Attribute Causality in Environmental Choice Modelling

When selecting attributes in environmentalChoice Modelling studies, preference should begiven to those attributes that aredemand-relevant, policy-relevant, andmeasurable. The use of these criteria willoften result in a short list of environmentalattributes of which some are causally related. The inclusion of attributes that have a``cause-effect'' relationship may stimulate somerespondents to seek to understand the causalrelations among attributes in order to assigngreater meaning to the alternatives, andpotentially, simplify the decision makingprocess. This may have implications for theweights they assign to each of the attributeswhen identifying the preferred alternatives,and subsequently for the implicit prices and/orwelfare estimates. A test of the impact ofincluding an attribute that causesimpacts on ecosystem health as well as anattribute relating to ecosystem health effectson parameter estimates, implicitprices and welfare estimates is conducted. Twoquestionnaires are developed, one with the`causal' attribute included and one without. Acomparison of results indicates that when the`causal' attribute is included in the vector ofchoice attributes, the implicit value of asingle endangered species falls by 34 per centwhilst no significant difference is detected inthe parameter estimates. Importantly, however,estimates of compensating surplus for a givenpolicy package do not differsignificantly across the two treatments. Thisimplies that to the extent that the inclusionof a `causal' attribute reduces the implicitprices for one or more of the `effect'attributes, the associated loss in utility isapproximately offset by the utility nowassociated with the new attribute. Copyright Kluwer Academic Publishers 2002