This paper examines the validity of the use of existence value in cost-benefit analysis, blending together tools of economic theory and metaethical value analysis.

University Degree Business and Administrative studies

Project Analysis Aly-Khan Hirani

154-525 110248412

11th April 2003

Existence Value

The Theoretical Case for and Against its Use in Cost Benefit Analysis

11th April 2003

Abstract

This paper examines the validity of the use of existence value in cost-benefit analysis, blending together tools of economic theory and metaethical value analysis. The technical issues of definition and the theoretical justification for using the contingent valuation method is discussed. This is followed by a consideration of what existence value is trying to categorize, and whether the ‘welfarist’ approach embedded in neoclassical welfare analysis is an appropriate framework for understanding such values. Finally, arguments for and against the use of existence value in cont-benefit analysis are assessed, concluding in a ‘third way’ that accepts the it as a valid category of values to be considered in the project evaluation process, but asserts that the total valuation framework is not an appropriate method for assessing existence value for both economic and metaethical reasons.

Aly-Khan Hirani

McGill University

110248412

Course 154-525

Professor R. Cairns

1. Introduction: the current debate

“Whether or not nonuse values fit neatly into some regimented, atheoretic, cost-benefit manual is irrelevant in a sense. The fact is that nonuse values are consistent with neoclassical welfare theory and this theory provides a solid foundation upon which to help evaluate the desirability of many public policies”

R. J. Kopp

“Calculations of existence value in essence seek to employ formal economic methods to resolve matters of cultural symbolism and social ideology… it is a mistake for economists now to try to resolve such basic social disagreements in a quantitative way through cost-benefit and other formal analytic methods.”

D. H. Rosenthal and R. H. Nelson

The concept of existence value was first raised in one of the most influential papers in environmental economics by John Krutilla (1967), where he advocated a move away from the Pigouvian approach to conservation of natural resources in terms of the optimal intertemporal utilization of natural resource stocks to an appreciation of several unique features of such resources. These included non-renewability, the possibility of option value, and existence, or nonuse, value. Krutilla was advocating nothing less than pushing back the boundaries of the discipline: economists should consider more than just the valuation of common tangibles. The controversy that this assertion caused remains today, and divides most economists into two camps. Those such as Rosenthal and Nelson suggest that existence values are not to be treated merely as another type of good, as they represent something quite different, while those such as Kopp believe that what existence value is is not as important as the fact that we can envision it as consistent with neoclassical theory, and therefore are warranted in measuring it in a total valuation framework.

To assess the validity of these competing claims, the appropriate domain of discussion requires more than just the tools of economic analysis. As Portney (1994) writes: “…the critical scrutiny directed at the contingent valuation method has led some economists to think more deeply about cognitive processes, rationality, and the nature of preferences for all goods, public or private”. This paper is primarily concerned with the last area that Portney identifies: the nature of preferences and whether existence value generates preferences that can validly be compared with preferences for more conventional economic goods. Preferences are derived from the relative valuations that individual agents hold for various goods. This takes our discussion into the domain of value judgment: a subject more commonly attributed to metaethics. This paper blends the approaches of analysis of both economic theory and metaethics for a definite purpose: while this is an economic issue, discussions of motivation, agency and value are more fruitfully analyzed using a combination of tools from both disciplines.

Krutilla suggested that a person’s wellbeing is not merely dependent on the physical goods and services they consume, but also on the various states of knowledge they have of the existence of social and physical characteristics of the world. This position was subsequently developed: implicitly, consumers should be willing to pay something for this form of consumption, so there must be some way of measuring these values in dollar terms. More precisely, modern definitions of existence value (such as Kopp, 1992) suggest that existence value is really about the enjoyment of a nonconsumptive service without engaging in any observable behavior. Nonconsumptive services, which yield nonuse value, are non-excludable, non-rivalrous, and often can be enjoyed without any monetary expenditure. Hence existence or nonuse values can be considered values for pure public goods. Since economists such as Samuelson have believed that public goods should be treated in the same welfare theoretic structure as private goods, it is reasonable to believe that a nonconsumptive service that is provided by a natural resource can be treated in the same welfare framework, giving rise to the motivation to measure existence value in dollar terms using some form of compensating or equivalent variation measure. If this is true, then we do not need to rethink our descriptions of consumption to justify existence values, and instead can focus on how to describe the public good services underlying nonuse values. Amongst the majority economists that practice cost-benefit analysis, this is the position held today.

Initially, little attention was paid to the development of the concept in the early 1970s. However, as potential uses grew and the policy stakes rose, a debate has arisen in the economics professions as to the appropriateness of measuring existence value and whether it is possible to measure it. More recently, the Exxon Corporation, faced with large potential damage assessments as a result of the Exxon Valdez oil spill of the coast of Alaska and concerned that a damage assessment grounded in the possible existence values across various states of nature for the Prince William Sound region, hired a group of leading economists to assess the appropriateness of including existence value. Their evaluation was overall negative, while the State of Alaska and the U.S. Federal government’s group of economists took a more positive view. Shortly afterwards, the National Oceanographic and Atmospheric Association (NOAA) convened a panel of leading economists, including Kenneth Arrow and Robert Solow, to review the issue. The panel ended in a compromise: although great care should be taken to avoid misuse, existence value should be included among the tools available to government analysts (Arrow et al. 1993). The issue remains unresolved, and debate continues.

This paper provides an overview of the issue of existence value and the validity of its inclusion in cost-benefit analysis. A consensus is emerging along the lines of the NOAA position. This paper challenges that position using techniques both from economic and meta-ethical theory.

The aim here is not to demonstrate that existence value is not validly measured in the total valuation method, but rather it attempts to change the terms of the debate by showing that the confidence placed in the dominant methodology is misplaced. There is a belief implicit in the notion of total valuation that the object of cost-benefit analysis should be to provide a complete measure of all the costs and benefits of any given proposal. It shall be argued that this is not the case, that existence value supervenes on ethical beliefs and that therefore existence value cannot be evaluated in the inherently ‘welfarist’ approach to be found in the neoclassical view that treats individuals as rational egoists. Nevertheless, existence value exists, and it should be taken into consideration. It shall be shown that the economic understanding of value is instrumental: goods have value inasmuch as individuals attribute value to them. Instrumentality of value inherently requires that prices are relative prices: if a good has no inherent value, it is nonsensical to suggest that it has a price that can be defined abstracted from other goods is nonsensical. If it can be shown that existence values can be inherent values, then it is possible that existence values for goods are noncomparable and therefore not coherently expressed as relative values, and therefore it is possible that using the total valuation method may be inappropriate. The onus is therefore on the advocates of using existence value to show that this is not the case.

There are both technical problems of measurement and substantive problems of the genus of values that can be adequately be assessed in a ‘welfarist’ framework such as economics. In section two, technical issues are dealt with: the appropriate definition of existence value, methods of measurement (i.e. contingent valuation methods – CVM), problems with CVM and the possibility of using indirect methods to estimate existence value are discussed. In section three, the meta-ethical side issues of the importance of motive on the assessment of existence value, the validity of psychological egoism implicit in neoclassical economics are discussed and the case against considering existence value as a conventional pure public good is made.

This paper is directed towards the informed economist, rather than the philosopher, and the mode of analysis shall be skewed accordingly. For brevity, a working knowledge of cost-benefit analysis, and in particular the Contingent Valuation Method (CVM) is assumed, as is a good knowledge of welfare economics, in particular Hicksian and Marshallian demands and Arrow’s Impossibility Theorem. Conversely, the metaethical analysis should be accessible to the general interested reader: no undefined terminology or symbolic logic is used.

2. Definitions and Measurement of Existence Value

2.1 A Definition of Existence Value

Economic theory is inclusive about possible conceptions of value or benefits: what counts as a benefit is not delimited by the act of use or by the presence of a commercial transaction. Instead, an action has benefits or value in the economic sense if that action increases the availability of something that is scarce at the margin, given that that ‘something’ is desired by someone. In other words, so long as something could conceivably be expressed as an argument in a utility function, it can be considered an economic benefit. In order to define existence value, we therefore need to show how existence value enters an individual’s utility function.

The most common route (e.g. V. Kerry Smith, 1993; D. M. Larson, 1993; McConnell, 1983) starts from the notion of total value. Total economic value (TV) is uniquely defined, and decomposes into existence value and various categories of existence value.

Consider an individual with the utility function:

[1]

where x is environmental services, Q is the state of the environment, z is a vector of ordinary goods and services. The subscripts denote: e, existence; s, site experience; 1, activity 1 at the site (i.e. a type of use value); and 2, activity 2.

The solution to the problem:

[2]

s.t.

is the expenditure function:

[3]

Here the price vector p is taken as exogenous, and is therefore implicit in [3]. We can now define TV in terms of Hicksian compensating variation:

[4]

p* represents the choke price, po is the baseline price, and Qo is the baseline level of resource quality. Since existence values are unpriced, p*e is a shadow price (the price that would be efficient where the efficient price is absent) so that a virtual market in the existence value would clear if efficient markets existed.

Equation [4] is a one-shot, holistic measure of total value that does not disaggregate its different components. ...

This is a preview of the whole essay

is the expenditure function:

[3]

Here the price vector p is taken as exogenous, and is therefore implicit in [3]. We can now define TV in terms of Hicksian compensating variation:

[4]

Equation [4] is a one-shot, holistic measure of total value that does not disaggregate its different components. If we want to isolate each component, this can be done as:

[5a]

[5b]

[5c]

[5d]

where line [5a] captures existence value, [5b] defines site experience value given existence value and lines [5c] and [5d] define the use values. Here, total value is the sum of existence and use values evaluated sequentially. The conventional view that existence value is some residual between use value and total value is true on the condition that the components are evaluated in a valid sequence.

A common error is to evaluate total value by aggregating components estimated independently, termed independent valuation and summation (IVS). The IVS method is analogous to [5]:

[6a]

[6b]

[6c]

[6d]

The existence value and each use value is evaluated independently. This is not consistent with [5]. In general, IVS is not equivalent to TV, which is to be expected, since the consumer surplus value of a bundle of goods is generally not the same as the sum of consumer surpluses of individual goods. Further, TV is unique, whether evaluated holistically or sequentially. IVS systematically overestimates the benefits of a policy change, and non-beneficial policy changes may be misidentified as beneficial. The problem is caused by the fact that an IVS estimation procedure does not take proper account of resource scarcity and the interactions (substitution and complementarity) among different kinds of resource services.

An important problem is that the value of each component is not unique, but depends on the position in the valuation sequence. If the sequencing matters, then we should establish a protocol for the appropriate ordering of all values, including existence value. While most studies consider existence value to be something that should be added at the end (possibly as a misinterpretation of the notion of a “warm glow” effect), some, such as Randall (1991) take it as self-evident that existence value should be the first item in the sequence. Establishing a protocol is not the central aim of this paper, but inasmuch as existence value is considered the residual between use value and total value, it seems that there is a prima facie case for including it as the first in the sequence; this view is buttressed by the notion that existence value involves knowledge of a state of the world, and that use (in many cases) requires that the agent has preexisting knowledge of the good they are using, hence existence value is at least prior to use, and may be considered a precondition to use value. Whether or not these arguments are used, it may be prudent to estimate existence value first and then last in the sequence to establish an upper and lower bound for its value.

2.2 Non-observability of preferences and measuring existence value

The utility function in [1] is just one way of defining existence value, and it does not capture the notion of existence value as a public good. A special structure of preferences – weak separability of the environmental good from the market goods) gives rise to what has been termed “pure existence value” and a special case of pure existence value arising from strong separability has been termed the “hopeless case” (Larson 1993). When the preferences take these forms, the environmental good leaves no trace in a Marshallian demand system, so that existence value cannot be observed indirectly from market demands.

The result of this problem is that, given that there is no possibility for estimating existence values which, by definition, leave no trace, the only way to them is through direct methods such as CVM.

The match between the total valuation framework and CVM is close: CVM has the flexibility to value a wide range of plausible, constructed policy scenarios, and CVM scenarios constructed to implement a total value approach can, in theory, be as plausible as the total valuation framework.

Strongly separable preferences: the ‘hopeless case’

The use of strong separability attempts to capture the intuition that the values people hold for environmental goods and services may not be reflected in their demands for market goods and services. Let E denote the public environmental good, such as environmental quality (which might, but does not necessarily, correspond to xe in the previous section). Let E be strongly separable in preferences from all market goods x. We can represent this in a utility function of the form:

[7]

where U’, g’, and fi’, I=1,…,k are all positive. From inspection it is clear that the marginal rates of substitution between market goods are independent of E, so that E does not appear in Marshallian demand functions.

Therefore, there is no foundation for an inference of a change in welfare with respect to changes in E from market behavior, since there is no change in the observable demand functions that represent individual behavior.

This yields the property that market goods, taken collectively, are inessential. Consider the indirect utility function:

[7’]

where the optimal Marshallian demands from maximizing [7] subject to the constraint m = px have been substituted in. Nonessentiality of market goods means that there exists an Ê for any m such that:

[8]

Therefore there is a level of Ê that substitutes for m = 0 provided that g(E) is unbounded from above. The strong separability of [7] implies that utility from money and utility from environmental quality are perfect substitutes.

Nonessentiality is clearly not a realistic assumption. It turns out that assuming that market goods are essential (a much more reasonable proposition) still leaves the issue mathematically tractable. We assume that, market goods have collective essentiality, so that for a given level of utility, at some minimum level of income no amount of income can be given up as E increases to maintain utility:

[9]

so that for a given indifference curve:

[10]

where is the marginal utility of income. Equation [10] shows that for any change in the environmental good , the essentiality condition (that ) implies that , which in turn implies that , some constant. Therefore, the essentiality of market goods at some level rules out the strongly separable form of utility function in [7].

Weakly separable preferences: pure existence value

A more general case than the one above is provided by a utility function of the form:

[11]

where and . Just as above, the marginal rates of substitution between all pairs of market goods are independent of E, so that E is not an argument in any Marshallian demand functions.

The essentiality condition for a weakly separable utility function is:

[12]

This time varies with p and m as well as E: this means that weak separability of preferences does not rule out the essentiality of market goods. can be zero at and positive at higher levels of income.

2.3 Possibilities of indirect valuation of Existence Value

Although the possibility of utility functions that are separable in preferences yields the problem that existence value is not observable in Marshallian demand functions, it has been shown that introducing the plausible assumption of essentiality of market goods eliminates at least the ‘hopeless case’. However, a weakly separable utility function is compatible with essentiality. Does this mean that existence value will fail to leave any behavioral traces in practice?

The weakly separable utility function in [11] is a very restrictive specification: it suggests that E is separable from all market goods. If any of the markets are in the weakly separable group with E, then all market demands will depend on E, and there is a possibility of a behavioral trace in markets. All that is required, then, is the nonseparability of E from at least one market good in practice, and it does not seem to require a stretch of the imagination to conceive of such a state of affairs as obtaining.

We can examine the case of at least one complementary market good with E in the following manner. Using the equation from [7], we assume that good is complementary to E, which could be direct uses such as bird watching or fishing, or indirect uses such as the purchase of wildlife art. As E changes there is a change in use value, which can be assessed from the change in areas between Hicksian demands as E changes. Using the compensating variation measure of total value from section 1.1, we can measure the total value of a change from E0 to E1 to be:

[13]

we know that the Hicksian and Marshallian demand functions are related by the identity:

[14]

from which, by differentiating the first expression with respect to E, we can get the Slutsky-Hicks equation:

[15]

where are observable coefficients for the Marshallian demand function, and is the marginal change in expenditure. Adapting the notation from section 1, and letting p*1 be the choke price of good 1 (the complementary good) we can express the total value of a change in the level of E from E0 to E1 as:

[16]

By integrating back from the Marshallian demand function (see also footnote 1) we can recover the use value, as a measure of the change in value of x1 as E changes. However, we do not know ∂e/∂E, and we must therefore make some assumptions about preferences. However, there is a range of possible assumptions that can be made (see, for example, Mäler 1974, Bockstael and McConnell 1983). Here just one is investigated: the case of hicks-neutrality.

Good x1 is Hicks-neutral to E if , which means that demand for x1 is independent of the level of environmental quality. In other words, the expenditure function is strongly separable in E and the price of good 1, p1. This means that, for all prices p:

[17]

Clearly, we can observe the two derivatives on the right hand side, and we can therefore estimate ∂e/∂E by evaluating the two derivatives on the basis of observable demand coefficients.

Hicks-neutrality of preferences may offer a way out of the non-observability problem, but is it an attractive, or even defensible assumption in this context? Two points should be noted. Firstly, Hicks-neutrality may be attractive since the resulting formula in [17] identifies the willingness to pay solely for a change in E solely for existence value; since demand for x1 is independent of environmental quality , the Hicksian demand function does not shift on the basis of changes in E. The key point, however, is that this is just one possible assumption about the structure of preferences that could be employed, and if Hicks-neutrality is thought unjustifiable, there are other options available. Given the considerations above, it may be premature to conclude that we can only use CVM to estimate existence value, and that we cannot use market-based methods.

3. Existence Value and Competing Theories of Ethical Agency

3.1 Why Should Motivation Matter to Economists?

The initial insight into the possibility of existence value is commonly attributed to Krutilla, but he did not provide anything more than an analogy to make the point: some people may care about preservation of the wilderness even if they may not want to actually ever visit it themselves. This was part of a new framework for dealing with the value of natural resources; existence value served as just one of may of the different kinds of services that natural resources could provide to increase the well-being of individuals. Existence value was easily overlooked previously, because other types of more direct consumption involve observable behavior, but existence value does not.

This account was vague, and early discussion attempted to develop this concept in three distinct moves. The first was to establish whether or not existence value was a value in the economic sense at all. Economic theory takes the position that the ‘economic value’ of goods is not inherent to the goods themselves: goods have value as designated by economic agents. Value is a mind-world relation, where values are from agents for various goods. Goods do not possess intrinsic value. This is a notion more familiar to economists in the form ordinal non-comparability of preferences in Arrow’s Impossibility Theorem. The first, and possibly the boldest, conceptual move was to class existence value in the same way: the natural environment has value in the same mind-world relational sense. Natural resources, just like other goods, do not possess intrinsic value.

This move was the most important, and deserves more discussion. The assertion that goods do not possess intrinsic value is not so much a metaethical theory produced by economic theorists as much as it is a workable assumption that tried to abstract from the common philosophical debates about the nature of value. Goods possess value as given to them by individuals, and in this sense can be considered ‘instrumental’ value, as goods are instrumental in reaching the objective of utility maximization. Two points can be made at this point. Firstly, this is a good assumption for dealing with the most common subject matter of economics: it is hard to justify that goods such diamonds and pearls are good in any sense beyond their ability to give us utility. To be good in any sense that goes beyond this would be to be good in an ethical sense. Secondly, it is the instrumentality of value that gives rise to relative prices. The price of any good is its price in terms of (all) other goods. All economic goods have a propensity to yield utility, but are relatively different in their propensity to help us maximize utility. In the absence of an objective anchor for values of goods, prices must reflect the relative instrumental value of each good, relative to every other good. Different people have different tastes, and the neoclassical notion of value is amenable to such subjectivity: all individuals can have their own sets of relative prices. It is the fact that there is one goal that gives meaning to the notions of substitutability and complementarity amongst goods, and it is the fact that value is defined instrumentally in terms of satisfying this goal that gives rise to relative prices. It can be seen that this move, while bold, is necessary: to be able to measure existence value, we have to be able to put a price on it, and to do this, we have to assume that it is a non-inherent value.

The second move was to appreciate that, although existence value involves the same type of relation between agents and the world as use value, the relation is different in that it did not require use of a resource to obtain. This is reflected in the analytical treatment of existence value in section 2.1, where it is treated as an additional component of total value. This move, however, lacks substance: it provides a definition of what existence value is not, but does not fill out a description of what existence value is.

Finally, in response to this last problem, economists attempted to establish what the motives were underlying existence value: it was believed that, if we knew where these values came from, we might understand what existence value is. However, more recent views on this debate have suggested that the focus on motives such as bequest and stewardship are misguided (McConnell 1983; V. Kerry Smith 1993). If we can make the case that existence values arise from pure public good services provided by environmental goods, then existence value becomes an immediately tractable concept within the framework of neoclassical economics. We do not need to discover where the motivation for the attributions of existence values comes from or why.

The neoclassical view states that the economic value of an existence ‘good’ depends entirely on the satisfaction that the utility-maximizer gets from having it protected, preserved or enhanced: it is in this sense that environmental goods do not have intrinsic value. Economic man is an egoist, and the neoclassical interpretation of his motivations for all action, including those that seem to be for the benefit of others, is a form of psychological egoism. Psychological egoism (more commonly referred to in the economics literature as “welfarism”) is a description of human behavior that asserts that all action is motivated by self interest or personal welfare, and that apparently altruistic motives can be validly redescribed in terms of self-interested motives. For example, an agent’s existence value for the Grand Canyon, while apparently coming from altruistic motives for the benefit of those who may visit it and enjoy its natural beauty, can be redescribed in terms of wanting to enjoy the satisfaction of knowing that the Grand Canyon is there to be enjoyed. Hence, all motivation can be redescribed in terms of psychologically egoistic motivation, so that the entire discussion is a red herring: wherever existence value comes from, if it is a value that comes from agents for goods, it is compatible with the neoclassical view. The implication is that, since the value-relation between agents and ‘existence-goods’ is of the same kind as ‘use-goods’, then an existence good may be a substitute for other goods that are valued by utility-maximizing agents. By adjusting income to adjust the amount of material goods that can be bought, the agent can be compensated for changes in the level of an existence good, underlying the approach in section 2 of measuring existence value with Hicksian compensating and equivalent variation. This is an unpopular theory of agency, since it is commonly alleged that psychological egoism confuses motivation with consequences: even if we derive benefit from the consequences of our actions, this is not necessarily the reason why we acted.

Many economists might take a more subtle position than advocating strong psychological egoism as the best theory of ethical agency: there may be genuinely altruistic motives for acts, but that is not the domain of economic discussion. The neoclassical view of motivation is mathematically tractable and is compatible with observed behavior. The discussion is about human motivation, which is private, and so cannot be empirically tested. Therefore, psychological egoism is a closed theory: a theory that is non-verifiable and non-falsifiable. While this may indicate that it is a theory built on insecure foundations, this is a metaethical issue, not an economic one. Psychological egoism works, and is therefore, on pragmatic criteria, a valid assumption.

This view misses the central concern: by assuming that all values are reducible to the values of a psychological egoist, neoclassical economics can assert that goods are possibly substitutable, which underlies the Hicksian welfare measures used to define existence value. But if psychological egoism is not valid, and existence value refers to a different kind of agent-world relation, then the use of compensating and equivalence variation is invalid. Where this second position makes a false move is to contend that, whether or not strong psychological egoism is true, weak psychological egoism is true by definition, and since these two theories are virtually the same, then the matter requires no further discussion.

This move is invalid since where the two theories diverge is exactly the point at issue. Weak psychological egoism states that, although truly altruistic behavior is possible, whenever an agent acts, the act is the one they wanted to do at that point. If A supports the preservation of the Grand Canyon, it must be the case that A wants to support the preservation of the Grand Canyon. This theory is inherently problematic, as it is tautological (an agent is motivated by their motivation), but more importantly, it moves away from describing what the motivation generating the value-relation between agents and ‘existence-goods’ is, to what conditions they satisfy. This is a misinterpretation of the first insight that the value of an existence good is a relation between an agent and the existence good: an existence good is not valuable because it satisfies this condition; it satisfies this condition because it is valued. The subtlety of this distinction belies its importance: an adoption of a truism such as weak psychological egoism means that the neoclassical view is attempting to aggregate two potentially non-comparable things in a cost-benefit analysis, and Hicksian methods of measuring existence value may be unjustified. Where existence value comes from, and what motivates it, then becomes a crucial issue.

3.2 Technical Problems with Views Compatible with Psychological Egoism

The most popular way of reconciling the apparent pervasiveness of altruistic motivation and psychological egoism is to invoke some form of reciprocal altruism, where we act benevolently in the expectation that others will react benevolently towards us. Madariaga and McConnell (1987) show that this approach is not without its difficulties. Here they make the distinction between intrinsic and altruistic motives: intrinsic motives are ethical concerns about “how the world should be ordered”, while altruistic motives are subdivided into individualistic altruism (where agents gain value from the enhanced wellbeing of others) and paternalistic altruism (where agents gain value from the use of a particular good or service by others).

Individualistic altruism can be formalized in a two person, A and B, model of utility as follows:

[18] UA = UA(YA,UB(YB,E)

[19] UB = UB(YB,E)

where Ui and Yi are utility and income levels of person i. A unit increase in E yields existence value to person A when:

[20] (∂UA/∂UB).(∂UB/∂E) > 0

Any good that yields value to person B yields existence value to person A.

If the motive is paternalistic altruism, then A’s utility function in [18] can be rewritten as:

[21] UA = UA(YA,E)

so that

∂UA/∂E > 0 ∂UA/∂YB = 0

Now consider a project that increases E from E1 to E2 and costs C t be paid by the user, B. B’s surplus from the change SB is given by:

[22] UB(YB – SB, E2) = UB(YB,E1)

If we suppose that C > SB then cost-benefit analysis may not give the correct answer if it does not include the benefits to nonusers. We need to establish how much surplus A gets from the project when motivated by individualistic altruism. This is given by:

[23] UA(YA – SA, UB(YB – C, E2) = UA(YA, UB(YB, E1))

Since we know that the utility of B is greater if the project does not go ahead, since the costs outweigh the surplus, we know that A must also be compensated, since they will receive nonuse disutility from B’s loss in utility. Thus the aggregate benefits remain less than costs with the inclusion of existence value:

[24] SA + SB < C

Therefore, when individualistic altruism prevails and the user pays all the costs, adding in existence value does not change the cost benefit outcome. We can add a caveat that if there are N users, and A cares more about a restricted group in particular, then there may be a different outcome. For example, in an income distribution problem, if there are three agents and one agent has individualistic altruism for one of the other agents, then the optimum distribution involves giving the cared-for agent more income than the others to maximize aggregate utility. And as can be seen from inspection of [22], the traditional view – that existence value can make what appears to be a non-worthwhile project worthwhile – is possible if the motive is paternalistic altruism, provided that the existence value is large enough. Therefore, that there is reciprocal altruism is not enough: the type of altruism is also important.

3.3 Alternative Explanations of Existence Value and their Implications for Cost-Benefit Analysis

Section 3.1 does not propose that the metaethical position underlying neoclassical economics is wrong (although it is not a particularly popular one), it merely shows that the claims that motivation is not relevant to the discussion is incorrect. There are a wide variety of positions on the nature of human agency, and not all of them have detrimental effects for the account of existence value used in the neoclassical mode.

‘Humean’ altruism is the reciprocal altruism used by Madariaga and McConnell, and is the insight underlying the Folk Theorem in game theory that explains the common occurrence of norms of behavior that avoid prisoners’ dilemma-style suboptimal equilibria: seemingly altruistic behavior is explained through the redirection of enlightened self-interest. This view is compatible with the psychological egoism underlying rational choice theory, and therefore allows us to treat existence value as in section 2.

Other view of altruistic acts suggests that altruistic motives contain no self-referential interests: they are completely other regarding. If this is true, and if this is the type of motive that underlies existence value, then it would seem that they might not be accurately measured using Hicksian welfare measures that use self-referential measures of compensating and equivalent variation.

One approach of this type is Sen’s (1977, 1987), or Griffin’s (2001), where agents are said to hold-self regarding interests, but, since they recognize that other agents are agents as they themselves are, deserve that we hold some degree of interest in their goals and interests also. Interest in other agents is termed ‘commitment’ by Sen:

“…commitment does involve, in a very real sense, counterpreferential choice, destroying the crucial assumption that a chosen alternative must be better than (or at least as good as) the others for the person choosing it…[I]t drives a wedge between personal choice and personal welfare, and much of traditional economic theory relies on the identity of the two.”

(Sen 1977, 328-9)

This identity Sen refers to is generated by the adoption of psychological egoism above. On this account, an altruist is someone whose commitment to others is independent of self-interest, compensation and substitution. It is, as Sen notes, “closely related to one’s morals.” This is the view that drives many economists fears that attempting to measure existence value is an attempt to measure cultural symbolism and social ideology in dollar terms.

To conclude, it should be noted that this discussion does not attempt to show that psychological egoism cannot be the right description of human agency. Egoism is not inconsistent, and it is coherent. It may be right. But there are other accounts of human agency that also appear to be coherent and consistent, and provide a much more appealing description of human motivation. If morals matter to the extent suggested in theories such as Sen’s, then the theoretical basis for including minimum wiliness to pay as a measure of existence value is questionable at best. To include existence value in a total value assessment may be an attempt to aggregate apples and oranges.

Conclusion:

Should Existence Value Be Used in Cost-Benefit Analysis?

In the literature on the appropriateness of using existence values, there appears to be a general division between those who believe that, despite its difficulties as a legitimate economic concept, existence value should included in cost-benefit analyses and the smaller camp that argues that existence value elicits moral concerns that should not be included in our analysis. In this section, the arguments for both sides will be reviewed, but the case is made for a third view. It is true that the ethical considerations that are generated by concern for natural resources cannot be captured in existence value as defined in the neoclassical view, but these ethical issues must be taken into consideration. Therefore, as Quiggin (1993) argues, cost-benefit analysis “is inherently incomplete.”

The case for including existence value is normally motivated by reacting to points of contention made by those against its inclusion, of which there are three main strands. Firstly, it is alleged that existence value is not an economic value, leading to claims that they reference altruism (Madariaga and McConnell, 1987), social ideology and cultural symbolism (Rosenthal and Nelson, 1992) and moral values (Sen 1977, 1987; Edwards 1992). It has been argued in section 3 that there is no reason to accept that the neoclassical view of existence value is right a priori, but also that there is no sound argument against this position.

The second case suggests that the neoclassical view of existence value underdefines the concept, so that there is a potentially limitless range of existence values arising from changes in our perceptions of the state of the world (Rosenthal and Nelson, 1992; Portney, 1994; Diamond and Hausman, 1994). Existence value refers to any nonuse related change in the state of the world that affects the utility of at least one agent, therefore anything that affects the perceptions of agents of the state of the world can potentially affect existence values. Krutilla’s arguments were intended for the analysis of natural resources, although he accepted that other things such as works of art could possess existence value. Yet there does not seem to be any prima facie case for stopping there: once we accept that a change in the state of the world can affect existence value, we can make the case that an all-inclusive cost-benefit analysis will quantify anything that affects the worldview of anyone, anywhere. Some economists claim that this would be a “vast task” (Rosenthal and Nelson, 1992) to which others reply that this would merely “increase the work load of benefit-cost analysts” (Kopp, 1992), and that this is a weak argument to use against the inclusion of nonuse values. Along similar lines, the argument that “some number is better than no number” is invoked. However, this seems to miss the real force of this argument; the amount of work that may be required to accurately assess existence value may be potentially limitless, as the knock on effects of a project yield positive and negative effects, each of which have different existence values for different contingencies. There may even be existence values that supervene on other existence values. The argument is summarized clearly by Portney:

“Imagine now the difficulty of doing applied benefit-cost analysis when virtually every citizen in the United States is potentially benefited or injured by virtually every possible program. In principle, at least, it will become extraordinarily difficult to draw bounds around those likely to gain and lose so as to facilitate valuation.”

(Portney, 1994; p. 14)

The only practical way around this difficulty is to assume that existence values for many cases are insignificantly small, and that we can therefore ignore them. However, this presupposes that we have a reliable method of evaluating the magnitude of existence values, and it is not certain that we do (as shall be argued shortly). Alternatively, we can restrict our consideration to a small set of existence values that we believe we have good reason to treat as more significant. This view puts economists in the position of making normative judgments about some sources of utility being more important than others.

The final case against existence value makes the point that nonuse values cannot be measured reliably. This point is related to the debate about CVM, a detailed treatment of which is beyond the scope of this essay. Four key issues are most relevant: knowledge about the good in question; the context or ‘frame’ in which the good is described in a CVM questionnaire; embedding; and symbolism.

The first issue is that the existence value of a natural resource for an individual seems to depend on an individual having prior knowledge of the existence of the resource. This seems to be trivially true, but nevertheless it seems legitimate to ask whether a natural resource really has existence value only when an individual has knowledge of its existence. While it seems untroubling to conceive of the value of market goods as contingent on the preferences of economic agents, we seem to have a prejudice towards attributing inherent value to natural resources such as biodiversity. However, some, such as Kopp (1992) would assert that we are asking the wrong question here: what matters is the conditional existence value to individuals: if I were to suffer a utility loss from the damage to a natural resource were I to be informed, then that is all that is required: whether I actually know about the natural resource should not matter. This appears to be an important departure from the neoclassical view, however: it is not clear that we cannot regress this argument to actual welfare (rather than just counterfactual welfare) so that our actual welfare changes on the basis of changes in existence goods that we know nothing about.

The second issue is framing: it is suggested that existence values should be stable, but Kahneman (1986) has found frequent occasions when the values yielded in CVM questionnaires have been affected by the way the questions have been asked. However, this may not be as surprising as it sounds: a change in context can be considered a change in the good that an individual is being asked to value: this is the motivation for advertising and sales tactics, such as those used by car salesmen. On this view, if context didn’t matter, this would constitute a violation of neoclassical theory.

The embedding phenomenon refers to the fact that people’s valuations of resources do not seem to be responsive to the quantity of the environmental good: for example people’s valuations for the preservation of 200, 2000 and 20,000 birds of an endangered species may hardly differ at all. Embedding may also occur across environmental goods: my valuation for the preservation for one species may be almost the same as the preservation of two. A response to this problem is that we are adding substitute goods to the choice set of a budget constrained individual, so that the value of individual components should decrease.

Finally, the issue of symbolic bias refers to the observed tendency for people faced with CVM questionnaires to take the good in question as representative of their valuation for an entire class of goods, rather than the specific good in question. For example, when asked the value of preserving a national park, we may state the value of preserving national parks in general, rather than the value of preserving that particular park. This is responded to by saying that careful questionnaire design should be used to avoid the problem.

It is clear that there are a number of controversies surrounding CVM, and they are not discussed in full here. This is especially so, because both sides of the debate seem to miss the central point: there is an implicit view that the method of total valuation should provide a complete measure of all the costs and benefits of any given project. It is not sufficient, however, to argue that the ethical concerns inherent in existence value are simply a particular class of public good. We cannot compensate someone who is dissatisfied with the ethical state of the world with use goods: the damage to someone’s welfare from a feeling of injustice about an act of pollution that damages non-replaceable natural capital cannot be eliminated by substituting sufficient five star holidays in its place.

Problems with the neoclassical reduction of altruistic motives using psychological egoism are numerous. For example, if my ethical objections are based on individualistic altruism for other members of society, then, if this altruism is universal and impartial, this altruism makes no difference to the cost-benefit analysis, as it merely reinforces the conclusion that would have been reached anyway (see section 3.2). If my altruism is partial, however, then my concern for a restricted group of society will affect the analysis: such concerns may yield an efficiency justification for redistributive projects to benefit particular groups. The more partial my concerns, the more influential they are.

Another problem with the general terms of debate is that they focus on the measurement of the utility/disutility of the outcome. However, the process by which states of the world come about is of key concern in ethical beliefs. If the damage from the Exxon Valdez accident had been caused, instead, by a natural disaster, then our valuation of the damages may well be substantially different. We may feel more pain at having a sum of money stolen from us than from losing the same sum through our own actions. However, the conventional view seems to suggest that the state of the world, rather than the causal process, is primary. To place ethical beliefs as arguments into a utility function does not allow an appreciation for the causal process that caused that belief, and it is not clear how we could codify such causal processes even if we desired to do so.

An alternative strategy is to accept the method of total valuation as an ethical criterion, so that we base our moral code on the efficiency criterion. However, there is nothing in economics that states that we must do so, and there appears to be no compelling reason to do so outside the discipline either. To suggest that positive economics provides a normative criterion for what we should do is to derive an ‘ought’ from an ‘is’, the fallacy first pointed out by Hume. For example, it may appear, on the basis of cost-benefit analysis, that the availability of alcohol is worse for consumer surplus than prohibition. However, this does not commit me to a particular moral view. I am free to calculate a negative consumer surplus from alcohol use, and decide not to ban it on moral grounds. What is required is an appreciation that consumer surplus (or, more accurately, Hicksian measures of welfare) is a summary statistic for revealed preference over a bundle of goods, and is therefore a description of observed behavior. This is a positive result, not normative, and therefore has no moral implications. It does not warrant the use of economic welfare measures as an evaluation of the relative merits of different ethical beliefs.

This yields the problem of whether existence value should be used at all. The position taken in this paper suggests that this is not an empirical issue: the debate is not about whether we can get stable and accurate estimates of existence values, but whether the methods used are appropriate. There is no reason to think that they are. As such, estimates of existence value are at best harmless, but at worst they bias the judicial process of establishing appropriate damages, and the policy-making process, in an adverse manner. Economists are to be commended for identifying existence value, and thinking deeply about its potential significance. However, we should be wary of overextending the domain of our discipline.

References

Arrow, K. J., Sen, A. K. and Suzumura, K., Handbook of Social Choice and Welfare (2002)

Bishop, R. C., Champ, P. A. and Mullarkey, D. J., “Contingent Valuation” in Bromley, D W. (ed.), Handbook of Environmental Economics (1995)

Diamond, Peter A. and Hausman, Jerry A., “Contingent Valuation: Is Some Number better than No Number?” The Journal of Economic Perspectives, Vol. 8, No. 4. (Autumn, 1994), pp. 45-64. Stable URL: http://links.jstor.org/sici?sici=0895-3309%28199423%298%3A4%3C45 %3ACVISNB%3E2.0.CO% 3B2-O

Edwards, Stephen F., “Rethinking Existence Values,” Land Economics, 68, no. 1 (1992) 120-122

Hanemann, W. Michael, “Valuing the Environment Through Contingent Valuation,” The Journal of Economic Perspectives, Vol. 8, No. 4. (Autumn, 1994), pp. 19-43. Stable URL: http://links.jstor.org/sici?sici=0895-3309%28199423%298%3A4%3C19%3AVTETCV%3E 2.0.CO%3B2-5

Hume, David, A Treatise of Human Nature, (ed. P. H. Nidditch), Oxford, (1739)

Kahneman, D. and A. Tversky, (1986) “Rational Choice and the Framing of Decisions,” Journal of Business, vol. 59,

Kerry Smith, V., “Nonmarket Valuation of Environmental Resources,” Land Economics, 69, no. 1 (1993), 1-26

Kopp, R. J. and V. K. Smith, “Natural Resource Damage Assessment and Resource economics: Introducing the Issues” in R. J. Kopp and V. K. Smith (eds.) Valuing Natural Assets

Kopp, Raymond J., “Why Existence Value Should Be Used in Cost-Benefit Analysis”, Journal of Policy Analysis and Management, 11, no. 1 (Winter 1992), 123-130

Krutilla, John V., “Conservation Reconsidered,” American Economic Review, 57, no. 4 (1967), 777-786

Larson, Douglas M., “On Measuring Existence Value,” Land Economics, 69, no. 4 (1993), 377-388

Madariaga, B. and McConnell, K. E., “Exploring Existence Value,” Water Resources Research, 23, no. 5 (1987), 936-942

Nelson, Robert H., “Does ‘Existence Value’ Exist?” The Independent Review, 1, no. 4 (Spring 1997) 499-521

Portney, Paul R., “The Contingent Valuation Debate: Why Economists Should Care,” The Journal of Economic Perspectives, Vol. 8, No. 4. (Autumn, 1994), pp. 3-17. Stable URL: http://links.jstor.org/sici?sici=0895-3309%28199423%298%3A4%3C3%3ATCVDWE %3E2. 0.CO%3B2-9

Randall, A., “Total and Nonuse Values,” in John B. Braden and Charles D. Kolstad (eds.) Measuring the Demand for Environmental Quality (1991)

Rosenthal, Donald H. and Nelson, Robert H., “Why Existence value Should Not Be Used in Cost-Benefit Analysis,” Journal of Policy Analysis and Management, 11, no. 1 (Winter 1992), 116-122

Sen, A. K., “Rational Fools: A Critique of the Behavioral Foundations of Economic Theory.” Philosophy and Public Affairs, vol. 6, (1977) 317-44

Sen, A. K., On Ethics and Economics, New York (1987)

Page of

Throughout this paper, bold characters represent vectors

This might be clearer if expressed using integrals:

[16’] so that we can recover the use value by evaluating the integrals in the first set of parentheses ( the Marshallian demand) and recover the nonuse or existence value by evaluating the final integral between E0 and E1.

And, as the appreciation of consumer and producer surplus shows, value is not the same as price

A more rigorous way of stating this point is to say that, if the value relation is considered to be a kind with a real essence, then existence value is of the same kind, and possesses the same real essence. This is a key move, which is motivated by the assumption implicit in psychological egoism, as explained in the subsequent argument.

This is not meant in the pejorative sense, but in the ethical sense that the motivation for all action is self-interest.

This view may be refined by invoking essentiality of goods (either market goods or environmental goods) in the manner of section 2.2, but the essence of substitutability between various types of values remains.

Here and important supporting argument about methodology and descriptions is omitted for clarity. Stated in the philosophical jargon: existence values have a real essence, and therefore the method of analysis that tries to identify this real essence via criteria makes a mistake if it treats the essence and the conditions it satisfies as one and the same thing, since the analysans would fail to be distinct from the analysandum. This is not a universal view, but it does not necessarily follow from the substitutitivity principle (that the expression standing for the analysandum and the one standing for the analysans are mutually substitutable) should imply that the essence of the value can be substituted by the conditions which it satisfied, since the thing itself and the conditions it satisfies are different kinds. Treating them as substitutable denies this, and leads to problems such as the paradox of analysis.

As Hume states: “It is only a sense of common interest; which sense all the members of society express to one another, and induces them to regulate their conduct by certain rules…the actions of each of us have reference to those of the other, and are perform’d upon the supposition, that something is to be perform’d on the other part” (Hume 1739, 490)

There are strong arguments against this position, however. For example, J. Butler argues that psychological egoism cannot work, because I must desire things other than my own welfare. For example, if I derive welfare from the benefit of others, we must have a pre-existing desire that others do well, otherwise we would not derive benefit from their doing well. Welfare results from my action, but cannot be the aim of my action. Nevertheless, the psychological egoist can reply that this is not an accurate description of the pre-existing desire: we may simply have a pre-existing desire that we know that others do well, so that whether they actually do well or not is not the real issue. This quickly retreats into a competition of describing and redescribing motives and desires in an altruistic versus an egoistic fashion.