SUBJECT: Approaches to the Estimation of the Values of Biodiversity: Non-market and Market Approaches
AUTHORS: Timo Goeschl, Andreas Kontoleon and Timothy Swanson
DATE: 11th April 2003

KEYWORDS: biodiversity values, non-anthropogenic valuation, stated preference, revealed preference, production function, search theoretic approach.

SUMMARY: Several approaches have been developed to address the problem of valuing biodiversity. Most are based on the marketed goods and services derived from biodiversity (e.g. information, insurance) and the values that society places on these sorts of goods and services. Some others have recognised that many of biodiversity's unique values flow outside of the market, and attempt to estimate them without reference to marketed goods and services. Others have gone so far as to attempt to value biodiversity without reference to human society, a non-anthropocentric approach. In this piece we describe each of these approaches to the valuation of biodiversity in turn.

1. Non-anthropocentric Valuation of Biodiversity: Diversity Metrics

One approach to biodiversity valuation has attempted to value diversity as an end in itself. The Weitzman (1992) model is the foremost example of this literature (see also Weitzman 1993, Solow and Polasky 1992). In all of these papers, the assumption is that biodiversity has an important role to play, and that it has this value irrespective of the existence of a society or market that values it.

Refining the pioneering work by Solow et al. (1993), Weitzman (1992, 1993) offers the most sophisticated approach to translating biologists' analysis of taxonomy into an optimization framework. Here, diversity is rigorously defined as a quantitative variable elated to taxonomic concepts of relatedness. Weitzman assumes that there is perfect information about the genetic make-up of each species, and that we are then able to rank the "relatedness" of a given set of species.

In short, the Weitzman approach takes genetic similarity as the common metric for valuing biodiversity. To the extent that one set of genetic resources has more dissimilarity than another set, this feature alone is enough to give it a greater value. The construct of a measure of value absent any sort of social or utilitarian justification is the essence of this approach. It is based in a belief system that provides that in the case of biodiversity it is possible that values might exist irrespective of the existence of human society.


2. Preference Based Methods: Non-market Valuation Techniques

If it is concluded that human preferences are the appropriate metric for assessing the value of biodiversity, substantial problems remain with the determination of the method for assessing that value. Many if not most of biodiversity's goods and services do not flow through any market or other social institution. Most of the many millions of species that exist on earth do not have a known or noticeable impact on human society, but many individuals would nonetheless recognise the rights of these species to exist. Most approaches to valuation attempt to register these human preferences for other species' existence, with or without the use of markets. Non-Market valuation techniques are classified into stated and revealed preference techniques.


Stated Preference techniques (including contingent valuation, choice experiments, and contingent ranking) are used in situations where non-market based values need to be estimated and/or when no surrogate market exists from which environmental (use) value can be deduced. These techniques use questionnaires to develop a hypothetical market through which they elicit values (both use and non-use) for the environmental good under investigation. Stated preference techniques do not suffer from the same technical limitations as revealed preference based approaches (see below) and can also be applied to non-use values. Yet, the hypothetical nature of the market constructed has raised numerous questions regarding the validity of the estimates (See Bateman et al., 2003 for a review).

Table 1 gives an example of a set of stated preference studies that have been used to estimate the willingness to pay (WTP) for a range of different endangered species.

3. Market based Estimation: Surrogate Markets Approach

Revealed preference valuation techniques (including travel costs, hedonic pricing and wage differential approaches) rely on information from individual consumption/ purchasing behaviour occurring in markets related to the environmental resource in question (surrogate markets). The price differential of the good (purchased in the surrogate market), once all other variables that affect choice apart from environmental quality have been controlled for, will reflect the purchaser's valuation of that particular level of environmental quality. These methods have the appeal of relying on actual/observed behaviour but their fundamental drawbacks are the inability to estimate non-use values and the dependence of the estimated values on the assumptions made on the relationship between the environmental good and the surrogate market good.

Using the "travel cost approach", for example, it has been possible to estimate the value of various forms of parks and protected areas. The idea is that the costs of travel act as surrogates for the non-marketed good, i.e. the biodiversity within the park or protected area that is the reason for the travel. This assumption enables the approximation of a demand curve, and the estimation of values placed on the non-marketed values.
Table 1 summarizes the results from a set of travel cost studies that have estimated visitor consumer surplus for various national parks.

4. Market Based Estimation: Production Function Approach

An approach related to the surrogate market approach is the production function approach. This method derives from the assumption that the non-marketed good or service is an important input into the production of a marketed good or service, such as the role of clean air as an input into the production of human health.

Evenson (1995) has used this approach to estimate the contribution of genetic resources to plant breeding. This is done by specifying an "R&D production function", and then estimating the extent to which its various component parts have contributed to the past production of new information. An R&D production function in the context of plant breeding, for example, would have to consist of at least: i) the scientific input (human capital); ii) the technological input (physical capital); iii) the genetic resource input (natural capital). The theory of a production function states that increases in these various inputs would result in increases in the desired output: new modern plant varieties. (Evenson and Gollin 1991)

(Evenson 1995) applies this theoretical framework to conduct an empirical study which attempts to estimate the relative contribution of genetic resources in the R&D process in plant breeding. Here the R&D production function of new plant varieties N is specified as: N= f (LKG)
where L: level of input from human capital (scientists)
K: level of input from physical capital (technology, machinery)
G: level of input from genetic capital (biological diversity)

The empirical study is based upon the record of plant breeding at the International Rice Research Institute since 1960, and estimates the extent to which new varieties of rice were attributable to the various forms of investments. This study estimated that approximately 35% of the production of modern new rice varieties has been attributable to the genetic resource input into the R&D function. This implies that the inputs supplied by plant breeders in rice breeding (human and technological) generated no more than 65% of the useful information within modern plant varieties. The imputed present value of a single landrace accession according to this study was $86-272 million. The imputed present value of one thousand accessions with no known history of use was $100-350 million. Given that the initial stock of rice germplasm (in 1960) was 20,000 accessions, the added stock of germplasm since that time (about three times as many accessions) have been estimated to be responsible for fully 20% of the green revolution in rice production (Evenson 1995). In the context of rice production, diverse germplasm contributes 35% of the "total input" required for the production of a new plant variety.

5. Market based estimation: Search-theoretic Approach

In an influential article on the valuation of genetic resources, Simpson et al. (1996) develop a search-theoretic perspective on the problem that is inspired by (Brown and Goldstein 1984). They ground the value of biodiversity in the activity of "biodiversity prospecting" by an R&D intensive industry and deduce the marginal willingness to pay for an additional sample to be prospected when screening of samples is costly. The aim of their work is to quantify the willingness to invest of private firms in the conservation of biodiversity when the value of each sample is the outcome of a Bernoulli trial (the screen). In other words, they evaluate genetic resources from the vantage point of expected private profits from research.

The typical model features a fixed probability p of identifying a valuable trait in a sample where valuable traits give rise to a product with fixed revenue R through a process of further R&D. The cost of screening a sample is fixed at level c. The expected value of a search over n samples can then be expressed as V(n) which is
V(n) = pR-c+ (1-p)(pR-c) + (1-p)2 (pR-c) + .....
The marginal value of the nth sample is then
v(n) = (pR-c)(1-p)n

The empirical problem with the formulation in this equation is that the probability of a 'hit', p, is the most important parameter for estimating v(n), but that data on p is notoriously difficult to obtain. Simpson et al. solve this dilemma by evaluating the expected value of the marginal species under the most optimistic conditions. One interesting finding is that the function mapping the probability of success in any single trial to the value of the marginal species is single-peaked and strongly skewed to the right. This means that once the probability of a successful trial is such that the expected marginal value of a trial exceeds the cost of the trial, the value will rise very rapidly to its maximum value and then decrease again rapidly. This observation is crucial as it shows several points: Sampling costs are an essential determinant of the marginal value, and studies that do not take these costs into account are bound to overestimate the marginal value significantly. Secondly, the fact that the marginal value of the species is not a monotonously increasing function of the probability of success brings an issue to the fore that had previously been overlooked by many researchers, namely the presence of substitutability between species.

If substitutability is very scarce, i.e. the probability of success is very low, then the marginal value is depressed since the expected revenue from the marginal trial is too low to warrant a high volume of trials. If substitutability is not scarce, then the expected revenue from the marginal trial is too low to warrant a high volume since it is very likely that a success has occurred already. In other words, if there is a high level of redundancy within the stock of samples, a significant proportion of the samples can be discarded prior to screening with little loss of expected revenue since it is very likely that a success will be found within the remaining portion.

Based on a number of reasonable assumptions regarding the market value of a product and other parameters, Simpson et al. derive an upper bound for the willingness to pay for the marginal sample and translate this into an per-area WTP for conservation using the common MacArthur-Wilson approach of relating habitat size to the extant stock of biodiversity. Based on this computations, the maximal willingness to pay for a hectare of biodiverse lands in Western Ecuador, one of the "biodiversity hot spots", is US$20,63. The rainforests of the Amazon elicit only US$2,59 per hectare. This implies that most areas with even extraordinary biodiversity do not justify significant payments from the pharmaceutical industry for their preservation. The conclusion of Simpson et al. Is that there is little reason to expect that the industrial use of genetic resources will result in their preservation by private investors.

The problems with this approach to valuation of biodiversity as an R&D input are well-studied. First, if there is prior information about which areas are more likely to produce information on which problems, the values of marginal biodiversity are altered significantly. (Rausser and Small 2000) Secondly, if there is a belief that problems will continue to reemerge on account of selection and resistance, then the social value of biodiversity (as opposed to the private patent-based values) are much greater. (Goeschl and Swanson 2002).

Conclusion

Numerous sorts of approaches have been taken to estimating the value of biodiversity. The problem is complicated by the fact that the value of biodiversity is both a fundamental philosophical question concerning the relationship between human society and the biological world, and a difficult methodological question concerning the nature of the values that are to be estimated.

The problem is further complicated by the fact that, in the most fundamental sense, the value of biodiversity is boundless. One interesting paper has conceptualised the value of biodiversity as the value of catastrophe-avoidance. (Weitzman 2000). And another researcher has stated the belief that any numerical value placed on the ecosystem services delivered by biodiversity "is a serious underestimate of infinity". (Toman 2000) Without some amount of biodiversity it is generally accepted that the world as we know it could not function, and this fundamental notion undermines any attempt to estimate partial values of biodiversity.

Nevertheless, it remains a useful exercise to apply scientific methodologies to the ascertainment of various parts of the value of biodiversity. The anthropocentric values of biodiversity range from clear use values, such as insurance and information, to very abstract non-use values, such as the existence values of endangered species. Careful construction of valuation methodologies is required to capture any of these values. In no case is there a direct market-based method for deriving the value of any of biodiversity's goods and services. These estimates must be inferred from the application of methodologies based on various surrogate goods and markets, and even from stated preference techniques.

This survey has demonstrated the range of methods available for use in valuing these various parts of biodiversity's goods and services. Although these studies can only estimate small parts of the total value of biodiversity's goods and services, each one demonstrates that there is emerging a set of scientific methods capable of careful estimation of some of these values. Although together they must always represent "a serious underestimate of infinity", they do provide some guidance to policy making for biodiversity conservation. (Kontoleon, Macrory and Swanson, 2002).

References:
Bateman, Ian Richard T. Carson, Brett Day, Michael Hanemann, Nick Hanley, Tannis Hett, Michael Jones-Lee, Graham Loomes, Susana Mourato, Ece Özdemiroglu, David W. Pearce, Robert Sugden and John Swanson (2003), Economic Valuation With Stated Preference Techniques: A Manual (In Association With the DTLR and DEFRA), Edward Elgar
Brown, G.M. and J.H. Goldstein (1984): A Model for Valuing Endangered Species. Journal of Environmental Economics and Management 11, 303-309.
Evenson, R., Gollin, D. (1991): Priority Setting for Genetic Improvement Research. Los Banos, Laguna, Philippines: International Rice Research Institute
Evenson, R.E. (1995): The Valuation of Cop Genetic Resource Preservation, Conservation and Use. Paper prepared for the Commission on Plant Genetic Resources, Rome.
Gillespie R (1998) Economic Value of Recreation and Tourism in Forests in the Eden RFA, A Report Undertaken for the NSW CRA/RFA Steering Committee, Project Number NE 33/ES
Goeschl, T. and T. Swanson (2002): The Social Value of Biodiversity for R&D. Environmental and Resource Economics 22(4) (Aug. 2002), 477-504
Kontoleon, A (2003) Essays on non-market valuation of environmental resources: policy and technical explorations, PHD Thesis, Department of Economics, University College London
Kontoleon, A., R. Macrory, and T. Swanson (2002) 'Individual Preference Based Values and Environmental Decision Making: Should Valuation have its day in Court?', Research in Law and Economics, 20, 177-214
Rausser, G., and A. Small. 2000. Valuing Research Leads: Bioprospecting and the Conservation of Genetic Resources. Journal of Political Economy 108(1): 173-206.
Simpson, R.D., Sedjo, R.A., Reid, J.W. (1996): Valuing Biodiversity for Use in Pharmaceutical Research. Journal of Political Economy 104(1): 163-185
Solow, A., Polasky, S., Broadus, J.M. (1993): On the Measurement of Biological Diversity," Journal of Environmental Economics and Management 24, 60-68.
Toman, Michael A. (1998) "Why Not Calculate the Value of the World's Ecosystem Services and Natural Capital." Ecological Economics, Vol. 25, 57-60.
Weitzman, M. 2000. Economic Profitability versus Ecological Entropy. Quarterly Journal of Economics 115(1): 237-63.
Weitzman,M. (1993): What to Preserve? An Application of Diversity Theory to Crane Conservation, Quarterly Journal of Economics 111, pp. 157-83.
Weitzman,M. (1993): On Diversity, Quarterly Journal of Economics 107, 363-406.


A contribution by:

Timo Goeschl
Faculty of Economics and Politics
University of Cambridge
Cambridge CB3 9DD, UK

Andreas Kontoleon and Timothy Swanson
Department of Economics
University College London
Gower Street,
London WC1E 6BT, UK

swanson tables.doc
Swanson contribution tables (26,624 bytes)