Choices of Ecosystem Capital without Discounting and Prices

Abstract:. In the last decade, neo-classical economists have used the Consumption-Growth Based Utility (CGBU) framework to suggest an appropriate discount rate for eco-system capital. The CGBU framework, being based on the simplicity and single-valued nature of man-made capital, is subject to many limitations. Three general limitations of the framework, irrespective of its use for man-made capital or ecosystem capital, paradox of consumption growth, expectations’ insensitive utility specification, and insensitivity of the framework to the nature of returns are discussed. With respect to its use for ecosystem capital, the framework has two limitations. First, ecosystems cannot be commoditized. Second, ecosystem capital satisfies differentiated needs, and, hence, gross substitution between different components of ecosystem capital or between ecosystem capital and man-made capital is not possible. Therefore, the rate of discount for ecosystem capital suggested on the basis of the framework is arbitrary. In the context of international security, choices of ecosystem capital are similar to the choices of military decisions. Similar to military decisions, ecosystem choices should be made without discounting and prices. The neo-classical framework does not have appropriate tools for such decision analyses, and other streams of economics, specifically post-Keynesian economics, may help assist decision- makers in analysing ecosystem choices without prices and discounting.

Key Words: Commodity, Development, Expectations, post-Keynesian Economics, and Utility theory.

Acknowledgements: I am thankful to Yaoqi Zhang, Susan Lee, Julie Casimirri and three referees for their helpful comments. Research funding from the Social Sciences and Humanities Research Council of Canada (Grant No.410990343), Natural Sciences and Engineering Research Council of Canada (Grant No. 203032-98 RGPIN), and Connaught Fund, University of Toronto, are greatly acknowledged.

Choices of Ecosystem Capital without Discounting and Prices

1.0 Introduction

The world's ecosystems are capital assets yielding a flow of vital goods and services. Since the beginning of industrialisation, economists have emphasised the value of man-made capital, and the wealth of a nation or society has been measured by the quantity of man-made capital. On the other hand, ecosystem capital has been treated as a free good, and services from ecosystems have been assumed to be available without any cost. This approach might have been safe enough in the sort-run, in the period of abundance of ecosystem capital, but has proved to be wrong in the long-run. Alarmed by the increasing scarcity of ecosystem capital, in the last two decades many economists have attempted to assign a monetary value to ecosystem capital. In terms of the number and scope of the studies on the valuation of ecosystem capital, an impressive progress has been made in development of concepts in this area, but economists need not be enthusiastic to claim that these valuation techniques have improved the efficient allocation of ecosystem capital. In the case of a multi-dimensional and multi-valued resource such as ecosystem capital, the problems of cognition, incongruity, and composition result in information losses, and if hypothetical prices, assigned by these valuation techniques, do not convey the information pertinent to particular choices, there is no way these prices can contribute to efficient allocation of ecosystem capital (Vatn and Bromley, 1995). Hence, hypothetical valuation techniques carry no more normative significance than that of claims made by different interest groups, and there is nothing in neo-classical economics in general – or in hypothetical valuation in particular – that can address the socially optimal level of land devoted to parks and wilderness or ecosystem capital (Bromley, 1989; Vatn and Bromley, 1995).

Valuation, no doubt, is one of the most critical elements of resource allocation (Dally et al., 2000), but discounting is also as important as valuation, if not more, for inter-temporal resource allocation. The choice of an appropriate discount rate has troubled economists for many decades, and environmental goods and ecosystem capital have further magnified the troubles of economists. In the 1990s, many economists, such as Kolb and Scheraga (1990), and Weitzman (1994) have addressed the issues of discounting of environmental expenditures, and other economists, such as Arrow et al. (1995), Arrow et al. (1996), and Azar and Sterner (1996) have examined the issue of discounting for climate change. These recent discussions on discounting of environmental goods and ecosystem capital have moved from the traditional position, of corrected market rate of interest as a proxy for the social discount rate (SDR) or the social rate of time preference (SRTP), to an argument that the future should be discounted because we will be better off in the future. We will term this argument as consumption growth based discounting (CGBD) framework. The results of consumption-growth based discounting have been derived using two-period utility models. The main argument of these discussions is that the social rate of time preference will be equal to the sum of "the rate of pure time preference" and a discount element due to the expected declining marginal utility of additional consumption. Using this formulation, Arrow et al. (1995) argued that the environmental literature and some modern philosophers, who are critical of cost-benefit analysis, do not recognise the distinction between a zero rate of pure time preference and a zero social rate of time preference, and as long as consumption growth is positive there will be a nonzero social rate of time preference. However, Arrow et al. (1995) and other economists who have put forward similar arguments do not recognise multi-dimensional limitations of the consumption-growth-based utility (CGBU) framework of discounting, which can prove fatal in inter-temporal allocation of ecosystem capital. In this paper, we will highlight and discuss two broad categories of these limitations - general limitations of the framework and specific limitations of the framework with respect to ecosystem capital[1].

The CGBU framework of discounting has three general limitations, irrespective of its use for man-made capital or ecosystem capital. First, the main argument of the framework is that the marginal utility of additional consumption in future will be less, but the resultant discount factor is used to discount the total consumption in future. The results obtained from the discounting of the total future consumption place a question mark on the need for consumption growth in the future. Second, in the framework, expectations of a higher standard of living in the future are recognised, but these expectations are not incorporated in the utility functions used to derive the results. Third, the framework is based on an analysis of discrete one-time returns, and, hence, is irrelevant to a continuous stream of returns.

The foundation of CGBU framework is in the simplicity and single-valued nature of man-made capital, which leads to its specific limitations with respect to discounting of multi-attribute and multi-valued ecosystem capital. In the case of ecosystem capital, the main question is - what is being discounted, physical units or monetary units? In the System of Integrated Economic and Environmental Accounting (SEEA) inadequacies of monetary accounting of environmental goods and ecosystem capital are very well recognised, and physical accounts are an integral part of SEEA (United Nations, 1993). Despite this many neo-classical economists argue that the economic value of physical units of ecosystem capital should be discounted. But, as Vatn and Bromley (1995) argues that due to the problems of cognition, incongruity, and composition, economic value of ecosystem capital obtained from the hypothetical valuation methods is as arbitrary as a value claimed by any non-economists group. Hence, the question remains as to what is being discounted - an arbitrary number termed as economic value of ecosystem capital?

Our argument is that due to the complexity of ecosystem capital, created by multiple-attributes leading to multiple economic as well as non-economic values, the CGBU framework has many specific limitations with respect to its use for ecosystem capital[2]. First, the CGBU framework is based on commoditization of ecosystem capital, while ecosystem capital is non-reducible to the venerable fiction of economic commodities. Hence, it is not possible to assign a single monetary number that captures the total value of ecosystem capital. Second, in the CGBU framework, similar to the neo-classical utility framework, wants are undifferentiated, wants and needs are not distinguished, needs are also undifferentiated, and gross substitution is possible between all goods of a utility bundle. However, contributions by different attributes of ecosystem capital correspond to differentiated needs, and there may not be a gross substitution between different levels of needs and different needs at the same level. Hence, the discussion of the social rate of time preference of ecosystem capital in current literature follows the same path as followed by valuation techniques of ecosystem capital. As a result, the numerical measures suggested by this discussion are as arbitrary as the prices assigned to ecosystem capital by valuation techniques. Discounting of an arbitrarily fixed economic value of ecosystem capital by an arbitrary discount factor cannot improve the efficiency of inter-temporal allocations of ecosystem capital. In fact, there are no tools in the neo-classical economic theory to find an appropriate discount rate for ecosystem capital. In the present economic systems, many decisions, such as the allocation of resources for national defence (military), are made without using a discount rate that is suggested by the neo-classical economic analysis. Similar to the decisions of national security, many decisions regarding allocations of ecosystem capital are the decisions for the international security. Hence, these decisions should also be made without prices and without an economically determined rate of discount – and without apologies to neo-classical economists[3].

In support of our arguments, we first provide an overview of the existing literature on discounting, specifically on the CGBU framework of discounting. Second, we discuss three general limitations associated with the CGBU framework. Third, we discuss two specific limitations of the CGBU framework with respect to ecosystem capital. Fourth, we compare the decisions related to ecosystem capital with the decisions of national defence. Finally, we conclude that the neo-classical economics has no tools for economic analyses without prices and discounting, and the concepts from other streams of economics, specifically post-Keynesian economics, could be useful in such analyses.

2.0 An Overview of Economic Literature on the Social Rate of Discount

Fisher (1930, pp. 104-106) postulated that under conditions of a perfect capital market, an individual's rate of discount would be, at the margin, equal to the real rate of interest in the market. In the case of public projects, the social rate of discount (SRD), also referred to as the social rate of time preference (SRTP), replaces an individual's rate of discount. Many economists, such as Dasgupta et al. (1972), Little and Mirrlees (1974), Sen (1967), and Stiglitz (1982), have discussed the issue of the social rate of discount, and the focus of these discussions is on factors, such as market distortions due to taxes, responsible for the difference between the market rates of interest and the SRD. It has been commonly agreed that in assessing public programs, benefits and costs that accrue as consumption must be valued differently from benefits and costs that accrue as investment, and, generally, two methods are in practice to accommodate this proposition. In the first method, simplified models of the economy, that incorporate relevant market imperfections, are used to solve for the optimal policy, and then to solve for the social discount rate that will produce the optimal policy (Lind, 1990). The most common outcome of these models is the suggestion that the social rate of discount should be some weighted average of the marginal rate of return on capital and the consumer's rate of interest. In the second method, which is known as the "shadow price of capital", all costs and benefits are converted to consumption equivalents, and then the consumers’ rate of interest is used as a single rate of discount (Lind, 1990).

In the 1990's, environmental issues specifically the issue of global climate change entered into the debate on social rate of discount. Many social scientists, such as Kolb and Scheraga (1990), Broome (1992), Cline (1992), Nordhaus (1993), Schelling (1993), Weitzman (1994), Arrow et al. (1995), Arrow et al. (1996), and Azar and Sterner (1996) have contributed to these discussions. Some economists, such as Schelling (1993) and Arrow et al. (1995) have called these discussions confusing. However, an alternative theoretical framework, which was referred to as the CGBU framework earlier, and d'Arge et al. (1982) and Sen (1982) termed as Benthamite welfarism, has emerged prominently through these discussions[4]. In this framework, the SRTP is composed of the two components: (i) the rate of “pure” time preferences, and (ii) a component due to the expectation that society will be richer in the future. The rate of pure time preference is rationalised in terms of impatience and uncertainty about the future. The second component is based on the decreasing marginal utility of additional consumption in the future. Hence, if income is expected to grow over time, the decreasing marginal utility implies that an additional unit of consumption in the future is worth less than what it would be if it was consumed today. In this formulation, the SRTP is expressed as:

SRTP = r +q.g (1).

where r denotes the discount rate arising solely from myopia or impatience and is called the "rate of pure time preference," and q.g represents the discount element due to the expected declining marginal utility, and is equal to the product of the absolute value of the elasticity of marginal utility (q) and the growth rate of per capita consumption (g). This formulation has been used to argue that even if we assign a zero value to the rate of pure time preference[5], for inter-generation equity, we have to discount an income or cost because we expect to be richer at the time when it occurs[6].

Kneese and Schulze (1985) analysed the case of nuclear waste disposal, and demonstrated that even a Utilitarian ethic would require discounting only if compensation could actually be paid to future generations. Since, it is unreasonable to assume that compensation can be paid to future generations, their analysis demonstrates the use of a zero discount rate for the egalitarian specification of the Utilitarian ethic, and the outright rejection of nuclear waste storage for a Libertarian ethic. Even in the presence of these results of Kneese and Schultze, also derived from the Utility theory, the CGBU framework of discounting is fairly common and dominant among economists discussing discounting of climate change. An extensive use of the framework clearly demonstrates that these economists, at least implicitly, agree that the market determined rate of return on capital, the consumer's rate of interest, and their weighted average are not appropriate measures for the SRTP for climate change or ecosystem capital. The results of the CGBU framework have been derived without considering any specific features of climate change or ecosystem capital. Hence, we do not see any improvement in the arguments based on the CGBU framework over the previous market-based arguments for discounting of ecosystem capital. In addition, the CGBU framework has some limitations that question its appropriateness even for the discounting of common economic commodities or man-made capital. We discuss these general limitations next.

3. General Limitations of CGBU Framework of Discounting

Three general limitations of the framework are: (i) the paradox of consumption growth; (ii) the specification of the utility function being insensitive to the expectations of the consumers; and (iii) the insensitivity of the framework to the nature of returns from economic goods.

3.1 Paradox of Consumption Growth: In the CGBU framework, the future should be discounted because the additional consumption in future will be of, so-called, lesser inherent value. It is argued that even if we reject the impatience component of the SRTP (r = 0), the expectation of higher standards of living and, therefore, lower a marginal utility of incremental consumption in the future will be the sole factor contributing to the SRTP (Cline 1992, p. 249). This discount rate should reflect the expected diminishing marginal utility of consumption associated with the expected rise in per capita income over time (Cline 1992, p. 249). Hence, even though economists using the utility approach have recognised that it will be the additional consumption in the future that will have lesser inherent value, it is the discount rate derived from this approach that is used to discount the total consumption in the future. Price (1999) has rightly called this a fatal error in the literature. This error gives rise to a phenomenon termed the “paradox of consumption growth”. Suppose for a given country, per capital consumption in the base (zero) year is $G(0), and per capita consumption grows at an exogenously determined annual rate of growth of g for n years. The per capita consumption in year n will then be G(n) = G(0)[1+g]ⁿ, and the discounted present value of nth year per capita consumption, using a discount rate of q.g, will be G(n₀) = G(0)[(1+g)/ (1+ q.g)]ⁿ . In such cases where the absolute value of elasticity of marginal utility is greater than one, the present value of the nth year per capita consumption (G(n₀)) will be less than the value of the per capita consumption in the base year (G(0)) for a positive rate of consumption growth, and G(n₀) will be greater than G(0) for a negative rate of consumption growth As a result, negative consumption growth will be welfare-enhancing, and the countries which are working hard to achieve higher consumption growth will be working to reduce their welfare!

Given the annual rate of consumption growth equal to g and the per capita consumption of one dollar in the base year, the factor [(1+g)/(1+ q.g)]ⁿwill give the present discounted value of per capita consumption in year n. Let us call this a Conversion Factor. The values of the Conversion Factor for different number of years and for positive and negative rates of consumption growth are given in Figures 1 and 2, respectively. These figures indicate that as we move further into the future (higher value of n), the discounted value of per capita annual future consumption (G(n₀)) will decrease for positive consumption growth, and increase for negative consumption growth. The figures also indicate that the countries that grow at a faster rate will be worse off than those which grow at a slower rate. Countries that are unable to show a positive growth rate will be better off than those countries that do show a positive growth rate. Countries that show a higher negative growth rate will be better off than those that show a lower negative rate of growth. Hence, according to the CGBU framework of discounting and the practise in which it is being used, countries, which want to be better off in the future in terms of discounted present value of consumption in the future years, should aim for a higher negative rate of consumption growth. The simplest solution of this paradox of consumption growth is to ensure that the discounting factor derived from this approach is used only for discounting the incremental consumption, and not the total consumption. However, another limitation of the CGBU framework, discussed next, limits the usefulness of the simplest solution.

3.2 Utility Specifications Insensitive to Expectations: In the CGBU framework, economists have recognised that economic agents will have expectations of high living standards in future, but the preferences incorporated in the utility function remain static. In the framework, the utility derived from the unit X of the consumption bundle in the future is assumed to be the same as in the present period, irrespective of the present and future expectations of the economic agent or society. Thus, utility specifications are insensitive to expectations, and this is the second error of the framework. The degree of satisfaction (utility) is a relative concept, and must not be treated as absolute. Thus, the utility derived from the same unit (say X) of consumption item may be different during two time periods if the expectations of consumers are different in these periods. For example, suppose in the base time period T(0), the economic agent expects to consume ten units, while in the next time period T(1), the agent expects to consume fifteen units of the same consumption item. The utility derived from the first unit of the consumption item or any other unit of the consumption item in the base time period (T(0)) will not be the same as the utility derived from the same unit in the next time period.. The utility from any unit of the consumption item will depend upon its relative contribution to total expectations. In the first period, the terminal point of the utility scale of the agent corresponds to the total utility from ten units, while in the second period the terminal point corresponds to the total utility from fifteen units. Hence, different units of the consumption item cannot be placed on the same points on two different utility scales, one scale corresponding to each time period. In some cases, the marginal utility of the eleventh to the fifteenth unit of the item may be zero in the base time period, but in the next time period, T(1), all these units may have some positive marginal utility. Hence, the argument of decreasing marginal utility (in an absolute utility sense) of additional consumption may not hold in view of the expectations of high standard of living in the future.

The idea of relative utility is not really strange of new. The Greek philosopher Epicurus (341-270 BC) postulated that the degree of satisfaction (utility) depends upon the proportion of attainment and expectations (Rescher 1980, p.12), and many scholars, such as Arnold (1943), and Campbell et al. (1976) have provided empirical support for this proposition. Normally, people’s expectations are geared to their past experiences, and their attainments are subject to the limits of finitude that are generally prevalent in human affairs. Inevitably, our expectations surpass our attainments. Objectively, the per capita consumption may increase due to consumption growth, but it does not increase fast enough to meet subjective expectations. Rescher (1980, p.19) called this phenomenon the Fundamental Paradox of Progress: where “progress produces dissatisfaction because it inflates expectations faster than it can actually meet them, and this is virtually inevitable because the faster the expectations actually are met, the faster they escalate.” In view of this relationship of utility with attainment and expectations, utility functions used in the CGBU framework must be modified accordingly.

The CGBU framework has used a constant relative risk aversion (CRRA) utility function (U = C^{(1 +}^q)/ (1+q), where q is the constant elasticity of marginal utility which is also the negative of the coefficient of risk aversion). In this utility function C is the absolute quantity of the consumption item. However, as argued earlier, utility depends on the proportion of attainment and expectations of the agent, and not on the absolute quantity of the attainment. Hence, an appropriate formulation of utility function will be in the terms of proportion of the quantity of the consumption item with respect to the expectations of the agent. Suppose, in the base time period T(0), the units consumed are expressed as C₀ and the expectations of the agent are of C₀₀ units, which we assumed to be constant for a given period. CRRA utility function will be expressed as U₀ = [(C₀ /C₀₀ ) ^{(1 +}^q)]/ (1+q). Similarly, if the units consumed are C_n and the expectations of the agent are C_nn in the nth time period (T(n)), CRRA utility function will be expressed as U_n = [(C_n /C_nn ) ^{(1 +}^q)]/ (1+q). Since, the expectations for a given period (C₀₀ and C_nn) are constant, the mathematical features of the utility function - the elasticity of marginal utility - remains unchanged, and equal to q.

Now, we postulate an exogenously determined future path of per capita consumption, similar to the CGBU framework, rather than seek to identify the growth path after taking into account investment opportunities. Suppose that in this outlook, per capita consumption grows at the annual rate of g, and the expectations of the people for per capita consumption grow at the annual rate of g_e , which will normally be higher than g.

Let b_n be the ratio of the number of additional units of consumption in the nth period to the number of additional units of consumption in the base period required to generate identical increases in utility in the two years. Then

b_n =U^¢₀ / U^¢_n , where U^¢ is the marginal utility. (2).

In the case of CRRA utility function, marginal utilities for time period (0) and time period (n) are C₀^q/ C₀₀ ⁽¹⁺^q)and C_n^q/ C_nn ⁽¹⁺^q) , respectively. Hence,

b_n =(U^¢₀ / U^¢_n)= (C_n /C₀)^-^q (C₀₀/C_nn ) ^–(1+^q) (3).

Now, since consumption grows at the annual rate of g and the expectations grow at the annual rate of g_e,,

C_n = C₀ (1+g)ⁿ , and C_nn= C₀₀ (1+g_e )ⁿ

Therefore,

b_n =((1+g)ⁿ)^-^q (1/(1+g_e)ⁿ) ^–(1+^q) (4).

The SRTP, or R, then, must be such that it equates b_n extra units of consumption in time period (n) to 1 extra unit of consumption in the base period (0). Thus

b_n = (1+R)^n. (5).

From Equations 4 and 5, we have

1+ R = ((1+g))^-^q (1/(1+g_e)) ^–(1+^q) , and on simplification[7], it gives

R = - q(g - g_e) + g_e = -- qg + g_e (1+ q) (6).

Equation 6 indicates that the SRTP will depend upon the rate of growth of consumption as well as the rate of growth of expectations of the agents, and this result is different from that suggested by the CGBU framework. The values of SRTP for two cases – rate of growth of consumption (g) being equal and not equal to the rate of growth of expectations (g_e) ) are discussed next.

Case 1: Rate of growth of consumption (g) is equal to the rate of growth of expectations (g_e): The SRTP will be equal to the rate of growth of consumption or the rate of growth of expectations, irrespective of the value of the elasticity of marginal utility. In this case the SRTP will be less than the value of SRTP given by the CGBU framework, which is equal to the product of the argument of q and g, and argument of q is greater than one. If we use the same values of g equal to 1 and q equal to -1.5, as used by Cline[8], the SRTP will be equal to 1%, and not 1.5%. However, it is an ideal case, and as argued earlier the expectations grow at a faster rate than the growth rate of consumption.

Case 2: Rate of growth of consumption (g) is not equal to the rate of growth of expectations (g_e): The elasticity of marginal utility (q) is negative, and economic literature suggests that its absolute value is greater than one. In Equation 6, therefore, the component corresponding to the rate of consumption growth will be positive and the second component corresponding to the rate of growth of expectations will be negative. Thus, in all the cases the SRTP will be less than the SRTP suggested by the CGBU framework. If the rate of growth of expectations is higher than (q/(1+q)) times the rate of growth of consumption, the SRTP will be negative. For example, if we assume the same value of q (-1.5), the SRTP will be equal to 1.5g - 0.5g_e). Now if the rate of growth of consumption is 1%, and the rate of growth of expectations is 2%, the SRTP will be equal to 0.5%, and if the rate of growth of expectations is greater than 3%, the SRTP will be negative. Similarly, the SRTP will be negative if g_e>2.67g for q = -1.6, g_e>2.15g for q = -1.8, and g_e>2g for q = -2.0. Hence, as the elasticity of marginal utility increases the SRTP becomes negative at the lower ratio of the growth of expectations to the growth of consumption (g_e/g).

The discussion of the inclusion of expectations in the utility function demonstrates that the SRTP suggested by the CGBU framework is not an appropriate discount factor. Our results, however, are also subject to the form of the utility function - CRRA - being used. The selection of other utility functions such as the Constant Absolute Risk Aversion utility function will result in a different rate of discount. Hence, even the inclusion of expectations of economic agents will not result in a universally appropriate single rate of discount. In addition, the inclusion of expectations in the utility function cannot account for the nature of returns from economic goods, as discussed in the following section.

3.3 Insensitivity of the Framework to the Nature of Returns from Economic Goods: In the CGBU framework, and even in the above-proposed formulation that incorporates expectations, the rate of time preference is conceptualised as the marginal rate of substitution between one-time current consumption and one-time future consumption. But, the generic form of individual's behaviour involves a person with a total amount of resources or gratification (rewards and punishments) that are potentially available for him/her within a given time-period. Hence, an individual’s decision problem is to allocate those resources or gratification over a given amount of time or series of time periods in the future such that the outcome is most pleasing (Jungermann, 1988). In this formulation, the decision problem is to choose between alternative sequences of outcomes and not to choose between one-point outcomes. As soon as an inter-temporal trade-off is embedded in the context of alternative sequences of outcomes, the psychological perspective or "frame" shifts, and individuals become more farsighted, usually wishing to postpone the better outcome to the end that reflects a negative rate of time preference (Loewenstein and Prelec, 1991). Several empirical studies by economists have confirmed a negative rate of time preference for a sequence of outcomes. For example, many life-cycle studies reveal that consumption tends to increase over time until retirement, and in the absence of binding borrowing constraints, this pattern is consistent with the life-cycle theory only if people have a negative rate of time preference (Courant et al, 1986). The preference by the majority of tax payers in Canada and the USA for returns from revenue agencies and the preference of school teachers for their annual pay in twelve monthly instalments (September-August) instead of nine (September-June) confirm the negative rate of time preference for a sequence of returns (Loewenstein and Thaler, 1989). In a study by Varey and Kahneman (1990), subjects preferred brief sequences of decreasing discomfort even at the cost of experiencing an overall greater discomfort. Similarly, in a study of museum visitors by Loewenstein and Scherman (1991), a majority of visitors indicated a preference for increasing wage profiles over those that are flat or declining over time (holding total value constant).

We also conducted a survey among graduate students at the Faculty of Forestry, University of Toronto. Forty students participated in the survey. The survey included two questions on food items and one on shelter facilities. In the first question on food items, respondents were asked to choose preferred combinations of three meals (breakfast, lunch, and dinner) - one, two, or all three meals on a single day - over a period of thirty days such that the total number of meals should not exceed thirty. Seventy percent of respondents preferred two meals per day for the last fifteen days, and ten percent of respondents preferred three meals per day for the last ten days. In the second question on food items, the number of meals was increased to ninety and the total period was of three months. Seventy percent of respondents preferred one meal for all ninety days, and ten percent of respondent preferred a daily two meals for the last forty-five days. In the question on shelter facilities, ninety percent of respondents preferred accommodation in a condominium for their whole life compared with five years of accommodation in a one-star hotel, or one year of accommodation in a three-star hotel, or one month of accommodation in a five star hotel. The results indicate that once the choice set consists of sequences of returns, then respondents prefer either a sequence with uniform returns over longer periods (ninety days with food items and a whole life with shelter facilities) or better returns towards the end of the sequence. These results suggest the existence of a negative rate of time preference in a sequence of returns.

In an individual’s life, normally, the members of his/her long-term choice set will be sequences of returns while the members of short-term choice set may be one-time returns. It, therefore, seems reasonable to believe that the SRTP, calculated by the modified CGBU framework in Section 3.2, will be appropriate only for short-term decisions that involve one-time returns. The use of SRTP, calculated by the CGBU or the modified CGBU framework, in long-term decisions that involve a sequence of returns would be an error, and the results obtained would be erroneous.

4.0 Specific Limitations of the CGBU Framework

In addition to these three general limitations, two features of the CGBU framework - commoditization of ecosystem capital and undifferentiated needs and gross substitution - lead to inappropriateness of the CGBU discounting framework for ecosystem capital.

4.1 Commoditization of Ecosystem Capital: The utility framework of neo-classical economics is based on demarcation of certain aspects of reality as commodities. Polanyi (1944, pp.72-73) argued that commoditization is entirely arbitrary, and termed it as commodity fiction. Commoditization results in drawing conceptual and definitional, though arbitrary, boundaries around objects. Hence, commoditization converts systems into disaggregated and discrete units, these units are treated as separable, and it is assumed that there are no functional relationships between different discrete units of a commodity or units of different commodities. These discrete units can be transacted through exchange mechanisms, and the market exchange mechanism assigns a value equal to the market price to a discrete unit. In the economic arguments of valuation and discounting of ecosystem capital, an assumption that ecosystem capital is a ”commodity” is implicit, which is not obvious to many scholars, specifically ecologists or ecological economists, other than neo-classical economists.

The ecosystem capital is valuable to human society not only for the products, which may be thought of as commodities, that it contributes to the economic system but also for its functional contributions to the well-being of humanity. Regulation, production, carrier, and information have been identified as functions of ecosystem capital (De Groot, 1994). Examples of the regulation functions are the regulation of climate and energy balances, and the storage and recycling of organic matter, nutrients, and human waste. Production of oxygen, water, food, fuel, fodder, and timber are examples of the production function. Similarly, the provision of space and a suitable substrate for human habitation, recreation and tourism, and energy conversion are examples of the carrier function of ecosystem capital while aesthetic, spiritual, religious, historic, and scientific information, and cultural and artistic inspiration are examples of the information function (Prugh et al., 1995, p.57). Thus, most of the contributions of ecosystem capital are derived by keeping its different components working in their existing functional relationship as a fully functionalised system, and not separately from disjointed discrete units of the ecosystem. In a fully functionalised system, each part is as “valuable” as the whole, and hence the value of any single component cannot be understood separately from its contribution to the whole. The value of many components of ecosystem capital is derived from the very act of keeping them working in their existing functional relation (Vatn and Bromley, 1995). But, neo-classical economists have commoditized ecosystem capital, and this has led them to suggest that the total economic value of ecosystem capital is composed of five different, separable, components of use values and non-use values: (i) commercial use value; (ii) onsite recreation value; (iii) an option value resulting from maintaining the potential to visit the resource in future; (iv) an existence value derived simply from knowing the resource exists; and (v) a bequest value derived by individuals from knowing that future generations will be able to enjoy the existence or use of a resource (Pearce and Turner, 1990; Loomis et al., 1991, p. 412-413; and Prugh et al., 1995). However, none of these components is directed specifically toward the functional value of the ecosystem capital. In addition, all the five components of values need not be totally separable as proposed by neo-classical economists, and the functional value will be overlapping over all other values. Hence, economic values of ecosystem capital obtained through the process of commoditization are totally arbitrary.

Neo-classical economists defend commoditization of ecosystem capital in the name of analytical sufficiency and objectivity that are guaranteed by converting a real world into a sum of atomistic agents that are acting on atomistic objects. Hence, holism of an ecosystem is discarded in the name of objectivity. However, the greatest degree of subjectivity is introduced by converting an integrated system into discrete atomistic units, and this subjectivity is totally overlooked by these economists. It can, thus, be seen that the problem in neo-classical analysis of the ecosystem is not related to undervaluing certain factors of ecosystem capital, but it is of reducing entire problems to terms that misstate their underlying structure (Tribe, 1972, p.97). As a result, the process of commoditization leads to three main problems with respect to discounting of the ecosystem capital.

First, the concept of commoditization itself is contrary to the concept of a system (ecosystem) or holism, and is a contrived artefact. It overlooks the functional aspects and functional values of ecosystem capital. Discounting by a factor based on consumption growth implies that the value of many functions such as climate regulation, storage of organic matter, and nutrients etc will reduce to almost nothing after a few centuries. This is contrary to what many international conventions such as the Convention on Climate Change, and the Biodiversity Convention seem to indicate, according to these conventions society is placing higher values on future returns from the functional aspects of ecosystem capital.

Second, many attributes of ecosystem capital such as life and the quality of life of wild animals and birds are dominated by moral dimensions (Opaluch and Segerson, 1989; Stevens et al., 1991; and Kahneman and Knetsch, 1992). These are the attributes where social norms will restrict or reject commoditization. Similarly aesthetic, spiritual, religious, and cultural attributes are not subject to commoditization. The non-commoditized attributes will be along the dimensions that are orthogonal to the dimensions of commoditized attributes such as consumption items - timber and non-timber products. This orthogonal attachment - incongruous nature - will restrict the aggregation of all attributes to a single economic (monetary) measure. Thus, the use of a discounting factor (SRTP), based on diminishing marginal utility of only consumption attributes, to all attributes including those that are orthogonal to consumption attributes of the ecosystem capital will be arbitrary, and thus will not improve resource allocation by any standard.

Third, commoditization either neglects continuous returns or converts them into one-time returns. For example, functional aspects and functional values of the ecosystem capital are of continuous nature, and similarly aesthetic, cultural, spiritual, and religious attributes are of continuous nature, but are ignored in the commoditization process. Returns from several consumption products from ecosystem are also a continuous stream instead of a one-time return. Hence, once the continuous nature of ecosystem values is incorporated in the decision problems, outcomes will be much different than these suggested by the CGBU framework as discussed in Section 3.3.

4.2 Needs, Wants, and Gross Substitution: In the consumer theory of neo-classical economics, all goods and services are covered under a single-dimensional measure - utility, and, therefore, utilities derived from different goods and services are additive. Everything needed by a consumer corresponds to a want, and all needs can be subsumed into wants, and all wants are comparable. Although, ranking of wants is possible, the rank will depend solely on the utility derived from the want. In the end, one good can compensate totally for the loss of another or there is a gross substitutability of a good which meets the requirements of one want, and implicitly one need, by another good which meets the requirement of another want, and another need. Thus, according to the neo-classical consumer theory, there must be some price ratio which will induce a maximizing rational economic agent to forego (substitute totally) one good, say food items, by another good or service, say sex or aesthetic goods, whichever is comparatively cheaper. For neo-classical economists, there is no difference in the utility derived from food items and sex, and food can be totally substituted by sex or vice-versa, and it is known as the axiom of gross substitutability. Hence, three critical features of neoclassical consumer theory are undifferentiated wants, no distinction between needs and wants leading to undifferentiated needs, and gross substitution. However, some classical economists, post-Keynesian economists, socio-economists, and ecological economists have been sensitive to the absolute abstraction and gross substitution in the utility bundle of consumers.

Karl Menger proposed a list of needs, ranging from the essential to less essential: water, food, clothing, lodging, transportation, and tobacco (Lutz and Lux, 1979, p.18). Marshall identified a variety of wants, some being more basic than others (Lavoie, 1992, p. 65). Haines (1982, p. 111) proposed a hierarchy of needs: biological needs (food, clothing, shelter, variety); health, education, and security; friendship, affection, belonging, conformity with social customs; distinction, excellence, and morality. Haines hierarchy of needs is similar to Maslow’s pyramid of needs that is constituted by five sets of needs, from the basic to the highest: physiological needs (air, water, food, sex, sleep); safety needs (health, education, shelter, stability, protection); social needs (subdivided into two sets: belongingness and love needs, and self-respect and the esteem of others); and finally self-actualisation; that is, the search for truth (Lutz and Lux, 1979, p.11). Recent researchers, such as Lee et al. (1987, p.146) and Chambers et al. (1989, p.13), have clustered the needs in either two levels (lower level or higher level) or three levels (subsistence, security, and self-respect). Irrespective of the actual number of levels, it seems well established that, outside the boundaries of neo-classical economics, there is a hierarchy of needs, and individuals respond to differentiated needs, rather than to un-differentiated wants.

The famous rat experiment, reported by Kagel et al. (1975), has been used by post-Keynesians to demonstrate differentiated needs. In this experiment, rats had to pay, in terms of pressing a lever a certain number of times, to obtain drinks or solid food. In the experiment, the price of food relative to water was increased four times, the rats refused to drink more and eat less (Lutz and Lux, 1979, p.69). It shows that drinking and eating are two different needs which cannot compensate for each other, and hence food items cannot totally substitute drinks. However, various types of foods – Chinese, French, Indian, Italian, or Thai - represent wants specific to the need of food, they can be substituted for one another (Lavoie 1992, p. 67). Hence, needs and wants are different, and there is a hierarchy of needs, some needs are more basic than others, and basic needs should be fulfilled first. Needs are the main driving force behind consumer behaviour, while wants evolve from needs and constitute “ the various preferences within a common category or level of need” (Lutz and Lux, 1979, p.21). Wants may lend themselves to a quantitative measurement while needs are of qualitative nature. Thus, needs lack commensurability, and there is no single dimensional measure of the utility that these irreducible needs can generate (Lavoie, 1992, p.67). Further, there may not be any substitution between different types of needs, while there may be some substitution between some sub-needs, and substitution is possible between different wants related to the same need.

The three pillars of neoclassical consumer theory – undifferentiated wants, no distinction between needs and wants leading to undifferentiated needs, and gross substitution – are also the backbone of the CGBU framework of discounting. It is by not questioning these pillars that Arrow et al. (1995) and others, who argue that as long as consumption growth is positive, ecosystem capital should be discounted at a positive rate of discount determined by the rate of consumption growth, missed the point that different goods and services provided by the ecosystem capital may be meeting the requirements of needs at different levels of hierarchy. For example, the needs that are being satisfied by use values will be either physiological needs or safety needs of Maslow’s pyramid (lower-level needs), while the needs satisfied by existence value and option value will be the social needs of the pyramid (higher-level needs). The utilities, then, derived from different values of ecosystem capital for different levels of needs cannot be aggregated into a single number. Similarly, utilities derived from other forms of capital – man made capital - for lower-level needs (physiological or safety needs) cannot be aggregated with the utilities for a higher level needs derived from ecosystem capital. Therefore, there cannot be a gross substitution between ecosystem capital and man-made capital[9], and a discounting factor based only on man-made capital cannot be used to discount all the values of ecosystem capital.

In addition, the basic element behind discounting – diminishing marginal utility – is meaningful for the variety of wants and not to different needs. According to the hierarchy of needs, the basic needs should be fulfilled on a priority basis. Once the basic needs are fulfilled, individuals will put more emphasis on a higher level needs. Growth in consumption does not mean that the value of a higher level needs in the future would be less, but, in fact, it will be more. It is well established that with growth in consumption people place more emphasis on higher-level needs such as stewardship of natural resources, protection of wildlife, pollution-free or a less polluted environment, and sustainable management of forests etc. Therefore, these needs should not be discounted. Similarly, even at the lower-level importance of some needs such as forest-based outdoor recreation increases with consumption growth. It was predicted, in 1972, that recreation in the UK Forestry Commission's forests would increase at 10% per year for 20 years, and then at 3% per year indefinitely thereafter (Treasury, 1972, quoted in Price, 1999). Similar trends of forest-based recreation are observed in other developed countries. Discounting of forest recreation at the rate suggested by the CGBU framework will mean less value for future forest recreation. But then, why will people choose more visits to forests if they assigned less value that too probably at a greater cost? This means that the CGBU based discounting factor is not an appropriate factor to discount ecosystem capital.

5. International Security: Military Decisions versus Ecosystem Decisions

The emerging concept of international security is founded on a new philosophy encompassing a common institutionalised system of standards and shared values rather than concepts based on the balance of power (SIPRI, 1998). It is now commonly understood that security comprises much more than military security including ecological and environmental security. Numerous global conventions, in the last decade, such as the Agenda 21 1992, the Convention on Climate Change 1992, the Convention on Biodiversity 1992, and the Kyoto Protocol 1997 provide evidence for the extended international security system. The majority of nations, however, still places a high priority on the military dimension of security. In 1999, the total world military expenditure was roughly $780 billion (current dollars) and represents 2.6% of the world GNP (SIPRI, 2000). In some countries, military expenditure is as high as 15% of GNP. Globally, it is highly concentrated in a few countries. The 15 major spenders accounted for 80 percent of the world total in 1999; the USA accounted for 36 percent, followed by Japan and France with 7 percent each and Germany and the UK with 5 and 4 percent, respectively (SIPRI, 2000). Military decisions are strategic decisions, and are normally not subject to discounting or conventional benefit-cost analysis. But, lexicographic ordering plays a critical role in these decisions, and the decision process used is termed as non-compensatory filtering process by Earl (1986, p.183). Several forms of lexicographic ordering process, which can be used for these decisions, include naïve lexicographic process, behavioural lexicographic process (lexicographic ordering with saturation point, lexicographic ordering with indifference curves beyond threshold) (Earl, 1986, p. 233), and lexicographic ordering with Georgescu-Roegen's quasi-indifference curves (Georgescu-Roegen, 1954 and Encarnacion, 1964).

Military decisions are targeted to save humans from the arms and weapons created by other humans, and heavy military expenditures are the outcome of the desire of some human-beings to dominate others. No doubt that human-beings have proved themselves to be dangerous for other humans. Nature, however, can prove to be even more dangerous to mankind. The dangerous effects of several outcomes of destructive processes, such as depletion of the ozone layer, global warming, desertification, and deforestation are evident in the present world. Hence, ecosystem decisions are the decisions for international security and the security of humanity, and these should be treated at par with military decisions, which means ecosystem decisions have to be made without prices and discounting factors that have been fixed arbitrarily by neo-classical methods.

6. Summary and Conclusions

The choices of ecosystem capital are complex and problematic precisely because these entail systems (holistic) aspects that defy reduction to the venerable fiction of commodities and gross substitution across differentiated needs. There are no simple answers for the social rate of discount of ecosystem capital or for the inter-temporal allocation of ecosystem capital. These results may be frustrating for those who seek simple answers, but such are not to be found. The decision on the appropriate rate of discount or allocation of ecosystem capital would entail judgments concerning the relevant context and constraints. Hence, it does not mean that choices that reject the commodity fiction and gross substitution among differentiated needs are ill-informed, inconsistent, or not in the interest of efficiency. The neo-classical framework, however, is not broad enough to accommodate the decision process of such choices, and decision makers have to search for appropriate decision tools in other streams of economics.

Socio-economists, specifically Etzioni (1988), extended the boundaries of the neo-classical consumer theory in many dimensions. First, single utility is replaced by the two irreducible utilities –pleasure and moral values. Second, instead of maximizing their pleasure, people try to balance the service for two major purposes – to advance their well-being and to act morally. Third, rational assumption is replaced by the assumption that people select means, not just goals, first and foremost on the basis of their values and emotions. However, this extension has fallen short of putting up a full-fledged socio-economic theory of consumer behaviour. Similarly ecological economists, such as Herman Daly, Robert Costanza and Richard Norgaard, introduced many useful concepts such as complementarity between ecosystem capital and man-made capital, ecosystems as sustainable systems, green GDP, and ecological taxes etc., but there is no full-fledged consumer choice theory in ecological economics either. However, post-Keynesian economics has a full consumer choice theory, equivalent to the neo-classical consumer choice theory. Post-Keynesian consumer theory is based on six principles - the principles of procedural rationality, satiable wants, non-independence, subordination of needs, irreducibility of needs, and growth of needs (Lavoie, 1992, p.63). Procedural rationality consists of means to avoid complex calculations and considerations, and of procedures of decision-making even in the absence of full or accurate information. Some of these procedures are rules and habits. The principle of satiation means that eventually more of one good or more of a characteristic will bring less supplementary satisfaction. The satiation, in post-Keynesian economics, often arises with positive prices and finite income while in neo-classical economics, it arises only if prices are null or incomes are infinite. The principle of non-independence means that decisions and preferences are not made independently of those of other agents. The principle of the subordination of needs is related to hierarchy of needs. The principle of irreducibility states that some needs cannot be substituted for others, and as a consequence, everything does not have a price. The principle of the growth of needs means that the growth of real incomes will induce new needs.

In summary, post-Keynesian consumer theory is founded on the separation between needs and wants, a hierarchy of needs, the absence of gross substitution but the possibility of some substitution between wants corresponding to the same need, lexicographic ordering of wants, and possibility of choices without prices – by habits and rules. Hence, post-Keynesian consumer theory seems to have some tools that could be used by decision-makers to analyze the choices of ecosystem capital.

Literature Cited

Arnold, Thomsen.: 1943, 'Expectations in relation to achievement and happiness', Journal of Abnormal Psychology 38, 58-73.

Arrow, K. J., W. R. Cline, K-G. Maler, M. Munasinghe, and J. E. Stiglitz.: 1995, ‘Intertemporal equity and discounting’,. in Mohan Munasinghe (ed.) Global Climate Change: Economic and Policy Issues, Washington, DC, The World Bank, pp. 1-32.

Arrow, K. J., W. R. Cline, K-G. maler, M. Munasinghe, R, Squitieri, and J. E. Stiglitz. 1996, 'Intertemoral equaity, disocunting, and economic efficiency', in James P Bruce, Hoesung Lee and Erik F Haites (eds) Climate Change 1995: Economic and Social Dimensions of Climate Change, Cambridge, Cambridge University Press, pp. 128-144.

Azar, C. and T Sterner.: 1996, 'Discounting and distributional considerations in the context of global warming', Ecological Economics 19, 169-184.

Becker, G. S. and C. B. Mulligan.: 1997, 'The endogenous determination of time preference', Quarterly Journal of Economics CXII(3), 728-758.

Bishop, R. C. and T. A. Heberlein.: 1979, 'Measuring values of extramarket goods: are indirect measures biased?' American Journal of Agricultural Economics, 61, 926-930.

Bromley, D. W.: 1989, Economic Interests and Institutions: The Conceptual Foundations of Public Policy, Oxford, Basil Blackwell.

Broome, J.: 1992, Counting the Costs of Global Warming, Cambridge, MA, White Horse Press.

Campbell, A., P. E. Converse and W. L. Rodgers.: 1976, The Quality of American Life. Perceptions, Evaluations, and Satisfactions, New York, Russell Sage Foundation. 583 p.

Chambers, R., N. C. Saxena and T. Shah.: 1989. To the Hands of the Poor: Water and Trees, New Delhi, Oxford and IBH Publishing Company.

Cline, W. R.: 1992, The Economics of Global Warming, Washington, DC, Institute for International Economics.

Costanza, R. and H. E. Daly.: 1992. 'Natural capital and sustainable development.' Conservation Biology 6(1), 37-36.

Courant, Paul, Edward Gramlich and John Laitner.: 1986, ‘A dynamic micro estimate of life-cycle model’,. in Aaron, Henry G., and Gary Burtless (eds.) Retirement and Economic Behaviour, Washington DC, Brookings Institution.

Dally, G. C., T Soderqvist, S Anlyar, K Arrow, P Dasgupta, P. R Ehrlich, C Folke, A Jansson, B Jansson, N Kautsky, S Levin, J Lubchenco, K Maler, D Simpson, D Starrett, D Tilman and B Walker. 2000, 'The value of nature and the nature of value', Science 289, 21 July, 2000, pp.395-396.

Dasgupta, P., S. A. Marglin and A. K. Sen.: 1972, Guidelines for Project Evaluation, New York, UNIDO.