3 Policy use and practices for value transfer

Practical tools for value transfer in Denmark – guidelines and an example

3.1 Policy use
3.2 Databases of valuation studies
3.3 Guidelines for primary studies and value transfer
3.4 Acceptable transfer errors

3.1 Policy use

Value transfer increases the uncertainty in the estimated environmental value, and a crucial question becomes: What level of accuracy is acceptable, and how does the need for accuracy vary with the policy use of the value?

Environmental valuation has four main types of policy use:

i) Cost-benefit analysis (CBA) of investment projects with environmental impacts, natural resources preservation and restoration projects, and environmental policies (e.g. the EU Water Framework Directive)

ii) Environmental accounting at the national level to construct “Green” Gross Domestic Product (GDP) (i.e. GDP corrected for changes in natural resources and environmental quality)

iii) Environmental costing in terms of marginal environmental and health damages of economic activities, which can be used as basis for e.g. setting the optimal size of environmental charges

iv) Natural Resource Damage Assessment (NRDA)/Liability for environmental damages; i.e. compensation payments for natural resource injuries from accidental oil spills and other pollution incidents.

For NRDAs, and partly also for environmental accounting and costing, there seems to be a more direct link between the outcome of the analysis and policy impact, and the group affected is more well-defined than in a CBA. Thus, the need for accuracy of the economic values increases, and thus the applicability of value transfer techniques decreases, as we move down the above list of potential policy uses (Navrud and Pruckner 1997). However, even in CBA the need for accuracy is higher when costs and benefits are very close (and the scale is large), and the risk of making the wrong decision is large when using value transfer.

Environmental costing exercises based on the DFA and value transfer have been performed for externalities of electricity production, both in the US and Europe; see e.g. Rowe et al (1995), Desvousges et al. (1998) and European Commission – DG XII (1995, 1999). However, very few of the environmental impacts we focus on in this report have been valued in these exercises. New York State and a few other US states have used these environmental costing exercises to construct "adders" to their electricity prices. Adders are increments added to the private marginal costs that allow you to get closer to full marginal social cost of electricity production. In this case they were used to make more rational decisions for electricity generation by using marginal social rather than marginal private costs (Brennan et al 1996).

In Europe, however, in only a few instances these environmental costing exercises have been used to construct environmental charges. The Swiss heavy vehicle charging scheme has been based solely on a careful analysis of externalities using DFA and value transfer (and some new original valuation studies). In the UK, the landfill tax was also based on a careful value transfer exercise, and the aggregate tax (from mining operations) was based on a comprehensive, new CV study.

Value transfer has also been used in green national accounting exercises; see e.g. the Green Accounting Research Project (GARP) of the European Commission (Tamborra 1999, GARP II 1999). The UN´s statistical division UNSTAT has actively supported the development of resource accounting systems (e.g. the Handbook on Integrated Environmental Economic Accounts).

In the US CV studies have been used in NRDAs to estimate lost use and non-use values of oil spills in marine and coastal areas (see e.g. Carson et al. 2003), and value transfer has been used for affected recreational activities e.g. beach recreation. However, currently many NRDAs apply the Habitat Equivalency Analysis (HEA), which applies a replacement cost approach to value lost ecosystem services in terms of discounted service-ha-years for beach recreation and biomass loss. With the EU Environmental Liability Directive in place we can expect to see both HEA and valuation studies and transfers used in NRDAs in Europe.

However, environmental valuation and value transfer methods are mostly used in CBAs. In Europe, the Directorate General (DG) Environment of the European Commission (EC) regularly performs CBAs of new directives using value transfer; see e.g. their recent CBA for the CAFE programme http://europa.eu.int/comm/environment/air/cafe/activities/cba.htm (based on the Clean Air for Europe – CAFE Strategy) and other air quality directives (see; http://europa.eu.int/comm/environment/enveco/studies2.htm.). However, the EU Water Framework Directive (WFD) is the first directive to explicitly ask for economic analyses to be undertaken in terms of identifying program of cost-effectives measures to achieve ”good ecological status” in water bodies, and to justify exceptions to reaching this goal by showing disproportionate cost. This implicitly means some form of cost-benefit analysis, and a need both for new valuation studies and massive value transfer to cover all water bodies. Recognizing the need for benefit transfer in implementing the WFD, Hanley and Wright (2005) performed value transfer validity test of CE data^{^{^[7]}} applied to the implementation of the WFD in the UK, and question the transferability of CE data as they observed transfer errors of 60-70 % between rivers in the UK. Similar national CBAs of the adaptation to the WFD are planned in e.g. France and Norway. EU member countries, notably the UK, have also conducted full CBAs of national implementation of new EU directives (and strengthening of existing ones) in order to inform their negotiating stance. In the UK, a House of Lords committee questioned whether the benefits of coastal water quality improvements mandated by the Bathing Waters Directive (and moves to strengthen these standards) were large enough to justify the cost, and work was commissioned to undertake this CBA; see Georgiou and Bateman (2005).

In Norway, the Ministry of Finance (2005) recently updated their guidelines (from 2000) on economic analyses of public projects. This CBA guide contains chapters on economic valuation of environmental goods and health impacts, which conclude that value transfers have acceptable transfer errors for use in CBAs if the studies values are transferred from are of high quality, of a similar good, and is close in space and time (see citation in Norwegian below):

I en del tilfeller er det svært dyrt å gjennomføre originale og tilfredsstillende verdsettingsstudier. Et alternativ vil da være å overføre verdiestimater fra tidligere studier ved hjelp av etablerte teknikker for verdioverføring. Imidlertid medfører slike overføringer større usikkerhet enn en original verdsettingsstudie av en spesifikk miljø- eller helseeffekt. Tester av slike overføringer tyder på at usikkerheten er til å godta for bruk i samfunnsøkonomiske analyser. Dette gjelder særlig hvis de underliggende verdsettingsstudiene er av høy kvalitet, er gjort av tilsvarende miljøgoder og er foretatt nært i tid og rom. (Ministry of Finance 2005, p. 96)

According to the Swedish Ministry of Finance and Naturvårdsverket, their general guidelines for CBA do not address environmental valuation and value transfer methods in similar detail. This confirms the findings three years ago by Frykblom and Helgesson (2002) that the use of CBA was then rather limited in Sweden. Sundberg and Söderquist (2004) also found that examining the field “Used in CBA/policy” in ValueBase ^SWE only turns up five valuation studies that were part of a cost-benefit analysis or had otherwise been used in decision-making. However, they note that since this result is based only upon the information actually given in the valuation studies, it excludes primary studies that at a later stage have been used as input in a CBA or provided the basis for designing a policy.

However, all Nordic countries have CBA guidelines for transportation projects that use unit values for selected environmental and health impacts (and noise), typically based on quite rough unit value transfer procedures. The Norwegian Directorate for Public Roads (Statens Vegvesen - Vegdirektoratet) is, however, now re-considering this practice as they would like to base their valuation of environmental and health impacts on the DFA, and plan to conduct new empirical valuation studies which are constructed for value transfer.

3.2 Databases of valuation studies

On the whole, the empirical evidence shows that international benefit transfer is as valid as intra-country transfer (see chapter 2.3). Thus, databases of valuation studies in other countries are a potentially useful source for value transfer.

Appendices C and D review and evaluate existing databases of original valuation studies with regard to their potential use for value transfer in Denmark. The Environmental Valuation Reference Inventory (EVRI www.evri.ca) is the most comprehensive database of valuation studies in terms of the number of valuation studies worldwide, and seems to be the most useful database for Denmark due to its coverage both geographically and with regards to environmental goods. However, it should be populated with more Danish studies, including unpublished, “grey” literature (i.e. studies that do not make it to the scientific journals, but contain information useful for value transfer). Denmark should aim for the development of a Nordic database of valuation studies (which the Nordic Council of Ministers has submitted a call for) that should also be included in EVRI. All Nordic countries should join the EVRI Club in order to secure access and use of the database for all Danish citizens. Meanwhile, the list of Danish valuation studies provided in Appendices A and B (and updated versions of this, as many new Danish valuation studies will be completed in the next few years) should be used together with the database for Swedish valuation studies (ValueBase^Swe), the UK list of valuation studies and similar lists in other Nordic and European countries, in order to identify valuation studies that can be used for value transfer exercises in Denmark. This recommendation is based on the assumption that transfer of valuation studies from countries which are closer geographically, culturally and institutionally will have smaller transfer errors.

3.3 Guidelines for primary studies and value transfer

There are several excellent guidelines and list of “best practice” criteria for performing Contingent Valuation (CV) studies and other Stated Preferences techniques; see especially Mitchell and Carson 1989, Arrow et al. 1993 (NOAA guidelines), Bateman and Willis (1999), Bishop (2003) and Carson (2000). For CV studies, the main recommendations in the NOAA Panel (see Appendix J) still hold but should be supplemented with the practical guide for assessing the quality of CV studies outlined in Carson (2000)^[8]. For a guide to the best practice in Choice Experiments (CE), we would recommend Bennett and Blamey (2001). Bateman et al. (2002) also provide practical guidelines to Stated Preferences surveys in general. For practical guidelines for both stated (SP) and revealed preferences (RP) methods see Champ et al. (2003). Based on most of the references listed above, Soutukorva and Söderqvist (2005) have developed lists of “best practice” criteria for both SP and RP methods; see the lists in their Appendix B1 (especially tables B.1.1 – B 1.6) and Appendix B2.

There are, however, few detailed guidelines on value transfer. In the US there exist guides that cover the key aspects of conducting a value transfer, notably Desvousges et al. (1998) aimed at transfer for valuing environmental and health impacts of air pollution from electricity production.

These guidelines outline the following seven steps:

1) Define the value(s) to be estimated at the policy site

2) Conduct a literature review to identify relevant valuation data

3) Assess the relevance of the study site values for transfer to the policy site

4) Assess the quality of the study site data

5) Select and summarize the data available from the study site(s)

6) Transfer benefit measures from the study site(s) to the policy site

7) Determine the “market” over which benefit estimates are to be aggregated

These guidelines have recently been applied also in Europe to value impacts from climate change in the UK (Metroeconomica 2004). For use values, Rosenberger and Loomis (2001) provide a rather detailed guide for transfer of values for recreational activities based on an extensive database they also helped to develop (see appendix C). Non-use value transfer in the US, however, seems to be more ad hoc. The same seems to be true for most use and non-use value transfers performed in Europe. Thus, Smith (1992)´s call for practical guidelines for value transfer still stands.

3.4 Acceptable transfer errors

Even if we cannot determine general levels of acceptable transfer errors for different policy uses, some general decision rules for how to determine the acceptable transfer errors in CBA can be outlined.

There are two main sources of error in estimated values from value transfer:

i) errors associated with estimation of the unit value/value function at the study site

ii) errors from transferring the study site value(s) to the policy site.

By using “best practice”-guidelines for original valuation studies we can minimize the first type of errors. The second type of errors arises because we usually would need to transfer estimates both in space and time. Results from validity tests of different value transfer procedures for different type environmental goods have shown that individual transfer errors in spatial value transfer vary from a few to several hundred percent (see appendix G). However, average transfer errors, both for national and international value transfer, seem to be about + 25 - 40% (see chapter 2.4). In many cases this would be an acceptable transfer error in CBA. Sensitivity analysis should be performed to see if this interval for the estimated values would influence the outcome of the CBA. The size of the critical transfer error, i.e. when Net Present Value (NPV) of the project is zero should also be calculated, especially in cases where we suspect the transfer errors could be larger. These cases include international value transfers of complex environmental goods from study sites that are quite different from the policy site in terms of magnitude and direction of change, initial level of environmental quality, availability of substitutes (scarcity), different size of affected areas, different type of population (locally most affected population versus the national population) etc.

In order to reduce the uncertainty and calibrate the transferred value estimate there is the option of conducting a small-scale study at the policy site in terms of a valuation workshop, focus group or a valuation study of a small sample. However, the costs of providing this additional information should be compared to the benefits in terms of reduced uncertainty (and the reduced risk of the CBA showing the wrong outcome). This could be done by adopting a Bayesian perspective to inform the decision on whether to conduct primary research at the policy site, and if so how much. Here, value estimates or functions from existing studies can be used to form a prior distribution on the value of the good at the policy site. Valuation research conducted at the policy site provides new information on the value of the good. An updated distribution of the value of the good at the policy site contains information from both previous studies conducted at other sites, and from the new research conducted at the policy site. Thus, the decision should be made based on a comparison of the expected value of the information to be gained and the cost of conducting new research. This Baysesian approach could also be adopted where information on the value of similar goods is available, but there is concern that the value at the policy site may be unique (see Atkinson et al. 1992 for a theoretical discussion, and León et al. 2002 for an application to a national park). However, currently the approach does not seem to be sufficiently developed and simplified to be applied on a routine basis in practical value transfer exercises.

One should be even more careful in using value transfer for policy uses where the demand for accuracy is high; especially environmental costing exercises aimed at determining the level of environmental charges and NRDAs aimed at calculating compensation payments to be paid by those that were responsible for the emissions causing the damage.

Footnotes

^[7] Most value transfer validity tests have been performed on CV data (see appendix G), and this is one of the very few such tests of CE data.

^[8] "The article is available at http://pubs.acs.org/journals/esthag/34/i08/pdf/es990728j.pdf. See also the site http://weatherhead.case.edu/econ/publicPolicy/contingent.cfm for further guidance."