The System of National Accounts (SNA) is an integrated system providing information on various aspects of an economy's performance (Inter-Secretariat Working Group on National Accounts 1993). Most familiar are the current accounts, which present information on gross domestic product (GDP) in three ways: as the sum of value added (revenue minus intermediate consumption) across all industries (the production account)1, as the sum of final consumption and savings (= gross investment; the use of income account)2, and as the sum of employee compensation and operating surplus (the distribution of income account). Production — "a physical process, carried out under the responsibility, control and management of an institutional unit, in which labour and assets are used to transform inputs of goods and services into outputs of other goods and services" (Inter-Secretariat Working Group 1993, p. 4) — is the fundamental concept determining which economic activities the SNA includes. By design, the SNA does not include all economic activities, or all goods and services that contribute to human well-being. With few exceptions, it focuses on goods and services that are produced using human labor and other factors of production and are bought and sold in markets.
The consequence of this production orientation is that macroeconomic indicators in the SNA, such as GDP, are not reliable measures of aggregate economic welfare.3 Documents on the SNA recognize this, emphasizing that the SNA aims at providing information on production, not welfare. Those documents also point out that GDP is not a measure of sustainable income. That is, GDP does not capture the impact of current production activities on future income, much less welfare.
The exclusion of considerations related to welfare and sustainability helps the SNA achieve its main goal, which is to generate accurate information on production activities regardless of the country applying the system. Moreover, the emphasis on production is understandable when one considers the dominant economic issues during the mid-century period when the accounts were initially designed: recovering from the Great Depression and reconstructing the world following World War II. Since that time, however, issues of welfare and sustainability have gained prominence. The absence of information on these aspects of economic activity within the SNA, and the lack of an alternative, institutionalized system to provide it, hamper the ability of policymakers to address these issues effectively.
In this situation, policymakers often resort to GDP as a proxy measure for national welfare, despite the warnings by national accountants. This is unlikely to lead to distorted policymaking when few economically significant benefits and costs fall outside the SNA production boundary. Then, GDP offers a reasonable first-order approximation to short-run national welfare, and value added of constituent industries provides a reasonable approximation to those industries' contribution toward the total. Problems are more likely when industries generate important benefits or costs ignored by the SNA. Then, policymakers whose perceptions of the economy are shaped by the SNA can misinterpret the welfare impacts of changes in the economy's performance and the consequences of policies intended to shape that performance.
Forestry is an example of an activity whose contribution to the economy in a welfare sense is unlikely to be measured well by value added in the production account.4 Forests contribute directly to welfare through the provision of amenity values, which may not satisfy the SNA's definition of "production." They also provide other industries with services, such as watershed protection, whose value the SNA records as part of the operating surplus of recipient industries instead of as services furnished by forests. For these reasons, the SNA likely understates the economic contribution of forests.
For industries unrelated to natural resources, the production account provides information on sustainability through the estimation of net value added, the difference between gross value added and the consumption of fixed (human-made, or physical) capital.5 The portion of gross value added generated by consuming capital is not sustainable, any more than a personal lifestyle financed by running down one's bank account is sustainable. Even net value added as reported in the SNA does not reflect sustainability in the case of forestry, however, because it ignores the consumption of natural capital that occurs when forests are harvested or converted to other uses. The depreciation of forest capital due to these processes can cause GDP in forest-rich countries to give an exaggerated impression of the increase in a country's income — that is, to understate the long-run costs of such depreciation. Documents on the SNA in fact warn that the neglect of resource depletion effects is one reason GDP is not a measure of sustainable income (Inter-Secretariat Working Group 1993, p. 41).
This report examines how to formulate better measures of the economic contribution of forests. It takes figures in the current and asset6 accounts of the SNA as the starting point, and it describes adjustments necessary to convert those figures to welfare terms. The various adjustments presented are surely not implementable in all countries, certainly not immediately. Moreover, they are unlikely to be accepted as standard accounting procedures by national accountants, as some of them violate the SNA definition of the production boundary. Nevertheless, laying out a conceptually sound system is the essential first step toward improved accounting of forest-related benefits. Pilot applications of such a framework should be able to contribute to improved planning and policymaking in countries with forest resources, even if the procedures are not immediately institutionalized and are applied by economic planners and policy analysts rather than by national accountants. We would hope, however, that national accountants would be involved in such applications, to ensure that data from the SNA are used appropriately and to promote a dialogue about the eventual possibility of integrating welfare and sustainability considerations related to forest resources into the SNA, either through a system of satellite accounts or as part of the SNA per se.
The report was commissioned by the Forestry Department of FAO, which intends to use it as one source of information for preparing a handbook on economic accounts for forest resources. That handbook will parallel a recent, similar one for agriculture (FAO 1996). The report has two principal objectives. The first is to present basic concepts that should guide the development of economic accounting procedures for forest resources. The report presents those concepts in two chapters. Chapter 2 identifies the principal adjustments that should be made to the current and asset accounts, and Chapter 3 discusses methods for making them. The second objective is to review empirical experience with such adjustments, with the intention of assessing their feasibility. To this end, we reviewed more than 30 studies from more than 20 countries (more than 100 countries if one includes countries included in regional and global studies). Chapter 4 summarizes the results of that review. The final chapter, Chapter 5, summarizes the findings of the report, compares, the proposed framework to the SNA, and highlights implications for the FAO manual.
We have attempted to structure the report to be accessible, and hopefully useful, to both specialists and nonspecialists. Part I, which consists of the five chapters just described, emphasizes key findings and the intuition behind them. It contains minimal mathematical detail. Only Chapter 3 contains much mathematical analysis, involving net present value calculations of the sort found in introductory forest economics textbooks. Part II contains the technical details supporting the analysis in Part I. It consists of 7 appendices.
It should be understood that this report focuses on economic accounting, not natural resource accounting. That is, it does not cover accounting systems whose principal goal is to organize ecological information on forests. Such systems are valuable for resource managers and policymakers, and they can facilitate the construction of economic accounts, but they are not our concern here, beyond a few comments we will make when we review existing studies. Readers interested in forestry accounting in the ecological sense may wish to consult OECD (1994, 1995) and a report prepared for the International Tropical Timber Organization by the IIED Forestry and Land Use Programme and The World Conservation Monitoring Centre (1994).
In the remainder of this chapter, we briefly discuss one of the principal rationales for adjusting national accounts for natural resources, i.e. to measure economic sustainability. Then, we review what the SNA and the related System of Integrated Environmental and Economic Accounts (SEEA; United Nations 1993, Bartelmus and van Tongeren 1994) have to say about adjustments for natural resources, in particular forest resources. Perhaps more important, we identify the principal differences in philosophy and purpose between those systems and the system presented in this report. The main point is that the SEEA and the system presented in this report are not two means to the same end, but rather systems that provide different types of information. Viewed this way, the two systems are complementary, not in competition or conflict.
Measuring economic sustainability
The idea of developing improved measures of welfare and sustainability by adjusting macroeconomic aggregates like GDP for changes in the quantity and quality of environmental resources is not new. In the early 1970s, Nordhaus and Tobin (1972) proposed adjusting GDP for the disamenities of urban life. In the early to mid-1980s, several studies examined issues related to natural resource depletion (Ward 1982, Landefeld and Hines 1985, Stauffer 1986). During this same period, academic researchers established the theoretical links between economic sustainability and depletion-adjusted measures of investment and domestic product (Weitzman 1976, Hartwick 1977, Dasgupta and Heal 1979, Solow 1986).
The 1989 publication of Wasting Assets, a study of Indonesia by the World Resources Institute (WRI; see Repetto et al. 1989), brought these concerns into mainstream discussions about the national accounts. Since that time, international organizations, in particular the World Bank and the United Nations (especially the United Nations Environment Program and the U.N. Statistical Division), have invested considerable effort in studying the issues involved. Volumes edited by Ahmad et al. (1989) and Lutz (1993) provide useful compilations of key studies related to those efforts.
In the following paragraphs, we discuss two theoretically equivalent procedures for analyzing an economy’s sustainability. Both are based on modified components of the SNA.
Testing for sustainability in the asset accounts
There is, of course, no way to predict with certainty the future course of a country's economy. However, the academic work cited above, and subsequent work by Hartwick (1990), Dasgupta and Mäler (1991), Mäler (1991), and others, has established that maintaining the value of a country's total capital stock is a necessary condition for enabling a country's population to be as economically well off in the future as in the current period.7 The total capital stock includes not just human-made capital (equipment, structures, infrastructure, etc.), but also natural capital (forests, subsoil assets, air and water quality, etc.) and human capital (human skills and ingenuity). Raising welfare over time requires that the value of the total capital stock increase, not merely remain constant.
Maintaining the total capital stock is a necessary condition for sustainability, but it may not be sufficient. It assumes some degree of substitutability among different types of capital, an assumption that has been labeled "weak sustainability." If substitution is not possible, then maintaining minimum stocks of certain types of capital might be necessary to avoid catastrophic losses associated with "threshold effects." Many ecologists have argued that threshold effects are likely in the case of environmental resources. Even if this is the case, however, note that it is not a reason to forgo monitoring the total capital stock. Rather, it suggests that policymakers should pay attention to not just a single number giving the aggregate value of the total capital stock, but also the values of its components.8 In cases where there is evidence of threshold effects, values of corresponding capital stocks should be interpreted with reference to physical indicators as well, to gain a deeper understanding of prospects for sustainability. Information expressed in both value and physical terms should be considered, rather than just one or the other.
Concerns about threshold effects aside, the sustainability test is therefore to check whether the value of the total capital stock is at least constant over time. Under certain circumstances, direct estimation of the change in value, or net accumulation, might be possible, instead of first calculating the value of the total capital stock at two points in time and then taking the difference between those estimates. The sustainability test is then to check whether net accumulation, broadly defined to include all additions and all subtractions, is nonnegative when summed across all forms of capital. Only if this sum is at least as great as zero does a country maintain the value of its total capital stock, and only if that occurs can the country's future consumption level (broadly defined) match its current consumption level.
The SNA does not furnish sufficient information to check whether the value of the total capital stock is at least constant over time. In practice, asset accounts typically include just financial assets and produced non-financial assets, although in principle they can include non-produced, non-financial assets as well. In Chapter 3, we will look closely at methods for estimating changes in the capital value of forests. Those methods are straightforward in conceptual terms, as they simply involve changes in the present value of future streams of benefits provided by forests. They are less straightforward in practice, as in principle they should reflect all forest-related benefits, not just easily measured commercial products like timber. When they can be applied, however, then the resulting estimates can be combined with estimates for other forms of capital to form an aggregate measure of the change in the value of the total capital stock.
Although the greatest value from estimating the capital value of forests probably lies in generating an input for estimating the total capital stock, and thus for analyzing overall sustainability prospects, it does have stand-alone value. It enables one to determine whether a country's forests are becoming more or less valuable as a source of current and future benefits, that is whether their economic contribution can be sustained. It reveals whether a country is becoming more forest-rich or more forest-poor in a quite literal sense.
This economic measure of forest sustainability does not necessarily coincide, however, with a physical measure such as the sustained-yield production level. For example, a country with declining timber stocks might experience an appreciation of its forest capital, if the rate of increase in stumpage values outstrips the decline in physical stocks. From an economic perspective, capital value, which is forward looking and weights goods and services according to their welfare contributions, has more meaning for sustainability than does physical stock, which provides information on current status in ecological terms. We emphasize, however, that estimates of forest capital that leave out important benefits (such as nonmarket ones) and are not interpreted with reference to physical indicators of threshold effects can be highly misleading indicators of sustainability.
Testing for sustainability in the current accounts
Maintaining the value of the total capital stock has a counterpart expression in "flow" terms. That is to check whether net domestic product (NDP), broadly defined to include all forms of capital, remains at least constant over time. NDP is ordinarily defined as the sum of consumption and net investment in human-made capital.9 To convert it to a measure of economic sustainability, one must include net accumulation instead of net investment (so that it reflects the change in value of capital, not just the value of the change), and include net accumulation for all forms of capital, not just human-made capital.
While this condition might appear different than the first, the two are in fact theoretically equivalent. As we will see in Chapter 2, however, in addition to net accumulation, there are some consumption-related adjustments that should be made to NDP in the case of forest resources. Although the two sustainability conditions remain theoretically equivalent when these additional adjustments are made (because they affect capital values too), analyzing changes in both capital value and NDP is useful if one wishes to shed as much light as possible on an economy’s sustainability. Analyzing both is advisable for a more practical reason, too: even though the two conditions are theoretically equivalent, in practice they can yield different predictions about sustainability. This can happen, for example, if high levels of investment in human-made capital cause estimates of net accumulation to be positive, but much of the investment turns out to be unproductive. In that case, NDP might start declining, and thus reveal the unsustainability of the economy, even while net accumulation still appears (incorrectly) to be positive.
Industry-specific vs. economy-wide sustainability
While one can certainly calculate capital values and net value added (NDP at the industry level) for just forestry, and should in fact do so to gain a more accurate understanding of forestry’s net economic contribution, one must be careful not to misinterpret the implications for sustainability. Considering timber alone, one should expect the economic contribution of forestry to decline from an initial point, when a country’s forest estate consists of old-growth forests, to a later point, when it consists of sustainably managed second-growth forests. Old-growth forest stands tend to be more valuable than second-growth stands, due to their higher timber volumes and larger trees. It follows that, even in the absence of deforestation, the conversion of old-growth forests to second-growth forests as logging proceeds in the former should register in adjusted asset accounts as a declining value of forest capital, and in adjusted current accounts as declining net value added for forestry. Incorporating nontimber values would probably strengthen the evidence of unsustainability, as old-growth forests tend to be higher in bi odiversity, provide better soil protection, and be more attractive for recreation than second-growth forests.
These indicators correctly signal unsustainability of forestry’s economic contribution: forests will not be able to furnish as great a value in the future as at present. Yet, unsustainability of forestry’s contribution can coincide with enhanced sustainability of the overall economy — the total capital stock and NDP are rising — if sufficient investment occurs in other industries. Unsustainability of forestry’s economic contribution therefore does not necessarily imply that overall economic well-being is unsustainable, any more than sustainability of the latter implies sustainability of the former. For this reason, we recommend relating forest-related adjustments to not just asset values and net value added for forestry, but to the total capital stock and NDP for the overall economy as well. Otherwise, the analysis is likely to exaggerate the economic consequences of forest-related unsustainability.
Forest-related aspects of the SNA and SEEA
Flexibility for making forest-related adjustments in the SNA appears to be greater in the asset accounts than the current accounts. The SNA defines economic assets as "assets over which ownership rights are enforced and which provide economic benefits to their owners" (Bartelmus and van Tongeren 1994, p. 5). These include both produced and non-produced assets. Produced assets can include natural assets, such as livestock for breeding and timber plantations. Non-produced assets include land and natural forests. Depending on the type of asset, asset accounts prepared according to SNA guidelines present information in some or all of the following categories:
In principle, asset accounts in the SNA explicitly and implicitly contain information on forest resources. For example, the economic appearance of non-produced assets includes virgin forests being added to economic reserves, economic disappearance includes timber harvesting, catastrophic losses include forest fires, and natural growth includes the annual timber increment in immature trees.
The SEEA proposes a multi-step process for creating satellite environmental accounts. As a first step, it proposes making existing environmental and resource-related information in the SNA more apparent by disaggregating the current and asset accounts, without changing their basic structure. Regarding the current accounts, it proposes splitting intermediate and final consumption transactions between expenditures on environmental protection and all other expenditures. Similarly, it proposes splitting gross capital formation and consumption of fixed capital between produced assets for environmental protection and all other produced assets. It presents a draft Classification of Environmental Protection Activities, which includes, among others, the forest-related "Protection of nature and landscape (protection of species, habitats; erosion, fire and avalanche protection, etc.)" (Bartelmus and van Tongeren 1994, Box 5).
Regarding the asset accounts, the SEEA proposes reclassifying the information on "Other changes in volume" for non-produced economic assets into four categories (Bartelmus and van Tongeren 1994, p. 7):
These modifications simply involve repackaging existing information in the SNA, without modifying the links between the current and asset accounts or expanding the amount of information they contain.
As a second step, the SEEA proposes shifting "Depletion," "Degradation," and "Other accumulation" within the asset accounts, and also including information on them in the current accounts. In the asset accounts, it proposes adjusting "Gross capital formation" of non-produced economic assets for "Depletion," "Degradation," and "Other accumulation," to yield "Net capital accumulation." In the current accounts, it proposes subtracting "Depletion" and "Degradation" from NDP to yield an adjusted estimate of NDP. It defines "Depletion" as the amount of production of a natural resource above the sustained-yield level.
As third step, the SEEA proposes introducing nonmarket environmental costs and assets, such as those associated with air and water quality and biodiversity. In the asset accounts, it proposes adding a third asset type, non-produced environmental assets. In the current accounts, it proposes subtracting "Depletion" and "Degradation" of non-produced environmental assets, as well as non-produced economic assets, from conventional NDP to yield "Environmentally adjusted NDP" (EDP). It also proposes including transboundary environmental impacts by following the principles that distinguish GDP and GNP. For example, damage caused to another country by transboundary pollution would be deducted from the polluting country’s GNP but not its GDP.
Implementing the second and third steps requires valuing natural and environmental assets. The SEEA proposes several methods for valuing stocks of depletable natural resources. In their operational guide to the SEEA, Bartelmus and van Tongeren (1994) review the net price method in the most detail, although they also mention two other methods, the present value method and El Serafy methods (we will discuss and evaluate all three methods in Chapter 3). Following Repetto et al. (1989), they suggest first establishing physical accounts that contain information on:
They note that the structure of physical accounts will vary from resource to resource. Again following Repetto et al., they suggest valuing "Opening stocks" by multiplying them by the net price of the resource at the beginning of the period, where net price is defined as market price of the extracted resource minus average total extraction cost (including a normal return to capital). Similarly, they suggest multiplying "Depletion" and "Other volume changes" by average net price during the period, and "Closing stocks" by net price at the end of the period. They suggest valuing "Degradation" and "Discharge and treatment of residuals" by directly observing changes in market values, to the extent that is possible. Finally, they suggest calculating a "Revaluation" term as the residual difference between the value of the closing stock on the one hand and the sum of the values of the opening stock and the four intervening terms on the other.
For valuing degradation of nonmarket environmental assets, the SEEA advocates using "maintenance costs": "the costs … which would have been incurred if the environment had been used in such a way that its future use would not have been affected" (Bartelmus and van Tongeren, p. 16). For example, if acid deposition from airborne pollutants changes the chemistry of forest soils, the maintenance cost approach involves estimating the abatement costs the polluting industries would incur if they reduced their pollution to levels that did not cause such damage. The SEEA refers to this as a "cost-caused" valuation approach, as opposed to "cost-borne" approaches that value pollution damages in affected industries or sectors.
This report vis-à-vis SEEA
The analysis in this report confirms the main thrust of the SEEA, namely that nonmarket environmental values and the depletion of forest-related assets should be reflected in national accounts. The accounting system presented in this report is not linked, however, provision-by-provision to the SEEA in the way that, say, the report of the Eurostat Task Force on Forest Accounting is (Newson and Gie 1996). In fact, doing so would be inappropriate, as its philosophy and purpose are fundamentally different from those of the SEEA.
The SEEA maintains the production focus of the SNA. Hence, the only forestry benefits it includes are ones related to forestry output within the SNA production boundary. Its primary purpose is to provide expanded information on the impacts of production on the environment, not the reverse. So, for example, it aims more at identifying the sources of industrial pollution than at quantifying how pollution deposition, regardless of the source, affects forestry production. Similarly, in the asset accounts, it is more concerned with identifying how production affects the values of environmental assets (with respect to "Depletion," "Degradation," etc.), than in quantifying the overall change in asset value for purposes of analyzing sustainability.
The production focus, coupled with the assumption of "strong" sustainability (stocks of environmental capital, not just the value of the total capital stock, must be maintained), helps explain the SEEA's preference for the maintenance cost valuation approach. This approach estimates the additional costs to producers (polluting industries) of using the environment in such a way that it is not degraded. It also helps explain the definition of "depletion" as production of a natural resource in excess of its sustained-yield level.
In contrast, the accounting framework presented in this report focuses on welfare. It considers all forestry benefits, regardless of the nature of the production process generating them, as long as they affect human welfare and are scarce (i.e., have value in an economic sense).10 It is concerned with the impacts of environmental changes on welfare, not the impacts of production on the environment, except insofar as the latter impacts have welfare implications. With welfare as the focus, identifying the industrial causes of certain forms of environmental degradation (e.g., the industrial sources of pollution that damages forests), or the causes of changes in the value of environmental assets (e.g., how harvest, growth, etc. affect the value of a timber production forest), is less important than expressing flow and stock effects consistently and correctly in economic terms.
For this reason, we propose benefits-based ("cost-borne"), not costs-based, valuation methods. Maintenance costs bear no necessary relationship to the welfare consequences of protecting the environment. The costs of maintaining environmental quality might well exceed the benefits of doing so. Under an assumption of strong sustainability, the cost of environmental protection estimated by "cost-caused" methods is implicitly assumed to be a justified expense for maintaining natural capital. For similar reasons, we value dynamic aspects of forest resources like depletion by changes in the present value of benefits, not by the costs of preventing depletion.
In sum, the SEEA is therefore quite consistent in following through on the implications of the SNA's production emphasis and the strong sustainability assumption. These characteristics, however, diminish its value for measuring the economic contribution of forest resources — which, in fairness, it does not purport to do. Our system is designed with the latter purpose in mind. Hence, it emphasizes welfare over production, and benefits-based valuation over cost-based valuation. It also assumes weak, not strong, sustainability. As both production and welfare are important for sound policymaking, as are both the costs and benefits of protecting the environment, and as the degree to which strong or weak sustainability concepts apply will always be a matter of considerable debate, we conclude that both systems can potentially make important contributions to improved forestry policymaking.