SIXTH FRAMEWORK PROGRAMME

Project no: 502687NEEDS

New Energy Externalities Developments for Sustainability

INTEGRATED PROJECTPriority 6.1: Sustainable Energy Systems and, more specifically,

Sub-priority 6.1.3.2.5: Socio-economic tools and concepts for energy strategy.

Deliverable n° 5.1 - RS 3a

“Green Accounting and Externalities. Review of Methodologies & Application in

selected countries”

Due date of deliverable.: August 2008Actual submission date: March 2009

Start date of project: 1 September 2004 Duration: 48 months

Organisation name for this deliverable: University of Bath (Bath), UK and Charles University Environment Center (CUEC), Prague , Czech Republic

Authors: Alistair Hunt (UBath) andd Milan Ščasný (CUEC)

Project co-funded by the European Commission within the Sixth Framework Programme (2004-2008)Dissemination Level

PU Public xPP Restricted to other programme participants (including the Commission Services)RE Restricted to a group specified by the consortium (including the Commission Services)CO Confidential, only for members of the consortium (including the Commission Services)

1

Contents

Part I. Conceptual background to the development of Green National Accounting.............3I.1 System of National Accounts....................................................................................3I.2 The shortcomings of the current SNA.......................................................................4I.3 Why welfare may not be proportional to GNP...........................................................5I.4 Depletion of capital stocks......................................................................................10I.5 Corrections due to sustainability.............................................................................12I.6 National Accounts Adjustments...............................................................................12

Part II. Applications of Green Accounting...........................................................................16II.1 The UN System for Economic and Environmental Accounting................................16II.1.1 Physical Flow accounts.......................................................................................17II.1.2 Economic data: Combining asset and flow accounts..........................................18II.1.3 Economic data: Asset accounts..........................................................................20II.1.4 Valuing degradation............................................................................................27II.1.5 Environmental adjustments to the flow accounts................................................28II.2 The GARP and GREENSTAMP projects................................................................31II.3 Genuine Saving.......................................................................................................33II.4 Alternative indicators of welfare..............................................................................36II.4.1 Introduction and Background: ISEW and MEW...................................................36II.4.2 Calculating the ISEW..........................................................................................38II.5 Greensense.............................................................................................................40II.6 Costs and Damages Associated With Climate Change..........................................42II.7 Conclusions on methods and a step-by-step procedure to include the external costs in national accounting framework.......................................................................................44II.7.1 Incorporating the external costs into the national accounting framework: A step-by-step Guide.....................................................................................................................44

Part III. Green Accounting Application: Extended Genuine Savings measure.....................51

Appendix................................................................................................................................. 67

2

Part I. Conceptual background to the development of Green National Accounting

I.1 System of National Accounts

The concept of Green National Accounting can be viewed as an extension of the original concept of national economic accounts formalised in the Standard National Accounts (SNA) framework. This originated from work done during the second world war by James Meade and Richard Stone, presented in the paper, “An analysis of the sources of war finance and an estimate of the national income and expenditure in 1938 and 1941”. 1 Stone subsequently chaired the League of Nations group responsible for the first edition of the SNA in 1952. The basic form of the national accounts as developed by Stone (1951) can be summarised as:

NNP = C + I – D + X – M Where:NNP = Net National ProductC = ConsumptionI = InvestmentD = DepreciationX = ExportsM = Imports

Thus, the NNP of an open economy consists of Consumption plus net investment (Investment minus Depreciation) plus net exports (Exports minus Imports). The theoretical basis for the framework is Keynesian, in that it was developed to be used for Keynesian macroeconomic analysis and stabilisation policy. However, during the post-war period the resulting indices, Net National Product, (NNP), and particularly Gross Domestic Product and Gross National Product, (GNP), have come to be regarded not simply as tools for economic management but also as indicators of economic performance and economic wellbeing, and of a country's "income".2 Although never intended by the economists and

1 We can find early roots of national accounts further back in human history. Political arithmeticians, namely W. Petty and G. King, tried to estimate a scale of national wealth and derive national income in the XVII century (see W. Petty, 1960, “Political Arithmetik”). Economic Tables developed by F. Quesnay (1694-1774) can be considered as the real origin of national accounts. Economic Tables presented the first model of national economy aimed at representing the mutual relationships between economic individuals.2 Note: the difference between Gross and Net is capital (man-made) depreciation; the difference between Domestic and National is characterised by the area on that the measure refers. Domestic Product measures all economic production produced in given country by all residents without considering their country of origin, while National Product includes also monetary flows attributable to economic entities which are respective country citizens but are economic active in another country. Specifically, production of British citizen who has been working in the Czech Republic (and being the Czech resident) contributes to the Czech Domestic Product (but not to the Czech National Product) as well as to National Product of the UK (but not to British GDP)..

3

accountants, the measure of GDP has become considered an indicator of a country’s progress, success or even quality of life.

I.2 The shortcomings of the current SNA

Never-ending story on SNA corrections

Since 1952 when the System of National Accounts was firstly standardised, accounting rules and approaches have been continuously revised resulting in its 1968 and 1993 SNA updates (see UN 1968; UN 1993).3 However, the system of national accounts and particularly its most well-known aggregate measure, i.e. GDP, have been criticised. Nordhaus and Tobin (1972) in their pioneering paper “Economic Growth: Is Growth Obsolete?” propose several early corrections of conventional Net National Product in order to develop new alternative measures based on adjusted system of national accounts. The resulting constructed ‘Measure of Economic Welfare’ adjusts items of national accounts due to i] reclassifications of final expenditures into consumption by either subtracting those items that can be considered as intermediate or ‘regrettable’ rather than a final product, or by re-classifying others between consumption and net investment, ii] considering services flowing from durables and public investments rather than capital investment itself, and including leisure time and unpaid work at homes, and iii] income corrections due to inconveniences related with urbanisation, estimated from the additional income needed to sustain people in highly-populated urban areas. After adjusting NNP, the Measure of Economic Welfare was smaller, (though still rising), during 1929-1965 in the USA, but showing a less strong growth rate than NNP did (summarised also in e.g. Dorfman et Dorfman, 1977, pp. 479-488).

In the last twenty years or so, there has also been growing public concern regarding the extent to which environmental damage may offset the positive effects of economic growth on wellbeing, and the extent to which economic growth may be unsustainable due to the depletion of productive environmental resources. The first concerted call for standard national accounts to be corrected to account for environmental issues came in 1989 with the publication of a World Bank symposium on "Environmental Accounting for Sustainable Development", edited by Ahmad, El Serafy and Lutz (1989).4 This publication reflected the concern among researchers that standard measures of national income provide neither an accurate reflection of citizens’ welfare, nor an accurate indicator of the long-term prospects of the economy. The shortcomings of the standard national accounts as a measure of wellbeing were discussed in the symposium, and are outlined in the next subsection.

These, and other, arguments as to whether and how the SNA should be adjusted, are well-known and thus will be discussed only briefly in this section. Considering all the arguments, one can identify three possible approaches to correct or adjust national

3 The world-wide standardised accounting system was embodied in so called European System of Accounts in 1979, then, revised in ESA-1995 (EC 1995).4 One should however add that an effort to incorporate environmental degradation yielded even earlier in modelling of natural resources accounts by Norwegian Government in 1974 or composing new satellite accounts for natural resources in physical units for Indonesia or adjusting NNP by damage due to emission that exceeded standard calculated by relevant abatement costs for Japan since 1972 (see Uno 1988; or Pearce 1991).

4

accounts. The first group of arguments promotes corrections due to proper measurement of welfare. The second group aims to address the longer term implications of economic activity and corrects for depreciation of productive stocks of capitals. Lastly, approaches have been developed in order to better deal with sustainability principles.

I.3 Why welfare may not be proportional to GNP

Proper measurement of welfare incorporates a wide range of adjustments and corrections to national accounts. These consist of factors that are not included in measured GNP but that contribute to wellbeing, and that may conflict with increased economic output. There are also social and environmental by-products of economic activity that detract from wellbeing, which, it is argued, should be deducted from GNP. It has also been suggested that there are factors that are included in GNP that do not contribute to welfare and thus should be deducted. These consist principally of expenditures to reduce the unwanted side-effects of economic activity, so-called defensive expenditures, and of so-called “intermediate expenditures”, such as the costs of commuting. These corrections might also include the service flow of durables rather than the capital expenditures on such durables. Additionally, there are human activities that might positively contribute to welfare such as unpaid work at homes or having leisure. All of these adjustments refer to an examination of welfare potential of the item that is included in the national accounts and that contributes to the income account and, thus, macro aggregates such as GNP.

Some authors include an extension of the welfare measure by including welfare effects in future generations. De Wit (1996, p. 22), for instance, suggests that an environmentally adjusted Gross National product would consist also of items that contribute to welfare of future generations such as net investments in man-made capital, net extraction of natural resources and environmental degradation that exceeds its absorption capacity.

One of the major issues in the development of Green National Accounts has been the identification of the factors that should be corrected for in adjusting NNP. Although the principal candidates for welfare-related environmental adjustment are environmental services, environmental damages and defensive expenditures, we discuss six specific items that are though important in defining the welfare-income relationship.

i) damage provided by economic activity

GNP measures economic activity in terms of the amount of final demand satisfied by economic output. The use of this figure as a measure of wellbeing rests on the assumption that wellbeing is proportionate to consumption of produced goods. However, economic activity has many negative external effects. These can include inter alia environmental externalities, a major example being the pollution (to air, water and soil) caused by production. Some of the damage caused by this pollution is reflected in reduced output, as in the case of reduced agricultural output due to polluted soils and water supplies. This damage is thus reflected in the national accounts, since output figures are lower than they would otherwise have been. This also holds for the possible negative impacts of air pollution on forestry and fishery.

However, environmental damage also directly affects human welfare and this loss is not reflected in the national accounts. There are several complex ways in which welfare can be

5

affected due to the impact of pollution on human health. Firstly, loss of welfare can be due to discomfort of breathing polluted air or ingesting toxic substances, i.e. heavy metals, as well as the increased risk of health damage that this entails. These impacts present non-market goods; either having pain, suffer or any kind of inconveniency associated with illnesses, or suffering premature mortality – both result in lower welfare of the individual. Secondly, emissions of heavy metals, specifically of mercury and lead, can lead to human development impairment that may consequently cause lower future earnings. Such a reduced flow of earnings has a direct effect on future incomes, and welfare as well. Earnings can, however, also be reduced due to sickness involved though by pollution. In this case, one can expect a short-term impact on income and GNP. Lastly, being ill can also involve medical treatment (mitigating) costs and/or averting expenditures. These costs might be incurred by the public health security system, the private health system or be in the form of off-pocket expenditures of the individual. All of them can, however, be considered a kind of defensive expenditures (see below).

Pollution can also reduce the recreational and aesthetic benefits that people derive from the environment. Thus, for any level of output, welfare can generally be assumed to be higher, the better is environmental quality. Standard national income does not reflect this and thus implies the same level of wellbeing for a given level of output, regardless of the welfare effects due to pollution.

Peskin (1989) suggests that environmental damages should be deducted from GNP, in order that net welfare derived from the economy and the environment should be accurately measured. Hamilton (1994) likewise argues that the value of environmental services should first be added, and then the value of environmental damages, valued at their marginal cost, be deducted. There are measurement difficulties inherent in this, and the literature is ambiguous as to whether the deduction should be made to GNP or only to NNP (see more in the part on SEEA, II.1.4 and II.1.5 below).

ii) welfare provision by environmental services

It has been suggested (e.g. by Peskin, 1989; Hamilton, 1994) that the value of environmental services should be included in GNP. Hamilton notes that whilst environmental services to production are already reflected in GNP, (their contribution constitutes part of the value of production), the direct contribution of environmental services to human welfare is not reflected in produced output. Thus, the suggestion that the value of these services should be included in GNP reflects the desire that GNP should be a measure of total welfare generated in the current period, including the welfare generated by environmental assets. The difficulty with this of course is in the measurement of the value of the contribution of the environment to human welfare.

Further problems include the fact that estimates of the utility value of environmental services include consumer surplus5, which national accounting estimates do not, and that including the value of environmental services changes the production and consumption boundaries of the national accounts. This implies that the most appropriate aim of national accounts is to measure the net contribution to welfare of economic activity during a given period. This is what practical applications of Green Accounting techniques have in fact

5 In the case of welfare measurement of changes in uncontrolled goods and services – non-market goods and services say – the utility is measured by compensating or equivalent surpluses derived from willingness-to-pay, or willingness-to-accept a compensation respectively (see e.g. Freeman, 2003).

6

aimed to do.

iii) Defensive expenditures

One of the most often-quoted anomalies of standard GNP is that not only are the negative side-effects of economic production not accounted for in the measurement of wellbeing, but the expenditure to reduce these side-effects is included and thus appears to inflate the measure of wellbeing. In the case of pollution, the measurement problem stems from the fact that some resources are devoted to mitigating its consequences. This so called defensive expenditure does not contribute to welfare, but because it is final expenditure, it is counted as part of GNP. It has been suggested that these defensive expenditures should be deducted from national income in order to estimate more accurately the net welfare created by economic activity.

To support this idea, for instance, Mishan (1982, p. 82) interprets total value of income as total value of satisfaction expected by the society during given year. In this case, the accounting measure should be designed to show how economic activity additionally contributes to society’s satisfaction and wellbeing.6 He considers the expenditures made in the course of economic activity as the additional input costs needed to sustain a given output or level of satisfaction rather than additional output (Mishan, 1993, p. 85).7

This is in line with proposed corrections by other authors. Nordhaus and Tobin (1972), for instance, brought this idea into their Measure of Economic Welfare by introducing ‘intermediate’ or ‘regrettable’ expenditures.8 Daly (1989) suggests that defensive expenditures should be deducted from NNP since they constitute not a net contribution to welfare but the avoidance of a decrease in welfare. Together with Cobb, he argues that these expenditures are needed to protect against undesired side effects of production (Daly and Cobb, 1989, p. 70) and they apply this idea more concretely by the inclusion of several items of their ISEW measure. Following this concept, Leipert (1989; 1995) found that the defensive expenditures for six items calculated for former Western Germany rose from 5.6% to 10% of its GDP during 1970-1985 and had a four-times higher growth rate than GDP itself. Bayer (1993) also found that the defensive expenditures in the area of water quality and forestry in Austria were as high as 1.5% of GDP.

However, Harrison (1989) and Vanoli (1995) argue against deducting these expenditures from national income. Harrison argues that such a deduction contravenes standard accounting practice, and that accounting cannot discriminate between items that clearly qualify as final expenditure in not being subject to further sale. Vanoli points out that it is moreover extremely difficult to identify those expenditures that are truly defensive. Moreover, this is in line with current accounting practise. As noted by SEEA-2003, “the

6 He illustrates this problem with changing social status of women in the (economic) society. According to him, women were mostly producing electric equipment for households in factories that were used for activities previously done by woman who stayed at home and kept for their house.7 Moreover, Mishan calls these expenditures as costly institutional lubricants required for smooth maintenance of society (ibid, p. 86).8 Nordhaus and Tobin describe nine types of regrettable expenditures: expenditures on external defence and internal safety, on household equipment and devices, government expenditures, expenditures on certain services provided by private sector, expenditures on commuting, on education, information media and expenditures to abate pollution. As noted by Mishan (1993), there is a methodological problem related with the choice of the reference year that will be used to compare current level of welfare with reference one in order to assess whether expenditures were defensive.

7

accounting system makes no judgement about the moral worth or physical necessity of production. It simply records those products which are the subject of transactions in the market place and ensures that when these transactions are completely and consistently covered, the three measures of GDP will give the same answer“ (UN et al., 2003, p. 439, para 10.113).

There has been considerable discussion of this point in the literature subsequent to the publication of Ahmad et al (1989). One of the early theoretical explorations of correcting the national accounts for environmental and other non-market goods was made by Mäler (1991), who provided a theoretical rationale for deducting certain items of expenditure from national accounts. This was done by constructing a model of an economy in which utility is dependent on a produced good as well as on environmental quality and leisure. Output is a function of invested capital, labour, environmental quality and the harvest from a natural stock. His model indicated that defensive expenditures should not be deducted from NNP, since to do so would be double counting. However, this is true only to the extent that the value of environmental services is included in the basic measure of NNP. He argued that if the measure of welfare involves adding to conventional NNP the welfare provided by environmental goods, and if actual damages can be estimated, then current defensive expenditures should not be deducted.

Although the issue of defensive expenditures is clear, there is no general consensus on this subject. A number of commentators suggest that defensive expenditure should simply be left out of national accounts calculations. Others suggest that it be reclassified from final consumption to intermediate consumption. A guide on treatment of defensive expenditures is provide in Integrated System of Environmental and Economic Accounting (SEEA) (UN, 2003). Taking an accounting perspective, the SEEA-2003 considers the problem differently. It argues that “…basically, it is not possible to delete an element from the accounting matrix, or reallocate it, without working through the consequences to ensure that the final system still balances for each row and column pair.“ (ibid, p. 61). Some suggestion as to how to resolve this problem is provided in the 2003 SEEA guidance we discuss below, specifically in part II.1.5 of this paper.

iv) Service flow from durables and capital investment

A fourth problem that has been highlighted, e.g. by Nordhaus and Tobin (1972), or by Daly and Cobb (1989), relates to expenditure on consumer durables, and on public infrastructure. These expenditures are counted as consumption in the year of purchase, but in fact yield a flow of services over their entire lifetime. The result of this is that as the lifetime of consumer durables and investment goods becomes less, GNP appears to rise as it is inflated by more regular purchases. The benefit gained from the goods does not necessarily rise proportionately. A related problem is the fact that the purchase of “positional goods”, i.e. those purchased so as to indicate a superior financial position, such as expensive cars, may not increase the general level of wellbeing.

Following this suggestion, Daly and Cobb (1989) suggest to subtract household expenditures on durables and add the service flow from the durable goods over its lifetime in their measure of ISEW. A key issue is that of the deprecation rate used in the estimation process. Originally, the service flow from durables was arbitrarily set out at 10%, assuming a linear deprecation rate. A revised US ISEW applied an annual depreciation rate at 15% and used an interest rate of 7.5% to account for the opportunity costs of these

8

expenditures.A similar depreciation rate was applied in the ISEW applications in the UK, Chile, or Austria. Moreover, a similar logic, with more arbitrary assumptions applied, was applied to calculate service flow from using roads and highways in a revised US ISEW, and recalculating public expenditures and services from public capital in an Australian ISEW (see Ščasný et al., 2002 for the ISEW applications review).

v) Un-paid work and leisure

The shortcomings of GNP as an estimate of wellbeing include the fact that unpaid work is not included, so that GNP does not reflect the full value of output. The SNA 1993 recommends compiling an income account for unpaid work in a household in the form of a satellite account of the national accounts. SNA’s reluctance about incorporating unpaid domestic work into the system of national accounts comes from the fact that a different economic assumption is used in imputing value for these services (compared with the approach usually applied for valuation of other items such as imputing value for own house). Unpaid work at homes enters, for instance, into the ISEW measure by multiplying estimated hours spent on house-keeping and average hourly compensation provided to paid domestic house cleaners. The Australian Genuine Progress Indicator – an alternative composite welfare index similar to the ISEW – also considers unpaid work for society, e.g. charity.

A further problem is the fact that leisure is not accounted for. The problem lies in the fact that the leisure that is given up in order to increase working hours, and thus production, has a welfare value. Because more leisure is generally considered to make people better off, a measure that does not account for the time required to produce a certain level of output does not accurately reflect the net benefit of that output. This means that an increase in measured output may not lead to a proportionate increase in wellbeing even in the absence of external costs of production. Likewise, an increase in output that also involves an increase in hours worked is more likely to appear to be socially beneficial if the associated loss of leisure time is not reflected in the measure of wellbeing.

vi) Equity and social issues

Further shortcomings of national income as a measure of welfare include social issues. For instance, job security may conflict with increased output, but almost certainly contributes to human wellbeing. Thus, an increase in output that is achieved at the expense of increased insecurity will not necessarily lead to a proportionate increase in wellbeing. The type of work that people do, and the extent to which it is fulfilling, is an important determinant of wellbeing that is not reflected by GNP.

Finally – and often noted – the use of average income per capita does not reflect the effect of income inequalities, and diminishing marginal utility of income; the greater the inequality of income distribution the lower social wellbeing, other things being equal. There are some alternative proposals that consider income equity in compiling alternative welfare measures such as the ISEW or GPI so that consumption or welfare measure is somehow weighted by Gini or Atkinson index.

9

I.4 Depletion of capital stocks

A second reason for the correction of national accounts comes from the fact that depreciation of various forms of capital may decrease the potential for economic production in future and thus result in an unsustainable form of economic performance. These corrections are supported by the fact that NNP is intended to reflect the longer term implications of economic activity each year, because it includes the value of investments in and depreciation of productive stocks.

There is general consensus on this subject in economic theory. The theory states that at any point in life of asset, its value should be equal to the value of future income streams to be provided by the asset discounted to the present period. This can be applied to all kind of assets. Whilst the asset is used, its value declines. The SNA measures the decline in value of man-made capital as the consumption of fixed capital. It is referred to as depreciation or amortisation that should be deducted from measures of income as long as sustainability is a concern. Thus, if consumption of fixed capital is higher than the acquisition of new fixed capital, the situation is not sustainable. As SEEA-1993 notes “the SNA supports this valuation and the calculation of the decline in the value of the natural resource but leaves the value of the decline in the other changes in assets account under a category of economic disappearance” (cit. in UN et al., 2003). Consequently if a natural asset appears, the corresponding increase in value would also be recorded in the same asset account within the SNA. Moreover, if there is a net decrease in the stock of environmental asset, this decrease should be recorded.

This means that, while the problems associated with the interpretation of GNP apply also to the interpretation of NNP, there is the additional problem that under the SNA, while estimates are made of the depreciation of manmade capital stocks, depreciation of natural capital stocks is not accounted for. Therefore, if a country depletes, for example, its forest stocks, its income is inflated by the proceeds, but there is no corresponding deduction to reflect the decrease in value of its asset. Similarly, stocks of social capital and human capital are not reflected. This means that if investment in, for example, pre-school education is reduced, the social and economic effects may be felt in future years, but the fact that human and social capital stocks are being depreciated is not reflected in GDP. Any proper accounting measurement should therefore focus on all kinds of capital, i.e. whether stocks of man-made capital, or natural resources, human and social capital.9

Regarding natural resources, the value of overall productive capital is reduced when stocks of productive natural capital, such as forests and fish stocks are depleted. The fact that standard national accounts do not reflect this is the major complaint of Repetto et al. (1989). The consensus on this issue is that the decrease in value of the asset is measured by the "resource rents"10 on the decrease in the resource stock (which need not be equal to the

9 Consideration of stock reduction differs according to the sustainability perspective followed. While the weak form of sustainability assumes full substitutability of capital and requires keeping aggregate capital non-negative over time, strong sustainability assumes only substitutability within its sub-groups, i.e. it requires non-negative net investment for each type of capital separately. Very strong sustainability, therefore, does not assume any substitutability for certain forms of capital.10 Resource rents refer to the profits that accrue to a resource-extracting firm due to the fact that part of the price of the resource can be considered to be compensation for the fact that a valuable asset (i.e. the resource stock) is being liquidated. Thus, resource rents accrue where a non-renewable resource is being exploited, and where a renewable resource is exploited unsustainably. Resource rents are also known as Hotelling rents.

10

level of extraction). This amount should be deducted in the calculation of NNP as representing part of the depreciation of the overall capital stock. This adjustment is made in calculating green NNP in the welfare-equivalent tradition. In a practical application of this adjustment, Repetto et al. showed for the case of Indonesia that while GDP grew by 7.1% between 1972 and 1984, accounting for the loss of forestry, oil and soils meant that their estimate of environmentally adjusted NDP grew by only 4% in that period.

From the theoretical perspective, this correction relates to the notion of sustainable national income that originates in Hicks’ (1948) work. Hicks defines sustainable income as the maximum value that a man can consume in one period without impoverishing himself. There are particularly two contributions in economic literature that follow up Hicks’ analysis. Weitzman (1976) shows equivalence between the net product and sustainable income, and Hartwick (1977) – based on Solow (1974) – makes clear that a non-decreasing consumption flow requires a non-negative net investment flow.

This is also in line with findings provided by Mäler’s (1991) theoretical explorations of correcting the national accounts. Based on his model, he concludes that the goods used to enhance the stock of environmental resources do not contribute to current welfare, and their value should be deducted. Likewise, the value of changes in the stock of environmental resources should be deducted in the same way as depreciation on manmade capital is deducted.

In a similar vein, Hartwick (1990) suggests netting out the value of decreases in the stocks of renewable and non-renewable resources. Both of these contributions involve using Weitzman's (1976) result that the Hamiltonian11 of an economy's optimal trajectory represents the utility level that, if kept constant, would constitute a utility path with the same net present value as the optimal path. Thus, using the assumption of a constant discount rate, the Hamiltonian represents the "return" on the economy's capital stock and thus in this sense can be thought of as the economy's income, or NNP. Hartwick transformed the Hamiltonian into a monetary measure of NNP by dividing through by the marginal utility of consumption.

El Serafy (1989) proposed a method for calculating the amount that should be deducted from NNP to account for the depletion of non-renewable resources. This consisted of the annual amount that, if placed in an interest-bearing account, would result at the end of the life of the asset in the accumulation of a fund of the same value as the original value of the resource. It is often thought (see Hartwick and Hageman,1993) and Hamilton,1995) that El Serafy's "user cost" is equivalent to resource rents when the latter are calculated in the same context used by El Serafy, i.e. a constant extraction path. In fact El Serafy's user cost can be shown to be less than resource rents, by the amount of the interest on the accumulating fund.

11 The Hamiltonian is the expression to be maximised in a dynamic optimisation problem. It consists of the current value of an economy’s objective function, plus the value of any changes in (physical, environmental or social) capital stocks.

11

I.5 Corrections due to sustainability

As argued by Withagen (1998) and others, the implications of correcting to sustainable (Hicksian) income require strong assumptions to be fulfilled relating to the treatment of multiple consumers, non-constant time preferences, technological change and distortionary taxes. To overcome these problems, ‘green’ income measures have been developed to capture better the sustainability perspective. Inter alia, Hueting’s Sustainable National Income Indicator (1980; 1995) is one of such examples. To avoid problems particularly related to uncertain future preferences, Hueting suggests assuming an absolute preference for conservation of the natural environment. It implies that the value of environmental degradation is equal to the costs to preserve the environment and to remove any existing environmental burden (in this sense, the SNI indicator resembles a maintenance cost approach). In this case, 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 (Gerlagh, 2001, p. 2).12

In line with these suggestions, De Wit (1996) reports the third approach to adjust the national accounts that targets reaching sustainable income. Degradation is quantified by marginal prices at social optimum in welfare approach, while hypothetical prices that would lead to sustainability are used in the sustainability approach; welfare losses of current generation due to pollution are considered in former, while a focus of the latter approach is on sustaining the stock of capital at certain – sustainable – levels. Moreover, as argued by de Wit, the adjustments stemming from adopting the sustainability perspective might include equity weighting or applying one since value in any applied valuations.

I.6 National Accounts Adjustments

In the literature, we can identify three possible approaches to adjusting national accounts (O`Connor, Steurer, Tamborra, 2001). First, one can compile (environmental) satellite accounts outside of the standard system of national accounts. There have been many attempts – starting in the early 1970’s – to compile particularly satellite accounts for natural resources in physical units (in tonnes, litres, hectars, m3 or even pieces). Accounting approaches and methods are described in detail particularly in SEEA-2003, Chapter 3 which covers natural resource and emission accounts, basic supply and use tables for physical flows, and the latest practical applications of physical flow accounts such as economy-wide material flow balances and accounts, and physical input-output tables.

12 As further noted by Gerlagh (2001), 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 over-exploitation 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-à-vis the NNI is of apparent relevance for actual environmental policy.

12

A second group of national account adjustments includes extensions of national accounts through the use of monetary and physical data within the system of national accounts. These extensions include a wide family of approaches such as compiling so-called hybrid accounts13 (see Chapter 4 of SEEA-2003), accounting for economic activities and products related to the environment (Chapter 5), for environmentally related transactions (Chapter 6), or asset accounts and valuation of natural resource stock (Chapter 7 and 8).

The last group of national account adjustments presents more complex accounting procedures that result in environmentally-adjusted national accounts and aggregated measures. O`Connor et al. (2001) classify these adjustments according to two criteria: the first is given by the accounting system boundary, i.e. whether the adjustment considers an actual set of produced economic assets (given by boundaries of standardised SNA), or extends the accounting scope to all assets including environmental ones; the second criterion is given by a scope of adjustments that may either cover the existing economy, or an economy with new production processes, new technologies, revised production and consumer activities due to new specific environmental standards introduced. This yields four combinations of adjustments further described.

Box 1: Classification of environmentally-adjusted national accounts.

System boundary

Actual set of produced economic assets

Extension by all (environmental) assets

Ref

eren

ce

econ

omy

Statistics for actual existing economy

„Unadjusted “ GDP, or NNP based on standard principles

of national accounts

Environmentally-adjusted domestic product or national

income

„Shadow aggregates“ for modelled economy with environmental standards

GDP/NDP for measurement of environmentally-adjusted economy – “GDP of green

economy”

Concept so far not developed.

Source: O`Connor, Steurer, Tamborra (2001).

The left upper quadrant describes standard aggregates – though unadjusted – of national accounts, i.e. GDP or NDP. The second type of approaches (right upper quadrant) keeps its focus on existing economy but extends its asset coverage. In this case, aggregate indicators of changes in economic assets for a given year and given country are defined as a sum of net change in assets (net savings) and consumption. The aggregates given from such adjustment are referred to as “environmnetally-adjusted”, e.g. GDP, NDP, or national income. The 2003 SEEA labels such results as e.g. eaNDP.

This approach, also known as the maintenance costing technique, belongs to a family of cost-based estimates of valuing degradation (SEEE-2003, page 2-39). It attempts to answer the question:

13 The hybrid flow account is used to denote a single matrix presentation containing both national accounts in monetary terms and physical flow accounts showing the absorption of natural resource and ecosystem inputs and the generation of residuals. Widespread use of these tables belongs to the system of NAMEA (National Accounting Matrix with Environmental Accounts). Socially-motivated adjustments used to be covered by SAMEA (Social Accounting Matrix with Environmental Accounts).

13

What would the value of net domestic product have been for the same level of activity if all the costs of environmental degradation had been incurred and internalised within market prices?

The problem emerges with significant changes in environmental standards when the resultant price rises are likely to bring about a change in behaviour that would affect the level of demand for those products and manifest itself either as a change in the level of output of those products or a technological changes in production that reduce dependence on the newly expensive products (SEEA-2003, page 2-39). For marginal changes in standards, this technique may be used to give an upper bound on the adjustments. Otherwise, resolution requires a modelling approach.

There are several applications of this approach, including the ‘Sustainable Social Net National Product’ by Daly, calculated as NNP minus defensive expenditures and degradation of natural capital, the ‘Consolidated National Income’ derived as GNP plus environmental services minus environmental damages by Peskin, ‘Sustainable Net National Expenditures’ and ‘Environmentally adjusted Net National Product’ suggested by Bartelmus, or ‘Environmentally adjusted Net Domestic Product’ by Skinberg. Measure of ‘Genuine Savings’ developed by World Bank is also belonging to this family. Need references.

The third type of approaches (left bottom quadrant) do not step out of the (accounting say) system boundary; however, they focus on ‘the new’ economy that is measured by shadow prices. The so-called “green economy modelling” attempts to resolve the problems raised by the maintenance cost approaches of the treatment of non-marginal changes in environmental standards. This approach also uses cost-based estimates, but does not aim to change the accounting system, rather change the economic system the accounts are trying to measure. No accounting techniques are required, because it is the measure of the economy itself which is being changed. This exercise leads to asking

What level of GDP could be achieved if producers and consumers faced a different set of relative prices in the economy due to the existence of actual prices for environmental functions?

or how much it would cost to avoid the generation of residuals by changing production and consumption patterns. It illustrates how the economy would be, if it was modified according to certain environmental standards, i.e. introducing emission limits or ceilings on resource extraction. Such aggregates – labelled by the prefix “ge” in the SEEA - measure economic output of hypothetical – modelled – economy with its new structure, assuming that environmental standards are reached in a cost-effective way. These measures are to be seen as forward looking estimates rather than ex post accounting measures.

An application of a greened economy model would be to determine levels of output which lead to levels of income that are sustainable over a given time period, rather than to determine a set of values for output. This type of measure attempts to try to answer the following question:

What level of income and environmental functions can be sustained indefinitely?

14

A measurement of the “sustainability gaps” identified from use of Hueting’s Sustainable National Income, an approach based on critical capital developed within the CRiTiNC project by Ekins, or calculation of costs to reach sustainability gaps for each sector by Rademacher, are applications of this method. The aggregates from this approach are referred to as Sustainable National Income, i.e. SNI and are unique in addressing the issue of long-term sustainability. So far, there is no concept developed that extends its focus to all assets and simultaneously targets a new modelled economy (right bottom quadrant).

There is, however, another, radically different family of approaches that targets damages rather than costs. Such measures are derived from the impact of actual residual generation. This approach thus starts with the residuals generated (emissions, say) and then estimates the impact on different receptors (the so called “impact-pathway approach” developed in the ExternE project series).

When the damage is inflicted on economic assets, the impact should be captured in measures of consumption of fixed capital and environmental depletion. This may not be the case in practice, specifically in the case of impacts on human health. In this case,, “damage-adjusted income” is derived from National Domestic Income by deducting all damages arising from the impact of residual generation, including those to health. This approach answers the question:

What is the impact on the level of NDI of environmental impacts on natural and man-made capital and on human health?

As argued by the 2003 SEEA, “damage-adjusted income” is a first step on the way to converting GDP type-measures to welfare indices.14

14 As further noted by SEEA handbook (page 2-40), many other aspects of welfare are deliberately ignored. In terms of current account transactions no attention is paid to consumer surplus, utility considerations, or the value of household services, for example. In terms of an extended measure of wealth, both natural capital and that aspect of human capital relating to health are covered, at least in terms of changes. However, no extension to social capital or a more comprehensive form of human capital including education is made. More importantly, no inclusion is made for the services which result from an undamaged environment or the remaining services that an already polluted environment still provides. The damage-adjustments to macro-aggregates suggested here are thus better seen as measures of the difference to macro-aggregates that inclusion of this one aspect of welfare would cause, rather than an alternative estimate of the level of those aggregates.

15

Part II. Applications of Green Accounting

This section provides an overview of the work on developing green national accounting frameworks, in particular the UN system for Economic and Environmental Accounting, and the World Bank's Genuine Savings methodology. We then go on to consider some alternative proposals for environmental accounting, in particular the applications of the Index of Sustainable Economic Welfare (ISEW), and those made in the EU within the GARP, GREENSTAMP and GREENSENSE projects.

II.1 The UN System for Economic and Environmental Accounting

The first version of the Integrated Environmental and Economic Accounting – SEEA – was published in December 1993. It is closely linked to the structure of the SNA, and rather than reflecting environmental issues in the core accounts, includes them in satellite accounts which can be used to adjust the final figures. El Serafy (1996) noted – “The objective has been to reflect environmental deterioration in the SNA to the extent that the SNA framework will allow.” Given that the SEEA is a satellite system, as Heal and Kristrom (2001) point out, one cannot expect it "to be consistent with theoretical measures" of income. Indeed, one of the main criticisms that Heal and Kristrom make of the SEEA is that it is not clear, either in the original version or in the current revisions of the framework, what exactly it aims to measure. That is, it clearly does not provide a measure of economic activity for economic management, but neither does it provide a measure of either economic wellbeing or sustainable income. Rather, the SEEA expands on SNA, separating expenditures relevant to environmental issues, and including detailed accounts of the interactions between the environment and the economy. The framework thus allows for the impact of economic activity on natural assets, in terms of both stocks and flows.

We summarise here the framework outlined in SEEA 200015 and SEEA 200316. The purpose of these up-dates is to extend the SEEA to develop a coherent, comprehensive accounting framework, to measure consistently the contribution of the environment to the economy and the impact of the economy on the environment. This includes developments since 1993, especially in the area of physical accounting. The functions of the environment are accounted for, these being in broad terms the provision of resources, the absorption of residuals, and ecosystem services. Precautionary principles can be devised for the use of these services, and the extent to which these principles are satisfied is assessed. Costs and benefits should be accounted for where there are economic consequences, but it is recognised that, particularly in the case of ecosystem services, estimating such costs is extremely difficult. However, the framework should allow the analysis of the effects of environmental policy on the economy, and economic policy on the environment.

The approach is designed to standardise the organisation and classification of environmental data, in terms of compiling balance sheets for environmental assets, and applying physical supply and use tables, links with economic information, and identification of responsibility for environmental impacts. Such accounts should provide a

15 SEEA 2000 revisions available at: www4.statcan.ca/citygrp/london/london.htm16 http://unstats.un.org/unsd/envAccounting/seea.htm

16

basis for sustainable development indicators, as well as environmentally adjusted macroeconomic measures.

The approach in terms of adjusting the SNA to account for environmental issues is that although the 1993 revision of the SNA paid more attention to capital stocks and flows, and to including natural resources as an economic asset, work on environmental accounting is still in progress. This means that the interface between the economy and the environment was, and is still now, kept to satellite accounts, rather than involving a fundamental re-working of the SNA system.

The SEEA accounts can be thought of as being in three sections, as follows: (i) Physical Flow accounts - supply and use tables.(ii) Economic data - juxtaposing the physical environmental accounts with

monetary accounts. This reflects how income is distributed and redistributed, and includes environmental protection expenditure and the value of natural resource stocks (asset accounts). Due to our specific interest in resource depletion in the external costs assessment, we devote a separate section to the subject of resource depletion accounting.

(iii) Valuing degradation - Extension of the framework to cover interactions not presently valued, in particular the causes and impacts of environmental degradation.

It is noted in the manual that the rationale for examining the monetary values of environmental depletion and degradation is that mainstream policy, into which action to address these problems must be integrated, is conducted in these terms. We provide here an overview of each of these aspects of the SEEA.

II.1.1 Physical Flow accounts

The physical flow accounts include four types of flow: products (produced in the economic sphere and used within it), natural resources (mineral, energy, biological), ecosystem inputs (air and water) and residuals (solid, effluent, emissions). Each of these accounts is expressed in terms of supply to, and use by, the economy. The tables thus represent the flows between the economy and the environment. An overview of these flows is given in Table 1.

An identity central to both the SEEA and SNA is that total supply and total demand for products must balance, where total supply is domestic production plus imports and total demand, or use, is intermediate consumption plus household final consumption plus government final consumption plus capital formation plus exports. Account must also be taken of changes in inventories, which counts as an aspect of capital formation. Capital formation is thus split into fixed capital formation and changes in inventories. Thus, the full identity, which must hold as long as units of measurement are consistent, is:

Domestic production + imports = Intermediate consumption + household final consumption + government final consumption + fixed capital formation + changes in inventories + exports.

17

In terms of flow accounts for natural resources, these appear as either intermediate or final consumption, and are supplied only by the environment. Capital goods are unlikely to include unprocessed natural goods. Flow accounts for ecosystem inputs are similar to those for natural resources, but supplies can be imported e.g. when units such as aircraft are operating in another territory. Flow accounts for residuals generally involve waste flowing from economy to environment as by-products of production and waste from consumption. Scrap sold for reprocessing is classified as a product, but residuals recycled without payment are classified as demand by producers for residuals. Waste into landfill is classified as demand by capital. Also shown are transfers into and out of the domestic environment via environmental media. These are important in measuring the accumulation of pollutants in the national environment. Any time lags, e.g. in the case of mining and nuclear power, may have to be accounted for as liabilities generating negative future effects.

Table 1. Origin & destination of flows in the physical & use tables

Table A1 and Table A2 in the Appendix show the physical flows within the economy and between the economy and the environment. In presenting the origins and destination of the flows it serves to give an overview of the system as a whole. However, to make it manageable there is some aggregation required within the categories.

II.1.2 Economic data: Combining asset and flow accounts

Hybrid Accounts

The physical flow accounts presented in the previous section have the principal objective of reflecting the extent to which the economy is dependent on particular environmental inputs and the sensitivity of the environment to particular economic activities. Presenting the physical data in monetary terms allows us to identify where economic value is concentrated and so to target better activities and policies that allow for the decoupling of the economy from the environment and so to move towards a more environmentally sustainable economy.

18

In terms of the monetised supply and use tables, the structure of the monetary accounts is much the same as in the physical accounts. The classification of products and industries is the same as physical ones, depending on the central product classification (CPC) and the international standard industrial classification (ISIC) in both cases. Classification of government and household consumption is also compatible; both draw on the classification of functions of government (COFOG) and the classification of individual consumption by purpose (COICOP). However, there are some important differences.

The first difference is that a Value Added term is required to balance the supply/use identity, since otherwise the value of use would be higher than the value of supply. Value added is a balancing item entered in the supply part of table to reflect the excess that a producer receives over input costs with which to pay labour and cover capital costs. Here, value added is a balancing item, explaining excess of use over supply. In the physical accounts the item that balances the excess of supply over use is residuals.

A second difference is in terms of services, in particular the fact that for industrialised economies, services are relatively insignificant in the physical supply tables, while dominating the monetary tables. For this reason, government consumption is also less evident in the physical tables. The opposite is true of, for example, building materials. In terms of prices, there are differences between basic prices (prices at factory gate) and market prices (including taxes and service/ profit margins). Generally, a supply/ use balance can be achieved at either price, although it is easier to use market prices. Another problem is that many services (e.g. electricity, airline tickets, telecoms) are sold at a range of prices.

The physical and monetary data can be presented in “Supply and Use Tables including Environmental Accounts (SUTEA). This is essentially the physical supply and use tables with monetary accounts attached and an extra sub-matrix for value added. Residuals, natural resources and ecosystem inputs are in physical terms only, while the others are in physical and monetary terms. An example is presented in Table A3. Implementations of SUTEAs are often referred to as NAMEAs (National Accounting Matrix including Environmental Accounts).

The Distribution of Income

There are several other accounts that are part of the SNA, and which also measure money flows as part of the SEEA. The Value Added across all production gives GDP at basic prices. To arrive at the usually-quoted figure of GDP in market prices we must add a measure of taxes paid. The account for the “primary distribution of income then shows how value added is distributed to employees, owners of financial capital and land (property income). Value added adjusted for net property income payable abroad gives the balance of primary income, or Gross National Income (GNI). GNI adjusted for payments and receipt of transfers (e.g. taxes) gives national disposable income. The financial account then shows how disposable income is divided into savings or investment. Saved income is either used either for acquisition of fixed capital or converted into financial assets if in surplus. The sectors of the national accounts are financial and non-financial enterprises, households, government, non-profit institutions serving households, and the rest of the world.

19

These accounts are relevant to the SEEA in that they highlight issues of taxation, relevant to eco-taxation, and they highlight property income, which can be relevant to the ownership and use of natural assets and subsoil assets. Moreover the relationship between income, consumption, saving and investment can be argued to relate to discussion of sustainable income, which is discussed in chapter 10 of the revised SEEA.

Environmental Protection

In terms of environmental protection, the aim is to measure what is being done to protect the environment through direct protection, management activities, and spending on products to protect the environment. The activities to be covered, e.g. investment in clean technology, restoration of polluted land, recycling, production of environmental goods and services, are defined in the context of the Environmental Protection Expenditure Accounts (EPEA). Three types of account are useful, namely the details of activities undertaken, the units which acquire environmentally friendly goods and services (these two can combine into supply and use tables) and simple national aggregates of environmental protection. Clean products may be valued not at their full cost, but at the difference in cost between a clean version and a dirty version.

II.1.3 Economic data: Asset accounts

i] Accounting principles

In the SNA an asset is defined as an entity over which ownership rights are enforced, and from which economic benefits may be derived by its owner. Many, but not all, relevant environmental assets are included. Natural assets are recorded under tangible non-produced assets under item AN.21.

The SEEA expands the asset boundary to all measurable environmental entities of interest, grouping them as Natural resources (Mineral and energy resources, soil resources, water resources, biological resources), Land and associated surface water, and Ecosystems.17 For each of these items it is possible and meaningful to talk of a stock of the asset measured in physical terms. This makes the set of assets considered by the SEEA somewhat wider than that considered by the SNA.

Although, there is no asset account for environmental resources in the SNA system, there are accumulation accounts that describe the changes between opening stock and closing stock in balance sheet by using several items.18 Accumulation accounts are included to

17 In neither system – SNA or SEEA – is any attempt made to consider air or sea water as an asset for which comprehensive measurement is possible or meaningful.18 The accumulation accounts consist of four accounts. The first two accounts – capital and financial accounts cover the transactions in non-financial assets or financial assets and liabilities respectively. The third, the other changes in asset account, itemises other changes in the assets account that results in change to the value of assets not due to transactions. It covers exceptional events such as natural disasters and valuation changes due to the effects of inflation. The balance sheet reports the economic (dis-)appearance. Economic appearance does not relate to a physical appearance but “rather the case where a pre-existing entity is drawn into the economic sphere by acquiring an economic value” (UN et al. 2003, p. 246). The economic disappearance covers the symmetric case when asset loses its value or leaves the economy.

20

show how capital and financial accounts interact with other changes in asset accounts, in terms of linking opening and closing balances in financial terms. The accounts consist of opening balance, changes in the capital account, changes in the financial account, changes in other assets, and the closing balance sheet. Entries in the capital account reflect net capital formation in the national accounts. Changes in other assets refer to changes in entities which the SEEA regards as assets, while the SNA does not.

The asset account is, then, explicitly included within the SEEA structure. The concept of economic (dis-)appearance is changed to “environmental” (dis-)appearance in the asset accounts. The environmental (dis-)appearance includes not only discoveries and extraction, but also reclassifications due to quality change (for instance while the reserve is changed from probable into proved) and reclassifications due to change of functions (e.g. agriculture land changed into build-up area).

One significant issue here is the valuation of environmental assets, discussed in Chapter 7 of the 2003 SEEA. The SNA recommends the use of market prices where possible. Where these are not available, alternative methods are required.

ii] The idea of placing a value on assets

The idea of valuing of stock of the assets can be expressed as following (cited from SEEA2003, p.56-57):

Assets provide capital services to the production process and these are remunerated in the gross operating surplus generated. Gross operating surplus is that part of value added that remains after deducting the compensation of employees and the other taxes less subsidies on production. This operating surplus can be partitioned to show how much is due to produced assets and how much to natural assets. The part due to natural assets is the resource rent. The other part we will call economic rent though to be accurate it should really be called other economic rent. The value of the stock of the assets, whether produced or non-produced, can be equated with the present discounted value of the rent they will yield over their effective life.

Economic rent can be partitioned into a part which represents the decline in the value of the asset (the cost of “using up’ the asset), and the remainder which represents the return to the owner of the asset... In the SNA, the decline in the value of the produced assets is described as the consumption of fixed capital and it is deducted from gross operating surplus to derive net operating surplus. Net operating surplus thus covers the return to the produced asset plus the whole of the resource rent.

The value of the capital service flows rendered by the natural resources, their share in gross operating surplus, is the value of the extraction, harvest or abstraction of natural resources. It is also referred to as the resource rent. The term equivalent to consumption of fixed capital is called depletion. Depletion denotes the total volume of extractions of natural resources times the realised price per unit. It is not considered as the net effect of extractions, once the return to natural resources has been taken into account. Depletion is used, as in the SNA, to mean the change in value of the stock of the resource due to extraction.

21

Box 2: The decline in the value of fixed capital and the income it generates.

Taken from SEEA2003; UN et al. 2003, p. 275.

iii] Valuation of resource rent

There are three possible ways of estimating resource rent. The first is based on actual transactions and may be called the appropriation method. The other two methods depend on estimating resource rent by partitioning the information on economic rent for all the assets of a firm into that part pertaining to its produced assets and the part relevant to the non-produced assets. These two last methods start with the assumption that there is information available on the gross operating surplus of a firm or industry and also figures for the net capital stock of the same unit.

a) The appropriation method

In many countries, governments are the primary owners of the nation’s natural resources. The approach is based on the idea that governments could in theory collect the entire rent derived from extraction of the resources they own. Resource rent can be collected by governments through fees, taxes and royalties levied on companies that carry out extraction. However, in practice, fees, taxes and royalties tend to understate resource rent as they may be set by governments with other priorities in mind, for instance implicit price subsidies to extractors that encourage employment in the industry (SEEA2003, p. 276). Although this method may not be the most appropriate, due to its crudeness, one can compare the values with those estimated by other approaches.

22

b) Resource rent derived from PIM calculations

This method starts by determining the value of an asset which is n years old by making assumptions about the rate of decline in its value over the n years since it was purchased. This decline in this value since the previous year is set equal to the consumption of fixed capital. Net operating surplus is calculated by deducting the consumption of fixed capital from gross operating surplus (from the production account); the return to capital is calculated using the value of capital stock determined by the PIM. The resource rent earned by the unit is derived at the end of this sequence of calculations as shown in the following figure.

Figure : Resource rent derived from PIM calculations.

Taken from SEEA2003; UN et al. 2003, p. 277.

The idea is as follows (SEEA2003, UN et al. 2003, p. 277): taking the economic rent for all assets, the gross operating surplus (GOS), and deducting the consumption of fixed capital (CFC) gives the return to produced and non-produced assets or net operating surplus (NOS). The return to produced capital is taken to be the discount rate (r) multiplied by the value of the produced capital stock at the start of the year (Vt). Deducting this from the net operating surplus gives the return to non-produced assets, or resource rent (RR).

This approach is also reflected in the Eurostat Guide (2002b). The resource rent is generated as follows:

Output (basic "well head" prices) + Specific taxes less subsidies on products– Intermediate consumption– Compensation of employees– Other non-specific taxes less subsidies on production– Consumption of fixed capital– Return to fixed capital= Resource rent

Most of the variables are standard national accounts variables, except the return to fixed capital and the division of taxes and subsidies into specific and non-specific.

Return to fixed capital is calculated by applying a normal real rate of return to the net stock of fixed capital in the extraction industry, valued at the beginning of the period.

23

For EU/EEA countries, an 8% real rate of return on fixed capital should be taken as the default value (Eurostat, 2002b).

Specific taxes and subsidies are those that apply only to the oil and gas extraction industry, while non-specific taxes and subsidies apply to other industries as well. Specific taxes are considered part of the resource rent (appropriated by government); see Eurostat (2002b, page 2).

c) Resource rent derived from capital service flow calculations

The third method uses the theory of capital service flows to determine how much of the gross operating surplus represents the capital services rendered by the stock of produced capital. What is left when this is deducted from gross operating surplus is the resource rent attributable to the non-produced assets in use.

This methodology starts by considering and modelling the decline in the service provided by the asset over its life rather than the decline in price. (A light bulb for example may shed the same light throughout its life even though its value declines as it ages because the length of time for which it is expected to function declines.) Such measures of capital service flows are used in productivity studies as well as in the calculation of net income flows. The value of the capital service flows (CS) estimated from the stock of capital is deducted from the total economic rent (GOS) as recorded in the production account. The result gives the resource rent (RR) directly (see SEEA2003, UN et al. 2003, p. 278).

The capital service flows can be derived as follows (cited from SEEA2003): „If the value of the assets at the start of the year is V and the discount rate is r, then the income element can be expressed as rV. For this reason, this income is regarded by economists as representing the return to the capital used by the firm. For the firm as a whole, this item is the net operating surplus. The decline in the value of the asset is referred to as the consumption of fixed capital and is the difference between the value of the capital service flows rendered (and thus used up) and the income element which arises in the same period.“

Figure : Resource rent derived from capital service flow calculations.

Taken from SEEA2003; UN et al. 2003, p. 278.

24

iv] Decline in the value of an asset and account adjustments

There is a general understanding that the use of natural resources could be measured in a way that is consistent with the use of fixed capital. The logic provided in SEEA is as follows: Just as with fixed capital, the stock value can be estimated as the net present value of the future stream of benefits coming from the use of the resource. The benefits are equated with economic rent which is embodied in gross operating surplus of an enterprise. This can be partitioned into two parts, one part relating to the economic rent coming from the use of produced assets (fixed capital) and the other part due to the use of non-produced assets (natural resources), i.e. resource rent.

The economic rent arising from the use of a fixed asset can be partitioned into an element representing the decline in value of the asset, i.e. the consumption of fixed capital, and the remaining element which is taken to be the income arising from the use of the asset which is incorporated in the net operating surplus of the unit. Similarly, the resource rent can be partitioned into an element showing the decline in value of the natural resource and the return to its use.

The resource rent in year t, RRt, then equals to the difference in the value of the asset (RV] between the start and end of the period plus the return to the capital asset in question that represents the income element in year t:

where r is discount rate (see SEEA-2003, page 10-5 for more details).

As stated, the resource rent is able to be partitioned into one part that represents the decline in the value of the asset and one part that represents the return to the use of the asset in production which is regarded as income. There are three possible views on this subject which we review here.

The SNA prior to the 1993 version implicitly assumed that natural resources were so abundant that there was no decline in the value of the stocks, and thus that the whole of the resource rent could be treated as income. This is equivalent to saying that the term

is zero, i.e. the value of asset in the start and end of the period is the same. As the consequence all resource rent represents income.

There is an opposing view which is principally supported by proponents of the strong sustainability principle which says that the whole of the resource rent should be taken as a decline in value of the stock of the resource and none of it regarded as income. This view is a consequence of assuming that the value of the resources in the future will be the same for future generations as for present generations today, implying a zero discount rate and setting the second term in the equation equal to zero. It also implies that all rent should be excluded from NDP.19

19 This last view is equivalent to saying that natural resources should not be regarded as factors of production in the same way that fixed capital and labour are, because they were not produced (see SEAA-2003, page 10-6).

25

The majority opinion supports a mid-point compromise between these positions. The issue of how to partition the resource rent into a perpetual income stream and a depletion element therefore emerges. The “user cost approach” developed by El Serafy (1989) is one way of such partitioning. This assumes that the resource would provide an equal economic rent for each of n years, so its value is the net present value over n years of the ressource rent, RR. The income element has to be such that the net present value of X over an infinite period has to be the same as the net present value of the resource. Then, the income part of each year’s resource rent is derived as (see more e.g. SEEA-2003, Chapter 10):

Other methods include those developed by Hotelling (1931) and Hartwick (1990). As noted by SEEA-2003, whichever method of splitting income and extraction is used, the longer the life length and the higher the interest rate, the greater will be the share of income. If a zero rate of interest is used as a social discount rate, then there is never an income component to the economic rent; it is all regarded as extraction.

Once the resource rent for an asset has been determined, three further pieces of information are necessary to determine the net present value of the asset:

- For how many more years into the future will the asset generate economic rent?

- What will be the pattern of decline (if any) in the economic rent?- What is the appropriate value of this discount rate?

(Note that the term ‘extraction‘ is used to denote the total volume of extractions of natural resources times the unit resource rent and is synonymous with resource rent (SEEA-2003, page 10-7). The term ‘depletion‘ represents the effect of extractions on the value of the stock of the resource, once the return to natural resources has been taken into account and mean the change in value of the stock of the resource).

26

II.1.4 Valuing degradation

The methods described in chapter 7 of the SEEA draft revisions for valuing natural resources are not generally suitable for valuing environmental services, waste and degradation.

In accounting for environmental degradation, the 2003 SEEA regards the environment as an asset which will last forever, providing the same services year after year. It would imply that the whole of the value of that service may be regarded as income and if there is deterioration in the asset, then the decline in the value of the asset must be regarded as a deduction from income.

Ideally, one would need to devise an eco-environmental aggregate which included the value of all economic products and all environmental functions. Impairment of environmental functions would then represent a deduction from a state where all environmental functions were preserved. Unfortunately, such information is not available. Therefore a somewhat simpler and less ambitious approach has to be formulated.

The 2003 SEEA suggest damage-based and cost-based techniques as the alternatives.

1. Cost-based methods: i.e. how much would it cost to avoid the generation of residuals by changing production and consumption patterns. Within this type of method, there are two main approaches:

To estimate the value of NDP if hypothetical environmental standards are met using current technologies and costs. This is a partial equilibrium-based approach and does not allow for synergies between different environmental actions, or for the effect of increased prices on the demand and supply of environmentally damaging products. Therefore, for marginal changes it provides an upper bound of the impact on NDP.

To engage in "greened economy modelling" using a general equilibrium-based approach in order to model the effects of following the least-cost approach to reach certain environmental standards on economic activity and output. The objective is to measure the level of GDP that would be compatible with meeting environmental standards. Since all prices and quantities are endogenous, all costs are fully internalised.

2. Damage-based methods: what is the value of the damage caused by current levels of residual generation. The aim here is to derive the impact of the residuals generated. Where these impact on economic activity, e.g. though damage to crops and depreciation of capital assets, they should already be reflected in lower production, the consumption of fixed capital and in environmental depletion (although this may not be the case in practice). The largest remaining impact is to human health, and thus "damage-adjusted income" is derived by deducting all damages, including those to health.

Many of the damages do not have market prices and so have to be measured using non-market valuation techniques. These techniques are summarised in Appendix (see e.g. Chapter 9 of SEEA-2003 for further details). A belief that increased uncertainty is attendant to the values derived from techniques that move from production function approaches, through revealed preference approaches to stated preference in some countries has led to guidance to recommend that, as data and resources allow, analysts should consider the use of the former approaches before considering the latter.

27

II.1.5 Environmental adjustments to the flow accounts

i] Depletion

The argument here in favour of adjusting the national accounts aggregates for natural resource exploitation is that a further deduction should be made from net operating surplus to allow for the decline in the value of the natural resource, i.e. the value of the extraction less the return to natural resources. Such an adjustment would give a figure for depletion adjusted operating surplus.

If a decline in the value of resource stock due to extraction is subtracted from net domestic product, extraction adjusted domestic product is determined. If a decline in the value of resource stock due to extractions net of discoveries is subtracted from net domestic product, depletion adjusted domestic product is determined.

In principle, these adjustments could be made and such an account compiled for an enterprise, an industry or for the economy as a whole.

ii] Defensive expenditures

Although, the scope of defensive expenditure is potentially extensive, the discussion in the SEEA-2003 handbook centres around the subject of environmental protection expenditure. Regarding this type of defensive expenditure, it seems relatively easy to consider excluding environmental protection expenditure by government since it can be identified within output and expenditure without problem (see page 438 of SEEA-2003).

Goods and services acquired by industry are not, however, themselves part of final demand. They are incorporated into other products which fulfil this function after one or more further steps. It is therefore also an inaccurate simplification to think that because a firm’s environmental protection expenditure is counted as intermediate consumption, it does not add to GDP. It is, however, difficult to make a case to exclude the wages and salaries earned by, say, street cleaners and to classify these workers as unemployed on an “environmentally adjusted” basis. Their earnings are used to acquire goods and services for their families and there is no reason or obvious accounting means of excluding these purchases and the subsequent consequences from the accounts. Simply omitting activities and preserving a closed accounting system is not an option. (SEEA, page 439, para 112)

As further argued by SEEA-2003 (p. 439, para 10-112 and 10-115), the suggestion to remove the easily identifiable elements of defensive expenditure from the expenditure side of the accounts is not tenable within a coherent accounting system. There is no way, if this were done, that exactly compensating deductions could be made within the other methods of calculating aggregate GDP. Nor would this remove the whole of the defensive activity. A symmetric treatment of environmental protection expenditure by government and industry cannot be achieved by simply omitting some part of the accounting system. However, a form of symmetry can be achieved by reclassifying some of the existing transactions.

To overcome this possible problem, the SEEA-2003 recalls the example in maintaining roads (see page 439, para 10-116). It reminds us that the 1968 SNA took the position that

28

repairs and maintenance would be sufficient to ensure a road lasted for ever and thus that there would be no allowance for consumption of fixed capital for roads in that version of the system. Gross capital formation was taken to be a measure of net capital formation and all repairs and maintenance were treated as intermediate consumption. As further noted by SEEA, there is another possible way to reach a similar position, one which was in use in a number of Scandinavian countries before they adopted the 1968 SNA. This is sometimes known as the “gross gross” method of recording capital formation. Under this, all repairs and maintenance were recorded as part of gross capital formation and that part which would otherwise be counted as current repairs were treated as a form of consumption of fixed capital thus eliminating the double counting just introduced.

The “gross gross” approach to recording of environmental protection expenditure is one way to achieve a symmetric treatment of such expenditure by government and industry. If the expenditure undertaken by an industry is treated as both capital formation and consumption of fixed capital, the level of output of the industry on its other products will not alter. GDP will increase by the amount of the environmental protection expenditure but net domestic product will not change. The change in classification of government current environmental protection expenditure to capital formation will not affect GDP, though some final consumption will now appear instead as fixed capital formation. On the other hand, net domestic product will fall by the amount of this expenditure just reclassified. In this way we have a symmetric recording of environmental protection expenditure between industry and government and the gap between GDP and NDP is increased by the whole amount of this expenditure, by increasing GDP by the current expenditure by industry and reducing NDP by the current expenditure by government (see p. 439, para 10.117, or para 2.170).

Households can also purchase environmental protection products. As noted by SEEA (p. 440, para 10-118), simply reclassifying this household consumption to be a form of capital formation seems to introduce a new concept in the SNA - the idea of capital formation without any associated production process. However, several types of collective consumption, including the environmental protection expenditure carried out by government, could be described as comprising capital formation not directly linked to production processes. At present in the SNA, only government undertakes collective consumption, but if the guidelines were relaxed to permit other sectors to finance collective consumption, this would allow household consumption on environmental protection expenditure to be reclassified as capital formation also (see also Harrison, 1999 who suggests a similar mechanism that is an alternative to “gross-gross” recording).

iii] Degradation

Damage from environmental degradation may affect either actual transactions which are recorded in the accounts leading up to calculation of GDP, or result in subsequent adjustments to reach depletion-adjusted domestic product. Damage thus affects the value of man-made assets and environmental assets, i.e. targets the asset boundary of the SNA or those defined by the SEEA. The 2003 SEEA guidance refers these to as damage-affected transactions. The effects of degradation can be then easy recognised in, for instance, deforestation due to acid rain and ecosystem degradation caused by acidification and eutrophication. Similarly, damage to produced assets can be found. Considering the air pollution, one can identify deteriorated building materials and/or shortened asset’s life-time due to soiling and corrosion caused oxide sulphides.

29

Environmental degradation may have – and indeed does – significant adverse effects on human health. The SNA has, however, no means of taking such damage into account, unless it considers introducing human capital into the balance sheet. In this case, the health damage would be another form of reduction in the nation’s net wealth. Ideally, one would need to estimate the change in human capital due to changes in health.

The alternative route, as suggested by the 2003 SEEA (page 10-33), is to revert to the current welfare measure. This approach assumes that a deterioration in health represents a deterioration in welfare that many people would be prepared to pay to avoid if possible. In this sense, from the income generated from production we should deduct not only the allowance for depreciation of fixed capital which is part of conventional national accounts, and the allowance for depletion of natural capital, but also a sum representing the decline in welfare caused by environmental damage (SEEA-2003, para 10.146).

This suggestion represents a major innovation to the actual SNA principles. As noted by the SEEA guidance, the benefits we receive from a good state of health are not recorded in NDP, yet we suggest recording a decline in those benefits due to environmental degradation as a decline in NDP. If the welfare equivalent of health benefit is estimated to be H, then the decline in welfare due to environmental degradation would be the decline in the original sum NDP+H. This is not, however, possible as long as we have no robust estimates of H. SEEA therefore advises us to track the annual changes in damage rather than to compare it with the absolute level of NDP.

The damage-based income measure coming from the SEEA as follows:

GDP (gross domestic product)less consumption of fixed capital

equals NDP (net domestic product)less any damage adjustments to asset valuation not included in consumption of fixed capitalless depletion of natural resources

equals dpNDP (depletion-adjusted NDP)equals depletion-adjusted national income

less pollution damage to human healthequals daNNI (damage-adjusted national income)

Similarly, one can derive “damage-adjusted savings” as “damage-adjusted income” less final consumption by households and government; see more page 10-35 of the SEEA-2003. One of the well-known applications of the damage-based methods of valuation is the GARP series of projects. The GREENSTAMP methodology, which is an example of a cost based method using "greened economy modelling", arose in part because of the theoretical and practical difficulties associated with damage-based valuation methods. The rationale behind the GREENSENSE project is that if the object is to estimate the effects of current environmental pressure on current welfare, then some form of damage-based valuation is required, although not necessarily in the framework of green national accounting. However, if the objective is to assess the implications of meeting sustainability standards on economic activity, cost-based methods also need to be employed. Damage-based valuation is required if we are to identify the most stringent of welfare-based and sustainability-based standards. There will be further discussion of this later in the report. The next section provides a brief comparison of the GARP and GREENSENSE projects.

30

II.2 The GARP and GREENSTAMP projects

This section discusses the development and application of two major methodologies whose aim is to assist in the measurement of environmentally adjusted national income, namely the GARP and GREENSTAMP methodologies. The GARP project is a major example of an empirical study that gathers the type of information that could be used to estimate a Green NNP, or could be provided as satellite accounts to be interpreted alongside standard national accounts. The aims of the first phase were to use an advanced methodology developed under the EXTERNE project to provide estimates of the monetary value of environmental damages for the EU countries of Germany, Italy, the Netherlands and the UK. The aims of the second phase were to extend the method in terms of the range of pollutants covered and the attribution of the damages to different pollutant sources, and to assess the levels of defensive expenditure on various sources of damage.

The aim of the GARP welfare-based approach is to develop a practically measurable estimate of the net welfare that an economy generates. This is based on the Weitzman foundation as described above. We have noted that the problems with this approach include the fact that this type of income in general does not measure sustainability even in theory, and that observable prices in any case would not allow us to do so. Moreover, some environmental assets provide essential services but are not traded in markets. Thus, the sustainability implications of depleting these assets are not reflected in a welfare-based valuation approach.

The theoretical basis for a second approach to calculating Sustainable National Income (SNI) was developed by Hueting (1989) in response to these difficulties, and in particular to the fact that welfare-based green national income is not a measure of sustainable income. Hueting's suggested methodology is built on the idea of strong sustainability, in contrast to the welfare-based approach, and takes environmental sustainability as a starting point. The basic premise of the SNI concept is that national income can only be sustainable if essential environmental services are themselves sustainable. Thus, sustainable income is defined as the maximum level of net economic output that is compatible with respecting scientifically identified standards of environmental sustainability. This, given publicly declared support for a policy of sustainable development, can be taken to be the "social demand" for environmental goods.

The Hueting method estimates an adjusted GDP, compatible with undertaking the abatement necessary to reach sustainability standards, using existing market prices. This partial equilibrium approach however does not account for the general equilibrium effects of the non-marginal changes in economic activity that would be associated with implementing environmental sustainability standards. It therefore represents the first of the cost-based methods outlined above. While the GREENSTAMP team were in sympathy with the aims of Hueting in measuring sustainable national income, they proposed a third methodology that would account for the total price and quantity effects of imposing sustainability standards.

The GREENSTAMP methodology, like the Hueting methodology, is based on a requirement for strong environmental sustainability. This approach rejects the monetisation of the benefits of environmental goods in order to apply cost benefit analysis to identify the most efficient allocation of resources at a macroeconomic level. Rather, according to this

31

approach, the appropriate way in which to express society's "demand" for environmental goods is by respecting environmental standards compatible with preserving the ecological base required for sustainability. The GREENSTAMP methodology avoids the problems associated with the Hueting methodology by estimating sustainable national income using an empirically calibrated multi-sector equilibrium model of a national economy to estimate the economic output that is consistent with respecting environmental sustainability standards. However, the major disadvantage with this approach is, arguably, that the complexity of a real national economy is so great that the inaccuracies associated with attempting to model an economy would be at least as great as those associated with the other two methods.

Information delivered by each approach

In the GARP approach the integration of scientific knowledge into the methodology is crucial, representing as it does the basis for the methodology. The core objective of GARP is to provide a scientifically robust method for the calculation of damages to human health, crops and materials. These are first calculated in physical units by applying the impact pathway approach, based on consolidated exposure-response functions, and are subsequently transformed into monetary terms. Thus, information is delivered to policy makers in the form of both physical and monetary impact estimates.

This information is particularly important from a policy perspective because it constitutes a basis on which scientifically and economically justifiable environmental standards can be identified. However, these standards do not necessarily represent the standards required for environmental (or economic) sustainability, and moreover they are based on the monetary valuation of environmental damages, which is often considered to be controversial.

The GREENSTAMP project provides policymakers with information on the costs to be borne by society in meeting given standards of environmental protection. The idea behind the GREENSTAMP project is to adopt a cost-effectiveness approach, initially developed by Baumol and Oates (1971) in the context of the implementation of environmental quality standards for pollution control. The cost-effectiveness approach "takes a performance standard as given and aims to identify the way of attaining it that involves the least expenditure". As Brower, O'Connor and Radermacher (1999) point out "the standard setting is partly dependent on the initial estimates about the probable magnitude of these costs". In GREENSTAMP environmental standards are assumed to arise from a political decision making process, and given the commitment of governments to follow a policy of sustainable development, are considered to correspond to environmental sustainability standards. Once the standards are set, the next step is to implement these standards within a decision making process, resulting in scenario modelling. These modelled scenarios in turn yield estimated abatement cost curves. The implementation of environmental standards is based on costs of abatement associated with the hypothesis envisaged in each scenario. The aim is for this approach to allow policy makers to choose cost-effective solutions to meeting scientifically defined objectives.

However, one of the basic assumptions of this method, namely that the awareness and discussion of the requirements of environmental and economic sustainability is advanced, and that society accepts the need to achieve strong sustainability standards, is in fact difficult to defend. Political and economic realities imply that the policy decisions

32

regarding environmental standards will be based on budget constraints rather than on scientifically defined sustainability targets.

The measurement objective of the two projects is of course different. Therefore, the amount that each methodology would deduct from the measure of national income, would also be different even if the efficient environmental standard happened to be the same as the sustainability standard. Under the GARP welfare-based approach this would be the total welfare damage at the current level of pollution. This is illustrated in Figure as the area under the damage cost function from the vertical axis up to the level of current damage. Under the sustainability standards approach the measured cost to be deducted from national income would be the cost of getting from the current to the sustainable level. This would be the area under the AC curve between the current level and the sustainable level of pollution.

Figure Comparison of the objective of measurement between GARP and GREENSTAMP Source: Adapted from Fig. 22 in: http://esl.jrc.it/envind/theory/handb_09.htm

II.3 Genuine Saving

It was noted above that it is unlikely to be feasible to provide a single indicator that measures both current welfare and long-term sustainability. Thus, it may be necessary to divide the reporting framework into an indicator (possibly disaggregated indicators) of current wellbeing on the one hand, and an indicator (possibly disaggregated indicators) of sustainability on the other. An example of the latter, which is separate from indicators of wellbeing, is the Genuine Saving framework.

Genuine Saving is a methodology that concentrates on the measurement of total wealth as a sustainability indicator. It builds on original work done by Pearce and Atkinson (1993). Hamilton and Clemens (1999) present a model explaining the theoretical basis of genuine saving. Genuine saving is defined as investment in produced assets and human capital less the value of depletion of natural resources and the value of accumulation of pollutants. Thus, it is in effect the value of the changes in the economy’s stocks. Negative genuine saving is shown by Hamilton and Clemens to correspond to unsustainability, given the assumptions of their model.

33

Hamilton and Clemens note that Vincent et al (1997) show that for a small resource-exporting country that is a price-taker, NNP should include the present value of capital gains on resource exports. Being a capital effect, this would also affect genuine savings. However, they maintain that given near-constant real resource prices, capital gains can in effect be assumed to be approximately zero.

Hamilton (1995) argues that current education expenditures are not consumption, and therefore should be considered as investment, or saving. Thus, the formula for calculating genuine saving from real data is:

G = GNP – C - K – n(R-g) - (e-d) + mWhereG = Genuine SavingGNP = Gross National ProductC = ConsumptionK = depreciation of produced capitaln = the value of resource depletionR = resource extractiong = renewable resource growthe = pollution emissions d = pollution assimilation = marginal social cost of net pollution accumulationm = expenditure on education

n(R-g) is the value of the net growth of renewables, which is deducted when negative, but is not added when positive, on the grounds that much biomass growth is not economically valuable.

Hamilton and Clemens provide Genuine Savings calculations for a broad range of countries, although the adjustments are limited to the valuation of non-renewable resource depletion and of forest depletion, and the marginal social costs of CO2 emissions. Resource rents on non-renewable resource depletion are calculated as the world price for the resource minus average cost. The authors note that in fact marginal rather than average cost ought to be used to estimate resource rents, but these are very difficult to estimate.

A further caveat noted is that where countries sell their resources internally at less than the world price, using world prices to value depletion will overstate rents. In terms of valuing deforestation, the values used are stumpage values, which accounts neither for the non-timber values of forests, nor for the value of land in its alternative use. Carbon dioxide emissions are valued at a global marginal social cost of $20 (US$1990), a figure taken from Fankhauser (1994). These damages are charged to emitting countries, on the basis of a global Polluter Pays Principle.

The calculations of genuine saving performed by the World Bank (e.g. Hamilton and Clemens, 1999) begin with the calculation of Gross Saving, the traditional measure of wealth accumulation reported by the World Bank's World Development Indicators. This is measured as:

Gross Saving = GDP - Public Consumption - Private Consumption

34

The next stage is to calculate net savings, a first step towards a sustainability indicator as it accounts for asset depletion, as:

Net Saving = Gross Saving - Depreciation of Produced Assets.

Finally, Genuine Saving is calculated as:

Genuine Saving = Net Saving - Value of Resource Depletion and Pollution Damages.

The results are calculated for the period 1970-1994 and show that accounting for resource depletion in this way does affect the indicators of sustainability based on national accounting data, since several countries have positive net savings, but negative Genuine Savings. Genuine Savings rates are particularly low, and often negative, for Sub-Saharan Africa and for the Middle East and North Africa. The results are presented in Figure A1. It is noted that this results in part from the overestimation of resource rents, due to average rather than (the higher) marginal costs being used.

Figure A2 in Appendix shows the components of genuine saving as shares of GNP for Tunisia, omitting human capital investment. The top curve is gross domestic investment. Net foreign borrowing is then subtracted from this to give gross saving – the difference between production and consumption over the year. The depreciation of produced assets is then deducted, yielding the curve for net saving. Finally, the bottom line is genuine saving, obtained by subtracting the value of resource depletion and pollution damages from net saving. A second version of Genuine Savings is calculated by including expenditure on education as investment in human capital. Including such expenditure as investment can have a significant effect on Genuine Savings. For example, it means that genuine savings rates in Sub-Saharan Africa are brought close to zero.

The rationale behind calculating Genuine Savings as the real level of an economy's investment, rather than calculating investment as part of Green NNP, is that the wealth issues captured in Genuine Saving provide a more direct indicator of sustainability than does NNP, which combines current wellbeing with investment. One of the major criticisms of Genuine Saving (e.g. Martinez-Alier 1995) is that it provides an indicator of only Weak Sustainability, assuming as it does that investment in produced capital can compensate for decreases in the value of natural resources. The answer to this given by Hamilton and Clemens is that the calculation is useful, in that while a positive Genuine Savings rate may not indicate sustainability, a negative rate is very likely to indicate unsustainability. This seems to be true, as long as one can be reasonably sure that the values imputed for resource depletion are not overestimates. If one accepts that they are not, then the policy implications of Genuine Saving are clear: persistently negative rates of Genuine Saving will lead to declining wellbeing in the future.

35

II.4 Alternative indicators of welfare

It was noted above that it may be necessary to construct separate indicators for current wellbeing and long-term sustainability. The last section described a framework which can be used as an indicator of (weak) sustainability, but that does not attempt to measure current wellbeing. A variety of suggestions have been made as to what should be included in a measure of economic wellbeing. One such suggestion, which has been applied in several countries including the USA, is the Genuine Progress Indicator. The GPI adjusts personal consumption for income inequality, adds the value of housework and voluntary work, and subtracts estimates of the costs imposed by social breakdown, including crime and family breakdown, and environmental damage. Loss of leisure time is deducted from the GPI, as are the costs of long term issues such as resource depletion, climate change, nuclear waste and the accumulation of foreign debt. Finally, the services of both private and public capital goods are added to the index, while their cost in the year of purchase is deducted. This is to avoid the index being inflated, as is GDP, by the more regular capital purchases necessitated by capital goods becoming less durable.

In this section we look in detail at another, related, indicator that has also been widely applied, the Index of Sustainable Economic Welfare. This index, as the discussion will show, focuses largely on correcting consumption for other issues that affect wellbeing. It does also include some factors that relate to long-term wellbeing, we argue here that this is not consistent. However, this index provides a framework that could be adjusted to provide an estimate of wellbeing within the framework of this project.

II.4.1 Introduction and Background: ISEW and MEW

Nordhaus and Tobin (1973) note that in the postwar period “growth” has become an accepted aim of economic policy. In neoclassical economics, this has become more than ensuring that the economy reaches its potential in terms of output, and has been associated with actively maximising the level of output by increasing the productivity of labour. However, many had become disillusioned with policies to maximise growth in GNP, asking among other things whether GNP is a good measure of economic wellbeing and whether growth involves “wasting” natural resources. In order to investigate these questions, Nordhaus and Tobin developed a Measure of Economic Welfare (MEW) to correct for some of the discrepancies between GNP and economic welfare, or utility. These adjustments fall under the following headings:

1. Reclassification of GNP final expenditures. This involves subtracting some items conventionally counted as final output, but which could in fact be considered as “instrumental”, or intermediate, in that they do not yield direct utility but are an indirect means to obtaining utility. Expenses involved in commuting to work could be deducted for this reason, as could some government expenditure, e.g. police, road maintenance, sanitation and national defence. In terms of capital consumption, NNP corrects for depreciation of capital stocks, but does not account for the role of durable goods. Nordhaus and Tobin note that including durable goods as capital could avoid the paradoxical situation whereby output rises the more perishable they become. One can also add a “growth requirement” to account for the amount by which output must increase if per-capita consumption is not to decrease, given levels of population growth.

36

2. Imputations for Capital Services, Leisure and Non-market Work. This involves firstly imputing returns on owned capital stocks that are not considered as such in the national accounts, in particular consumer durables and public investments. Secondly, it involves imputing a value for leisure time to correct for the fact that welfare could increase while NNP decreases if people choose to work less. However, the rate at which to value leisure time and non-market goods is debatable.

3. Disamenities of Urbanisation. This involves estimating the extent to which the higher incomes received by residents of more urban areas is to compensate them for the externalities connected with urbanisation and congestion, and therefore the proportion of income that should not be included as welfare. Nordhaus and Tobin estimate this by regression analysis, showing that the resulting premium is substantial, being around 5% of GNP.

The adjustments made by Nordhaus and Tobin show that although MEW grows more slowly than NNP over the period 1929 – 1965, it does grow, indicating that growth in output is not incompatible with growth in economic welfare.

One can argue that including certain products in GNP as an estimate of wellbeing is arbitrary, as many sources of utility are excluded. Juster (1973) argued that much of household consumption could be classified as intermediate. Leipert (1989) argued that defensive expenditures, i.e. those made to limit “damages and deterioration that industrial society’s process of growth has caused to living, working and environmental conditions”, should not be included in measures of wellbeing. So, for example, the costs of road building and maintenance, and health treatment due to road accidents, should not be included in GNP. The issue of inequality is central to the calculation of the ISEW. Questions on who benefits from net output raise issues such as poverty, sexism and ethnic discrimination. Sen (1993) argues that not only the level of per capita income, but also the way in which it is distributed determines society’s level of wellbeing. This is reflected in the UNDP’s Human Development Index (HDI), which reflects human development in terms of improved health, knowledge and skills, and the use people make of their potential, in terms of leisure, productivity and social, political and cultural participation. A variation on the Atkinson index20 is used to convert national income into a measure of social welfare, W, as follows:

where y is the country’s income, and relates the nation’s income relative to the global average. If per capita GDP is below the global average, = 0 and W = y, while if it is above the global average, additional income contributes relatively little to human development. England (1998) suggests that although the UNDP claims that the HDI can be applied to all countries, the index is better suited to highlighting distinctions between the development paths of different developing countries, among which it shows dramatic differences.

II.4.2 Calculating the ISEW

20 The Atkinson Index estimates the proportion of a country’s total income that, if incomes were equally distributed, would provide the same level of social welfare as is currently achieved.

37

We have seen that a variety of suggestions as to the appropriate adjustments to indicators of wellbeing have been made. One well-known contribution which attempts to integrate these into a single measure is the Index of Sustainable Economic Welfare (ISEW), built on an original contribution by Nordhaus and Tobin, first proposed by Daly and Cobb (1989), and updated in the 1994 edition of the book.

The ISEW, as noted, attempts to measure current wellbeing, adjusted to account for issues relating to sustainability. The proponents of the ISEW argue that current welfare should be measured as the current flow of services from all sources, rather than current output of marketed goods. They estimate this flow for the USA by adjusting consumption as follows (the figures correspond to the authors’ calculations for the year 1990, in US$ (1970)):

ISEW = Personal consumption adjusted for income inequality 1164(a) + Services of household labour 520(b) + Services of consumer durables 225

+ Services of highways and streets 18+Consumption portion of public spending on health and education 45

(b) - Spending on consumer durables 235(c) - Defensive private spending on health and education 63

- Costs of commuting and auto accidents 67- Costs of personal pollution control 5 - Cost of air, water and noise pollution 39

(d) - Loss of wetlands and farmland 58(e) - Depletion of non-renewable resources 313

(f) - Long-term damage from nuclear wastes, greenhouse gases and ozone depletion 371

+ Net capital growth 29 Change in net international investment position 34

= $818(a) Valued at the wage rate of domestic workers(b) Since consumer durables provide services throughout their lives and not solely in the period of purchase, the index deducts their cost and adds an imputation for their services each period. In Daly and Cobb’s calculation for the year 1990, these two roughly cancel out.(c) Daly and Cobb maintain that education serves mainly to ration jobs according to the commodity of qualifications, rather than adding directly to wellbeing, or to capital. This has been criticised by Eisner (1994) in that it ignores the contribution of human capital to productivity.(d) This refers to the annual loss of productive services due to current and past conversion of wetlands and agricultural lands to urban lands. This entails increasing marginal costs of land conversion over time.(e) Rather than using a change in value, or user cost, approach to value depletion, the authors value depletion of reserves at the value of a renewable substitute, namely ethanol. (f) This adjustment acknowledges the costs that current economic activity places on future wellbeing by contributing to damaging stocks of, for example, atmospheric methane, CO2, stratospheric chlorine and nuclear waste. However, for greenhouse gases the adjustment is made on the basis of US rather than global emissions, and energy-related pollutants are assumed to persist infinitely.

38

Daly and Cobb’s initial estimates for the USA have been followed up by an application for the USA by Cobb and Cobb (1994) and estimates for Germany (Diefenbacher 1994) Italy (Guenno and Tiezzi 1996) Sweden (Jackson and Stymne 1996) and the UK (Jackson et al 1997). Updates to the methodology have been made by Jackson et al (1997) and Stockhammer et al (1997) in their application to Austria. The conclusion for all these studies is that sustainable economic welfare has risen much more slowly than GNP, and may even have fallen since 1980. In the case of the UK study, the decrease in real wellbeing is seen largely to follow from “environmental degradation (in particular depletion of non-renewable resources and long-term environmental damage) and income inequality” (Mayo et al 1997). This has led to a “threshold hypothesis" (Max-Neef 1995) to the effect that economic growth increases wellbeing but only up to a point, beyond which it results in declining wellbeing.

Stockhammer et al note that the ISEW goes much further than the SEEA, in that it aims to account for social costs, future costs and distributional costs, as well as depletion and environmental costs. They note that there are still significant problems with the ISEW including data availability, the monetisation of environmental damages, the need for arbitrary assumptions, and the treatment of distribution. The fact that leisure is not included is considered problematic, although measuring and valuing leisure would be difficult. The most important critique, in their opinion, is the fact that a multi-dimensional issue like welfare (and sustainable welfare at that) should be reduced to a single, monetary, indicator, which ignores its qualitative aspects.

England (1998) notes that a problem with the methodology in terms of Daly and Cobb’s original study is that over half of the difference between per-capita GNP and the ISEW is accounted for by long-term damage from non-renewable energy and CFC use. The valuation of these factors is somewhat speculative, and therefore the calculations and framework should be taken as a base for future work, rather than as definitive estimates.

Neumayer (1998) provides further criticisms of the ISEW methodology. Firstly, he notes that the ISEW lacks sound theoretical foundations. The fact that it is not derived from an optimal control model means that it does not provide a theoretically sound measure of welfare. Part of this criticism is that the items that are included or excluded are somewhat arbitrary. The case of defensive expenditures is clearest, in that it is difficult to know where to draw the line as to what is a defensive expenditure and what contributes to utility. For instance, he notes, are food and drink not defensive measures against hunger and thirst?

Secondly, he notes the point made by Stockhammer et al and England that the results are highly dependent on certain, rather arbitrary, assumptions. Regarding the weighting of income distribution, this has a significant effect on the difference between ISEW and GNP, and therefore care should be taken in making such an adjustment – widely differing conclusions can be drawn depending on the assumptions used. Furthermore, the valuation of non-renewable resource depletion is arbitrary, as is the valuation and long-term environmental damage.

Finally, he notes that one cannot measure current economic welfare using factors that pertain to future wellbeing and sustainability, and likewise an indicator of sustainability should not include factors pertaining to current welfare. Therefore, he maintains that one needs at least two indicators to measure these two distinct issues. Furthermore, the reduction of the issue of sustainability to a single indicator seems to imply substitutability

39

in consumption and production – an assumption that the advocates of strong sustainability would usually dispute. Therefore, the major problem that Neumayer identifies with ISEW is not the imperfections of its components since “in some way or other every social indicator is imperfect”. The problem rather is that it promises to measure something, namely current welfare and sustainability, that cannot reliably be measured in one indicator”, and his basic conclusion is that it is better to stick with GNP, while acknowledging that it is not intended as a measure of welfare.

II.5 Greensense

The GREENSENSE project therefore attempted to develop a framework for economic-environmental accounting that addresses the criticisms that have been made of the existing frameworks, while capitalising on their strengths. The framework developed is known as the Index of Consumption Corrected for Environmental Damage (ICCED). The major objectives of the reporting framework are to reflect:

(a) How increases in wellbeing are reduced when (certain categories of) environmental impacts are accounted for.

(b) How far current levels of environmental impacts are from (some definition of) sustainable impacts, and what dynamic policy targets could bring environmental impacts down to sustainable levels.

(c) The net effect on projected (short-term) future wellbeing of implementing the policies identified in (b).

The categories addressed in the project were Air Pollution, Climate Change, Biodiversity Loss, Natural Resource Depletion, Toxic Substances, Urban Environmental Problems (specifically noise), Waste and Water Pollution. These constitute eight of the ten main categories of the EUROSTAT pressure indices; the ninth ‘Marine Environment & Coastal Zones’ is excluded here, as the focus is on land area of the EU, and the tenth ‘Ozone Layer Depletion’, were excluded as this problem must be (and is being) addressed at the global level.

The two sets of information - on environmental damage costs and costs of meeting sustainability targets are combined in a country’s ICCED table on a per-capita basis. The ICCED table for Germany is presented in . The first and second columns in the table show consumption corrected for environmental damage for the years 1990 and 1998. The third and fourth columns show consumption corrected for environmental damage for projected economic growth and environmental damage, under two separate environmental policy assumptions. These are; intermediate sustainability, that is meeting the standards envisaged in current EU legislation, and strong sustainability, that is of putting the economic-environmental system onto a path that involves bringing the damage and depletion of the environment to within the assimilative and regenerative capacity of the environment before damage is done to critical natural capital. The sustainability standards used are presented in Table 12. Air pollution damages dominate the numbers although this is partly due to the incompleteness of some of the other impact categories - most notable water pollution - following from the lack of data available, or robust exposure-response functions in the case of toxic substances. Second, we note the fall in damages between 1990 and 1998, in Germany, where they have fallen by 56 percent respectively.

40

Table 2: The ICCED for Germany

The sustainability side of the framework, which shows the estimated avoidance costs of meeting the sustainability standards in 2006 for Germany, are given in . Again the data are most complete for air where robust technical cost data were supplied by IIASA. (however no information is available for the PM10 reduction). Data for the avoidance costs are also not available for biodiversity, toxic substances and noise and partially available for waste (only for the intermediate standard) and water pollution.

Table 3. Costs of meeting Sustainability Standards: Germany

Reporting Stock Values for Weak SustainabiltyTable 13 shows the estimates of the indicator of weak sustainability for the capital stocks for which this was possible.

Table 4: Weak Sustainability measures

41

The data show how dominant the figures for man-made capital are in these three countries. Valuation of ecosystems hardly registers on the scale of values, although this is partly due to problems in valuing these resources in a credible manner. Other researchers have come up with very high values but the methodology they have used is simply not credible. On the other end of the scale, the calculations made here come up with values that are intuitively too low. For natural resources, the valuations include only mineral resources relevant to energy, which now make up a very small part of the capital stock in these countries.

Given the difficulties in calculating measures of natural capital, even in advanced countries such as these, the prospects for a credible weak sustainability measure must be considered as poor. The same applies for the genuine savings measure, for which differences in these stocks are needed. Subtracting two large, uncertain numbers does not give a more accurate smaller number.

II.6 Costs and Damages Associated With Climate Change

This sub-section is taken from Tol and Heinzow, in Markandya et. al., (2007). shows the estimated marginal damage costs of climate change. Marginal damage costs represent weak sustainability. Climate change impacts are discounted with a pure rate of time preference of 1%, or a discount rate of about 3%. Dollar impacts are summed without weighting. also shows the estimated avoidance costs of climate change. Avoidance costs represent intermediate and strong sustainability. Under intermediate sustainability, carbon dioxide concentrations are limited to 550 ppm, roughly a doubling of pre-industrial concentrations. Under strong sustainability, carbon dioxide concentrations are limited to 450 ppm, and emissions are forced to zero by 2200. Costs are average costs, calculated as the ratio of the net present consumption losses and the net present emission reductions. The pure rate of time preference is 1%. Emissions and emission reduction costs are summed without weighting. Under intermediate sustainability, all regions with an income above $2500/person/year reduce their emission; under strong sustainability, this is lowered to $2000. Emission allocations are such that each region faces the same initial relative emission reduction – emission reduction increases such that the present marginal emission reduction costs is constant, corrected for the differential carbon cycle effects. The emission allocations are the basis for international trade in emission permits in which all regions participate, regardless of their income.

shows that the marginal damage costs are about $4/tC. The avoidance costs under intermediate sustainability are two orders of magnitude larger, around $400/tC. Under strong sustainability, the costs double again, to roughly $800/tC.

Table 5. Estimated damage and sustainability costs of carbon dioxide emissions.

Name Description Monetary valueWeak sustainability Marginal damages of climate change $4/tCIntermediate sustainability Limit CO2 concentrations to 550 ppm $396/tC

Strong sustainability Limit CO2 concentrations to 450 ppm, zero emissions by 2200 $819/tC

Source: Tol, in Markandya et. al. 2007

42

The figures are quite startling. Damage estimates are very much smaller than the costs of meeting either intermediate or strong sustainability targets and the costs of the strong sustainability target is an order of magnitude greater than that for the intermediate target. What does this imply for policy? Unfortunately, given the great uncertainty in the damage estimates, the conclusions are weaker than the numbers might suggest. Tentatively one might say that an intermediate target is justified on precautionary grounds but that aiming for the strong sustainability target would be difficult to justify. What is clear, however, is that there will be significant returns in policy making terms to getting more information on the damages.

Looking at non-renewable resources from a climate change perspective one can conclude that the use of these resources would be limited by climate change sustainability criteria rather than by availability. In other words, it is not that the planet will run out of non-renewable fossil fuel resources but rather that the climate change targets will limit the use of these resources over the next 100 years. The analysis also showed, under a number of energy use scenarios, that the present renewable targets set by the EU are in line with those required under the most environmentally-friendly socio-economic scenarios to meet the intermediate climate change objectives.

43

II.7 Conclusions on methods and a step-by-step procedure to include the external costs in national accounting framework

The conventional GNP national accounting measure does not include negative effects on welfare from environmental pollution, nor does it give any indication as to whether the country’s economic activity fulfils any criteria of environmental sustainability, however defined. Various recent research efforts have attempted to address these concerns and develop more comprehensive measures of welfare and sustainability – the EC GREENSENSE project being the most recent. The UNSEEA work also continues to consider how these measures should best be used with the conventional SNA framework that measures GNP; UNSEEA (2003) provides a comprehensive guide to the principal methodological issues involved. At present it is perhaps fair to say that the lack of environmental data availability somewhat limits the empirical results for policy purposes at present, though the magnitude of the estimates made in the various research initiatives indicate that welfare and sustainability issues associated with the environment should remain high on the policy agenda.

II.7.1 Incorporating the external costs into the national accounting framework: A step-by-step Guide

We consider here three general families of national accounts adjustments: resource depletion, defensive expenditures and environmental degradation.

i) resource depletion

It should also be possible – following Greensense - for some countries to estimate energy resource depletion over a time period and compare this with a change in man-made capital over a similar time period i.e. apply a weak sustainability criterion based on Genuine Savings. This will require derivation of resource rent estimates and depletion-adjusted operating surplus. Doing so, we suggest following four computational steps.

Step One

One needs to compile the stock account for subsoil asset (e.g. coal, gas, oil) in physical units. Possible sources for the basic data on volumes include geological survey institutions, relevant ministries and national statistical offices, or specific energy statistics departments and extraction companies.

Although there is no established international standard classification for subsoil assets, as suggested by EUROSTAT Guidance (2002b), one may use the McKelvey box which illustrates the classification of resources based on geological and economic criteria. While the geological view classifies the resources according to the degree of certainty that is determined by the result of exploration and development activity, the economic view

44

classifies the resources according to the profitability of exploration that varies with changes in prices and extraction technology.

Table l: The McKelvey Box

Physical resource baseDiscovered Undiscovered

Established Possible Hypothetical SpeculativeProven Probable

Economic Developed X - - - -Non-developed X X X X X

Sub-economic X X X X XNon-economic X X X X X

Following the Eurostat Guidance, according to ESA § 7.41, subsoil assets are “proven reserves of mineral deposits located on or below the earth’s surface that are economically exploitable given current technology and relative prices….”. As also noted, since the cost of identifying new reserves is often very high, oil companies only identify the volume necessary for a limited time of extraction, typically 5 to 10 years. Therefore, the volume of proven reserves is not representative of the overall volume of reserves of oil and gas present in the economic territory of EU/EEA countries. The EUROSTAT Task Force therefore decided to consider not only proven reserves, but also probable and possible reserves, as well as undiscovered reserves.

For monetary valuation, only “economically recoverable” resources have a non-zero value in balance sheets. Therefore it was decided – as a best estimate – to include in the subsoil asset accounts the expected, or probability weighted, level of discovered and undiscovered reserves. If reserve data weighted by probability is not readily available in the country, the Guidance suggests using the default 1.0 for proven reserves, 0.5 for possible reserves and 0.1 for probable reserves. Alternatively, the non-weighted sum of proven and probable reserves can be used as a second best estimate for the expected level of discovered reserves and the expected mean and the median will coincide.

Step Two

One needs to compile data on physical stocks and flows, as reported in the table below, in physical units for a given period, e.g. a year. This table might be used for weighted and for unweighted data on reserves. Moreover, one might wish to compile time-series of stock and stock changes for the respective reserves.

Table : Physical stocks and flows

DiscoveredUn-

discovered TotalProvenProbable PossibleDeveloped Non-

developedOpening stocksExtractionOther changes in volume Of which discoveriesClosing stocks

45

Step Three

The resource rent is calculated as the net operation surplus less return to the fixed capital. One can also alternatively derive the resource rent by following the capital service flow calculation. Application of the “appropriation method” might be also used to derive the resource rent, though it may result in underestimation of the resource rent.

In the resource rent calculation, following the Eurostat Guidance (2002b), one can assume the following:

- The period during which the asset generates rent corresponds with lifespan of the reserve,

- the current pattern of economic rent will remain constant (ceteris paribus condition),

- 4% discount rate. This is close to the average real rate of return on government bonds.

- 8% real rate of return on fixed capital should be taken as the default value for EU/EEA countries.

Moreover the Eurostat Guidance (2002c) suggests using the following rates in conducting sensitivity analysis:

discount rate of 0%, 2%, 4%, 6%, and 8%, and 6% and 10% of rate of return to fixed capital, using current year to describe the extraction pattern, and 3-year moving

average.

Table : Value of closing stock of oil and gas, based on different assumptions for unit resource rent, rate of return and discount rate

Natural Resource, e.g. coalUnit rent

3-year moving average Current yearRate of return Rate of return

Discount rate 6% 8% 10% 8%0%2%4%6%8%

Source: Eurostat 2002

One then needs to apply one of the three options for partitioning the resource rent into income and depletion components. We here suggest, as a conservative option, assuming the El Serafy “user cost” approach to be used. As an alternative, we suggest assuming that the depletion component is represented by the total resource rent.

The basic valuation formula for the stock, Vt, was introduced by Eurostat Task Force (2002b) and is as follows:

46

where RR is the per unit resource rent for the current year t, QT denotes the total quantity of reserves at the end of the year, n is the length of the reserves, where n can equal to Q T/Qt

with Qt the quantity extracted during the year and r the assumed discount rate.

The Task force also suggests the following classification and corresponding valuation methods for the changes in the monetary value of subsoil assets (Eurostat, 2002b, page 31).

Step Four

After the change in the value of stock is calculated, one can easily derive the extraction-adjusted, or depletion-adjusted account of national income.

ii) environmental degradation

As noted in Part I, GNP measures economic activity in terms of the amount of final demand satisfied by economic output. The economic activity has, however, many negative external effects (damages). A major example of such negative effects is the airborne pollution associated with energy generation or transport.

Some of the damage caused by airborne pollution is reflected in reduced output and thus it is already reflected in the aggregates of national accounts such as GDP. It includes reduced forestry output and agricultural crops due to polluted soils and water supplies. If more rapid depreciation of soiled and/or corroded construction materials and building was reflected in consumption of fixed assets, the degradation associated with acid rain would be directly reflected in net domestic/national income. However, shortened life-time of these assets are in reality unlikely to be measured. Thus, one can identify the externality associated with damage on building materials as environmental degradation and subtract it from income. It might be the case that such damage due to acid rain can be remediated and restored by reparation and maintenance expenditures. In this case, such expenditures represent defensive expenditures and one should be careful to subtract them from the national accounts (see below on this issue).

47

Environmental damage can also directly affect human health and thus welfare. There are several ways in which this can be the case.

Firstly, loss of welfare can be due to the discomfort of being ill as a consequence of breathing polluted air or ingested toxic substances. Moreover, loss in welfare may be caused by premature death caused by pollution or occupational accidents, as measured in the external costs assessment. Such impacts represent non-market goods and are not included in the standard national accounts. As suggested earlier, under certain assumptions, this externality can be subtracted from national accounts to compile the “damage-adjusted national income”.

Secondly, emissions of heavy metals, specifically of mercury and lead, can lead to human development impairment that consequently result in lower future earnings. A reduced flow of earnings has a direct effect on future incomes, and thus welfare. These reductions are not, and will not be, captured in the national income. Our suggestion would be to add them to current national income giving a hypothetical “without damage” level of income. Such damage is calculated as present value of the stream in earnings’s reductions.21

Thirdly, earnings can be also reduced due to sickness caused by pollution. In this case, one can expect a short-term impact on income and GNP. As in the case of human development impairment, it results in lower income.

Lastly, being ill can also involve medical treatment (mitigating) costs and/or averting expenditures. These costs might be borne by the public health security system, the private health system or by out-of-pocket expenditures of the individual. All represent, however, a kind of defensive expenditure (see below).

Pollution can also reduce the recreational and aesthetic benefits that people derive from the environment. Moreover, there could be other forms of dis-amenity such as those associated with noise or visual intrusion. Neither one is included in the current measure of income. Thus, for any level of output, welfare can be generally assumed to be higher, the better is environmental quality.

We therefore cluster these effects in two groups – they may either present damage that might be subtracted from income, or damage that should be added to the current level of national income in order to get a new hypothetical level of national income that would be reached if no damage were to occur.

21 Caveat: whilst the hypothetical reduction in earnings will last up to the year when the victim exits the labour market, the adjustment is made only in the year when the damage occurs.

48

Table : Treatment of environmental damage associated with energy externalities.

Damage that should be subtracted

Current level of national income

Damage that should be added

Fastened depreciation due to deviations from prospected lifetime of asset (buildings) Repair and maintenance costs of buildings

Diswelfare due to mortality Increased yield in forestry and in agriculture

Diswelfare due to inconveniences associated with morbidity

Loss in earnings due to human development impairment

Disamenities (noise, visual intrusion, recreation, aesthetical values] Loss in earnings due to sickness leave

Climate change

Given the energy externality focus of the NEEDS project, air pollution emissions can be obtained from EcoSense per installation and/or per kWh / GJ generated. This externality measure can be also obtained from the EcoSense report for each pollutant and each impact category. This allows us to dis-aggregate value into the two groups we have described above.

The Ecosense modelling of air pollution impacts also produces welfare cost estimates at a national level that can then be compared to the national consumption, according to the Greensense ICCED, in the countries we are interested in.

iii] defensive expenditures

It is not straightforward to subtract defensive expenditures from national income. The 2003 SEEA suggests applying the “gross gross” approach that allows us to achieve a symmetric treatment of expenditures by government and industry.

If the expenditure undertaken by an industry is treated as both capital formation and consumption of fixed capital, the level of output of the industry associated with its other products will not alter. GDP will increase by the amount of the environmental protection expenditure but net domestic product will not change.

The change in classification of government current environmental protection expenditure to capital formation will not affect GDP, though some final consumption will now appear instead as fixed capital formation. On the other hand, net domestic product will fall by the amount of this expenditure. In this way – as noted in the 2003 SEEA – we have a symmetric recording of environmental protection expenditure between industry and government and by increasing GDP by the current expenditure by industry and reducing NDP by the current expenditure by government the gap between GDP and NDP is increased by the whole amount of this expenditure,.

There is, however, no capital formation in the household sector. If the guidelines were relaxed to permit other sectors to finance collective consumption, this would allow household consumption on environmental protection expenditure to be reclassified as capital formation also.

49

We therefore suggest the following:

Externalities associated with impacts on building materials that are measured by repair costs should be treated as both capital formation and consumption of fixed capital. It implies that the level of output, and GDP, will consist of these expenditures, but net domestic product will not change.

Externalities associated with human health impacts and including medical treatment costs paid by the public health system should be considered as capital (human capital) formation. Then, these expenditures and thus externality would remain in the GDP measure. However, they should be subtracted in compiling the measure of (adjusted) net national income.

Externalities associated with human health impacts and including out-of-pocket health expenditures spent by individuals should be treated similarly to government defensive expenditures, as long as we permit the household sector to finance collective consumption. In this case, these expenditures can be reclassified as capital formation as well. They would be embodied in GDP, but should be subtracted in adjusted net national income measurement.

50

Part III. Green Accounting Application: Extended Genuine Savings measure

This report earlier outlined the methodological issues that are involved in incorporating the environment in aggregate measures of economic welfare and sustainability. A focus has been to describe how traditional measures of economic welfare such as Gross Domestic Product (GDP) or Gross National Product (GNP) can be supplemented to present a more comprehensive picture of welfare and sustainability. More specifically, welfare is impacted by: i) environmental degradation from pollution and other waste from economic activities, and; ii) costs of environmental protection reflected sometimes in defensive expenditures. Accounting for these can give a measure of Green GDP, though for the purpose of transparency, they are most often presented as satellite accounts.

Sustainability is defined by changes in capital stocks. Thus, environmental sustainability implies measurement of changes in natural resource capital stocks, as well as changes in produced capital measured in Net National Product (NNP).

Environmental sustainability can be defined as “weak” (natural capital is substitutable) or “strong” (natural capital is non-substitutable). Thus, in the simplest terms, the sustainability rule differs between that such as measured by the Genuine Savings method where an economy is judged sustainable if the overall capital stock is constant or increasing and unsustainable if the stock is decreasing, and the absolute physical constraints imposed by strong sustainability where if the physical stock decreases then the economy is judged unsustainable. Thus, when the economic costs of meeting such standards are subtracted from GNP a measure of what is sometimes called Sustainable National Income results.

In practice, whilst Genuine Savings measures are increasingly being utilised to guide policy, progress towards such measures at the national accounting scale is in its infancy. However, the underpinnings of these measures lie in the supply and use tables described above such as the Dutch NAMEA that map out in matrix form the linkages between economic activity and the environment in physical and monetary terms. Their value is primarily in signposting economic activities that are most strongly linked with environmental consequences e.g. pollution or natural resource depletion. Subsequent national accounting exercises may then wish to focus in more depth on those linkages that are most deleterious.

The Green Accounting applications undertaken in the newly associated countries and the Mediterranean Partner countries as part of the NEEDS project are therefore focussed on those issues that are of relative priority, and where data is likely to be relatively easily available. As highlighted above, the Genuine Savings measure fits the key criteria in selecting a method to apply; we therefore implement this method here. We also use the opportunity to explore the importance of uncertainty in policy design resulting from such accounting applications.

Our second application being performed within the NEEDS project, through which we document the use of the external costs quantification presents a derivation of sustainable development indicator; specifically, we suggest and compose the indicator on aggregated

51

external costs attributable to power sector for the entire EU27 as well as its each Member State. This application is reported in the stream 1d.

The Genuine Savings measure – also known as Adjusted Net Saving – is an established indicator used in the assessment of sustainability. It originates from the idea that if wealth – defined as the value of all assets in the economy – is the basis of future welfare, then changes in this stock of wealth will have implications for future welfare. Thus, a fall in the value of the capital stock would imply a fall in future economic welfare. Atkinson and Hamilton (2007) therefore define Genuine Savings as:

(3.1)

where: Ki are physical and non-material asset stocks and pi are their shadow prices.

The principal accounting innovation made by the Genuine Savings measure is to recognise that assets not only include man-made capital, but also labour and natural capital and pollution stocks, the latter having a negative value. The Genuine Savings measure has its value in showing whether the traditional measure of saving used in standard national accounting (SNA) procedures is presenting a true picture of future economic welfare possibilities, or whether, by including environmental assets in the accounting framework, the prospects for future welfare appears to change. If, for example, a country’s sub-soil assets such as oil are included in the accounts when the Genuine Savings measure is used and it is realised that these are being depleted and traded – and its proceeds consumed - without being replaced by another form of capital asset, e.g. a factory, then the conventional measure savings gives an overly optimistic view of the prospects for future welfare. This is therefore potentially important information for those with responsibility for the country’s economic management.

One interpretation of the Genuine Savings measure is that if it is positive then the economy satisfies a simple criterion of sustainability whilst if it is negative the implication is that the current level of economic welfare may not be able to be maintained in future time periods. The real merit of the Genuine Savings measure is that the basic concept that underlies it is relatively simple to understand. Its potential limitation lies in the fact that not all natural capital can be seen as equally substitutable with regard to alternative forms of capital. For example, the increased stocks of Greenhouse Gases in the atmosphere that result in climate change, having a negative value, may not be seen as being substitutable with man-made assets that have a positive value.

The remainder of this paper provides an application of the Genuine Savings measure in the UK, Bulgaria, the Czech Republic, Tunisia, Morocco and Egypt and so provides coverage from Northern Europe, Central and Eastern Europe, and North Africa. The primary data source is the World Bank which compiles national measures of Genuine Savings. The added value of our research, however, is to demonstrate that there is a considerable degree of uncertainty in the measurement and valuation of the environmental assets considered. We then explore whether the uncertainties are sufficient for the interpretation of the Genuine Savings measure to be clear, and how the results should then be used in policy analysis and how subsequent research methods should be focussed to reduce the impact of uncertainty.

52

Method

The formula for calculating genuine saving from real data is:

G = GNP – C - K – n(R-g) - (e-d) + m (3.2)

whereG = Genuine SavingGNP = Gross National ProductC = ConsumptionK = depreciation of produced capitaln = the value of resource depletionR = resource extractiong = renewable resource growthe = pollution emissions d = pollution assimilation = marginal social cost of net pollution accumulationm = expenditure on education

n(R-g) is the value of the net growth of renewables, which is deducted when negative, but is not added when positive, on the grounds that much biomass growth is not economically valuable.

Genuine Savings: a state-of-the-art application

The World Bank provides us with our base data for these savings categories. Figure 1 shows how, for one country – UK - the individual categories contribute in adjusting the aggregate savings measure from Net National Savings to Genuine Savings since 1990.

Figure 1. Historical Genuine Savings - UK

UK Genuine Savingc components (1990-2006)

0.002.004.006.008.00

10.0012.0014.00

1990

1992

1994

1996

1998

2000

2002

2004

2006

Year

GS

as %

of G

ross

Na

tiona

l Inc

ome

NNS

EDE

NFD

END

MID

CO2

PM10

NNS= Net National Savings; EDE = Education Expenditure (NNS + EDE); NFD = Net Forest Depletion (EDE – NFD); END = Energy Depletion (NFD-END); MID = Mineral Depletion (END-MID); CO2 = Carbon Dioxide (MID-CO2); PM10=Air Pollution from large particulates (CO2-PM10)

53

Under resources, the World Bank calculations include energy, minerals and forestry. Pollution includes air pollution and CO2. Each line in the figure shows the incremental adjustment from NNS through to the line labelled PM10, which is the final Genuine Savings line. The figure shows that though the shape of the Genuine Savings curve is primarily dictated by the Net National Savings, (Gross National Savings minus capital depreciation), the subsequent of inclusion of investment into education into the savings function has a significant positive effect on the GS curve, the increase subsequently reduced to some degree by the impacts of natural resource depletion and air pollution and climate change caused by carbon dioxide. The final GS values do, however, remain positive across the whole time period.

Figure 2 shows, for the six countries being considered, how the NNS differs from the GS since 1990. Whilst the GS tends to broadly shadow the NNS it is noticeable that in Egypt exploitation of natural capital has not been compensated by investment in other capital forms, resulting in a negative GS.

Figure 2. GS and NNS comparison for six countries – 1990 - 2006

C hart T itle

-15-10

-50

510

152025

30

19901991

19921993

19941995

19961997

19981999

20002001

20022003

20042005

2006

Axi

s Ti

tle

B ulgaria NNS B ulgaria G S C z ech R ep. NNSC z ech R ep. G S E gypt NNS E gypt G SMorocc o NNS Morocc o G S Tunis ia NNSTunis ia G S UK NNS UK G S

Table 1 presents the data for the individual savings categories published by the World Bank. The table shows that of the six countries, Egypt is the only one with a negative GS, primarily reflecting the fact that it has depleted its energy resources (oil) without compensating investment in another form of capital. Conversely, GS proves to be higher than NNS in the UK, Czech Republic and Morocco, where investment in education has more than balanced any depletion of natural capital stocks.

54

Table 6. Net National Savings and Genuine Savings for 2006 (% of Gross National Income)

Country Gross National

Saving

Consumption of Fixed Capital

Net National Saving

Education Expenditure

Energy Depletion

Mineral Depletion

Net Forest Depletion

CO2 damage

PM10 damage

Genuine Savings

Bulgaria 15.55 11.92 3.62 4.24 0.94 2.03 0.00 1.21 1.55 2.13Czech Republic 25.42 13.71 11.70 4.21 0.32 0.00 0.05 0.73 0.14 14.67Egypt 22.08 9.81 12.27 4.41 24.42 0.16 0.21 1.08 0.98 -10.17Morocco 34.97 10.49 24.48 6.47 0.19 0.76 0.00 0.50 0.09 29.40Tunisia 26.89 11.42 15.47 6.67 7.35 0.39 0.09 0.60 0.27 13.43United Kingdom 14.17 10.21 3.96 5.33 2.18 0.00 0.00 0.18 0.04 6.89

The categories that we subject to additional uncertainty analysis are energy and CO2. This analysis is described in the following sub-sections.

Genuine Savings: NEEDS extension

Calculation of genuine savings – as shown in eq. 3.2 – requires to calculate damage associated with pollution that is basically given as a product of emission level exceeding assimilation capacity of the environment and marginal social cost of net pollution accumulation. So far, genuine savings measure included damage associated with emission of particulate matters.

Our contribution from the NEEDS is twofold: first, we are using proper monetary values of damage as calculated by EcoSense tool being updated within NEEDS research streams 1b and RS1d and, then, generalised by using a parametrised dispersion model in WP1 of stream 3a by Philipp Preiss; secondly, we are considering damage for more pollutants, specifically we are able to derive damage attributable to one ton of NOx, SO2, NMVOC, NH3, two fractions of particulate matters (2.5ppm and between 2.5ppm and 10ppm), and several micro pollutants such as emission of cadmium, arsenic, nickel, lead, mercury, chromium, formaldehyde and dioxins (see report of WP1 of RS3a for the details).

Results of our approach are documented in Table 7. NEEDS results for damage attributable to power sector only are displayed in the lines marked “power”, while damage attributable to total emissions released by entire economy are shown in lines “econ_all”. GS components as derived by World Bank are reported in the line WB and are expressed as the share of GNI. In each case, in order to calculate damage due to climate change, CO2 emissions are multiplied by marginal damage of carbon that is in WB report assumed to be 20 USD1995 per ton of carbon22. While WB calculation of damage due to pollution covers PM10 only, we calculate damage for both PM fractions as well as for the other classical and micro pollutants. Adding other pollutants into GS makes GS measures smaller, in the case of for Bulgaria, Slovakia or Portugal GS becomes negative compared with their positive values if calculated by original WB approach.

22 Having data on GDP in Euro2005 and CO2, we are therefore recalculating marginal damage in Euro2005

per ton of CO2 (deflated by OECD CPI) that is 5.62 Euro2005 per ton of CO2.

55

Table 7 NEEDS extensions of Genuine Savings measure.

Country Name

Gross NS

Net NS

Damages due to pollution and climate change

ANSclimate change

particulate

matters

SO2+NOx

NH3+VOC

micro poll

Czech Republic WB

25.42

11.70 0.73 0.14  n.a. n.a.  14.67

  power     0.05 1.92 0.02 13.10  econ_all     0.61 6.8 0.06 7.07United Kingdom WB

14.17 3.96 0.18 0.04  n.a. n.a.  6.89

  power     0.00 0.225 0.001 6.80  econ_all     0.17 1.07 0.01 5.60Bulgaria WB

15.55 3.62 1.21 1.55  n.a. n.a.  2.13

  power   1.35 20.65 0.09 -18.24  econ_all       0.00 34.35 0.34 -31.21Slovak Republic WB

21.17

-0.76 0.61 0.01  n.a. n.a.  2.24

  power   0.02 1.03 0.01 1.62  econ_all       2.30 6.54 0.17 -6.99Portugal WB

12.66

-4.74 0.23 0.43  n.a. n.a.  0.10

  power   0.00 0.33 0.01 0.29  econ_all       1.43 1.15 0.05 -2.28

Damage due to climate change

The treatment of carbon dioxide as a stock pollutant that results in climate change is a relatively recent addition to the components of Genuine Savings measures, reflecting the much higher recent degree of certainty that has been developed with regard to the existence of human induced climate change. However, whilst this scientific uncertainty has been reduced, the uncertainty surrounding the physical and monetary quantification of the linkage between CO2, (and other), gas emissions and consequent impacts of climate change remains substantial. This uncertainty is reflected in the findings of NEEDS Deliverable no. 5.4 in work stream 1b, which tests alternative assumptions, primarily related to the monetary valuation of impacts.

In this exercise we illustrate how the uncertainty in the marginal damage cost of a ton of carbon emission – also known as the shadow price of carbon (SPC) – affects the estimated GS in our six countries. We utilise a range of SPC values derived from the NEEDS Deliverable 5.4. The values used follow the lead given by the selection of central unit values (Preiss, pers. comm.) in NEEDS, which uses the SPC of $102.4. This value assumes: 2005 emissions, discounted to 2005; a 1% pure time preference rate; a 1% trimmed mean; a world-average equity weighting; a conversion from carbon to CO2 in the ratio of 44:12, and; an exchange rate of Euro 1: $1.35. The range of values is defined on the

56

low side by adopting a no equity-weighting assumption with other assumptions kept constant, and defined on the high side by an EU equity weighting assumption, other assumptions kept constant. There is no basis for defining the range of values to be used in this way other than that it spans some, (but by no means all), of the uncertainty presented in the Deliverable results. The resulting range is 7, 21 and 98 Euro per ton of CO2 emitted.

When these values are applied to the estimation of GS, the effects are significant as shown in Table 8. As would be expected, higher values result in lower GS. The most powerful result is that for Bulgaria which shows that a GS which is positive for CO2 unit values of 7 and 21, becomes negative with a unit value of 98 used.

Table 8 Genuine savings – sensitivity analysis for the value of damage due to climate change

Country Name  

Gross Nation

al Saving

Net Nation

al Saving

Adjusted Net (genuine) Savings5.6€/ t

CO2

7€ / t CO2

21€ / t CO2

98€ / t CO2

Czech Republic 

WB 25.42 11.70 14.67 14.56 12.86 3.55NEEDS-all   7.35 14.56 12.86 3.55

United Kingdom

WB 14.17 3.96 6.89 6.85 6.42 4.05NEEDS-all     5.68 6.85 6.42 4.05

Bulgaria

WB 15.55 3.62 2.13 2.16 -0.21 -13.21NEEDS-all     -31.01 2.16 -0.21 -13.21

Slovak Republic

WB 21.17 -0.76 2.24 2.14 0.73 -7.05NEEDS-all     -6.76 2.14 0.73 -7.05

Portugal

WB 12.66 -4.74 0.10 -0.01 -0.68 -4.38NEEDS-all     -2.10 -0.01 -0.68 -4.38

Note: damage due to climate change is expressed in €2005 per 1 ton of CO2. The value of damage 5.6 €2005

per t CO2 is an equivalent of 20 USD1995 per ton of carbon.

57

Energy

The uncertainty explored in the context of accounting for energy resource depletion is methodological, though the uncertainty that arises from the use of alternative methodological approaches stems originally from the problem of how to deal with data uncertainties. The fundamental issue to be addressed is the estimation of the change in value of a capital stock such as oil where its value is determined by the sum of revenues minus costs discounted over the period until the resource is exhausted. The challenge is therefore how to best incorporate forecasts of both quantities and prices over the projected remaining lifetime of the asset stock. There are five principal approaches – described in further detail in e.g. Atkinson and Hamilton (2007) – that can be utilised. These approaches are summarised in Table 9.

Table 9. Alternative measures of resource depletion costs

_________________________________________________________________________Method Formula Assumptions

Total rent pqt - ċqt Constant unit extraction cost ċ, Hotelling rule

Marginal rent pqt – ć(qt)qt Constant price, increasing marginal extraction cost; Hotelling rule

Exhaustion pqt/(1+r)N Constant price, increasing marginal extraction cost; Hotelling rule

Simple PV pq - c/(1+r)N Constant total rent

Quasi optimal Ɛ(pq – c)/1+(Ɛ-1)(1+r)N Constant price, iso-elastic cost function with increasing marginal costs

_________________________________________________________________________Where p = price; q = quantity; ċ = constant extraction cost; ć = increasing marginal extraction cost; r = discount rate; N = no. of years until exhaustion; elasticity of extaction cost increase. Based on Table 2 in Atkinson and Hamilton (2007)

The interest for us in this exercise is not in the differences and relative merits of the alternative measures, in themselves. Rather, it is the fact that application of these alternative methods may have an effect on the interpretation of the Genuine Savings measure. In order to test this hypothesis, we compare the GS measure using the different resource depletion measures. Atkinson and Hamilton compare the measure for oil in 2000; we transfer their results for the two countries that are common to our, and their, analysis – Egypt and UK – and assume that the differences found in the measures, as a percentage of GNI, applied to oil are the same as for energy resource depletion more generally. Since oil is the dominant energy source in both countries this seems to be reasonable assumption.

Table 10 shows the results for the value of oil depletion in Egypt and UK in 2000. It shows that the alternative measures give quite different values relative to each other. In absolute terms, the difference in value is large - between 0.4% of GNI and 1.3% of GNI is equivalent to $13bn in 2006. However, the pattern of differences is not the same for the two countries, primarily reflecting differences in the lifetime assumed for the asset. When

58

these proportional differences are applied to energy resources as a whole and those differences are applied to the GS measure, we derive the range of GS estimates presented in Table 10 below. Whilst the range of GS for the UK is relatively narrow – 5.9 to 7.6 – and consistently suggesting that the country has sustainable savings patterns, the same cannot be said for Egypt. In Egypt, the range of GS values varies from -10.2 to + 2.6. The uncertainty in methodological approaches for the valuation of energy depletion therefore results in the indicator signalling either unsustainable or sustainable savings patterns. It is worth noting, then, that whilst the World Bank uses the Total Rent method – resulting in the most negative results – the evaluation by Atkinson and Hamilton tentatively suggests that the Quasi-optimal approach – which gives the most positive results to Egypt – is the most robust.

Table 10. Depletion estimates for oil in year 2000

  % of GNI

 Total Rent

Marginal Rent Exhaustion

Simple PV

Quasi optimal

Egypt 4 3.6 4 2.3 3.7UK 0.6 0.4 1.3 0.4 0.6

Table 11. Energy Depreciation values and resulting GS – 2006.

Total Rent

Marginal Rent Exhaustion

Simple PV

Quasi optimal

EgyptEnergy Deprecn. 24.42 21.98 21.98 12.64 11.69GS -10.17 -7.72 -7.72 1.62 2.56

UKEnergy Deprecn. 2.18 1.45 3.15 2.10 2.10GS 6.89 7.62 5.92 6.97 6.97

Genuine Savings for Hungary 2000-2006

Another example illustrates Genuine Savings measure derived with or without damages due to airborne pollution. Although the trend in GS is given by net national savings, thanks to damage due to airborne pollution, genuine savings almost approaches zero in 2003. In both cases, we can conclude for the period 2000-2006 that economic development was not sustainable in Hungary during 2000-2003.

59

Figure 3 Genuine savings for Hungary 2000-2006 with/without damage due to airborne pollution

…with damage due to pollution …without damage due to pollution

0

5

10

15

20

25

30

35

2000 2001 2002 2003 2004 2005 2006

EducationCFC

energy depletion

mineral depletionPM damageclassical pollutants

CO2(21€/tCO2)ANS 0

5

10

15

20

25

30

35

2000 2001 2002 2003 2004 2005 2006

Education

CFC

energy depletion

mineral depletion

CO2(21€/tCO2)

ANS

Conclusion

This report presents the results of simple applications of uncertainty in the valuation of energy resource depletion and CO2 emissions to the estimation of one widely used indicator of sustainability, Genuine Savings. The results serve to show that whilst uncertainty has always been recognised as being important in the estimation of external costs in marginal pricing policy, the uncertainty can also be important in the application of sustainability measures in green accounting exercises. This latter fact has not been highlighted in the green accounting literature to date. It should also be highlighted that future research exercises could usefully combine these uncertainties with those that exist in the estimation of air pollution externalities from e.g. human health exposure-response functions and the valuation of mortality impacts.

60

References

Ahmad, Y.J., El Serafy, S. and Lutz, E. (eds) (1989) "Environmental Accounting for Sustainable Development" Washington: The World Bank.

Arrow, K, Bolin, B., Costanza, R. Dasgupta, P., Folke, C., Holling, C.S., Jansson, B-O., Levin, S., Maler, K-G, Perrings, C., Pimentel, D. (1995) “Economic growth, carrying capacity, and the environment.” Ecological Economics 15(2) pp 91-95.

Asheim, G.B. (2000) "Green National Accounting, why and how?" Environment and Development Economics, 5(1-2) pp 25-48

Asheim, G.B. and Weitzman, M.L. (2001) "Does NNP Growth Indicate Welfare Improvement?" Working Paper

Barbier, E.B. (1987) "The Concept of Sustainable Economic Development." Environmental Conservation 14(2) pp 101-110.

Barnett, H.J. & Morse, C. (1963) “Scarcity and Growth: The Economics of Natural Resource Availability” Resources for the Future, John Hopkins Press, Baltimore.

Baumol, W. J. and Oates, W.E. (1971) "The use of standards and prices for protection of the environment", Swedish Journal of Economics, Vol. 73, pp. 42-54.

Bayer, K.: Konzept einer Umweltdefensivkostenrechnung für Österreich, in Ökologische Gesamtrechnung 6/1993, Bundesministerium für Umwelt, Jugend und Familie, Wien 1993.

Beckerman, W. (1994) “Sustainable Development- is it a useful concept?” Environmental Values 3 pp 191-209.

Brekke, K.A. (1997) "Economic Growth and the Environment: on the Measurement of Income and Welfare" Edward Elgar, Cheltenham, UK.

Brower, R., O'Connor, M. and Radermacher, W. (1999) "GREEned National STAtistical and Modelling Procedure: the GREENSTAMP approach to the calculation of environmentally adjusted national income figures", International Journal of Sustainable Development, Vol. 2, No. 1, pp. 7-31.

Cobb, C. and Cobb, J. (eds) "The Green National Product" University Press of America, Lanham.

Costanza, R. (1991) “Assuring Sustainability of Ecological Systems.” in Costanza, R (ed.) “Ecological Economics: the Science and Management of Sustainability.” Columbia University Press, New York.

Daly (1989) "Toward a Measure of Sustainable Social Net National Product", in Ahmad et al. (1989).

61

Daly, Herman. (1990) “Towards Some Operational Principles of Sustainable Development” Ecological Economics 2(1) pp 1-6

Daly, H.E. & Cobb, J.B. (1989) “For the Common Good : “ Redirecting the Economy Towards Community, the Environment and a Sustainable Future. “Green Print, London.

de Wit, G. (1996), Policy Relevance of Various Proposals for Greening GDP. Centrum voor energiebesparing en schone technologie, Delft.

Diefenbacher, H. (1994) “The index of sustainable economic welfare: a case study of the Federal Republic of Germany” in Cobb, C.S., Cobb, J.B. (Eds.) The Green National Product: A Proposed Index of Sustainable Economic Welfare, University Press of America, Lanham, pp 215-245.

Dorfman, R. et Dorfman, N., S. (eds.) (1977), ”Economics of the Environment: Selected Readings.” Second edition, W. W. Norton & Comp., Inc., New York.

Easterlin, R.A. (1974) “Does Economic Growth Improve the Human Lot?” in “Nations and Households in Economic Growth” (P.A. David & R.M. Wener, Eds.) Academic Press, New York.

Ehrlich, P.R. and Ehrlich, A.H. (1981) "Extinction: the causes and consequences of the diappearance of species." Random House, New York.

Eisner, R. (1989) "The Total Incomes System of Accounts" University of Chicago Press, Chicago.

Eisner, R. (1994) "The index of sustainable welfare: Comment" in Cobb, C. and Cobb, J. (eds) "The Green National Product" University Press of America, Lanham.

El Serafy, S. (1989) "The Proper Calculation of Income from Depletable Natural Resources" in Amad et al. op.cit.

El Serafy, S. (1996) "Environmental Accounting for Sustainable Development: Conceptual and theoretical Issues Relating to the System of Integrated Environmental and Economic Accounting (SEEA)" Journal of Economic Cooperation among Islamic Countries, 17, 1-2, 87-107.

England, R.W. (1998) "Measurement of social well-being: alternatives to gross domestic product" Ecological Economics 25 pp 89-103

EUROSTAT (1995), European System of Accounts – ESA 1995. Europeran Commission, Luxembourgh.

EUROSTAT (2002), Natural Resource Accounts for Oil and Gas: Eurostat proposals for further harmonisation of monetary estimates. B1 – National accounts methodology, statistics for own resources. 11 July 2002.

62

EUROSTAT (2002a), Natural Resource Accounts for Oil and Gas 1980-2000“. Theme 2. European Communities. Luxembourg, 2002.

EUROSTAT (2002b), Accounts for Subsoil Assets Results of Pilot Studies in European Countries“. Theme 2. European Communities. Luxembourg, 2002.

Fankhauser, S. (1994) "Evaluating the Social Costs of Greenhouse Gas Emissions" The Energy Journal, 15(2) 157-184

Fautin, D.G. (1995) "Sustainability, more than a Buzz-Word?" Annual Review of Ecological Systems, Vol 26, pp v-x.

Freeman III, A.M. (2003): The Measurement of Environmentala nd Ressource Values. Theory and Methods. 2nd edition (1st edition 1993). Ressource for the Future, Washington, DC.

Goodland, R. (1995) "The Concept of Environmental Sustainability." Annual Review of Ecological Systematics 26 pp 1-24.

Goodland, R. & Ledec, C. (1987) “Neoclassical Economics and Principles of Sustainable Development.” Ecological Modelling, 38 pp 29-46. GREENSTAMP Project (1997) "Methodological Problems in the Calculation of Environmentally Adjusted National Income Figures" Final Report to the European commission DGXII.

Guenno, G., and Tiezzi, L. (1998) “An index of sustainable economic welfare for Italy” Working paper 5/98 Fondazione Eni Enrico Mattei, Milano.

Hamilton, K. (1994) "Green Adjustments to GDP" Resources Policy 20(3) 1550168.

Hamilton, K. (1995) “Sustainable Development, the Hartwick Rule and Optimal Growth.” Environmental and Resource Economics 5 pp 393-411.

Hamilton, K. and Clemens, M. (1999) "Genuine Savings Rates in Developing Countries" The World Bank Economic Review, 13(2) pp 333-356.

Harrison, A. (1989) "Introducing Natural Capital into the SNA". In Ahmad et al. (1989).

Harrison, A. (1999). Making Services Visible, Paper No. STD/NA(99)26 presented at the OECD national accounts meeting, Paris.

Hartwick, J.M. (1977) “Intergenerational Equity and the Investing of Rents from Exhaustible Resources.” American Economic Review 67(5), pp 972-974.

Hartwick, J.M. (1978) “Investing Returns from Depleting Renewable Resource Stocks and Intergenerational Equity.” Economic Letters 1. pp 85-88.

63

Hartwick, J.M. (1990) "Natural Resources, National Accounting and Economic Depreciation". Journal of Political Economy 43 pp 291-304.

Hartwick, J. & Hageman, A.P. (1993) “Economic Depreciation of Mineral Stocks and the Contribution of El Serafy.” In Ernst Lutz (ed). “Toward Improved Accounting for the Environment.” UNSTAT-World Bank Symposium, Washington DC.

Heal, G. and Kristrom, B. (2001) "National Income and the Environment", in Maler & Vincent (eds.), The Handbook of Environmental Economics (forthcoming), Elsevier, North Holland. .Hicks, J.R. (1946) “Value and Capital” 2nd ed. Oxford University Press, Oxford.

Hirsch, F. (1976) “The Social Limits to Growth” Harvard University Press, Cambridge.

Hueting, R. (1980) “New Scarcity and Economic Growth: More Welfare through Less Production?” North Holland, Amsterdam.

Hueting, R. (1989) “Correcting National Income for Environmental Losses: Toward a Practical Solution.” In Ahmad et al (1989)

Hueting R. 1995. Estimating sustainable national income in Taking nature into account; towards a sustainable national income. In W.v. Dieren (Editor), Ch. 13, Springer-Verlag, New York.

Jackson, T., Laing, F., MacGillivray, A. et al (1997) “An Index of Sustainable Economic Welfare for the UK 1950-1996” University of Surrey Centre for Environmental Strategy, Guilford.

Juster, F.T. (1973) "A framework for the measurement of economic and social performance" in Moss, M. (ed) "The Measurement of Economic and Social Performance" Columbia University Press, New York.

Kneese, A.V., Ayres, R.V. and D'Arge, R.C. (1970) "Economics and the Environment: A Materials Balance Approach." The John Hopkins University Press, Baltimore.

Leipert, C. (1989) "Social costs of the economic process and antional accounts: The example of defensive expenditures" Journal of Interdisciplinary Economics 3, 27-46.

Leipert, C. (1989), Die heimlichen Kosten des Fortschritts. Frankfurt am mein.

Leipert, C. (1995), Defensive Expenditures, in: Van Dieren (eds.), Taking Nature into Account, Springer-Verlag, New York.

Mäler, Karl-Goran, (1991) "National Accounts and Environmental Resources” Environmental and Resource Economics 1 (1) pp 1-15.

64

Martinez-Alier (1995) "The environment as a luxury good or "too poor to be green"?", Ecological Economics 13 (1) pp 1-10.

Max-Neef, M. (1995) “Economic growth and quality of life: a threshold hypothesis” Ecological Economics 15 pp 115-118.

Mayo, E., MacGillivray, A. and McLaren, D. (1997) “More isn’t always better: a special briefing on growth and quality of life in the UK.” New Economics Foundation and Friends of the Earth. London. http://www.foe.co.uk/resource/briefings/ more_isnt_better.pdf

Mishan, E., J. (1993), ”The Costs of Economic Growth.” Revised edition (Firsts published in 1967). Weidenfeld and Nicolson, London.

Mishan, E., J. (1993), Spor o ekonomický růst. Praha, SLON.

Neumayer, E. (1998) "The ISEW: Not an index of sustainable economic welfare" Social Indic. Res. 48 pp 77-101

Nordhaus, W.D. and Tobin, J. (1973) "Is Growth Obsolete?" Income and Wealth vol. 38

Norgaard, R.D. (1990) “Economic Indicators of Resource Scarcity: A Critical Essay.” Journal of Environmental and Resource Management 19(1) pp 19-25.

Pearce, D.W. and Atkinson, G. (1993) "Capital Theory and the Measurement of Sustainable Development: An Indicator of Weak Sustainability" Ecological Economics 8 103-8.

Perrings, C. (1991) “Reserved Rationality and the Precautionary Principle; Technological Change, Time and Uncertainty in Environmental Decision Making.” in R. Costanza ed. “Ecological Economics: the Science and Management of Sustainability.” Columbia University Press, New York.

Peskin, (1989) "Environmental and Non market Accounting in Developing Countries". In Ahmad et al. (1989).

Rennings, K & Wiggering, H. (1997) “Steps towards indicators of Sustainable Development: linking economic and ecological concepts.” Ecological Economics 20(1) pp 25-36.

Repetto, Robert, William Macgrath, Michael Wells, Christine Beer, and Fabrizio Rossini. (1989) “Wasting Assets: Natural Reources in the National Income Accounts.” Washington, D.C.: World Resources Institute.

Sen, A. (1993) "The Economics of Life and Death" Scientific American 265(5) pp 40-47

Stockhammer, El, Hochreiter, H., Obermeyer, B., Steiner, K. (1997) “The index of sustainable economic welfare (ISEW) as an alternative to GDP in measuring

65

economic welfare: the results of the Austrian (revised) ISEW calculation 1955-1992” Ecological Economics 21 pp 19-34

Stone, R.T. (1951) "Functions and Criteria in a System of Social Accounts" Income and Wealth, Series I. p. 1-74, Cambridge, Bowes and Bowes.

Ščasný, M., Kopecký, O, Cudlínová, E. (2002), HDP Alternatives – evaluation of precondition and use of the Index of sustainable economic welfare (ISEW) in the Czech Republic. In: Towards to Sustainability in the Czech Republic. Part 4 – Education, Information, Indicators. Charles University Environment Center. (in Czech)

Ščasný, M. (2004), Application of environmental accounting on subsoil asset. The Case study for the Czech Republic. Presented at the 56th Annual Congress on European Association of Environmental Accounting (EEA-2004 ), Prague, 1-3 April 2004. Abstract in: Mejzlík, L. a kol. Programme and Collected Abstracts. European Accounting Association and Vysoká škola ekonomická v Praze, Oeconomica. ISBN: 80-245-0686-6, pp. 336-337.

Tol, R. and T. Heinzow, (2007) Estimates of the external and sustainability costs of climate change. In Markandya, A. Hunt A. and N Dale: .

UNSEEA (2003). Integrated Environmental and Economic Accounting 2003. United Nations.

Vanoli (1995) "Reflections on Environmental Accounting Issues", Journal of the International Association for Research in Income and Wealth, Series 41, 1995.

Vincent, J., Panayotou, T. and Hartwick, J.M. (1997) "Resource depletion and Sustainability in Small Open Economies" Journal of Environmental Economics and Management 33, 274-286.

WCED (World Commission on Environment and Development).(1987) “Our Common Future.” Oxford University Press, Oxford.

Weitzman, M.L. (1976) "On the Welfare Significance of National Product in a Dynamic Economy" Quarterly Journal of Economics (90)

Weitzman, M.L. (1998) "Sustainability and Green Accounting" AERE address in Annapolis, MD, June 2.

Withagen C. 1998. Untested hypotheses in non-renewable resource economics. Environmental and Resource Economics 11: 623-634.

66

Appendix.

Table A1. Matrix of physical flows between economy and environment

67

Table A2. Matrix of physical flows between economy and environment (continued)

68

Table A3. Simple hybrid supply and use table (Source: UN SEEA, 2003)

69

Valuation techniques for measuring degradation

Production function approachesProduction function approaches estimate the value of a given non-market good/service from the measurement of changes in marketed output as a consequence of changes in the provision of (usually environmental) inputs in the production of the output. In the dose-response, or exposure-response, method the physical output change as a result of a change in environmental quality is multiplied by the market price of the affected good to estimate an economic (use) value of the good. An example is the impact of low-level ozone caused by road transport on crop yields. Assuming that the loss of yield can be quantified using an exposure-response function, the quantity of crop lost can be multiplied by the market price of the crop. The great advantage of this method is that it relies on the use of market prices to derive values rather than having to infer values through indirect means.

The replacement or restoration cost method assumes that the economic cost of a non-market good can be estimated by the market price of a substitute market good that can replace or restore the original quantity or quality level of the non-market good. For example, a habitat may be disturbed in the construction of rail infrastructure, but its original condition may be restored by expenditure on the import of certain plant species. This expenditure may therefore be seen as a proxy for the value of this aspect of the habitat. If the expenditure is made it is, at best, a lower bound on true willingness to pay (WTP). If the expenditure is not made, it may be seen as an upper bound on true WTP.

Indirect revealed preference techniques These techniques use models of relationships between marketed goods/services and the non-market good/service of interest, assuming that there is some kind of substitute or complementary relationship between the two goods. The advantage of this group of techniques is that they make use of information about people’s actual behaviour and related personal preferences. Their principal disadvantage is that the statistical models used to isolate the value of interest from other influences are sensitive to the specification and functional form assumed. A number of econometric issues are generally involved in the estimation of the value of the desired attribute and it is a resource intensive exercise to make these estimates.

The travel cost method estimates recreational use values through the analysis of travel expenditures incurred by consumers to enjoy recreational activities.

Avertive/abatement costs, or defensive/preventive expenditures, assumes that individuals spend money on certain activities that reduce their risks (e.g. impact of pollution, risks of accidents) and that these activities are pursued to the point where their marginal cost, (i.e. the expenditure on the last unit purchased), equals their marginal value of reduced impact. Averting goods related to pollution include e.g. air filters, water purifiers, and noise insulation, while averting goods that reduce risks of death may include seat belts and fire detectors.

Hedonic price analysis refers to the estimation of non-market values by deriving prices for individual attributes of a market commodity that are implicit when environmental goods/services can be viewed as attributes of a market commodity, such as properties or wages. Thus, the hedonic price model provides the basis for deriving welfare change measures from observed differences in properties’ prices or wages offered in the job

70

market. For example, differences in ambient noise levels in two areas, and the values individuals place on these, may be reflected in relative house prices in the two areas.

Stated preference techniquesStated preference is a generic name for a variety of techniques including the contingent valuation and choice experiments including contingent ranking, contingent choice and conjoint analysis. Using these techniques, researchers pose contingent or hypothetical questions to respondents, inducing responses that trade off improvements in public goods and services for money. From the responses, preferences for the hypothetical good or the value of changes in provision of the hypothetical good can be inferred. Values are derived from preferences made in relation to (hypothetical) prices, or via trade-offs with other attributes. The major advantage of the technique is that the questions put to respondents can define exactly what needs to be valued. The main limitation is that the method provides hypothetical answers to hypothetical questions, which means no real payment is undertaken, so that no real commitment is made.

71

Figure A1. Genuine Savings Rates by region (Source: Hamilton & Clemens, 1999)

Figure A2. Genuine savings for Tunisia, as % of GDP (Source: Hamilton & Clemens, 1999)

72

Table 12. Sustainability targets analysed under the GREENSENSE project

Environmental Impact

Weak Sustainability Intermediate sustainability target Strong sustainability target

Air pollution Invest the value of damage to capital stocks due to air pollution.

Current legislation with Emission Ceilings Medium Ambition GAP Closure + Emission Ceilings/ Maximum Technical Feasible Solution

Climate Change Invest the NPV of the cost of current carbon emissions ($4/tonne current estimate)

550 ppmv by 2120 450 ppmv by 2120

Biodiversity Invest the value of damage to capital stocks due to biodiversity loss

Natura 2000 network to be preservedNo further wetland loss or degradation15% of agricultural area under management contractsNo further deterioration of natural and semi-natural forests

20% of all land to be preserved in natural condition

Natural resources Energy: Invest % of resource rents Invest value of future price IncreasesForestry: Invest value of future price increases

12% energy from renewables by 2010 16-19% energy from renewables by 2010 (current estimate)

Toxic Substances Invest the value of damage to capital stocks due to Toxic substances

Concentration levels of lead and cadmium given in EU Directives

Future steady-state concentrations of lead and cadmium

Urban Environmental Problems (Noise)

Not applicable since only current welfare impacts

Not applicable since only current welfare impacts

Not applicable since only current welfare impacts

Waste Invest the value of damage (e.g. land converted for landfill) due to waste

Landfill max. 35% of household waste; Recycle 25%

Land space availability

Water Pollution Invest the value of any decline in water resource stocks.

Satisfaction of the EC Water Framework Directive

Satisfaction of the EC Water Framework Directive

73

73

74

74