A measure of sustainable national income for the Netherlands

Ecological Economics 41 (2002) 157– 174 This article is also available online at: www.elsevier.com/locate/ecolecon

ANALYSIS

A measure of sustainable national income for the Netherlands Reyer Gerlagh a,*, Rob Dellink b,1, Marjan Hofkes a, Harmen Verbruggen a a

IVM/VU, Institute for En6ironmental Studies, Vrije Uni6ersiteit, De Boelelaan 1115, 1081 HV Amsterdam, The Netherlands b En6ironmental Economics Group, Wageningen Uni6ersity, Hollandseweg 1, 6706 KN Wageningen, The Netherlands Received 20 August 2001; received in revised form 7 January 2002; accepted 4 February 2002

Abstract We present calculations on the sustainable national income (SNI) indicator, first proposed by Hueting, which corrects net national income (NNI) for the costs to bring back environmental resource use to a ‘sustainable’ level. Using an applied general equilibrium (AGE) model specifying 27 production sectors, we calculate different variants of SNI for the Netherlands in 1990, given a set of pre-determined sustainability standards. The AGE model is extended with emissions and abatement cost curves, based on large data sets for nine environmental themes. The model combines the advantages of a top-down approach (the AGE model) with the information of a bottom-up approach (the environmental data and data on emissions reductions costs). The presented numerical results show that in 1990 Dutch SNI is about 50% below NNI, though many uncertainties are still present in the data and the model. The enhanced greenhouse effect is the most expensive environmental theme. © 2002 Published by Elsevier Science B.V. Keywords: Accounting; Economic welfare; Green GDP; Sustainable income

1. Introduction Modern national accounting is based on the equality of total income (wages, rents, interest, and profits) and total expenditures (private and public consumption, investments, taxes, and net export spending). In principle, all measurements * Corresponding author. Tel.: + 31-20-4449555; fax: + 3120-4449553; www.vu.nl/ivm. E-mail addresses: [email protected] (R. Gerlagh), [email protected] (R. Dellink). 1 Tel.: + 31-317-482009; fax: + 31-317-484933; http:// www.sls.wau.nl/me/staff/dellink/dellink.htm

involved should be based on observable transactions in the market. However, there is a long history of critique, which argues that aggregates such as Gross Domestic Product (GDP) and Net National Product (NNP) are often used as measures of welfare and that such use is incorrect and may provide policy makers with the wrong signals. In a famous paper, Nordhaus and Tobin (1972) discuss several corrections of conventional NNP in an early quantitative attempt to develop a more correct Measure of Economic Welfare (MEW). Most importantly, they find that nonmarket activities such as leisure time substantially

0921-8009/02/$ - see front matter © 2002 Published by Elsevier Science B.V. PII: S 0 9 2 1 - 8 0 0 9 ( 0 2 ) 0 0 0 2 1 - 6

158

R. Gerlagh et al. / Ecological Economics 41 (2002) 157–174

contribute to welfare but do not show up in the accounting aggregates. Of special relevance for this paper, they also correct income for environmental damage costs, and more explicitly for the disamenities of urbanization. Nordhaus and Tobin find that for the period 1929– 1965 per capita MEW grew by 1.1% per year, 0.6%-points slower than per capita NNP. They conclude that the NNP is an acceptable measure, not precisely fitting welfare, but describing the magnitude of growth correctly. Daly and Cobb (1989) further elaborated upon the concept to take better account of environmental damages and social inequity, presenting the Index of Sustainable Economic Welfare (ISEW). They find for the period 1951– 1986 an increase of the per capita ISEW of 0.53% per year, as opposed to a GDP growth of 1.9% per year. Notably, the ISEW declined during the 1980s, whereas the GDP level continued its growth. Thereby, Daly and Cobb make the point, contrary to Nordhaus and Tobin, that common aggregates such as GDP are not good measures of welfare and are misleading when used for evaluating economic growth. Since then, there have been many environmental accounts, most of them focusing on valuing environmental damages; we mention Repetto et al. (1989), Castan˜ eda, 1999; Cruz and Repetto (1991, 1992) and more recently, Thampapillai and Uhlin (1997), Liu (1998), Torras (1999). In these studies, a typical finding is that numerical results strongly depend on assumptions on both the future valuation of environmental resources and the future substitutability between man-made goods and environmental goods.2 As it stands, there is still no set of commonly accepted assumptions that provide the researchers with guidelines for empirical analyses in this field. A number of authors have examined the assessment of sustainable national income (SNI) from a theoretical perspective. Often, a reference is made to Hicks (1948) early definition of sustainable income as the maximum value that a man can consume in one period without impoverishing himself. Two famous contributions following up on Hicks’ analysis are Weitzman (1976), who showed 2 See Gerlagh and van der Zwaan (2000) for an analytical study on the topic of substitutability and environmental valuation.

equivalence between the net product (consumption plus net investments) and sustainable income, and Hartwick (1977), based on Solow (1974), who made clear that a non-decreasing consumption flow requires a non-negative net investment flow. However, their results require strong assumptions to be fulfilled and break down with multiple consumers, non-constant time preferences, technological change and distortionary taxes (Withagen, 1998). This is something of a paradox. The analytical studies show us how to calculate sustainable income in a perfect competitive world without externalities, but the very reason for constructing a sustainable income measure is that it is believed that the current economy is far from such an efficient path (Aaheim and Nyborg, 1995). Uncertainty adds to the problem. Most of the analyzes assume perfect foresight, that is, all relevant aspects of the future are known today with certainty. Yet even if there were a complete understanding of the economic and physical mechanisms at play, no unique future and no unique sustainable income would exist. For example, variations in the distribution of property rights for goods that are currently not in the market can cause substantial variations in equilibrium allocations and prices, see Howarth and Norgaard (1992), Gerlagh and Keyzer (2001), Gerlagh and van der Zwaan (2001). In short, these formal results are less useful when policy makers ask for applied numerical calculations that can help them in developing sustainability policies (Hueting and de Boer, 2001). Further comments on green accounting estimates can be found in (Aronsson and Lofgren 1998; Holub et al., 1999; Nordhaus et al., 2000; Van Ierland et al., 2001). In this paper, we present calculations for Hueting’s (1980, 1992, 1995) SNI indicator, a ‘green’ income measure that avoids problems related to an uncertain future, and specifically to uncertain future preferences. Hueting suggests that we should assume an absolute preference for conservation of the natural environment. He argues that under this assumption, the value of environmental degradation is equal to the conservation costs, i.e. the costs to preserve the environment and remove existing environmental burden. In this sense, the SNI indicator resembles the

R. Gerlagh et al. / Ecological Economics 41 (2002) 157–174

maintenance cost approach (UN, 1993). The emphasis on resource conservation, and the determination not to count revenues from resource depletion as income, gives the SNI indicator a flavor of the strong sustainability paradigm. In the context of accounting systems, El Serafy (2001) opposes this position. Alternatively, he asserts that revenues from resource depletion are to be counted as income, but that this income should be distributed over present and future generations equally, an approach labeled the ‘user cost method’ (Ahmad et al., 1989). In this paper, we will not give a full account of the differences in underlying presumptions, but we confine ourselves to the note that the SNI indicator will typically calculate a lower income value then the user cost method, the difference in value dependent on the speed of resource depletion. The gap between the NNI and the SNI level measures the dependence of the economy on that part of its natural resource use that exceeds the sustainable exploitation levels. If the NNI level increases substantially while the SNI level increases less, that is if the gap between the two measures increases over time, it follows that the basis for economic growth is unsustainable. Growth is then mainly driven by an increase in natural resource use, and the dependence of the economy on overexploitation of natural resources increases. On the other hand, if the gap between the NNI level and the SNI level decreases over time, this points to a decrease in the economy’s over-dependence on natural resources. For policy makers, who are mainly interested in the economic and political feasibility of environmental regulation, an increase in the gap signifies that an increasing effort will be required to implement actual sustainability measures, while a closing of the gap indicates a decrease in the economy’s dependence on natural resources. In this sense, the dynamics of the SNI vis-a`-vis the NNI is of apparent relevance for actual environmental policy.3

3

To be sure, we do not intend to discuss the fundamental question whether continued economic growth can go together with sustainable resource use, or that this is essentially impossible. In this paper, we do not take a position on this subject.

159

From a methodological point of view, we can distinguish two steps in making the calculations of a SNI indicator. First, sustainable resource use is defined and compared with actual resource use. This procedure is based on insights from natural sciences (Hueting and Reijnders, 1998).4 Second, direct and indirect changes in income caused by the required changes in resource use are calculated by use of an economic model such as an Applied General Equilibrium (AGE) model. The second step is familiar to many economic analysts. However, it requires an extension of the standard AGE model to account for the specific data on the interaction between environmental and economic variables. And in this respect, our analysis also contributes to the literature as we describe and use a new methodology for incorporating extensive bottom-up data on abatement costs for various environmental themes within a top-down general equilibrium model. Yet we emphasize that our motivation and interpretation of results differs in an important way from common exercises with AGE models. In general, AGE models are used to calculate economy-wide consequences of specific policy instruments, for example energy taxes or carbon emission taxes, which aim at achieving certain environmental objectives, e.g. Jorgenson and Wilcoxen (1993a,b), Boyd and Uri (1991). In such a context, the AGE model is used to test the economic feasibility of specific policy instruments. Accordingly, a change in gross output of 1 or 2% of GDP is considered substantial. If we were to interpret our calculations as a standard policy analysis, we would say that we simulate an economy that moves towards a strong sustainability policy where resource use is substantially reduced, and we are not to be surprised if gross output and income decrease by 10, 20%, or even more. But we do not exercise a standard environmental policy analysis; we do not suggest that the immediate reduction of resource use that lies at the basis of 4 It should be noted that the assessment of the sustainable level of resource use, which we label sustainability standards, falls outside the scope of this study. We consider the sustainability standards as given and thus exogenous to our model calculations.

160

R. Gerlagh et al. / Ecological Economics 41 (2002) 157–174

our calculations is politically feasible. Instead, we use the model to calculate a measure of adjusted income, the SNI according to Hueting. In Section 2, we present the model. It has 27 sectors, it describes nine environmental themes, and it includes abatement cost curves, based on large data sets and describing the direct costs per sector of resource reduction, for all environmental themes. In Section 3, we elaborate on the methodological choices and modeling assumptions that are required. In Section 4, we present the numerical results. We analyze the sensitivity of our results regarding several assumptions, and we present a break up of the sustainable income along lines of sectors, production factors, and expenditures. In Section 5, we conclude. A comprehensive description of the model, its assumptions and calibration, and the results can be found in Gerlagh et al. (2001), Dellink et al. (2001) and Verbruggen et al. (2001), respectively.

2. Model set up and calibration

2.1. Description of the model Initially, Hueting (1980) expected that the calculation of his sustainable income measure could be achieved by use of purely statistical methods. He gathered that it was sufficient to collect data on the opportunities to reduce environmental resource use, either through technical measures or through the reduction of through-put, which he labeled volume measures. However, the magnitudes of changes in allocations are too substantial for such a statistical approach. If environmental losses have to be prevented or restored, an entirely different economic structure results. We then envisage a hypothetical sustainable economy with a hypothetical income. This can only be approached through model calculations. For that purpose, a static AGE model for the Dutch economy has been constructed to calculate a SNI indicator. The model has 27 sectors, and is extended to account for the use and abatement of nine environmental themes. The SNI– AGE model identifies domestically produced goods by the sectors

where these goods are produced. There are two primary production factors, labor and capital.5 The model distinguishes three consumers: the private households, the government, and the Rest of the World (ROW). In addition to these producers and consumers, there are several auxiliary agents that are necessary to shape specific features of the model. In order to capture non-unitary income elasticities in the model, the consumption of the private households is split into a ‘subsistence’ and a ‘luxury’ part. There is an ‘investor’ who demands investment goods necessary for economic growth, and a ‘capital sector’ which fabricates the composite capital good. Trade is modeled using the Armington specification for imports and a Constant Elasticity of Transformation (CET) production structure for sectors producing for both the domestic and the world market.6 Besides the model elements mentioned above, common to many other AGE models, the model distinguishes environmental themes such as the enhanced greenhouse effect and acidification. With each of the environmental themes, emission units are associated, e.g. greenhouse gas emissions are expressed in CO2 equivalents. The overview of the relationships in the model is presented in Fig. 1. In the figure, black arrows represent commodity flows that are balanced by inverse income flows; gray arrows represent pure income transfers that are not balanced by commodity flows. Demand and supply meet on the markets for goods and factors. The private consumers supply endowments (labor), which are used as inputs for the producers. The producers supply output of produced goods, which balances consumption by

5 In fact, capital is produced. The model accounts for maintenance costs and net investments. 6 The CET production function is used for production processes with multiple output goods. In analogy to the CES production function, it is assumed that the relative change in output for the various output goods is proportional to the relative change in prices. For example, if there are two goods and their initial output levels are the same, then if the price of the first good increase by 1%, the relative output level of the first good will increase with |%, where | is the elasticity of transformation parameter.

R. Gerlagh et al. / Ecological Economics 41 (2002) 157–174

161

Fig. 1. Overview of SNI – AGE model.

the private and public consumer and inputs for gross investments. Part of these investments reflects the depreciation of the capital stock, the remaining part, net investments, is used to sustain economic growth in the next period. The figure also shows the market for emission units, supplied by the government in an amount that is consistent with the sustainability standards. Hence, the revenues from the sale of emission units enter the government budget. The government levies taxes on consumption (VAT), the supply of endowments (labor income tax), and capital use (profit income tax). These public revenues balance, together with revenues from the sale of emission units, the public expenditures that consist of public consumption and lump sum subsidies for social security. Consumers spend their income from the sale of endowments and lump sum subsidies on consumption and net savings. Net savings are transferred to the ‘investor’, who spends it on the consumption of capital goods (thus: savings equal investments). Production technologies are assumed to have constant returns to scale, which implies that profits, apart from a rate of return on capital, are zero, and hence, that the value of inputs is equal to the value of outputs. In Fig. 1, this is visualized by placing a gray box around the agents, over which the net income and expenditure flows sum to zero. The same applies to clearing markets,

where (the value of) total supply matches total demand. This is visualized by a gray ellipse. By a careful examination of the income flows in Fig. 1, we find that the budgets close, except for the budget balances of the private and public consumers. This is due to the omission of international trade from the figure. For the domestic economy as an entity, the budget surplus is equal to the surplus on the trade balance, represented through the well-known identity Y= C + I+ (X− M), where Y − C− I is the income surplus of the consumers compared to the expenditures on consumption and investments, and (X− M) is the surplus of export compared to the imports. Of course, in case of a budget deficit the opposite holds.

2.2. Calibration of the model The model is calibrated using historical data for 1990 for the Netherlands, provided by De Boer (2002). The main data source is the NAMEA accounting system (Keuning, 1993), which captures both the economic and environmental accounts. To save space, we present a condensed Social Accounting Matrix (SAM) in Table 1. Net National Income (NNI) at market prices amounts to 457 billion guilders. The value of goods produced by the agricultural sectors amount to 37 billion guilders, of which 21 billion guilders accounts for the value of inputs, and 16 billion guilders (3+12+ 1) is labeled value added. The

R. Gerlagh et al. / Ecological Economics 41 (2002) 157–174

162

Table 1 Reference SAM for 1990 for the Netherlands (billion guilders) Agriculture

Industry

Agriculture Industries Services Capital

37 −12 −4 −5

−27 243 −66 −19

−1 −63 373 −36

Labor Profits Taxes

−3 −12 −1

−68 −44 −18

−158 −83 −31

−10

0

0

0

0

Sum

Services

Capital

Trade Balance

−0 −84 −17 111

−5 3 −14

Net investments

Consumption Endowments Sum

0 0 0 0

−4 −86 −272 −51

−16

−51

−27

230 139 88

0 0 0

−389

457

0

Each row represents a good identified by the sector producing it, or a production factor, respectively. For each row, the table entries should sum to zero representing the commodity balances. Columns represent agents. For each producer, table entries should sum to zero representing the zero-profit condition. The sum of table entries for other columns (not summing to zero) represents income and expenditures.

agricultural sector has a minor share in value added of about 4% of NNI (16 billion guilders divided by 457 billion guilders). Yet, the Netherlands owes one-third of its trade surplus (5 billion guilders divided by the overall trade surplus of 16 billion guilders) to the agricultural sector. The industries have a share of about 31% in value added, and the services produce 65% of value added. Gross investments amount to 111 billion guilders, more than half of which is for maintenance; net investments amount to 51 billion guilders. Capital goods are mainly produced by industry. About half of total income (after taxes) is attributed to labor, taxes account for nearly 20% of income, and capital returns receive the remaining 30%. As mentioned above, the AGE model includes nine environmental themes: climate change, depletion of the ozone layer, acidification, eutrophication, fine particles (PM10) in air, smog formation through volatile organic compounds (VOC), dispersion of heavy metals and PAKs/PCBs to water, dehydration of land, and soil contamination. For all these themes, data have been collected on actual emission/pollution levels7 and on the costs of available technical measures to prevent the 7 We use the terms emissions and pollution interchangeably to indicate the annual burden on the environment, even though we realize this terminology is not entirely correct.

environmental problems from happening or to restore the environmental quality. These data are described in abatement cost curves. Methodological issues related to these curves (e.g. on discounting) are extensively described in Dellink et al. (1997). From a modeling perspective, the inclusion of abatement measures within an AGE model is the major extension of our analysis compared to the literature. Recall that emission units are treated as production factors, similar to labor and capital, since an enforced reduction of emissions decreases output. A typical production

Fig. 2. Nested CES production structure with abatable and non-abatable emissions. For convenience, in the CES structure, we have not drawn the multiple lines associated with the environmental themes separately. For each environmental theme, a distinction is made between abatable and unabatable emissions, and abatement measures are treated as a substitute for abatable emissions.

R. Gerlagh et al. / Ecological Economics 41 (2002) 157–174

Fig. 3. Iso-output curve: the trade off between abatement and emissions.

structure employed in the model is presented in Fig. 2. Capital and labor are combined in a capital –labor composite good, with substitution elasticity |2. The values for the elasticities are based on the Dutch TaxInc model (Keller, 1980; Statistics Netherlands, 1990) and are sector specific. The values for |2 range from 0 to 0.9. Intermediates are combined into one composite intermediate good, with substitution elasticity |3, the value of which is typically in the same range as |2. The capital– labor composite and composite intermediate good are combined to produce the sector specific output good, with elasticity of substitution |1, the value of which lies in the range between 0 and 2. Some fraction of sectoral emissions cannot be reduced through technical measures. These ‘non-abatable’ emissions are proportional to output, and enter the nested CES structure at the highest level, at which there are no substitution possibilities. The remaining part of emissions is ‘abatable’, that is, they decrease if the input of abatement goods is increased. In modeling terms, abatement measures are a substitute for the abatable emissions; the substitution elasticity is denoted by |4. Viewed from another perspective, we can draw the so-called iso-output curve that portrays the trade off between abatement measures and abatable emissions, given a fixed output level (Fig. 3). By definition, the mirror image of the iso-output curve is the abatement cost curve. Abatement costs increase if the emission level has to decrease. The slope of the curve represents the marginal costs of technical options that are open to the agents for reducing their emission levels.

163

The CES approximation of the abatement cost curve turns out to fit the data well. As an example, we describe the cost curve for greenhouse gases (GHG). The GHG that cause climatic change are mainly carbon dioxide (CO2), methane, nitrous oxides, and CFCs and halons. The effects of these substances on climate change, as well as the duration of their effects, vary. The way in which these GHGs can be aggregated into CO2 equivalents is not unambiguous, but depending on the mix of emissions (and emission reductions). The coefficients that were chosen to aggregate the GHGs into CO2 equivalents are based on longterm Global Warming Potentials and are described in De Boer and Bosch (1995). Technical measures and costs to reduce fossil fuel use, and thus CO2 emissions, were taken from the ICARUS database (Blok, 1991; Blok et al., 1991), which comprises about 300 measures, ranging from more efficient energy use and co-generation to local solar power systems, and from the MARKAL model (Okken, 1991; Okken et al., 1992). Measures to reduce methane emissions were collected from various sources (see De Boer and Bosch, 1995) and comprise changes in the composition of animal fodder, more efficient use of manure, measures in the production and distribution of natural gas, and measures at waste dumps. The measures of changing animal fodder and of more efficient management of manure are also effective for reduction of nitrous oxides. And the measures to reduce CFCs and halons consist of replacing them by HCFCs (with much lower warming potential) or by other substances (VROM, 1994, 1998).

Fig. 4. Total annual costs of reduction of GHG.

R. Gerlagh et al. / Ecological Economics 41 (2002) 157–174

164

Table 2 Sustainability standards for the Netherlands, 1990 Environmental theme

Unit

Base 1990

Sustainability standard

Reduction (%)

Greenhouse effect Ozone depletion Acidification Eutrophication Fine particles in air Smog formation Dispersion to water Dehydration Soil contamination

bln kg CO2-eq mln kg CFC11-eq bln Acid eq mln kg P−eq mln kg PM10 mln kg NMVOS billion AETP eq % affected area c contaminated sites

251 10.4 38.4 312 440 44 194 100 600,000

53 0.5 10.0 128 240 20 73 0 0

198 (−79%) 9.8 (−94%) 28.4 (−74%) 184 (−59%) 200 (−46%) 24 (−55%) 121 (−62%) 100 (−100%) 600,000 (−100%)

Fig. 4 shows the resulting abatement costs curve. Data collected from the database of technical measures are plotted as diamonds and an estimated approximation through a CES curve is plotted as a solid line.8 As the figure shows, the approximation fits well. For the other environmental themes, a similar procedure is followed. Table 2 presents the sustainability standards for the various environmental themes which function as a reference for the alternative SNI calculations. Comparing the required reduction with available reduction measures as presented in Fig. 4 for GHG, we see that of the 198 bln kg CO2 equivalents, only about 90 bln kg can be realized through technical measures. The remainder of the reduction has to be realized through a restructuring of the economy.

3. Methodological choices and assumptions Given the AGE model, calculation of the sustainable income follows the same procedure as a classic policy analysis, in which one studies the consequences of a policy that strictly observes environmental sustainability standards. It is then necessary to make assumptions as to the time scale (e.g. static versus dynamic modeling), transition costs, labor market, international trade, emission reduction measures, ‘double counting’, private consumption and government budgets. 8 The database contains 332 measures; each tenth is plotted as a diamond.

Below, we briefly explicate the choices made. We have to be aware that results may significantly depend on the actual assumptions. It is thus not possible to speak of the result as the unique SNI; preferably, we speak of a SNI calculation. Correcting NNI for environmental losses is meant to be a strictly static approach and we therefore use a static general equilibrium model. The SNI calculations are not burdened with other costs than environment-related loss of functions. To arrive at a sustainable economy in the real world, a drastic restructuring and reallocation of economic activities has to take place. This inevitably involves a premature write-off of capital goods (transition or adaptation costs). However, these non-environment-related costs do not enter the SNI. In a way of speaking, it is assumed that the change to a sustainable economy is foreseen in advance, long enough that economic agents are able to integrate this transition in the planning of their investment decisions. By this way of reasoning, it is implicitly assumed that the early announcement enhances the substitution possibilities in the economy. This, in turn, should be expressed by applying medium to long-term substitution elasticities in the model calculations, instead of short-term elasticities, which are common in static modeling. However, long-term substitution elasticities are not readily available for the Dutch economy. As it presently stands, elasticities of a rather short to medium-term nature are applied. The already mentioned premise that in the calculation of a SNI only environmental losses are considered as relevant corrections also means that

R. Gerlagh et al. / Ecological Economics 41 (2002) 157–174

influences from the labor market on the SNI, be it positive or negative, should be neglected. According to Hueting, a sustainable economy will certainly not worsen the employment situation. We assume employment neutrality, through an exogenously given, inelastic, labor supply, and clearing markets through an adjusting wage rate. To calculate a SNI for a particular country, assumptions have to be made with respect to policies in the rest of the world. This is especially relevant for a small and open economy such as the Netherlands, as a unilateral sustainability policy could cause a major international reallocation of relatively environment-intensive production activities. We assume that similar sustainability standards are applied all over the world, taking due account of local differences in environmental conditions. However, it is not feasible to estimate the resulting costs and changes in relative prices in other countries. Instead, we have to make some simplifying assumptions, and in the results presented in this paper, we calculate two variants. The first variant abstracts from changes in prices on the world market. As relative prices in the Netherlands change, it becomes feasible for the Netherlands to partly reach its sustainability standards by importing relatively environment-intensive products, whose cost of production increase relatively much in the Netherlands, and by exporting less environment-intensive products, whose cost of production will relatively decrease in the Netherlands. The second variant assumes price changes on the world market proportional to price changes in the Netherlands. This variant implies a more stringent restructuring of the Dutch economy, as shifting environmental problems abroad is no longer possible. In the same international context, we have to specify an assumption concerning the trade balance. In the AGE model, the standard macro-economic balance equations apply so that the sum of the public and private savings surpluses (or deficits) equals the trade balance surplus (or deficit). The savings surplus is assumed to constitute a constant share of NNI. This, in turn, determines the trade balance through adjusting the exchange rate.

165

In addition to correcting NNI for the cost of technical and volume measures to meet the sustainability standards, NNI should also be corrected for so-called double counting. Double counting refers to the expenditures on compensatory, restoratory and preventive measures to re-establish or maintain environmental functions, sometimes denoted as defensive measures. According to Hueting and many others, these expenditures wrongly enter NNI as value added: the earlier loss of environmental functions was not written off, whereas restoration is written up. This line of reasoning can indeed be maintained in case defensive measures are taken in the sphere of consumption, not entering a production process as intermediate input. In our SNI calculations, the cost to reduce dehydration and the clean up of contaminated soils are double counting cases. For cleaning up soils that are contaminated in the past, it seems fair to adjust past income for the costs, and not to adjust current income. It is not obvious which procedure to follow, and we decided to use an ad-hoc solution. It is assumed that the total cost of soil clean up (estimated to amount to 408 billion guilders) is borne by the government. This can in all fairness not be charged to one particular year. Therefore, it is assumed that the soil clean-up activities are spread over a 20 years period. Each year, 5% of the total cost for soil clean up is contracted out and entered in the SNI calculations as a yearly deduction of government income. The reduction cost of dehydration are also assumed to be financed out of, and likewise deducted from, the government budget and amount to 550 million guilders on a yearly basis. Since sustainable income will be substantially below current income, and prices will substantially change, assumptions have to be made about the economic behavior of consumers. In the model, the effects of lower overall income levels are approached by the use of different income elasticities for different consumer goods. Demand for agricultural products decreases less than proportional, demand for services decreases more than proportional, and the demand elasticity for manufactured products depends on the stage of income. In this way, consumption is thought of as

166

R. Gerlagh et al. / Ecological Economics 41 (2002) 157–174

consisting of necessary goods for subsistence and luxury goods. If income falls, the consumption of necessary goods will remain relatively stable, which is compensated by a more than proportional decrease in the consumption of luxury goods. For each consumption good in the model, an income elasticity is specified. Moreover, relative price changes will affect consumption patterns, which will become more sustainable. In addition to income substitution effects, the model includes price elasticities. In general, the consumption of environment-intensive goods will decrease, whereas less environment-intensive goods and services will show an increase in relative consumption levels. It is assumed that private consumers have more substitution possibilities than the public consumer (the government), whose demand is determined by public services that have to be supplied. In line with the neglect of transition costs and employment neutrality, the government is supposed to obey budget neutrality. It is assumed that the government is the owner of the environmental functions, constraining their use to a sustainable level. The use of these functions should be paid for. Emissions to the environment are thus considered as public endowments, and as these emissions are constrained by sustainability standards, the value that is imputed in the context of the modeling exercise entirely accrues to the government. Put differently, the government sells emission permits of which the price is endogenously determined in the model. To guarantee budget neutrality, the revenues from the sale of emission permits are recycled by a linearly homogeneous reduction of taxes. In case revenues from emission permits exceed the government budget, the surplus will be redistributed to private households through a lump sum subsidy. Finally, we have to explicate the use of prices for income measurement. This is not so much a modeling assumption, as well a matter of presentation. In a statistical calculation of sustainable income, the correction of NNI is expressed in market or shadow prices. If, however, SNI calculations are made with the help of an AGE model, relative prices change, i.e. prices of environmentintensive products increase compared to other

products. In all figures and tables below, we use the Paasche price indexing. Values are calculated by using prices of the new equilibrium, and prices are scaled such that the value of consumption in the reference case measured in new prices equals the value of consumption at old prices.

4. Numerical results Table 3 presents the macro-economic results of the two SNI variants in comparison with the reference NNI situation. The table shows how NNI can be broken up in three different ways. First, expenditures are partitioned in private and public consumption, net investments, and the trade balance. This break up corresponds to the various columns in the right half of the SAM (Tables 1, 4 and 5). Second, national product is partitioned along the different sectors contributing to the value added. This break up corresponds to the various columns at the left half of the SAM. The row entry ‘other’ in Table 3 corresponds to the capital and abatement sector, and to taxes and emission permits paid for directly by consumers. Third, NNI is partitioned by income source: labor, capital, income from taxes, and income from the sale of emission permits. This break up corresponds to the various rows in the lower half of the SAM. The last two divisions of SNI have to be corrected for double counting (see Section 3). SNI variant 1 (with constant relative world market prices) is 47% below NNI in the base situation, whereas SNI variant 2 (with world market prices changing proportionally to domestic prices) is 56% lower. The extent to which SNI drops is thus significantly determined by the specification of international trade. The sustainable economy shows drastic compositional changes. This is particularly due to the imputed prices for emission permits. The consumption of private households decreases substantially in both variants, indicating that the sustainability policy not only affects the composition of the economy, but also the overall level of activity. The drop in private consumption is however smaller than the drop in national

R. Gerlagh et al. / Ecological Economics 41 (2002) 157–174

income, as the private households have some possibility to redirect their expenditures to less environment-intensive products. The decrease in government consumption is larger than for the private households, even though the assumption is made that the relative decrease in the size of the government is equal to the relative decrease in the consumption of private households. The larger decrease for the government is caused by the changes in relative prices, which affect the government (with its large demand share of services) differently from the private households. Another common finding for both variants is that, of the distinguished expenditure categories presented in Table 3, net investments, i.e. investments in addition to replacement investments, decrease most sharply. This can be explained by a reallocation of production from relatively environment-intensive sectors, which are on average also relatively capital-intensive, to cleaner and more labor-intensive sectors, such as services. The lower net investment share in SNI implies that the upward pressure on capital demand stemming from increased abatement activities is more than offset by a fall in capital demand due to this reallocation. All this results in a decreasing capital stock. Another way to look at this is that con-

167

sumers shift part of their income from savings to consumption in order to maximize their utility. Perhaps remarkably, we find the sharpest decrease in value added in the services sector, which is the sector with lowest emission intensity. In contrast, the agricultural sector shows virtually no decrease in its value added, despite its intensive use of natural resources. This finding is the result of two opposite effects. Though the volume of output in agriculture decreases sharply (by 60%, not presented in this paper), its price level increases by roughly the same amount since agricultural goods have a low income elasticity. We also notice that almost the entire value added in the agricultural sector stems from the use of emission permits, which is considered a primary production factor (Tables 4 and 5). In the same line of reasoning, we can explain that the value of consumption of agricultural goods increases from 4 to 6 billion guilders (compare Table 1 with Table 4 below). The opposite effects occur for services, where we find modest decreases in output levels combined with strongly decreasing relative prices. For completeness, we also notice that the last row, labeled ‘double counting’, describes that part of value added that the government spends on the reduction of dehydration and soil clean up.

Table 3 Macro-economic results (billion guilders)

Expenditures Private households consumption Government consumption Net investments Trade balance Value added Agricultural production Industrial production Services production Other Double counting Income sources Labor Capital Income from taxes Emission permits Double counting

NNI

SNI, variant 1 change (%)

SNI, variant 2 change (%)

456.7 314.0 75.1 51.5 16.2 456.7 15.9 130.8 273.1 36.9 0.0 456.7 229.6 139.5 87.7 0.0 0.0

241.4 (−47%) 187.9 (−40%) 30.2 (−60%) 14.2 (−72%) 9.0 (−44%) 241.4 (−47%) 15.9 (0%) 95.7 (−27%) 118.2 (−57%) 36.3 (−1%) −24.8 241.4 (−47%) 56.9 (−75%) 38.5 (−72%) 0.0 (−100%) 170.7 −24.8

201.4 (−56%) 159.4 (−49%) 23.5 (−69%) 10.9 (−79%) 7.7 (−53%) 201.4 (−56%) 23.8 (+49%) 91.5 (−30%) 81.2 (−70%) 29.1 (−21%) −24.2 201.4 (−56%) 30.7 (−87%) 29.5 (−79%) 0.0 (−100%) 165.4 −24.2

0

0

−0.0

−0.0 −0.0

0

−0.0

−0.0

−52.0

−42 −25

−1 −21 153 −11 −2

Services

0

−0 −26 −5 31

Capital

0

−1

31

−22 −7

Abatement

−9

2 14 −25

Trade Balance

−14

−14

Net investments

−243

−0.0

−0.0

−35.0

−26

−6 −74 −102

Consumption

266

170.5 0.0 0.2 0.0 0.0 0.0 0.0

57 39

Endowments

0

0 0 0 0 0 0 0

0 0 0

0 0 0 0 0

Sum

In comparison with the reference SAM (Table 1), the SAM for the SNI has additional rows for abatement activities and for the emission permits. When comparing the SAM with the results presented in Table 3, notice that because of double counting, the expenditures and the value of endowments in the SAM amount to 266 billion guilders, that is 241 plus 25 billion guilders as defensive expenditures are captured as part of the consumption column (notably the consumption of abatement goods). Empty cells reflect either cells that are absent or zero; cell entries ‘−0.0’ indicate small non-zero values. For the list of environmental themes, see Table 2.

Sum

−70.0

−14 −12

−17 134 −14 −5 −2

Industry

−13.5

−0 −2

Labor Profits Taxes

GHGs Ozone Acidification Eutrophication PM10 Smog Dispersion

22 −4 −0 −1 −0

Agriculture Industries Services Capital Abatement

Agriculture

Table 4 SAM for SNI variant 1 (billion guilders)

168 R. Gerlagh et al. / Ecological Economics 41 (2002) 157–174

−0.0 −0.0

0

0

−44.5

For the list of environmental themes, see Table 2.

0

−0.0

−0.1

−0.1 −0.0

Sum

−0.1

−71.9

−20.1

−22 −14

GHGs Ozone Acidification Eutrophication PM10 Smog Dispersion

−8 −12

−0 −3

−1 −16 105 −6 −2

Labor Profits Taxes

−26 136 −11 −5 −2

Services

34 −7 −1 −1 −1

Industry

Agriculture Industries Services Capital Abatement

Agriculture

Table 5 SAM for SNI variant 2 (billion guilders)

0

−0 −20 −3 24

Capital

0

−1

30

−23 −6

Abatement

−8

−2 −0 −6

Trade Balance

−11

−11

Net investments

−207

−0.0

−0.0

−28.4

−25

−6 −70 −78

Consumption

226

164.9 0.0 0.3 0.1 0.0 0.1 0.1

31 30

Endowments

0

0 0 0 0 0 0 0

0 0 0

0 0 0 0 0

Sum

R. Gerlagh et al. / Ecological Economics 41 (2002) 157–174 169

170

R. Gerlagh et al. / Ecological Economics 41 (2002) 157–174

Fig. 5. Break up of income in expenditure categories, from NNI to SNI.

Fig. 5 presents changes in the break-up of national income per expenditure category on the way from NNI to variant 1 of a SNI in steps of one-tenth of compliance with the sustainability standards. We see that expenditures smoothly adjust to the decrease in income. The most noticeable feature that can be learned from Fig. 5 is that SNI starts to drop substantially only after 70% of the sustainability standards are met. The policy implication that can be drawn from this figure is that a less stringent, but still substantial, environmental policy can be implemented with relatively limited economic costs. This result is in line with earlier studies, e.g. in the field of climate economics. Fig. 6 presents changes in the break-up of national income per value added category on the way from NNI to a SNI, variant 1. The transition is less smooth than for the expenditures: the first part of the process leads to only minor changes in the components, and after 70% of the required emissions reductions are reached, the value added produced by the industry and services rapidly declines. As mentioned above, the agricultural sector can maintain its value added (due to the higher prices). The pattern for the ‘others’ sector is caused by a combination of factors, a decreasing capital production and an increasing share of the abatement sector.

Fig. 6. Break up of production in sectors, from NNI to SNI.

Fig. 7 presents the partitioning of national income in various income sources. Labor income decreases by 75%, and the share of labor in total income decreases from 50 to 25% under SNI variant 1, and to about 15% under SNI variant 2. Income from capital also decreases substantially. The most striking result is the complete greening of the tax system. The government revenues from the sale of emission permits exceed government expenditure when about 70% of the sustainability standards are met. At that point revenues from emission permits replace all existing taxes. The

Fig. 7. Break up of income into income sources, from NNI to SNI.

R. Gerlagh et al. / Ecological Economics 41 (2002) 157–174

excess revenues are redistributed to private households as lump-sum payments. The value of these emission permits constitutes no less than about two-thirds of the SNI. Now, we will briefly further investigate the changes in the economic structure as presented in the SAMs for the two SNI variants (Tables 4 and 5). As already pointed out above, value added in the agricultural sector is mainly made up of the value of emission permits. Production of agricultural goods becomes rather expensive, and the Dutch economy becomes a net importer of agricultural goods, in contrast with its strong exporting position in the reference case. Apparently, under sustainability standards, the agricultural sector cannot continue its bulk production in the Netherlands if the rest of the world is capable of producing agricultural goods in a sustainable manner. The SAM shows a specialization of the Dutch economy into services, which require less emission permits per guilder of value added, compared to the other production sectors. The exports of services increase substantially, in contrast to the overall decline of the Dutch economy. The tables also clearly show the uneven significance of different environmental themes. In both SNI variants, most expenditures on emission permits go to climate change. The huge costs of greenhouse gas emission permits can be explained by the strict sustainability standard, especially in comparison to the amount of emissions that can be avoided through technical measures. Consequently, costly volume measures have to be taken to reduce greenhouse gas emissions. Once the economy has sufficiently restructured to satisfy the climate change sustainability constraint, other environmental constraints are satisfied without the need for very costly measures. This feature is partly explained by the fact that GHGs are associated with virtually all economic activity (especially through energy use), and consequently, the restructuring of the economy affects all sectors. Note that the value of the emission permits for the other themes are not exactly zero, but just too small to be expressed in the table: they amount to tens to hundreds million guilders.

171

Under variant 2, which assumes price changes on the world market proportional to price changes in the Netherlands, the SAM as presented in Table 5 shows a different pattern of production, consumption, and trade. In this variant, it is not possible to realize Dutch sustainability standards by importing relatively environment-intensive products, and exporting less environment-intensive products. Consequently, the Dutch economy cannot fully specialize towards services; agricultural remains a substantial exporter, using a substantial part of the emission permits. Though climate change is still the dominant environmental theme, the costs associated with the other themes are now somewhat higher than in the previous variant. This shows that these environmental problems can no longer be ‘exported’ to other countries, but have to be tackled by domestic measures. Being confronted with the sharp income reduction caused by a decrease of greenhouse gas emissions by 79%, it seems an obvious challenge to examine the model’s behavior when using it for a more standard policy analysis, and to compare our results with typical results in the literature. We calculate the costs, measured in loss of income, of a greenhouse gas emission tax that aims at reducing greenhouse gas emissions by 50%. Though this aim represents a rather stringent environmental policy, comparable calculations have been carried out in the literature, because of the understood urgency of the enhanced greenhouse effect. Boero et al. (1991) give an overview of AGE models that are used for this purpose, and find a decrease of income ranging from 1 to 4.5%. Similar to the calculations for the SNI, we have two basic variants, one with world market prices unchanged, and the other with world market prices changing proportionally to price changes in the Dutch economy. Table 6 presents the results, which roughly fall in the range found in the literature. Finally, we highlight the finding that costs of emission reductions sharply increase from about 4% of income for a 50% reduction to about 50% of income for a 79% reduction. Whereas the first half of emissions can be reduced through technical measures and by reallocating production fac-

R. Gerlagh et al. / Ecological Economics 41 (2002) 157–174

172

Table 6 Income effects of a 50% GHG emission reduction

National income

Base

50% reduction, variant 1 (change in %)

50% reduction, variant 2 (change in %)

456.7

434.5 (−4.9%)

441.2 (−3.2%)

tors across sectors, a truly deep cut in emissions requires an inevitable cutback of the overall activity levels. This finding stresses the consequence of assuming the present state of technology unchanged. If a policy were directing economic growth towards more sustainable production and consumption patterns, giving firms time to innovate and to adjust their production processes, it is not unlikely that substantial emission reduction measures may be reached while requiring less severe reductions in overall economic activity.

5. Concluding comments The main emphasis of the research was on the construction of an AGE model to calculate a SNI indicator for the Netherlands. We have been successful in the sense that we have shown that an AGE model can be extended with a list of environmental themes, and that abatement costs can be included explicitly in the model to account for the costs associated with emission reduction measures. Our calculations indicate that climate change, caused by emissions of GHG, is the most pressing environmental issue, in terms of costs involved to meet the sustainability standards. Whether the gap between NNI and the SNI level will increase or decrease over time largely depends on the GHG policies that will be implemented in the coming years. Energy is a necessary input for production and consumption, and the potential for reduction of carbon dioxide emissions (the major greenhouse gas) through energy savings is inherently limited. If climate change policies stimulate the potential of alternative carbon-free energy sources, such as various solar and wind energy sources or biomass, then emission reductions can be achieved through a transition towards these alternative energy sources. Costs of

emission reductions will substantially fall, and the gap between the NNI and the SNI level will decrease. In other words, another way of interpreting the dynamic development of the gap is that it measures whether the economy is developing its potentials to meet more stringent environmental standards, or not. It is therefore of apparent interest to extend the current analysis into a dynamic context. There are still many improvements, refinements and sophistications to be made to our modeling analysis. Many of these are already indicated in the text. Without being exhaustive, we recall some features that deserve special attention. First, the coverage of relevant environmental functions (themes) can be extended. For the Netherlands, land use and waste disposal ask for inclusion. Moreover, there is need for a discussion about what level of resource use can be called sustainable. We also note that the list of environmental themes included is biased to ‘sink’ functions of the environment, as opposed to the environmental ‘source’ functions associated with forests, mineral deposits, topsoil, fish stocks, and water resources. Our focus is typical for current Dutch environmental concerns. If the calculations would be repeated for other countries, one should consider including other natural resources. For many poorer countries, the sustainable use of the environmental ‘source’ functions is of more immediate importance then the sustainable use of the ‘sink’ functions. Second, the modeling of international trade can be further worked out. Third, the information on technical options of abatement and their costs needs to be kept up to date. Fourth, in the present model, emissions are linked to outputs. Part of the emissions can better be linked to certain types of inputs, for instance, CO2 emissions to fuel inputs. Modeling emissions through links with (fuel) inputs will allow a better reflection of sub-

R. Gerlagh et al. / Ecological Economics 41 (2002) 157–174

stitution possibilities. Fifth, the model might benefit from a differentiation of the abatement cost curves between sectors and a differentiation of the expenditure effects of technical abatement. In short, changing and modifying assumptions that underlie the SNI calculations means that a whole gamut of sensitivity analyzes can be done. Though sensitive to several assumptions and qualifications, we have a measure of the part of income that depends on an excessive use of resources. For the Netherlands, for 1990, we calculated that about half of its income could be attributed to resource use that exceeds a sustainable level. In the midst of the many theoretic analyzes on sustainable income levels, policy makers may find it useful to have a numerical result. It is now our responsibility to find attractive and understandable ways of communication to inform the public and politicians, as well as the scientific community, on the reached results, their meaning and their limitations.

Acknowledgements We are grateful to Roefie Hueting and Bart de Boer for sharing their views and comments. We also want to thank S. El Serafy for his critical review, which helped us to improve the paper. Finally, we acknowledge financial support from the Dutch Ministry of Economic Affairs and Ministry of Housing, Spatial Planning and the Environment, who funded part of the research.

References Ahmad, Y.J., El Serafy, S., Lutz, E. (Eds.), 1989. Environmental Accounting for Sustainable Development. World Bank, Washington. Aaheim, A., Nyborg, K., 1995. On the interpretation and application of a ‘green national product’. Review of Income and Wealth 41, 57 –71. Aronsson, T., Lofgren, K.G., 1998. Green accounting in imperfect market economies —a summary of recent research. Environmental and Resource Economics 11, 273 – 287. Blok, K. 1991., On the reduction of carbon dioxide emissions. Thesis, Rijksuniversiteit Utrecht, ISBN 90-73958-01-6. Blok, K., Worrell E., Albers, R.A.W., Cuelenaere, R.F.A., 1991. Data on energy conservation techniques for the

173

Netherlands. report no. W. 90008, Rijksuniversiteit Utrecht, Department Natuurwetenschap en Samenleving. Boero, G., Clarke, R., Winters, L.A., 1991. Controlling Greenhouse Gases: a survey of global macroeconomic studies. Discussion papers 91 – 25, University of Birmingham, Birmingham. Boyd, R., Uri, N.D., 1991. An assessment of the impacts of energy taxes. Resource and Energy 13, 349 – 379. Castan˜ eda, B.E., 1999. An index of sustainable economic welfare, ISEW, for Chile. Ecological Economics 28, 231 – 244. Cruz, W., Repetto, R. (Eds.), 1991. Accounts Overdue: Natural Resource Depreciation in Costa Rica. World Resources Institute, Washington DC. Cruz, W., Repetto, R., 1992. The Environmental Effects of Stabilization and Structural Adjustment Programs: The Philippines Case. World Resources Institute, Washington DC. Daly, H.E., Cobb, J.B. (Eds.), 1989. For the Common Good. Beacon Press, Boston. De Boer, B., Bosch, P., 1995. The greenhouse effect: an example of the prevention cost approach. Paper prepared for the second meeting of the London group on national accounts and the environment, Washington, March 15 – 17. De Boer, B., 2002. Economische structuren en milieugebruik; basisrapport van het project Duurzaam Nationaal Inkomen, concept. Statistics Netherlands, Voorburg. Dellink, R., Gerlagh, R., Hofkes, M.W., 2001. Calibration of the applied general equilibrium model for the Netherlands in 1990. IVM report W-01/17, Institute for Environmental Studies, Amsterdam, The Netherlands. Dellink, R., van der Woerd, F., de Boer, B., 1997. Kosteneffectiviteit van milieuthema’s. IVM-VUA report R-97/10, Institute for Environmental Studies, Amsterdam, The Netherlands. El Serafy, S., 2001. Steering the right compass: the quest for a better assessment of the national product. In: van Ierland, E.C., van der Straaten, J., Vollebergh, H.R.J. (Eds.), Economic Growth and Valuation of the Environment. Edward Elgar, Cheltenham, UK, pp. 189 – 210. Gerlagh, R., Dellink, R., Hofkes, M.W., Verbruggen, H., 2001. An applied general equilibrium model to calculate a Sustainable National Income for the Netherlands. IVM report W-01/16, Institute for Environmental Studies, Amsterdam, The Netherlands. Gerlagh, R., Keyzer, M.A., 2001. Sustainability and the intergenerational distribution of natural resource entitlements. Journal of Public Economics 79, 315 – 341. Gerlagh, R., van der Zwaan, B.C.C., 2000. Long-term substitutability between environmental and man-made goods. IVM report D00-18, Institute for Environmental Studies, Amsterdam, The Netherlands, accepted for publication in Journal of Environmental Economics and Management, May 15, 2001. Gerlagh, R., van der Zwaan, B.C.C., 2001. The effects of ageing and an environmental trust fund in an overlapping generations model on carbon emission reductions. Ecological Economics 36, 311 – 326.

174

R. Gerlagh et al. / Ecological Economics 41 (2002) 157–174

Hartwick, J.M., 1977. Intergenerational equity and the investing of rents from exhaustible resources. American Economic Review 67, 972 –974. Hicks, J.R., 1948. Value and Capital, 2nd ed. Clarendon Press, Oxford. Holub, H.W., Tappeiner, G., Tappeiner, U., 1999. Some remarks on the ‘System of Integrated Environmental and Economic Accounting’ of the United Nations. Ecological Economics 29 (3), 329 –336. Howarth, R.B., Norgaard, R.B., 1992. Environmental valuation under sustainable development. American Economic Review 82, 473 – 477. Hueting, R., 1980). New Scarcity and Economic Growth. North-Holland, Amsterdam. Hueting, R., 1992). Correcting national income for environmental losses: a practical solution for a theoretical dilemma. In: Krabbe, J.J., Heijman, W.J.M. (Eds.), National Income and Nature: Externalities, Growth and Steady State. Kluwer Academic Publishers, Dordrecht. Hueting, R., 1995. Estimating sustainable national income. In: van Dieren, W. (Ed), Taking nature into account; towards a sustainable national income. Springer-Verlag, New York, Chapter 13. Hueting, R., Reijnders, L., 1998. Sustainability is an objective concept. Ecological Economics 27, 139 – 147. Hueting, R., de Boer, B., 2001. The parable of the carpenter. International Journal of Environment and Pollution 15, 42–50. Jorgenson, D.W., Wilcoxen, P.J., 1993a. Reducing U.S. carbon dioxide emissions: an assessment of different instruments. Journal of Policy Modeling 15, 491 –520. Jorgenson, D.W., Wilcoxen, P.J., 1993b. Reducing US carbon emissions: an econometric general equilibrium assessment. Resource and Energy Economics 15, 7 –25. Keller, W.J., 1980. Tax Incidence. North-Holland, Amsterdam. Keuning, S.J., 1993. An information system for environmental indicators in relation to the national accounts. In: de Vries, W.F.M. (Ed.), The Value Added of National Accounting, Statistics Netherlands. SDU Publishers, The Hague. Liu, X.L., 1998. Adjusted forest accounts for China. Ecological Economics 27, 283 –298. Nordhaus, W.D., Tobin, J., 1972. Economic Growth; Is Growth Obsolete? Columbia University Press, New York. Nordhaus, W.D., Kokkelenberg, E.C., Woodward, R.T., 2000. Nature’s numbers: expanding the national economic accounts to include the environment. Land Economics 76, 486– 490.

Okken, P.A., 1991. CO2 reduction consensus? A conceptual framework for global CO2 reduction targets, the importance of energy technology development. Report no. ECNRX-91-093ECN, Petten. Okken, P.A., Lako, P., Gerbers, D., Kram, T., Ybema, J., 1992. CO2 removal in competition with other options for reducing CO2 remissions. Energy Conversion and Management 33, 6 – 10. Repetto, R., Magrath, W., Wells, M., et al., 1989. Wasting Assets; Natural Resources in the National Income Accounts. World Resources Insitute, Washington DC. Statistics Netherlands, 1990. Tax incidence in the Netherlands: accounting and simulations. Statistics Netherlands report M42, SDU Publishers, The Hague. Solow, R.M., 1974. Intergenerational equity and exhaustible resources. Review of Economic Studies 41, 29 – 45. Thampapillai, D.J., Uhlin, H.E., 1997. Environmental capital and sustainable income: basic concepts and empirical tests. Cambridge Journal of Economics 21, 379 – 394. Torras, M., 1999. Inequality, resource depletion, and welfare accounting: applications to Indonesia and Costa Rica. World Development 27, 1191 – 1202. UN, United Nations, 1993. Integrated Environmental and Economic Accounting. United Nations, New York. Van Ierland, E.C., Straaten, J.V.D., Vollebergh, H.R.J. (Eds.), 2001. Economic Growth and Valuation of the Environment. Edward Elgar Publishing Ltd, Cheltenham, UK. Verbruggen, H., Dellink, R.B., Gerlagh, R., Hofkes, M.W., Jansen, H.M.A., 2001. Alternative calculations of a sustainable national income for the Netherlands according to Hueting. In: van Ierland, E.C., van der Straaten, J., Vollebergh, H.R.J. (Eds.), Economic Growth and Valuation of the Environment. Edward Elgar, Cheltenham, UK, pp. 275 – 312. VROM (Ministry of housing, spatial planning and the environment), 1994. Methodiek Milieukosten. Publicatiereeks Milieubeheer 1994/1, The Hague. VROM, 1998. Kosten en baten in het milieubeleid, definities en berekeningsmethoden. Publicatiereeks Milieustrategie 1998/6, The Hague. Weitzman, M.L., 1976. On the welfare significance of national product in a dynamic economy. Quarterly Journal of Economics 90, 156 – 162. Withagen, C., 1998. Untested hypotheses in non-renewable resource economics. Environmental and Resource Economics 11, 623 – 634.