The impact of arbitrary constraints over the payment parameter on WTP: a case of a normally distributed random parameter

ABSTRACT

This paper aims to substantiate the impact of the ‘old’ practice in environmental valuation that arbitrarily imposes a constrained distribution on the payment parameter, and to offer a rule of thumb guidance for the accurate willingness-to-pay (WTP) estimation using choice experiments. A ratio of means approach and a mean of ratios approach were compared for estimation accuracy, involving eleven simulated datasets and two real world cases. Contrary to what is still practiced in the literature, incorrectly constrained payment parameters can in fact cause a systematic misrepresentation of true mean WTP values. The overall findings suggest for researchers using choice experiments, as a rule of thumb, that mean WTP values for environmental changes might be safely and accurately estimated using the ratio of means approach based on population moments with an unconstrained random payment parameter. In contrast, the mean of ratios approach based on conditional individual-specific WTP ratios might be neither accurate nor reliable.

KEYWORDS:

• Willingness to pay
• choice experiments
• random parameter
• arbitrary constraint
• simulated data
• reliability test

JEL CODES:

• Q51
• Q57

Acknowledgment

The idea was formulated and most analytical work was carried out at the University of Queensland.

Disclosure statement

No potential conflict of interest was reported by the author.

ORCID

Andy S. Choi http://orcid.org/0000-0001-6124-8999

Notes

1 This distributional form was chosen because the maximum likelihood estimates of the choice modelling parameters are expected to have an asymptotically normal distribution (Amemiya 1981), and this distributional assumption might be most widely used in discrete choice analysis.