|This page in the original is blank.|
Sustainability and Economic Accounting
In light of increasing environmental problems in many sectors, concerns have been raised about the sustainability of current patterns of economic activity in both developed and developing countries. What are the environmental, social, and economic implications of continuing "business as usual"? Will the current path of population, energy use, and growth of human settlements do irreversible harm to the natural ecosystems and life-support systems of the globe? Are we headed for economic overshoot and collapse if we continue to rely on today's technologies? In short, is our economy on a sustainable path?
The concept of sustainability was popularized by the report of the Brundtland Commission, which defined "sustainable development" as "development that meets the needs of the present without compromising the ability of future generations to meet their own needs" (World Commission on Environment and Development, 1987:43). The concern addressed by the Brundtland Commission was whether nations are wasting or abusing their natural endowments of minerals, soils, forests, and aquifers, along with vital environmental resources such as clean air and water, as well as the stock of genetic material.
A parallel effort among economists has been the development of measures of national income and output that take notions of sustainability into account. This appendix examines issues of sustainability from an economic point of view and describes the relationship between measures
of sustainable income and augmented national income accounting. It reviews alternative definitions of income and output; shows how net national product (NNP) is a measure of sustainable consumption under idealized conditions; and then demonstrates how the linkage between current output measures and measures of sustainable consumption breaks down in the presence of nonmarket consumption and investment in environmental, human, and technological capital.
Current Production Versus Sustainable Consumption
The origins of the concept of "social income" or "national income" go back centuries, but two fundamental approaches can be distinguished—one based on the idea of current production and one based on sustainable consumption. The former is the basis for modern national income accounting, while the latter is often used as the appropriate concept for the measurement of sustainable income.
Those who constructed the earliest national accounts were concerned with obtaining accurate measures of national output and national income. Particularly important was tracking current production so governments could take measures to stabilize the business cycle. In attempting to develop a careful definition of national income and output for production-based measures, economists have usually relied on the concept of Hicksian income, which defines net national output as the maximum amount that can be consumed while leaving capital intact (see also Chapter 2).1 In practice, this means national output is defined as consumption plus net investment. This concept is the standard definition of NNP used in the national income accounts of virtually all nations today. It is production based in the sense that it measures production in a given period in terms of market prices. Such a measure is not concerned with whether the economy is sustainable or not, whether production and consumption are growing or declining, or whether the economy is on a path toward prosperity or extinction. Rather, it measures what consumption would be if net investment were zero (that is, if the capital stock were kept intact), measured at the market prices of the economy. Given this definition of income and output, it is easy to understand the rationale of current approaches to augmented accounting. The purpose of these extensions is to expand the purview of the accounts to
include a broader definition of "capital." These studies augment conventional national income by including corrections for human capital; government capital; the stock of research and development; and natural capital such as forests, mineral resources, and environmental resources.
While standard production-based measures of income are useful tools for measuring current production, they do not address concerns about the sustainability of current decisions. An alternative approach, emphasizing sustainability, is provided by Solow (1992), who suggested in his talk of 1992 that a sustainable path for the national economy is one that allows every future generation the option of being as well off as its predecessors. Similarly, according to Repetto (1986:15-16), "The core of the idea of sustainability, then, is the concept that current decisions should not impair the prospects for maintaining or improving future living standards." For purposes of the present discussion, sustainable national income is defined as the maximum amount that can be consumed while ensuring that all future generations can have living standards that are at least as high as that of the current generation.
It should be emphasized that the sustainability-based approach—while deemed particularly useful and appropriate in the context of designing comprehensive national income accounts—is but one of many approaches to analyzing the sustainability of an economy or of the interactions between the economy and the environment. Literally dozens of definitions and approaches have been used in different contexts. It will be useful for present purposes to discuss one major distinction among the different approaches, which relates to the degree of specificity of the variables or objectives to be sustained.
The economic approach to sustainability considers only economic activities and excludes many important individual and collective activities and processes. Economic welfare in this view consists of per capita consumption of goods and services, both market and nonmarket. Living standards are measured on a per capita basis. Consumption includes market items such as food, shelter, and entertainment; in principle, it also includes nonmarket items such as home-cooked meals, along with recreational activities such as fishing or gardening. Consumption does not include many other important values, however. It excludes political and social values such as the importance of fairness, of freedom of speech or association, of religious convictions, and of happy families. Moreover, the values considered are ones that originate in human values. Thus, while human concerns and values about the natural environment are included, the feelings of animals or any intrinsic value of natural ecosystems, such as the existence of coral reefs or of species, are not. To exclude these latter measures is not to deny that they may have value; rather, our economic measures cannot go beyond the boundary of measurable economic activities.
Moreover, economic analyses of sustainability examine consumption or sustainability at the highest level of aggregation—the level of average consumption today and in the future. This level of aggregation masks a number of important ways of disaggregating the complex ensemble of economic and environmental activities. It omits details such as the sectoral or asset breakdown (for example, the separate trajectories of reproducible capital, stocks of subsoil minerals and timber, the quality of air and water, the health of salt marshes, and the value of stocks of genetic material). It assumes that within a sector or asset class, substitutes (including technology) will replace high-priced goods and services. It also overlooks the distribution of consumption among different groups within a country or among countries. It does not distinguish among different future generations and focuses only on the present generation versus the generalized future.
In addition, most treatments of sustainability do not deal with issues of uncertainty or risk. Must a path be sustainable with absolute confidence, or on average, or 90 percent of the time? Would we prefer a certainty of nondeclining consumption over an alternative that involves a robust growth in living standards plus a tiny chance of a small decline? How would we feel about a promising technology that offers a 99.9 percent probability of sustainability and a 0.1 percent chance of extinction? A short journey down the road of stochasticity raises numerous unanswered questions about the concept of sustainability.
In limiting the present analysis to this highly generalized and aggregated version of sustainability, it is recognized that many worthwhile goals will be overlooked. An alternative view of sustainability, for example, might hold that "maintaining capital intact" should apply at a more disaggregated level than the entire asset base of an economy. This narrower perspective might hope to protect certain assets or flows or subsystems—such as a suite of species or a group of important ecosystems, or even "natural capital" more generally—so that future generations could enjoy them at today's levels. Such a perspective depends on ''sector-specific" and "use-specific" definitions of sustainability. Defining sustainability in this narrower sense is often useful as a guide to policy making or as a practical shorthand way of expressing certain desirable conservation goals, but it generally is too narrow and subjective to serve as a principle for constructing measures of national income. Moreover, if taken literally, the injunction to keep "natural capital intact" is probably infeasible because human activities inevitably cause the levels of some natural assets somewhere to decline.
Sector-specific or use-specific definitions of sustainability raise other practical problems. One issue is selection of the list of assets to be maintained. Which specific set of resources is to be maintained? Who selects
this list? Who decides on global assets? Answers to these questions matter a great deal because, both literally and figuratively, what is sustainable for the forest is unsustainable for individual trees. Additionally, from a technical point of view, the sector-specific or use-specific approach assumes that no substitution is allowed between the particular resource or use chosen to be sustained and other resources not on the selected list. If killer whales are on the sustainability list while the porpoises they eat are not, not one more killer whale can be harvested, even if doing so would allow one thousand more porpoises to live.
This short discussion should help indicate both the usefulness and the pitfalls of alternative definitions of sustainability. On the whole, despite its shortcomings, the broad measure of sustainability is likely to be the best single criterion for defining sustainable national income. This measure defines sustainable income as the maximum amount that a nation can consume while ensuring that all future generations can have living standards that are at least as high as that of the current generation. Such a broad concept provides an intuitively appealing way of providing a simple summary answer to the complex question of where our economic growth and development are taking us.
Correspondence between Net National Product and Sustainable Income under Idealized Conditions
What is the relationship between concepts employed in the current National Income and Product Accounts (NIPA) and measures of sustainable income? To begin with, it should be noted that the most popular measure of output, gross domestic product (GDP), differs from a conceptually appropriate measure, NNP, in two ways. First, GDP includes capital consumption (see glossary), which leads to double counting of this source of income. Traditionally, output measures have emphasized gross rather than net product because depreciation is difficult to measure accurately. Second, domestic product excludes the net factor earnings abroad of domestic residents, which is included in national product. Inclusion of net factor earnings abroad is desirable if output is designed to measure the sustainable consumption of the nation. The recent switch in emphasis from national to domestic product came about because domestic product is more closely related to domestic output and employment. While the emphasis on GDP rather than NNP is understandable, the panel emphasizes that NNP is conceptually preferable as a measure of sustainable income.
However, there is a close relationship between traditional measures of output and ideal measures. This relationship, known as the output-sustainability correspondence principle, holds that under idealized condi-
tions, NNP and sustainable income are identical. More precisely, when the national accounts include all stocks of capital and other dynamic features that affect production and when markets accurately capture the social values of all inputs, NNP is an appropriate measure of sustainable income. In other words, the sum of total consumption and net capital formation is equivalent to the maximum sustainable amount of consumption an economy can indefinitely maintain. Hence under idealized conditions, including zero population growth, extending the NIPA toward a comprehensive measure of Hicksian income would make output and income more accurate indexes of sustainable income. 2
The balance of this section is devoted to explaining the output-sustainability correspondence principle; a number of qualifications to the principle are presented in the next section. We simplify the analysis by assuming that there is just one composite consumption index, which measures the real standard of living of the representative household. This generalized consumption index should be interpreted as being broader than a traditional index of real consumption because it comprises both market and nonmarket consumption, including not only food and concerts, but also wilderness experiences and highway congestion.
Consumption is produced by a large number of different kinds of capital goods. Some of these goods, such as equipment and structures, are included in the national accounts. Others are nonmarket capital, such as stocks of mineral deposits and fish, human capital, technological capital in the form of patents, and the like. Ideally, the list of capital goods should be as comprehensive as possible, subject to the limitation that the goods have meaningful units of measure and that scarcity prices—either actual market prices or imputed shadow prices—can be calculated to measure their values.
The major behavioral assumption here is that the outputs and prices of the economy are generated by an optimized economy for which there is a complete set of accounts and in which all spillovers are internalized. This means that consumption, investments, and prices are generated by a process in which (1) the accounting system is complete in the sense that all dynamic elements and capital stocks—natural, environmental, and technological—are included in the measure of income; (2) all transactions are "internalized," meaning that markets capture all the social costs and benefits of all economic activities; and (3) output, consumption, and investment result from social decisions that optimize a consistent intertemporal objective function.
The mathematical proof of the correspondence principle depends on a multisector generalization of the Ramsey-Cass-Koopmans optimal growth problem. In this approach, the objective is to maximize the present discounted value of the utility of consumption. The fundamental relationships are a utility function that represents society's intertemporal preferences over alternative consumption streams, a production function that indicates how consumption can be produced as a function of a wide array of capital stocks and autonomous dynamic factors, and a pure rate of time preference that indicates the relative priority of consumption of different periods or generations. For this purpose, it is not necessary to observe the utility function or the rate of pure time preference. Rather, it is assumed that the economy behaves as if it were the solution of an optimal growth problem. The present discounted utility specification has been axiomatically derived by Debreu and Koopmans as the appropriate intertemporal welfare function from postulates that encompass a general set of social objectives (for a full discussion of this approach, see the references in footnote 2).
Under these conditions, along with no growth in population, comprehensive net domestic product (NDP) is an appropriate measure of sustainable income. That is, national income as measured by current comprehensive consumption plus the sum of the values of the net accumulation of assets in different sectors is equivalent to the maximum level of consumption that can be indefinitely sustained. Moreover, under the stringent conditions of a complete, internalized, optimal growth path, this measure of sustainable income will be exactly captured in measured NNP. One important conclusion is that to the extent that the national accounts omit important components of consumption and of net capital accumulation, they may provide misleading measures of sustainable income.
Qualifications to the Output-Sustainability Correspondence Principle
The output-sustainability correspondence is of fundamental importance for guiding decisions about the design of the NIPA. However, there are important practical and theoretical qualifications to this principle that must be emphasized. These qualifications concern (1) the incompleteness of the consumption and asset categories; (2) the presence of autonomous technological and other processes; (3) revaluation issues; and (4) problems associated with imperfect markets, imperfect foresight, and other departures from the optimal-growth framework. Any of these four conditions generally implies that measured NNP will depart from the ideal measure of sustainable consumption.
In measuring comprehensive national income and output, we desire that our accounting system be as complete as possible in the sense of including as many components of consumption and net investment as is practical. To the extent that we omit certain items, this will lead to errors or residuals in our measure of national income. To clarify this point, we can write augmented NNP as:
where augNDPt is augmented NNP and is equal to consumption and net investment, CtI and CtII are consumption of types I and II, and ΔKtI and ΔKtII are net capital formation of types I and II. Suppose that type I consumption and capital formation refer to those flows as measured in the standard NIPA, while type II refers to nonmarket activities not captured in the accounts, such as the flows associated with forests, fisheries, and underground aquifers. Moreover, suppose these two sectors are all that matter for economic welfare. If the assumptions of the correspondence principle hold, one can accurately measure sustainable income by adding the appropriate values for sector II to conventional NDP. If, by contrast, sector II is omitted, sustainable income will differ from national output by a residual term as follows:
In equation (A.2), sustainable income can be measured as conventional NDP plus a residual (RtA), which is equal to omitted consumption and investment. The residual RtA in equation (A.2) reflects the omitted consumption and investment terms, RtA = CtII + ΔKtII. The residual element in RtA reflects elements of economic welfare that are not reflected in measured national output. Naturally, this residual may be negative or positive, depending on whether the sum of nonmarket consumption and net investment is negative or positive.
The major point here is that because of incompleteness in the consumption and investment categories, conventional measures of NDP will depart from sustainable income. Moreover, given the vast array of nonmarket activities—from leisure and home-based investments in human capital to nonmarket consumption and investments in environmentally sensitive sectors—there is a strong presumption that the residual may be significant and that current measures of national output do not adequately reflect sustainable income.
Technology, Institutions, and Other Autonomous Processes
In addition to market and nonmarket capital stocks, there are likely to be important social and technological processes that affect the trend of production. This point is well established in studies of economic growth and development, which have found that conventional capital formation explains only a small fraction of the growth of individual nations or the differences among nations. Examples of other important factors are the level of and improvement in technology; the institutional and legal structure, including the strength of tangible and intellectual property rights; the stability of the political system; the level of openness of the economy and the presence of pacific or warlike neighbors; and the honesty or corruptness of public and private transactions. It is likely that many environmental factors—such as the value of the biosphere or the climate—fall into this category.
These factors are in some sense "social capital" and clearly affect a nation's productivity and future income. But to call them "capital assets" is really a metaphor; they cannot be treated as such in any serious accounting sense. There is no metric for measuring many of these important social elements, nor is there an established methodology for valuing them. From the point of view of the national accounts, they are autonomous dynamic factors that may have a substantial impact on productivity.
These autonomous factors also lead to a divergence between sustainable income and measured national output. For example, if ongoing technological change leads to sustained productivity improvement, future generations will have higher levels of income than current generations, and current sustainable income is therefore higher. If, by contrast, current activities are leading to a general deterioration in the absorptive capacities of the environment or if climate change will lead to irreversible damage to the fertility of the earth, current sustainable income is lower.
Relatively little work has been done on the size or sign of the missing residuals, particularly those due to autonomous dynamic factors. Work on the sources of economic growth indicates that technological change has historically been a dominant factor in the growth of measured per capita output and living standards. Recent studies also point to the importance of institutional arrangements, such as the openness of the economy, the role of markets in resource allocation, and proximity to coastlines or large and growing markets. These findings, along with illustrative calculations, suggest that the autonomous residual may be positive and relatively large. Much work is needed in this area to obtain a better appraisal of the importance of these autonomous dynamic factors.
Revaluation effects, or real capital gains and losses, raise perplexing problems for measuring income and output (see also Chapter 3). Conventional measures of real national output involve only the level and changes in the weighted quantities of goods and services consumed and invested; revaluation effects are omitted from conventional income and product measures. A difficulty arises when there are changes in the relative prices of consumption goods across borders or over time. The simplest example occurs in the case of a change in a country's terms of trade. Suppose there is a 10 percent permanent and unexpected rise in world oil prices relative to consumption goods. An economy that produces only oil will have a 10 percent increase in its sustainable consumption level even though there has been no change in the time path of its physical oil production. Similar effects would occur if there were changes in real interest rates, which are in effect changes in the terms of trade between the present and future.
Changes in prices lead to revaluation effects, which are changes in the value of income or capital in terms of current consumption goods. Inclusion of revaluation effects in sustainable income is a major departure from traditional definitions of output in the national income accounts. In fact, from the point of view of a small open economy, price changes are often as important a determinant of consumption possibilities as changes in investment, in trade regimes, or in technology. In principle, current approaches to measuring sustainable income treat revaluation in a way that is parallel to the treatment of autonomous dynamic influences. (In practice, such revaluations are rarely done in national income accounting.)
Revaluation effects and other dynamic factors add another term to equation (A.2) that reflects the positive or negative contribution of price and interest-rate changes and other dynamic effects to the highest level of sustainable consumption. Equation (A.2) can be modified to incorporate revaluation effects and the influence of the autonomous elements as follows:3
where RtA reflects the residual terms in equation (A.2), and RtB is the residual impact on sustainable income due to dynamic autonomous factors and revaluation effects. These new elements cause particular diffi-
culties for sustainability accounting because they generally cannot be easily measured or found in marketable assets or consumption.4 In this most general expression, sustainable income is NDP plus the residual value provided by omitted consumption and capital formation plus the residual due to autonomous dynamic factors, plus the revaluation effects.
Departures from Efficient Decision Making
The present discussion of measurement of sustainable income depends on a stringent set of assumptions about the processes of social decision making. A set of qualifications concerns departures from the idealized assumptions of perfect competition, perfect foresight in asset markets, and an optimal allocation of resources over time. It is particularly important to understand that sustainable national income is equivalent to the maximum that can be consumed while leaving endowments sufficient to ensure equivalent living standards for the future. But to say that consumption can be sustained does not ensure that future living standards will be sustained. Consuming less than sustainable income provides the next generation the resources they need to keep the economy going at the same or improved living standards if they so choose. There is, however, no way to bind subsequent generations to serve as responsible trustees for their descendants. The current generation might leave a generous stock of oil, forests, and tangible capital, but some future generation might squander its inheritance through high living, bad judgments, or military adventures. But such behavior would not be reflected in current prices and quantities and therefore would not be reflected in current measures of sustainable income. This example illustrates one of the limitations of any income measure that relies on market prices and quantities.
Example for Petroleum Resources
The above approach can be illustrated with a specific example of how the sustainability framework can help in understanding real policy concerns. Consider the constraints on economic growth posed by the finiteness of petroleum stocks. What is the effect on sustainable income of the fact that we will run out of oil some day? Without an accounting frame-
work, it is difficult even to pose such a question meaningfully. Using the concept of sustainable income, it is possible to measure the impact of changing petroleum stocks on sustainable income by adjusting income with a new investment term that equals the market or scarcity price of petroleum reserves times the net change in those reserves. The fraction by which sustainability, and therefore future welfare, will be lowered as a result of running out of oil is in principle captured by the fraction of comprehensive product accounted for by the value of the change in net stocks of oil reserves. This measure, which is exactly the approach analyzed in Chapter 3, will provide a reasonably accurate estimate of the impact of depletions or additions on the sustainable consumption of the United States.
To better appreciate the power of these results, imagine for a moment that a dream team of world-class researchers is asked to analyze the important question of the impact on future living standards of the exhaustion of finite petroleum stocks. The researchers are provided with a large budget and told to make the best possible estimate. What would this dream research team do? They would project all relevant demand and supply response functions, estimate all sectoral rates of technological change, include all relevant elasticities of substitution between oil and everything else, project future technological advances, and so forth. They would then use thousands of linked parallel processors to simulate future trajectories with a massive, dynamic, computable general-equilibrium model. After this immense research project had been completed, the dream team would be asked to provide their best overall estimate of the impact of changing oil stocks on future living standards.
The surprising result is that estimates of the market value of oil depletions or additions probably offer the most accurate measure of the impact of changes in petroleum stocks on living standards. Moreover, the market's estimate is likely to be more credible than the dream team's estimate because it is based on an "invisible-hand" evaluation that is more reliable than the dream team's computer model. This invisible-hand model represents the judgment of the thousands of market participants who consider every relevant aspect of the problem treated by the dream team and additionally have their personal fortunes and livelihoods at risk if they make the wrong decisions.
There are many different approaches to sustainability. From the point of view of a national economy and national income accounting, a useful definition is that sustainable national income is the maximum amount a
nation can consume while ensuring that all future generations can have living standards at least as high as those of the current generation.
The NIPA have a close relationship to measures of sustainable income. The usual measure of NDP corresponds to the highest sustainable level of consumption under idealized conditions. The most important conditions underlying this correspondence are the inclusion of all consumption and net investment and the absence of technological change or other dynamic autonomous elements.
Measures of national income and output would be closer to the ideal measure of sustainable income if omitted consumption and net investment were included to obtain augmented income and output measures. Omitted items would include nonmarket consumption, such as home production, and final environmental services and nonmarket investment, such as changes in the value of resource stocks, along with investment in human capital.
When there is unmeasured consumption, investment, autonomous dynamic elements, or revaluation effects, residual terms must be added to reflect the contribution of these factors—positive or negative—to future income. Particularly important residual effects are due to technological change and institutional factors such as the nature of tangible and intellectual property rights and the degree of openness of the economy.