The Stated Preference Methods

The stated preference methods are usually used for the valuation of goods and services, which have no known existing markets, such as public goods (see Brookshire et al., 1982; Adamowicz, 2004; Lusk and Norwood, 2009; Wijnen et al., 2009). The main method from this group of economic impact methodologies is the CVM and, more recently, the conjoint analysis technique (Roe et al., 1996; Wijnen et al., 2009). However, the application of conjoint analysis technique has mainly been in the area of marketing research (Roe et al., 1996; Green et al., 2001). This discussion, therefore, focuses on the CVM.

4.2.3.1 The Contingent Valuation Method

The origin of CVM is traced to Ciriacy-Wantrup (1940) work on the benefits of prevention of soil erosion (Akwansivie et al., 2010). However, the earliest application of the method is credited to Davis (1963) who used the method to determine the value and wilderness lovers of a recreational area (Kula, 1997; Hammond, 2006; Akwansivie et al., 2010). The method is based on ‘value theory’ and works on the presumption that individuals value their own consumption in a rational manner. That is, they will seek to maximise consumption or utility and minimise their expenditure as best as possible subject to constraints like income and other socio-economic factors (Kula, 1997; Single et al., 2001; Abeka, 2005; Hammond, 2006). It, thus, seeks preference measurements from individuals who are affected by non-marketed goods based on the notion of compensating

Economic Evaluation of ULUP Page59 Economic Evaluation of ULUP 2012

and equivalent variations or the concept of WTP and WTA (Brookshire et al., 1982; Adler and Posner, 1999; Lusk and Norwood, 2009; Wijnen et al., 2009). Fundamentally, it is a process of eliciting people’s preference in terms of how much they are willing to pay for a satisfaction from a non-marketed good, seen as benefit or how much they are willing to accept for a loss in satisfaction from a non-marketed good. In the context of ULUP it means how much property owners/developers will be willing to pay for ULUP variables like sub-division planning scheme, formalised title, building permit and vice-versa when these variables are not available.

In eliciting responses for willingness to pay for a benefit or accept payment for a loss, the method, in the main, uses questionnaire survey (see Brookshire et al., 1982; Kula, 1997, Alberini et al., 1997; Akwansivie et al., 2010). The rationale is to simulate a market for the good, which has no market and generate its value based on the hypothetical market created and presented to respondents (see Bravi and Curto, 1997; Hammond, 2006; Akwansivie et al., 2010). As applied to this study, this may be describing vividly the ULUP variables and the applicable market conditions. The method requires respondents to be well informed about the good (Kula, 1997; Bravi and Curto, 1997; Akwansivie et al., 2010). Additionally, since the method works on the presumption that individuals seek to rationally maximise their satisfaction subject to their socio-economic constraints, the survey usually collects socio-economic characteristics of respondents to enhance the data analysis (Kula, 1997; Bravi and Curto, 1997; Akwansivie et al., 2010).

There are several approaches to eliciting bids from respondents on WTP and WTA. These include: the open ended elicitation method; bidding game and dichotomous choice method, which is sometimes referred to as the referendum method (see Bravi and Curto, 1997; 1998; Akwansivie et al., 2010). The open elicitation method has to do with precise assessment of individuals own reservation price based on introspection analysis. Example of the question to elicit this bid is:

How much will you pay for a standard 3-bedroom house in Kwabenya, Accra if it is covered by an approved building permit?

The bidding game also usually takes the form:

Will you pay X amount for a standard 3-bedroom house in Kwabenya, Accra if it is covered by an approved building permit? No…..will you rather pay Z amount?

This process will continue until the respondent submits a bid. The dichotomous choice method comes in two forms; single and double bound choices and has gained popularity among contingent valuation practitioners lately due to its simplicity in the use of collected data (Calia and Strazzera, 1998; Hammond, 2006; Akwansivie et al., 2010). With regards to the single bound method, respondents are required to answer yes or no if they are willing to pay a given amount for a non-marketed good. However, for the double bound method if the answer is in the negative, a further question is posed with a lower bid and vice-versa, until an acceptable bid is reached.

Following Davis (1963), CVM has gained wide application initially in the field of environmental economics (Brookshire et al., 1982; Akwansivie et al., 2010) and subsequently in the social policy arena (Bravi and Curto, 1997; McGranahan et al., 2001; Wijnen et al., 2009; Akwansivie et al., 2010). The method’s wide application is not limited to the developed world, but also in the developing world it has seen substantial application. For example, Abeka (2005) used the method to determine benefits individual household place on waste collection services in four depressed communities of Sukura, Nima, Ussher Town and Teshie in Accra, Ghana. More recently, Akwansivie et al. (2010) also used the method to estimate the willingness of residents of Kumasi and Accra, Ghana to pay for the cost of improving water quality in these areas.

This upsurge in the use of the CVM has been due to its rigorousness (see Brookshire et al., 1982) and versatility to incorporate different components of value of a good and make respondents aware prior to submitting bids (Bravi and Curto, 1997; Wijnen et al., 2009), a situation which is not possible under the hedonic model. Besides, the scope of the methodology is broad and does not depend on availability of data on peoples’ actual behaviour (Wijnen et al., 2009). That said, the method is said to suffer from hypothetical biases. This situation arises where there is a potential discrepancy between what people say they are willing to pay in a contingent market survey and what they actually pay when confronted with the real situation (Cummings et al., 1986; Kula, 1997; Bravi and Curto, 1997; Lusk and Norwood, 2009). In fact, it is observed that such behaviours are pervasive and, on average, people tend to overstate their willingness to pay by a factor of three in hypothetical settings compared to actual situation where money is involved (see List and Gallet, 2001; Little and Berrens, 2004; Murphy et al., 2005). Related to this problem are social desirability biases where respondents answer questions to please researchers or answer questions to conform to some social norms (Lusk and Norwood, 2009). Another problem with the method also has to do with the considerable resources involved in

Economic Evaluation of ULUP Page61 Economic Evaluation of ULUP 2012

carrying-out contingent valuation survey. Apart from the financial resources, the time required to brief respondents about the good and to fill questionnaires may be enormous. This in turn may not allow respondents to complete their decision-making process towards submitting a bid (Coursey et al., 1987; Abeka, 2005; Hammond, 2006).

It has, however, been argued that a well-designed and carefully executed contingent valuation survey can produce accurate and useful information on household preferences (Cummings et al.,1986; Abeka, 2005; Hammond, 2006; Akwansivie et al., 2010). As such, several solutions have been prescribed to address the problems with the methodology. For example, it is suggested that hypothetical and social desirability biases are due to strategic behaviour to free ride and derive utility respectively (see Hammond, 2006; Lusk and Norwood, 2009). Therefore, methods, such as framing appropriate questions, the adoption of ex post calibration, the use of cheap talks to make respondents aware of these biases, and making people to submit bids for others to avoid subjectiveness and biases are recommended (see Abeka, 2005; Hammond, 2006; Lusk and Norwood, 2009).