![Sample our Environment & Sustainability journals, sign in here to start your access, latest two full volumes FREE to you for 14 days]({"type":"image" , "src" : "/sda/5935/TOC-Subject-Sample-2021-Environment-Sustainability.jpg"})
ABSTRACT
Albert Einstein is reputed to have said ‘Everything should be made as simple as possible, but not simpler.’ This is good advice for valuing ecosystem services. The fundamental principle of economic valuation is simple, but powerful: value is determined on the margin. This means that context is crucial in estimating ecosystem service values and, therefore, ‘benefit transfer’ exercises that fail to account for location and relative abundance are, at best, meaningless, and at worst, counterproductive. I illustrate the principles of valuation with the example of water purification by riparian buffers. Values can differ greatly over even relatively small areas, and some ostensibly paradoxical results can arise.
Acknowledgments
This paper represents a focusing and condensation of ideas I have developed in earlier working papers. On that broader work I have received helpful comments from Heidi Albers, Edward Barbier, Scott Barrett, Linus Blomquist, James Boyd, Paul Ferraro, Nicholas Hanley, Steve Newbold, Laura Onofri, Erin Sills, Jeffrey Vincent, Lisa Wainger, Patrick Walsh and Martin Weitzman. An anonymous reviewer's careful reading has greatly improved the paper, and I am grateful to her/him for many helpful suggestions. I have also received helpful comments from seminar audiences at Resources for the Future and the 17th Annual Conference of the BioEcon Network. I am grateful to Ulö Mander for communications regarding natural science issues in riparian buffer performance. The usual disclaimer applies to all: I am responsible for any errors.
Disclosure statement
No potential conflict of interest was reported by the authors.
Notes
1. As is often the case of the purported utterances of great thinkers, there seems to be no reliable attribution that Einstein did, in fact, ever say this.
2. The credibility of the benefit transfer also depends crucially on the credibility of the original studies whose results are to be transferred from one context to another. It is not uncommon to find that the estimates of benefit transferred were themselves derived using flawed procedures: confusing average with marginal values, for example, or estimating values with replacement costs, which are often economically irrelevant.
3. For example, some natural systems support processes by which reactive nitrogen compounds are ‘denitrified,’ i.e., converted into elemental nitrogen.
4. This fact underscores the critical point that context is critical. Location, extent, and as this example illustrates, configuration all determine the value of ecosystem services.
5. The InVEST model of water purification and nutrient retention allows different sources to contribute loads in series. When there is a single source, the removal rate is constant, and as step sizes are reduced, the InVEST model reduces to Equation (3).
6. It is also worth pointing out that expression (5) is itself a sort of reduced form that could be further deconstructed. As noted earlier, estimation of marginal damages could require its own extensive data set and empirical procedures, and, while we might find figures for the removal rate in a particular area in the natural science literature, that literature also underscores that removal rates may themselves be complicated functions of topography, soil composition and other factors (Vidon and Hill Citation2004; Weissteiner, Bouraoui, and Aloe Citation2013).
7. This point can be made more formally. Differentiating marginal value, as given in Equation (5), with respect to the removal rate, φ, we have
where is the elasticity of marginal damage with respect to pollutant load times the fraction of pollutant load treated by the buffer. It is often reasonable to suppose that η > 0; that is, that marginal damage is increasing in pollution. Note that this means that the value of the marginal hectare would be lower than it would be if marginal damage were constant; heuristically, each additional kilogram of pollution removed causes less additional damage. Moreover, the pollution treated by a buffer strip in a particular area of interest often comprises only a small fraction of total load. If this were the case, η would be negligible. Thus, when the width, W, of a buffer is wide enough, the marginal value of a still wider buffer is likely to be declining in its effectiveness.
8. The reader might well respond that a wider buffer could provide other services – perhaps, it enhances recreational opportunity or sequesters carbon, and if it does, they should be considered. My only point here is that the sediment-trapping of any additional area of such a buffer would necessarily be minimal.
9. I derived this figure as follows. Stream density in the Chesapeake Bay watershed is about 2.2 km of streams per square kilometre of land (Baker, Weller, and Jordan Citation2007). That is, a kilometre of stream would drain an area of about 2.2− 1 = 0.45 km2 of land, or 45 hectares. Typical nitrogen runoff from a hectare of farmland might be about 28 kg ⋅ ha− 1 ⋅ yr− 1 (Ribaudo, Savage, and Aillery Citation2014). Therefore, kilometre of stream would receive 28 kg ⋅ ha− 1 ⋅ yr− 1 ⋅ 45 ha = 1260 kg ⋅ yr− 1 of reactive nitrogen. If the stream runs through the middle of the area it drains, half the load, 630 kg ⋅ yr− 1, would flow in from each side.
10. Removal rates varying between 0.0127 m-1 for the mid-Atlantic Piedmont portion of the watershed and 0.0315 m-1 for mid-Atlantic coastal plain may be inferred from Weller, Baker, and Jordan (Citation2011).
11. Jones et al. (Citation2010) report costs of a number of different practices, varying from over $400 kg− 1 ⋅ yr− 1 for urban stormwater controls to a few dollars for various agricultural practices. The $35 kg− 1 ⋅ yr− 1 is the estimated cost of reducing a kilogram of nitrogen loading from a municipal wastewater treatment plant.
12. Values vary by state as well as by locations within states. As one would expect, farmland in counties experiencing greater suburbanisation commands a higher price than land does in less densely populated areas. Moreover, pastureland typically sells for less than cropland.
13. I am, in the interest of simplicity, abstracting from another consideration. A 20 m wide buffer on both sides of a kilometre of stream would account for 2 (since it is on both sides) ⋅ 20 m ⋅ 1000 m = 40,000 m2 = 4 ha of land. Recall that a kilometre of stream is presumed to drain 45 ha of land; therefore, I am implicitly assuming that establishing riparian buffers would also take some not-unsubstantial area of land out of production. This would also, presumably, reduce overall loading.
14. Pollination is related to areas of natural habitat maintained, on the argument that wild pollinators benefit from the alternative sources of food that such areas maintain when crops are not in season, as well as the refuge they provide.
15. I might add – and am reminded by an anonymous but very helpful reviewer – that my findings are also supported in empirical work that does explicitly note the relationship between the extent of ecosystems and the services they provide. See, for example, Vincent et al. (Citation2015).