Attribute causality in environmental choice modelling

Abstract

When selecting attributes in environmental Choice Modelling studies, preference should be given to those attributes that are demand-relevant, policy-relevant, and measurable. The use of these criteria will often result in a short list of environmental attributes of which some are causally related. The inclusion of attributes that have a "cause-effect" relationship may stimulate some respondents to seek to understand the causal relations among attributes in order to assign greater meaning to the alternatives, and potentially, simplify the decision making process. This may have implications for the weights they assign to each of the attributes when identifying the preferred alternatives, and subsequently for the implicit prices and/or welfare estimates. A test of the impact of including an attribute that causes impacts on ecosystem health as well as an attribute relating to ecosystem health effects on parameter estimates, implicit prices and welfare estimates is conducted. Two questionnaires are developed, one with the 'causal' attribute included and one without. A comparison of results indicates that when the 'causal' attribute is included in the vector of choice attributes, the implicit value of a single endangered species falls by 34 per cent whilst no significant difference is detected in the parameter estimates. Importantly, however, estimates of compensating surplus for a given policy package do not differ significantly across the two treatments. This implies that to the extent that the inclusion of a 'causal' attribute reduces the implicit prices for one or more of the 'effect' attributes, the associated loss in utility is approximately offset by the utility now associated with the new attribute.