Session 3 Benefit transfer and uncertainty TEEB Training

Session 3: Benefit transfer and uncertainty TEEB Training

Outline q Benefits transfer – – What is benefit transfer? Approaches to benefit transfer Challenges Scaling-up values q Uncertainty in valuation TEEB Training

What is Benefit Transfer? q BT takes an existing value for an ecosystem or policy change (at the ‘study site’) and applies it to a new, similar, site (the ‘policy site’) q Why transfer values? – Ideally new ecological and economic studies would be commissioned – New valuation studies are expensive and time consuming q There are four categories of BT: – – Unit BT Adjusted unit BT Value function transfer Meta-analytic transfer TEEB Training

Unit BT q Mean WTP value is taken from study site and directly applied to policy site q Value person/household or per unit (e. g. hectare) is multiplied by population or units at study site q Quick and simple but ignores differences between study and policy sites q Adjusted unit BT – Simple adjustments made for differences between study and policy sites e. g. incomes or prices TEEB Training

Value function transfer q Value functions (e. g. regression equations) are estimated for study site q Parameter (coefficient) estimates for explanatory variables are applied to those variable for study site – i. e. variables such as income, age, attitudes etc. have the same effect on values at both study and policy sites q Takes into account wider range of differences between study and policy sites than simple adjustment of values TEEB Training

Meta-analytic function transfer q Multiple valuation studies are collected and value function is estimated q Allows greater variation of site and socio-economic characteristics to be considered TEEB Training

Complexity of BT approaches q These approaches show increasing complexity q Unit BT may provide the most accurate transfer value if primary study was high quality, robust and site characteristics have little variation q Meta-analytic transfer can be time consuming and expensive due to need to collect and code a database of primary studies q Not all values are expressed in per household/person terms (e. g. production function, net factor income) or benefiting population cannot be identified, area based units (e. g. $ or £/ha) are preferred TEEB Training

Challenges: Transfer errors q Transferred values may differ significantly from actual values q Three sources of error: – Errors in original measures of value at study site – Errors arising during transfer – Publication selection bias q Value function and meta-analytic transfers typically give lower transfer errors TEEB Training

Challenges: Aggregation q Aggregation is the multiplication of unit values by the quantity of those units to estimate total value – Per household/person values – need to know size of the ‘market’, i. e. how many people hold values for the resource – Per unit area values – need to know area of ecosystem to transfer to TEEB Training

Challenges: Aggregation cont… q Aggregation also refers to summing values of different ES supplied by an ecosystem – Potential double counting – ES may be jointly provided, mutually exclusive, interacting or integral, i. e. not independent – Aggregating across multiple ES may give implausibly large numbers (recall part-whole bias) TEEB Training

Challenges: Spatial scale q Differing spatial scales of ES provision add complexity to BT q Ecosystems differ in scale q ES provided at different scales q Demand arises at different scales – use and non-use values q Issues arise over heterogeneity of site and context characteristics, proximity of complimentary or substitute sites q GIS can help account for some of these spatial issues TEEB Training

Challenges: Variations in characteristics and context q Values are influenced by: – Characteristics of site: area, integrity, type of ecosystem – Beneficiaries: distance to site, number of beneficiaries, income, preferences, culture – Context: availability of substitute and complimentary sites and services TEEB Training

Challenges: Non-constant marginal values q Some ES values exhibit diminishing returns to scale q An additional hectare of forest added to a 10 ha site may be worth more than the same increase for a 100 ha site q Some ES may exhibit increasing returns to scale, e. g. area of ecosystem needed to support a large predator q Linear adjustments for changing ecosystem extent are likely to be inadequate: – non-linearities – step changes – thresholds TEEB Training

Challenges: Distance decay q Rate of distance decay likely to vary across ES q Direct use values likely to have strong distance effect q Non-use values may exhibit distance decay, – cultural or political boundaries may be more important q Some charismatic species (e. g. pandas, tigers, whales) may have zero rate of spatial discounting TEEB Training

Challenges: Equity weighting q Expectation that WTP is positively related to income – Incomes vary across sites and beneficiaries, so some adjustment is needed q But poorer people (esp. developing countries) are more reliant on ES and vulnerable to loss of ES q Welfare losses are higher for the poor than for the wealthy – Marginal utility of consumption is declining in consumption – a wealthy person with high consumption gains less from an additional unit than a poor person with lower overall consumption TEEB Training

Challenges: Availability of primary studies q Important to have sufficient primary studies of high quality for all relevant ecosystem types, ES, and socioeconomic and cultural contexts – well represented: wetlands and forest – under-represented: marine, grassland, mountain ecosystems – well represented: recreation, environmental amenities – under-represented: regulating services q Few valuation studies in developing countries TEEB Training

Scaling-up values of ecosystem services q We may be interested in estimating the value of the total stock of an ecosystem or of all ES within a large region, i. e. to scale-up values q We cannot simply add up all the estimated values from smaller sites q Appropriateness of marginal values depends on our position on the ES demand curve TEEB Training

Summary on Benefits Transfer q BT seeks to use existing values to avoid expense and time needed for primary valuation study q Value function and meta-analytic BT preferred to unit BT as accounts for differences between sites and beneficiaries – but can be expensive and time consuming to do robustly q BT is prone to errors due to poor primary studies, generalisation during transfer, and publication bias q A number of challenges arise due to complex nature of ecosystems, ES provision, context and differing socioeconomic factors q Scaling-up raises particular challenges due to changing marginal values and critical thresholds TEEB Training

Uncertainty: Outline 1. Supply uncertainty 2. Preference uncertainty 3. Technical uncertainty TEEB Training

Supply uncertainty q We might not have information on ecosystem functioning and how biodiversity supports provision of ecosystem services q What we can typically measure is quantity rather than quality: – value is applied to quantity of biomass or extent of ecosystem (e. g. hectares) rather than quality – Valuation may also focus on more identifiable resources (e. g. charismatic species) q Extent to which this is a problem depends on motivation for valuation – If we value per hectare but there is large variability per hectare then our assessments will be flawed TEEB Training

Supply uncertainty 1 q Where states of nature are identifiable and probability can be assigned then expected values may be estimated q Potential outcomes are weighted by probability of occurrence q For example: the expected level of carbon sequestration of a set of forest tree species is related to rainfall patterns (states of nature) TEEB Training

Supply uncertainty 2 q Barbier (2007) uses an estimated damage function (EDF) approach to values the storm protection benefits of coastal wetlands q Changes in wetland area affect the probability and severity of economically damaging storm events (states of nature) – WTP for expected damages resulting from changes in ecosystem stocks q Requires sufficient data on incidence of past events and changes in wetland area TEEB Training

Preference uncertainty q Valuation often assumes that preferences are known with certainty, i. e. they are aware of their WTP for a given ecosystem service q Evidence suggests otherwise: – Respondents adopt heuristic rather than systematic mode when processing information – Unfamiliar hypothetical market for an unfamiliar or intangible good q There are three methods for dealing with preference uncertainty in CVM TEEB Training

Preference uncertainty q Ask respondents to state how certain they are about their WTP answer (e. g. Loomis & Ekstrand 1998) – Does not solve problem of uncertainty per se – Investigates whether respondents’ perceptions and attitudes are correlated with self-reported ‘certainty scores’ TEEB Training

Preference uncertainty q Introduce uncertainty directly using multiple (bounded) WTP questions TEEB Training

Preference uncertainty q Ask respondents to give a range of values (e. g. Hanley et al. , 2009) TEEB Training

Technical uncertainty q Uncertainty arises from conceptual, methodological and technical aspects of all valuation methods q Credibility of stated preferences – SP methods assume that respondents answer truthfully q Do respondents only answer truthfully if it is their interests to do so? § Surveys often do not have an mechanism to constrain strategic behaviour § Quality of the survey will affect credibility § Sample size will affect margin of error TEEB Training