|| Checking for direct PDF access through Ovid
Seafood species vary in their health benefits (e.g., from omega-3 fatty acids) and risks (e.g., from methylmercury or polychlorobiphenyls). Reflecting these risks and benefits, multiple public and private organizations offer guidance to consumers on seafood consumption. The effect of this guidance is unknown; previous literature has been unable to disentangle the effects of messages with differing health information, provided by different sources, on demand for different types of seafood. The result is ambiguity regarding the drivers of observed changes in seafood demand. This study investigates the effect of health risk and benefit information on preferences for wild and farmed salmon and swordfish, three species targeted by consumer guidance. The analysis applies an experimental auction with seafood consumers informed by a Bayesian risk-learning model. The model provides a systematic way to disentangle effects on seafood demand, for example, by evaluating whether changes in demand for different species are due to information content or source. Using this approach, we test the effect of guidance provided by four different public and private groups in the United States. Difference-in-difference tests find no impact of health benefit information regardless of source or message, but find multiple effects of health risk information that vary across different types of guidance. These findings suggest that current guidance does not improve consumers' ability to balance health risks and benefits. We also identify potential avenues to improve the efficacy of this guidance.JEL codes: C93, D83, I1, Q18.