Opposite impacts of policy and payment consequentiality treatments on willingness-to-pay in a contingent valuation study

ABSTRACT

A growing number of studies investigated the effects of additional consequentiality scripts on respondents’ willingness-to-pay (WTP) and/or their beliefs in the survey arena. However, these works barely provide a clear distinction between policy and payment consequentiality, despite the possible varying effects of these two alternative beliefs. This study explored the impacts of additional policy and payment consequentiality scripts on respondents’ WTP and stated beliefs using a split-sample approach. Econometric analyses revealed that the policy and payment consequentiality scripts had significantly positive and negative impacts on respondents’ voting behaviours, respectively. Especially, the latter script was more influential than the former one. The payment consequentiality script was also found to be significantly effective in improving respondents’ stated payment consequentiality beliefs. These results suggest that emphasising payment consequentiality of a survey is important to encourage respondents to exhibit more careful attitudes towards a hypothetical scenario, thereby reducing some forms of bias in stated preference methods.

KEYWORDS:

• Consequentiality scripts
• contingent valuation
• payment consequentiality
• policy consequentiality
• single binary choice
• willingness-to-pay

Acknowledgement

This research was supported by the Environment Research and Technology Development Fund (S-15 Predicting and Assessing Natural Capital and Ecosystem Services (PANCES)) of the Ministry of the Environment, Japan.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 According to the Gibbard-Satterthwaite theorem, no response formats with more than two alternatives can be incentive compatible unless substantive additional restrictions are imposed on respondents’ preferences (Carson and Groves 2007).

2 Several studies also addressed this issue through laboratory or field experiments (e.g. Carson, Groves, and List 2014; Cummings et al. 1997; Landry and List 2007; Mitani and Flores 2014; Vossler and Evans 2009; Vossler, Doyon, and Rondeau 2012). However, lessons learned from the lab or field may be difficult to transfer to the survey arena (Herriges et al. 2010).

3 This payment vehicle follows the current local environmental taxation scheme (i.e. Yokohama green tax) and thus is deemed practically feasible. Also, from theoretical considerations, an annual payment scheme is more preferable than one-time payment because respondents are spared from performing complicated present value estimations and less likely to face binding mental accounting budget constraints (Egan, Corrigan, and Dwyer 2015).

4 In addition to the above notes, the following three scripts were also provided to the respondents: (1) all the collected tax would be spent for this purpose exclusively; (2) no payments will be required if such projects would not be initiated, but alternatively, the local coastal conditions would remain unchanged over the next five years; and (3) expected outcomes might not be achieved due to irregular weathers and natural disturbances.

5 Note that my internet-based survey did not allow participants to re-answer the previous questions. That is, the respondents, say, in the control group, were informed about these facts after they answered the value elicitation question and expressed their consequentiality beliefs over policy and payment.

6 With 29 valid samples and using payment cards varying from ¥0 to ¥5000, mean WTP was estimated around ¥893 (i.e. US$8.3). This value is quite similar to the current rate of Yokohama green tax (i.e. ¥900). A scope test with split sampling was also attempted in the piloting. Two combinations of the number of project sites and their total coverage were presented to each subsample, that is, 3 sites extending 3 ha in total (29 valid samples) and 10 sites extending 15 ha in total (18 valid samples), fully considering the potential local project sites in a practical sense. The WTP estimates were found to be scope sensitive (at the 15% significant level). Note, however, that the respondents facing the latter combination were more likely to choose don’t know to the value elicitation question. Thus, I decided to propose the former combination in the contingent scenario in the main survey.

7 It is a common practice to drop respondents who expressed ‘protest no’ or ‘yea say’ from the sample, but such identification requires the analyst to use subjective judgment often without any clear-cut decision rules or criteria (Johnston et al., 2017). Thus, all were treated as legitimate respondents and left in the sample.

8 The model is estimated by the command mvprobit (developed by Cappellari and Jenkins 2003) in STATA 16, which employs the Geweke-Hajivassiliou-Keane simulator to evaluate multivariate normal integrals, using 100 random draws.

Additional information

Funding

This work was supported by Ministry of the Environment, Japan [grant number: Environment Research and Technology Development Fund (S-15: JPMEERF16S11500).