1. Valuing Environmental Goods and Services: An Economic Perspective

The reasons that markets do not exist can vary. In many cases, the market failure arises because the environmental good or service is a public good. For example, air quality in a city is a pure public good because all individuals living in the city will benefit from an improvement in the city’s air quality and no one in the city can be excluded from enjoying the benefits of the improvement. Public goods suffer from the “free-rider” problem, which can impede the development of a market for the good. Markets can fail to exist for other reasons as well, including ill-defined property rights, information asymmetries, and difficulty in defining and monitoring tradable units.

Most theoretical discussions of economic values and nonmarket valuation methods focus on valuing a single change (e.g., a change in ambient air quality). However, most real-world valuation contexts involve changes in multiple environmental goods or services and multiple impacts on human well-being. The need to consider multiple changes or impacts raises questions about interconnectedness and aggregation and clearly complicates the valuation process. See National Research Council (2005, Chapter 5) for a useful discussion of valuing changes in multiple ecosystem services.

An important research question is whether alternative ways to define and measure assigned values yield consistent information about underlying preferences. The limited evidence that exists on this is mixed. See, for example, Cooper et al. (2004) and Spash (2006).

See Chap. 2 for a more formal definition.

In contrast, some psychologists believe that preferences are constructed through the elicitation process. See, for example, Lichtenstein and Slovic (2006).

However, see Common et al. (1997) for evidence suggesting that the assumption of substitutability that underlies the economic notion of existence value may not hold in all contexts.

Under some conditions, it is possible to value a change in output by aggregating the payments to the owners of the inputs used to produce the additional output. For example, with perfectly competitive markets and constant returns to scale, the value of additional agricultural output can be measured as the sum of the payments to the laborers, landowner, fertilizer producers, etc., who supply the inputs used to increase production. In this case, the income paid to farm workers can be used as part of the measure of benefits. However, including both the payments to the inputs and the revenue from the sale of the additional output as measures of benefits would be double counting. Benefits can be measured as the value of the inputs or the value of the output but not both.

The theoretical foundations of benefit-cost analysis as a basis for public policy are well established and described in detail in Just et al. (2004).

For an example of nonmarket valuation in this context, see Johnston et al. (2011).

For more detailed discussions of these and related issues, see, for example, Freeman et al. (2014).

As an alternative, preferences could be aggregated across individuals simply by counting the number of individuals who prefer one option to another. This approach underlies decision rules based on standard voting procedures. A key drawback to using votes as a means of aggregating preferences to determine outcomes is that the resulting decisions do not reflect the intensity of individual preferences for one option over another. For example, in a three-person vote, an option that is only slightly preferred by two individuals would win over an option that is strongly preferred by the third individual.

Occasionally, a weighted sum will be used in an effort to incorporate distributional concerns, i.e., to give more weight to the benefits that accrue to one group of individuals than to another (see, for example, Johannsen-Stenman 2005). However, most economists do not advocate the use of a weighted average as a means to incorporate distributional concerns; rather, they advocate providing decision-makers with an unweighted measure of total benefits along with information about the distribution of benefits across relevant subpopulations. See, for example, Arrow et al. (1996).

This is based on the compensation principle that underlies benefit-cost analysis. For a detailed discussion, see Just et al. (2004).

Note, however, that maximizing aggregate net benefits or aggregate income is not generally equivalent to maximizing the sum of utility across all individuals. The two will be the same if coupled with a redistribution of income that equates the marginal utility of income across all individuals. Absent that redistribution, a change for which benefits exceed costs could actually reduce aggregate utility. To see this, consider a choice x, where an increase in x generates an increase in income (i.e., a benefit) for Person 1 and a decrease in income (i.e., a cost) for Person 2. Let \(Y_{i} (x)\) be income for Person i (where i = 1 or 2) and let \(u_{i} (Y_{i} (x))\) be i’s utility. Then, if \(\frac{{\partial u_{1} }}{{\partial Y_{1} }} < \frac{{\partial u_{2} }}{{\partial Y_{2} }}\) and there is no actual compensation, it is possible to have \(\frac{{\partial Y_{1} }}{\partial x} + \frac{{\partial Y_{2} }}{\partial x} > 0\) (i.e., the gain to person 1 exceeds the loss to person 2) even though \(\frac{{\partial u_{1} }}{{\partial Y_{1} }}\frac{{\partial Y_{1} }}{\partial x} + \frac{{\partial u_{2} }}{{\partial Y_{2} }}\frac{{\partial Y_{2} }}{\partial x} < 0\) (i.e., aggregate utility decreases). While it might be reasonable to assume that direct mechanisms (such as taxes) are available to address income distribution within a given generation, this assumption is much more questionable for income redistribution across generations. For this reason, differences in the marginal utility of income across individuals are often not explicitly considered when aggregating benefits and costs at a given point in time, while differences in the marginal utility of consumption across generations play an important role in aggregation across time (see discussion in Sect. 1.8.3).

However, to the extent that current decisions affect future generations, the individuals in those generations are not around today to express the values they would assign to the relevant changes. Thus, in practice it is the current generation that must express values on behalf of future generations.

A simple example illustrating this result is provided in Cropper (2012). A declining discount rate is also consistent with some forms of nonstandard preferences, such as those that exhibit hyperbolic or quasi-hyperbolic discounting. See, for example, the discussion and references in Benhabib et al. (2010).

A negative discount rate might arise, for example, in the context of ecosystem services where those services are becoming scarcer (rather than more abundant) over time. See, for example, Heal (2005, 2009) and National Research Council (2005).

Although some authors distinguish between “risk” and “uncertainty” based on whether the probabilities of the possible outcomes can be quantified, in economics, the two terms are typically used interchangeably. We follow this convention here. For example, when talking about either the uncertainty associated with future preferences or the health risks from exposure to pollution, we assume that all possibilities can be identified and each can be assigned an objective (or possibly subjective) probability of occurring.

Such estimates are often expressed in terms of the “value of a statistical life” (VSL). However, this terminology has led to confusion and unnecessary controversy about the concept being measured, prompting some to argue that the term VSL should not be used (e.g., Cameron 2010).

For example, the U.S. Office of Management and Budget’s Circular A-4, which governs federal regulatory impact analyses, explicitly states: “Other things equal, you should prefer revealed preference data over stated preference data because revealed preference data are based on actual decisions, where market participants enjoy or suffer the consequences of their decisions”. See OMB Circular A-4, Section E (September 17, 2003), available at http://​www.​whitehouse.​gov/​omb/​circulars_​a004_​a-4#e.