This study conducts a multi-country survey on consumer acceptance of genetically modified (GM) foods in the United States and Taiwan and an experimental auction in Japan to estimate consumers' willingness to pay a premium for non-GM foods and/or their willingness to accept a discount on GM foods.
The first essay reports results from a U.S. national telephone survey. The survey questionnaire involves stated preference (SP) questions of vegetable oil, cornflakes, and salmon. American consumers are generally accepting of GM foods. Among the determinants of choice, subjective variables such as perception of benefits and risks and level of knowledge are important. The estimated mean non-GM premiums are, respectively, 44.5%, 35.7%, 38.1%, and 56.9% of the base price to avoid GM vegetable oil, GM cornflakes, GM-fed salmon, and GM salmon. The SP question takes account of indifferent respondents.
The second essay is based on the Taiwanese telephone survey. Taiwanese consumers indicate positive attitudes toward GM foods as compared to American consumers. The mean non-GM premium and the mean GM discount are estimated, and the equality between the two mean values is tested. We find no evidence against the equality, which indicates that Taiwanese consumers consider non-GM and GM foods to be substitutable. We also examine the effect of initial non-GM/GM choice on the propensity to switch alternatives given the price increase or discount. We find that the willingness to switch is different between the non-GM and GM choosers because the two groups have different choice criteria.
The third essay studies Japanese consumers' willingness to pay a premium for non-GM foods based on the experimental auctions. We find that Japanese consumers are not too opposed to GM foods, and the estimated GM discounts are 37% for canola oil and 44% for natto (a soy product), which indicates that Japanese consumers are willing to accept GM foods if they are sufficiently discounted. We test the equality between the mean willingness to accept from the experiment and that from a stated preference question but find no evidence against the equality to conclude that hypothetical bias does not always exist.