creation that results is, therefore, “indirect monetisation”.

The base run is our best guess for the trajectory of the South African economy over the next five years. The assumptions underpinning the base run are that restrictive monetary and fiscal policy will continue to dominate with the ratio of the public sector borrowing requirement to GDP falling from its current level of about 6% to 4,5% by 1999, the end of the forecast period. The nominal wage and SARB reaction functions will continue to behave as they have in the recent past, and private investment recovers to allow for a 3,3% rate of growth in 1995. Good rainfall is assumed to boost capacity growth in agriculture and capital inflows will prevent a severe depreciation of the rand.

TABLE 1 Macro-data for the base run

Source: model calculations

11990 = 100

Table 1 shows how the model behaves in the base run. The slight dip in GDP growth for 1997 is characteristic of the model in that it reflects the interaction of the multiplier and accelerator. A burst of investment causes capacity to rise and if growth in demand does not validate that capacity growth, investment will turn down again in the next period. Thus, the model will not predict smooth growth unless both capacity and demand grow at precisely the same rate. Since there is no mechanism to bring this about, the model typically produces some mild fluctuations in output.

Observe that GDP growth averages 3,3% over the period while inflation decreases to below 10%. The latter is the result of trade liberalisation and strict monetary policy which combine to slow the rate of depreciation of the rand. In the fifth column it can be seen that the real interest rate has increased significantly compared to the 1990-94 period, the result of the aggressive posture of the SARB as modelled in their reaction function. Note further that nevertheless the second-last column shows a slight increase in the real wage as a result of declining inflation. In this model, as in reality, restrictive monetary and fiscal policies are consistent with higher real wages for the employed and slower job growth for the unemployed. Consequently, the distribution of household income worsens during the forecast period, as confirmed by a rise in the Gini coefficient in the last column. Skilled labour benefits from the modest growth, but the latter is far from sufficient to compensate unskilled labour with higher employment.

The environmental block

In the simulations to follow, we track the output of a number of different environmental hazards. These are roughly grouped according to whether the externality is predominantly local or global. The electricity sector is a key contributor to South Africa’s industrial emissions (Booth 1994: 227). South Africa’s electrical power has been historically structured around coal-based stations burning low-quality bituminous coal. Given that electricity is an important input to many exports, particularly energy-intensive products, the global environmental externalities of the electricity sector may put a range of exports at risk from green trade restrictions. In order to evaluate the macroeconomic costs of green trade restrictions in an appropriate context it is necessary to examine the impact of reducing pollution generated by the electricity sector. The production externalities are clearly not limited to South African residents. The World Resources Institute places South Africa 17th on the list of its greenhouse index ranking (1991), contributing more than 1% of world greenhouse gas, while Eskom’s Annual Environmental Report 1995 claims that South Africa is the 12th largest producer of CO₂. Its per capita ranking for CO₂ is much lower, however, at 42nd. (World Resources Institute 1995: 201).

On the basis of a survey of 12 sectors of the South African economy, Bethlehem in this volume suggests that international environmental concerns are building up. While there is as yet no evidence of actual trade restrictions or barriers to foreign markets, the expectation is that international pressures are likely to become an important factor. Columbus and Alusaf, two primary processing operations, are installing state-of-the-art pollution control technology, anticipating that they will be forced to meet international standards (Finance Week, August, 17-23,1995: 6). Other, more local producers are following the lead (Deloitte & Touche 1994).

To quantify the interaction of macro and environmental variables, the database of the model is expanded to include emissions coefficients which are available for the year 1992. Table 2 shows harmful emissions identified in the model² with their respective total emissions³ and unit abatement costs in the electricity sector⁴ for 1992.

TABLE 2 Estimated total emissions for all production activities as well as unit abatement costs in the electricity sector

Source: DBSA unpublished data and DBSA model calculations

Suspended solids can be found in the waste product that remains after the screening and maceration of the sewage of domestic and industrial sources. As part of sewage sludge it would, in principle, be possible to undertake recycling were it not for the high concentration of harmful substances, such as heavy metals. Nevertheless, sewage sludge is increasingly being dumped. Suspended solids contained in sewage sludge that is used as fertiliser end up polluting vast tracks of agricultural land, gardens and parks (Booth 1994: 233). Suspended solids can be removed by means of sedimentation, flocculation and rapid sand filtration.

²Emissions not reported are hazardous waste, solid waste, nuclear waste and waste produced by households. Environmental emissions data is drawn from an unpublished DBSA database. This database is new and will need to be refined over time. Data parameters have, however, been confirmed by follow-up interviews.

³Total emissions are calculated by multiplying sectoral emission coefficients (per unit of gross value of production) from DBSA unpublished data with 1992 sectoral gross value of production from the DBSA model.

⁴ No published information was available for CO₂ abatement costs in South Africa. Abatement costs for CO₂ in China (World Bank 1994) appears to be around R200 per tonne. For our purposes it was decided to take the average of the abatement costs of CO₂ in China and as reported in the DBSA unpublished data.

Reducing the emission of suspended solids would reduce the quantity of contaminated sludge that must be absorbed into the environment thereby benefiting South Africans. Suspended solids are, therefore, exclusively local process and production method externalities and, under current standards and practice, should not induce green trade restrictions. But since the energy sector is responsible for a significant amount of sludge contamination, suspended solids are included in our list of negative local externalities.

Total dissolved salts contamination is the result of return flows polluted by industrialisation, urbanisation, irrigation and use of artificial pesticides and fertilisers in agriculture (Booth 1994: 239). The problem is worst in the arid and semi-arid regions of South Africa with its Karoo shales that are particularly vulnerable to groundwater seepage. A total dissolved salts concentration of more than 1000 mg per litre is regarded as a health risk (President’s Council 1991:41). However, lesser concentrations could still be detrimental to health if large doses of minerals such as sodium, chloride or magnesium are present. Crop yields may also decline as a result of high salinity (Booth 1994: 244). Apart from the unpleasant taste, high concentrations of sulphates can lead to gastro-intestinal and other long-term complications. Desalination using membrane technologies can be used to reduce total dissolved salts, but the costs are high.

Oxygen-demanding products and industrial and municipal wastewater (including township wastewater) also constitute a serious environmental problem. Aquatic life is severely restricted by organic waste which contains oxygen-demanding products since the growth of decomposer populations is accelerated. Emissions of oxygen-demanding products must be controlled with biofilters, trickle filters, oxygen dosing and maturation, that is, the biodegradation of waste that takes place naturally. Particulate matter—the result of solid fuel combustion and emission of industrial processes—is a highly visible type of air pollution and can lead to respiratory disease. Burning low-grade coal in township stoves also contributes to particulate contamination. This is, in part, the product of the lack of other energy options, but the use of coal stoves has also remained resistant to government efforts to promote the use of electricity for heating (Eberhard and van Horen 1995:166). The industrial areas on the Mpumalanga highveld, with its coal-fired electricity power stations, are main industrial contributors. Total dissolved salts, oxygen-demanding products and particulate matter are, clearly, local externalities.

There are a number of externalities which could trigger green trade restrictions due to their combined local and global character. Carbon monoxide is an example. Although co is poisonous its dispersal range is very limited and its major environmental significance is linked to the breakdown of hydroxyl, an atmospheric cleansing agent which is believed to contribute to the geographic spread of acid rain. Similarly, SO2 and NO2 contribute to acid rain. The Mpumalanga power stations are a main source of this type of pollution. Sulphur dioxide causes irritations of the respiratory passages and vulnerability to infections, and exacerbates asthma, emphysema and bronchitis. Acid rain reduces the biological productivity of waterways and dams as well as soil, affecting yields in timber and agriculture (Booth 1994: 230). Dry and wet scrubbers are used to trap acid-forming pollutants, such as SO2

The emissions most likely to trigger green trade restrictions are CO₂ and CH₄. These greenhouse gases lead to an increase in tropospheric temperatures which may result in global climate changes. There is a range of greenhouse gases, but it is widely accepted that CO₂, especially from fossil fuels, and CH₄ are major causes of concern. Not only industrialisation should be blamed; deforestation for household fuel has also played an important part in upsetting the global carbon balance. Methane emissions, caused by feedlots, rubbish dumps and coalmining activities, also contribute to creating a global thermal blanket.

Since it is extremely difficult to know how to weigh these various measures of environmental quality, we assume that each pollutant will be equally harmful. This option gives equal priority to the clean-up of local and global emissions. The obvious disadvantage of this strategy is that green trade restrictions are more likely to be imposed due to global rather than local environmental damage. From a domestic perspective, however, it would be somewhat unreasonable to select environmental policy strictly on the basis of its global impact. In what follows, we examine the local/global mix of environmental improvement in each simulation as a rough measure of the distribution of environmental benefits.

In the computable general equilibrium model, we use a quadratic extensive damage function which sums with equal weights all damage due to emissions. This function rises with output and decreases as firms clean up. Abatement costs for the electricity sector are shown in the last column of Table 2. These costs are assumed to be linear functions of output, simplifying the more likely underlying non-linear relationship. Table 3 provides some information on the pattern of emissions associated with the growth path of the base run. Significant environmental deterioration is clearly observable in the table, with all contaminants seen to rise in absolute value over the forecast period.

TABLE 3 Emissions by type for the base run (tonnes x 10^-6)

Source: model calculations

Note, finally, that the information of Tables 2 allows us to measure the change in extensive damage only. No attempt is made to estimate the additional decay which would be provoked by deterioration in the distribution of income. Clearly, if a given policy package registers the same value of the extensive damage function, but implies a deterioration in the distribution of income, there will be some collateral intensive damage. If properly measured, intensive environmental decay in South Africa would unquestionably increase the emissions of oxygen-demanding products and particulate matter through, for example, increased informal settlements with inadequate sanitation and energy services. It is less clear that emissions of suspended solids would increase, although it is likely. These intensive contaminants are strictly local and would affect local air and water quality. It is difficult to conceive of intensive decay contributing to significant global environmental risks in South Africa, such as CO₂ emissions, since it would most likely be the result of deforestation, which has largely already taken place. As noted, the principal cooking and heating fuel is coal rather than wood. There could conceivably be some small increase in the emissions of CH₄ due to intensive deterioration, but it is difficult to know how important this effect would be.

Simulations

Green trade restrictions will affect the economy by restricting export markets for both traditional and non-traditional exports, and the result will be a contraction in output and employment. To pre-empt the green trade restrictions, South Africa could clean up its coal-fired energy sector, which is an important input into both classes of exports. In the first of the simulations to follow, we ask what the impact on the macroeconomy would be if emissions by the electricity sector were significantly reduced.⁴ Clean-up costs, both foreign and domestic, will be seen to have a contractionary effect on output, due in part to the tightening of credit by the SARB as a reaction to the cost-push inflation. The benefit, of course, is a cleaner environment and the assumption is that the effort is sufficient to avoid restrictions on South African exports.

In two subsequent simulations, we quantify the costs of the alternative strategy, which is to say, avoiding the clean-up and then suffering from green trade restrictions. The first assumes that the green trade restrictions will be aimed exclusively at primary exports. In the last simulation, the green trade restrictions are imposed on non-traditional exports. As noted above, environmental decay as measured by an extensive damage function, declines with the clean-up of the energy sector. Green trade restrictions also reduce the value of the extensive environmental damage function. To compare the macroeconomic effects, we set the value of the damage function in the second and third simulations equal to its value in the first. We then backsolve for the level of green trade restrictions on traditional and non-traditional exports consistent with the level of environmental damage in the first simulation. In other words, given a target level of environmental quality, is it better to clean up or simply restrict traditional and non-traditional exports to meet the same target?

⁴ It should be noted, however, that the nature of the international requirements to reduce emissions is not known. We, therefore, model a hypothetical example based on a reduction of various emissions, including greenhouse gases. We acknowledge that some of the adjustments made in the model (such as a reduction in CO₂ ), are not easy to implement in the real world.

Cleaner electricity

The first experiment reduces all emissions per unit of output generated by the electricity sector by 35%, phased in an accelerated way over the five-year forecasting period. Given the abatement costs of Table 1, this will lead to higher energy costs and, with a fixed markup, higher prices. These higher costs are reflected in higher intermediate demand by the electricity sector for goods supplied by the manufacturing sector and imported goods due to higher maintenance of existing equipment. Similarly, new clean-up equipment purchases shift up investment demand by the electricity sector. It is further assumed that the composition of new investment goods reflects a higher proportion of imported versus locally supplied investment goods, as firms install state-of-the-art scrubbers, monitoring devices and other equipment not locally available.

Of course, if the mark-up rate in the energy sector remains constant, firm income increases with costs for the same level of output. Since the elasticity of demand for electricity is low, the mark-up was exogenously reduced by 7,5%, phased in over the period, to force this sector to internalise rather than pass through some of the costs of abatement. As a result, firm income in the electricity sector is more or less the same as in the base run. As a consequence, the price of electricity will still rise by more than 7% compared to the base run by the end of the period. Note that this conflicts with Eskom’s intention to lower the price of electricity to the municipal retailers (Eberhard and van Horen 1995). The macroeconomic impact of this experiment is evaluated in Table 4.

It can be seen in Table 4 that internalising environmental costs by the electricity sector is expected to have a slight contractionary impact on the South African economy. Gross domestic product is almost 2% lower compared to the base run at the end of the period. As just noted, rising costs are only partially passed on, and in the third column it can be seen that inflation is slightly higher than in the base run (0,1% on average). In spite of the slowdown in the economy, the reaction by the SARB is to tighten credit. Together with the drop in capacity utilisation, this will lead to lower investment.

In the fifth column it can be seen that public investment decreases slightly. As explained previously, government investment is assumed to be the residual of the public accounts. With constant growth rates assumed for government current expenditure and a contracting economy, the budget deficit will decline to respect the public sector borrowing requirement to income constraint. Consequently, less is available for government investment. In addition, tax revenues decline with the slowdown in economic activity, although this is to some degree counterbalanced by bracket creep due to the additional inflation.

The impact on the overall—that is, primary plus non-primary—export performance is approximately neutral. As a result of the export-clearing nature of the primary sectors

TABLE 4 Macro-data with cleaner electricity (ratios to base run)

Source: model computations

¹ Absolute differences from the base run (in percentage points).

and a slight contractionary impact on the macroeconomy, exports of the primary sector will increase as seen in Figure 1 and confirmed by Table 4. With a contraction in economic activity, agriculture and mining can export more. Exports of the non-traditional sectors, on the other hand, are linked to capacity and will, therefore, drop with GDP which, from data not shown, slows more rapidly than in traditional sectors. One of the reasons is that heavy manufacturing is more electricity-intensive than mining when backward linkages of the input-output submatrix of the model database are taken into account. A policy to penalise electricity intensive production activities will have a more negative impact on non-traditional sectors than on traditional sectors. Finally, it can be seen that imports contract so that foreign savings decline while exports remain roughly constant.

The impact on employment in shown in Table 5. It can be seen that employment in the primary sector is better protected during the contraction due to the export clearing nature of these sectors. Non-primary sectors show more of a decline, also due to their energy intensity, as discussed above. Higher inflation leads to a drop in real wages. Both skill categories appear to suffer a similar decline.

Note that virtually all the environmental improvement can be classified as extensive rather than intensive, due to the very small change in the distribution of income. The last column of Table 5 suggests only a microscopic improvement in the overall distribution of household income. The reason is that the economic contraction causes income from profits to fall and its distribution favours high-income households. The decline in the extensive damage function is probably a fairly accurate measure of the decline in the rate of environmental deterioration. The details of the change in the environmental measures are shown in the next two tables.

TABLE 5 Employment and wages with cleaner electricity (ratios to base run)

Source: model computations

¹Absolute differences from the base run.

Table 6 reveals that the contraction brings about a significant decline in levels of liquid emissions. Significant reduction in gaseous emission is shown in Table 7. Both toxic as well as greenhouse gases are about 10% lower at the end of the period compared to the base run. The main contribution to the lower emissions are made by SOz, NO2 and CO₂. Harmful particulates and CO report less substantial reductions, whereas the reduction in CH₄ is negligible.

We conclude that while environmental clean-up will trade off with output and employment, the results of this computable general equilibrium study suggest that the cost is not excessively high. For average loss of about 1% of GDP, locally harmful emissions can be reduced by an average of 6% and greenhouse gasses by 4%, where the implied elasticities refer to extensive deterioration. Moreover, the composition of the improvement favours domestic versus global residents. The latter benefit, of course, but the close link between domestic costs and domestic benefit would unquestionably enhance the political feasibility of the policy.

TABLE 6 Liquid contaminants with cleaner electricity (ratios to base run)

Source: model calculations

Green trade restrictions on traditional exports

Mining is the sector most vulnerable to trade restrictions based on environmental destruction, at least on a process and production method basis. Coal is also a major export commodity for South Africa and might incur green trade restrictions on the basis of consumption externalities. Direct regulation of mining to meet an environmental goal will be costly to the economy in foregone output, employment and foreign exchange earnings. As noted above, to measure the effects of green trade restrictions on the macro-performance of the economy, we reduce the output of mining to the point that the value of the damage function is the same as in the first experiment. The impact on the export sector of the South African economy is shown in Table 8, measured in terms of deviations from the base run.

Primary exports fall substantially, dropping to only 90,3% of its base run value by 1999. Surprisingly, non-traditional exports are also lower. As will be seen in the next table, the decline is attributable to lower economic activity and, through the accelerator, lower capacity growth. Since non-traditional exports are linked to capacity output, slower growth translates into lower non-traditional exports. The combined result is that total exports are about 7,5% lower compared to the base run at the end of the period. In the last column it can be seen that after the second year imports decline more rapidly. As a result, foreign savings are lower compared to the base run at the end of the period.

TABLE 7 Gaseous emissions with cleaner electricity (ratios to base run)

Source: model calculations

A decline in output of the primary goods sector sufficient to improve the environment to the level of the first simulation would be clearly harmful to the economy. As seen in the Table 8, GDP falls to 92,4% of the base run versus 98,1% in the first simulation. On average GDP falls by 4% rather than 0,9% with the electricity clean-up. With lower output, inflation also falls since the cost push is absent here and labour market pressures are reduced. Note further the dramatic drop in the rate of investment. This is in part a response to the decline in the level of activity, acting through the accelerator. The fall in investment would be worse were it not for lower interest rates.

The decline in economic activity has a dramatic impact on public finance. Public investment, which adjusts to the given public sector borrowing requirement to GDP ratio, falls to only 30,6% of the base run. One reason is the loss in mining output which causes a precipitous decline in tax revenues along similar lines discussed by Makgetla (1995). Lower inflation also reduces the indirect tax take. Furthermore, with constant government expenditure in real terms—despite lower real wages for government employees—there is less left over for public sector investment. The drop in GDP aggravates the decline since the public sector borrowing requirement must maintain the constant ratio to GDP. Moreover, since government investment crowds in private investment in the model, the latter collapses.

TABLE 8 Green trade restrictions on traditional exports (ratios to base run)

Source: model computations

¹ Absolute differences from the base run (in percentage points).

Table 9 shows that the blow to the mining sector causes economy-wide employment levels to suffer. Primary employment is hit worse than non-primary, but only marginally. Both skilled and unskilled labour share the pain of rising unemployment but, for those who keep their jobs, lower inflation results in a small increase in real wages. The real wages of skilled labour are more sensitive to excess capacity and, therefore, fall faster. The final column of the table shows a decline in the Gini coefficient which suggests that income is more equally distributed, albeit with a smaller pie. Thus, any intensive gains will add to the extensive improvement made here.

The general conclusion of Table 9 is that green trade restrictions on the primary sector must be avoided. To achieve a significant improvement in environmental quality by direct restrictions on mining exports would be very costly to South Africa. While the impact of green trade restrictions on primary exports would be far more contractionary than the electricity clean-up studied above, it remains to be seen whether some significant gains in local environmental quality would accompany the restrictions.⁶

Table 10 suggests that cutting back on mining production will not reduce total liquid contaminants production as much as in the electricity experiment. Although the production of total dissolved salts is lower, suspended solids and oxygen-demanding products are higher, affecting local water quality. Thus, on the one hand, emissions of total dissolved salts are much higher in mining than in electricity; on the other hand, the emissions of suspended solids and oxygen-demanding products are higher for electricity.⁷

⁶As noted, the overall impact is the same as in the first simulation.

TABLE 9 Employment and wages with green trade restrictions on traditional exports (ratios to base run)

Source: model computations

¹ Absolute differences form the base run.

Table 11 shows that restricting mining output fails to improve toxic gas emissions as much as with the clean-up of the electricity sector. Greenhouse effects are about the same. The emission of particulates, NO2 and CO, are reduced to a larger degree compared to the first simulation. However, emissions of SO2 and CO₂ are not reduced as much. On the other hand, primary sectors are major producers of CH₄, as is seen in the table. Given equal weights in the damage function, this results in the same reduction of overall environmental damage, with the ratio of benefits slightly more in favour of global versus local residents. The more rapid improvement in the Gini coefficient suggests that extensive and intensive environmental decay do not trade off in South Africa, but rather work in the same direction. Some minor improvement could then be expected in intensive deterioration.

These caveats aside, we conclude that green trade restrictions show no significant special advantages over the cleaner energy option with respect to the environment. While overall contamination remains constant by assumption, the benefit from the change in composition hardly seems worth the increased sacrifice required to achieve green trade.

⁷ This suggests that first-round or direct effects are dominating the final outcome with indirect emission effects playing a secondary role.

TABLE 10 Liquid contaminants with green trade restrictions on traditional exports (ratios to base run)

Source: model calculations

Green trade restrictions on non-traditional exports

As a non-tariff barrier, green trade restrictions might well target manufactured goods rather than raw materials. Primary exports could escape restrictions, while the more competitive goods are excluded. The previous section clearly shows that if green trade restrictions are imposed on primary goods, the effect on the economy would be contractionary. In this section we measure the impact of green trade restrictions on manufactured or non-traditional exports.

An average reduction of 12% of non-traditional exports is necessary to achieve the target value of the environmental damage function, as seen in Table 12. Note that the effect on GDP is slightly less contractionary than for traditional exports. This is due to the small rise in primary exports observed in the second column of the table. With the general contraction of the economy, intermediate demand for primary goods falls. Since the primary sectors always operate at full capacity, the backward linkage from the non-primary to the primary sector leaves a higher fraction of output available for primary exports. In the previous simulation, the restriction on primary output was not offset by a higher level of output in the non-primary sector since the loss of purchasing power in the former reduced demand for the latter. Green trade restrictions on the primary sector spilled over to the non-primary sector. Here the effect of the green trade restrictions is entirely borne by non-traditional exports, and this is the reason why the impact is not as contractionary. The structural character of the economy, with strong forward linkage running from the primary to the non-primary sector, but no linkage in the opposite direction, is principally responsible for this result.

TABLE 11 Gaseous emissions with green trade restrictions on traditional exports (ratios to base run)

Source: model calculations

The data of Table 12 indicate that the macroeconomy responds in much the same way as seen in the prior simulation. Inflation slows, but not by as much and the same for interest rates. The accelerator insures that investment will not fall by as much in the less contractionary environment. The public sector accounts behave in the same way as above. Public investment contracts with loss of tax revenue, but the effect is not quite as severe.

Table 13 shows that the strength of the forward linkages in that employment falls in the non-primary sectors by about the same amount as with green trade restrictions on traditional exports. The primary sectors, agriculture and mining, are less affected as they mainly have a powerful, unidirectional impact on the rest of the economy.⁸ As in the previous experiment, real wages rise. Again, unskilled workers benefit the most since their wages are more responsive to inflation and less to labour market conditions than skilled workers’. The last column of Table 13 shows that the income distribution has improved somewhat relative to the two previous simulations, further reducing the rate of intensive decay. This result is consistent with the rise in real wages for unskilled labour, which is not accompanied by a steep drop in unskilled employment.

⁸ For additional computable general equilibrium evidence of the power of the mining sector on the South African economy, see Gibson and van Seventer (1996a).

TABLE 12 Green trade restrictions on non-traditional exports (ratios to base run)

Source: model computations

¹ Absolute differences from the base run (in percentage points).

Again we ask if the composition of environmental damage is significantly altered by the green trade restrictions on non-traditional versus traditional exports. Non-traditional sectors clearly produce more liquid contaminants than their traditional counterparts. As can be seen in Table 14, the production of suspended solids and total dissolved salts is higher than in scenario 2, although it is counterbalanced by slightly lower oxygen-demanding products.

Compared to mining, non-traditional sectors produce on average more toxic gaseous emission when their exports are reduced. The reduction in greenhouse gases, on the other hand, is slightly lower but the CO₂ level is exactly the same. Methane production is higher compared to the mining simulation. Both mining and agriculture are significant CH₄producers, and when output is reduced, levels fall more dramatically.

Non-traditional exports are dirtier locally and cleaner globally than traditional exports and it is, therefore, clearly in the interests of South Africans that they should be cleaned up. Green trade restrictions on non-traditional exports will have less of an impact on the macroeconomy and will still reduce globally damaging emissions. Since they will inhibit competition in manufactured goods but not threaten the flow of raw materials, green trade restrictions are much more likely to be imposed on non-traditional exports (McKenzie and Foster 1995: 2). While green trade restrictions on non-traditional exports would clearly be less damaging than on traditional exports, the cost to South Africa is still high compared to the costs of cleaning up the energy sector.

TABLE 13 Employment and wages with green trade restrictions on non-traditional exports (ratios to base run)

Source: model computations

1 Absolute differences from the base run.

Conclusions

This paper shows that the interaction of the macroeconomy and the environment is a complex and subtle process that is only barely within the grasp of analysts. We call it complex because the macro-environmental interface is wide-ranging, highly diverse and uncertain.

Aggregating a fraction of environmental effects in a damage function, as done here, is clearly limited in its ability to convey the nature of environmental problems confronting South Africa. While the weights of the damage function appear to be in need of adjustment, few guidelines are available; it is not even clear how to weight local versus global environmental threats. We call it subtle because conventional analytical approaches seem inadequate to the task of properly evaluating the trade-offs involved. The relevant model must blend short and long-run considerations appropriately. An exclusively short-run framework will over-emphasise the macroeconomic costs of environmental preservation, while the traditional neo-classical analysis of the long-run places undue emphasis on individual utility maximisation, usually in a sterile informational and institutional setting (Heal 1984). The model of this paper tries to strike a balance between the short-run, in which the environment is “held constant”, and a long-run in which environmental destruction is fully anticipated and properly discounted.

TABLE 14 Liquid contaminants with green trade restrictions on non-traditional exports (ratios to base run)

Source: model calculations

Nonetheless, there are some serious shortcomings that must be acknowledged. Firstly, there is little feedback from environmental variables to macro-variables in the base run. Pollution matters, but the model never “sees” the effects on productivity, health, biodiversity or anything else. In this way it may be argued that the model is biased towards the short-run. Secondly, the model avoids the highly unrealistic assumption that all resources are fully utilised and, therefore, environmental protection bears a heavy opportunity cost. There is excess capacity and unemployment so that the principal constraints come from the institutional environment in which the economy functions. The paper shows that it is impossible to evaluate the effects of green trade restrictions on the economy independently of assumptions about the macroeconomic environment. Given the complexity of the institutional and associated policy-making framework currently governing South Africa, it should not be surprising that the opportunity cost of environmental protection is less than in the conventional neo-classical analysis.

These caveats aside, the message of the simulations in this paper is that the threat of green trade restrictions on the macroeconomy cannot be taken lightly. If green trade restrictions are imposed either on mining or non-traditional exports, the consequences for macroeconomic performance will be significant. It makes some, but not much, difference if the green trade restrictions are levied against raw materials versus manufactured goods. The message of this paper is that the cost of green trade restrictions to South Africa will be very high when compared to the costs of clean up. The simulations show that the differences in the production of greenhouse gases and, thus global warming, will be almost negligible, nor is there much to be gained in changing the composition of local versus global externalities.

TABLE 15 Gaseous emissions for green trade restrictions on non-traditional exports (ratios to base run)

Source: model calculations

While the macroeconomic effects of green trade restrictions are striking, it should be noted that only extensive environmental decay is measured here. Moreover, intensive degradation in South Africa is unquestionably local in its impact. Thus, to the extent that policies raise the Gini coefficient, the burden of environmental decay is not measured properly by the damage function. Typically, a deterioration in the distribution of income forces the poor and unemployed to increase the rate at which the environment is exploited. A contraction in output substitutes intensive for extensive environmental assault, and the environmental protection inherent in lower levels of economic activity is undermined. In the simulations above, changes in real wages overpowered the effects of unemployment and the distribution of income improved. It is unlikely, however, that this effect would persist if output were further reduced. Eventually, green trade restrictions would reduce employment to the point that intensive degradation, most likely of soil and water supplies, would increase dramatically. This observation further strengthens the general conclusion that South Africa must do everything within its power to minimise the risk of green trade restrictions being implemented, including cleaning up, before they become a real threat.

References

Booth, A., “State of the environment in southern Africa”, report by the Southern African Research and Documentation Centre, Zimbabwe, 1994.

Deloitte & Touche Consortium, “Environmental constraint scenarios for world trade”, Research Report 7, Project: The Use of Environmental Resource Economics in Environmental Impact Management, Department of Environmental Affairs and Tourism, Pretoria, 1994.

Eberhard, A. and van Horen, C., Poverty and power: energy and the South African state, Pluto Press, London, and UCT Press, Cape Town, 1995.

Gibson, B., and Van Seventer, D.E.N., “The DBSA macromodel”, DBSA Occasional Paper No. 120, Halfway House, Development Bank of Southern Africa, 1996.

Gibson B., and Van Seventer, D.E.N., “Trade, growth and distribution in the South African economy”, Development Southern Africa, Vol. 13 No. 4,1996.

Gibson, B., “The environmental consequences of stagnation in Nicaragua”, World Development, Vol. 24 No. 2,1996.

Heal, G.M., “Interaction between economy and climate: a framework for policy design under uncertainty”, in Smith, V K. and White, A.D. (eds), Advances in applied microeconomics, Greenwich, J.A.I. Press, 1984, pp. 151-158.

Karshenas, M., “Environment, employment and technology: towards a new definition of sustainable development”, Department of Economics, School of Oriental and African Studies, London, 1992.

McKenzie, C. and Foster, S., “International trade and the environment”, research report compiled for the Development Bank of Southern Africa, Halfway House, 1995.

Pezzey, J., “Sustainable development concepts: an economic analysis”, Mimeo (Washington, D.C.: World Bank, 1992.

President’s Council, “Report of three committees of the President’s Council on a national environmental management system”, Government Printers, Cape Town, 1991.

Seidman Makgetla, N., “Reconciling fiscal discipline with a Utopian vision”, Business Day, 28 April 1995.

Taylor, L., “Sustainable development: an Introduction”, World Development, Vol. 24, No. 2., 1996.

World Bank, “China: issues and options in greenhouse gas emissions control”, summary report prepared by a joint study team from the National Environmental Protection Agency of China, the State Planning Commission of China, the United Nations Development Programme and the World Bank, 1994.

World Resources Institute, World Resources: A Guide to the Global Economy, New York, Oxford University Press, 1995.

Document(s) 10 of 15