Worse is worse and better doesn't matter? The effects of favorable and unfavorable environmental information on consumers’ willingness to pay

Research output: Journal contributionsJournal articlesResearchpeer-review

Authors

Increasing consumers’ willingness to pay (WTP) for environmentally friendly products is a key challenge for sustainable development in market economies. Still, how consumers react to favorable and unfavorable environmental information of different quantitative extents is largely unknown. This research therefore uses prospect theory and competing theoretical foundations to derive pertinent hypotheses and test them by using a multi-level structural equation model. The analysis draws on a survey-based experiment conducted among a representative sample of the German population. Results confirm key assertions of prospect theory. The negative effect caused by unfavorable product carbon footprint information on WTP is stronger than the positive effect caused by respective favorable information. Besides this negativity bias, consumers tend to generally reward or punish deviations of a product's environmental performance from industry average instead of consistently accounting for the size of these deviations. From a sustainable development perspective, the observed patterns highlight a problematic contrast between the need for substantial environmental improvements and limited market incentives for companies. Consequently, political intervention is needed to introduce negative labeling, raise consumers’ reference points, set minimum industry standards, and subsidize companies for radical improvements.

Original languageEnglish
JournalJournal of Industrial Ecology
Volume25
Issue number5
Pages (from-to)1338-1356
Number of pages19
ISSN1088-1980
DOIs
Publication statusPublished - 10.2021

Bibliographical note

Publisher Copyright:
© 2021 The Authors. Journal of Industrial Ecology published by Wiley Periodicals LLC on behalf of Yale University

    Research areas

  • consumer behavior, industrial ecology, negativity bias, product carbon footprint information, prospect theory, willingness to pay (WTP)
  • Management studies

DOI

Recently viewed

Publications

  1. Modelling scenarios to identify a combined sediment-water management strategy for the large reservoirs of the Tuyamuyun hydro-complex
  2. End-to-End Active Speaker Detection
  3. Drafts in Action
  4. Self-Regulated Learning with Expository Texts as a Competence
  5. Proposing a social-ecological framework for successful grassland restoration in Germany—an overview and insights from the Grassworks project
  6. The Influence of Terrorism on Expatriate Performance: a Conceptual Approach
  7. Transfer of metacognitive skills in self-regulated learning
  8. Self-Compassion as a Facet of Neuroticism? A Reply to the Comments of Neff, Tóth-Király, and Colosimo (2018)
  9. Resource selection by sympatric wild equids in the Mongolian Gobi
  10. Value-based management in banking
  11. Design It!
  12. Empirische Erfassung eines „messy constructs“
  13. I Am Not A Hacker
  14. Kombinatorik mit Ziffernkarten
  15. Gemeinsam lernen mit Selbstlernprogrammen
  16. The challenge of managing multiple species at multiple scales
  17. Back to the future
  18. Temporary organizing and acceleration
  19. Portal als Experimentalsystem
  20. A review of FEM code accuracy for reliable extrusion process analysis
  21. Multimodal analysis of spatially heterogeneous microstructural refinement and softening mechanisms in three-pass friction stir processed Al-4Si alloy
  22. Leveling up? An inter-neighborhood experiment on parochialism and the efficiency of multi-level public goods provision
  23. Microstructure and properties of magnesium alloy Mg-1Zn-1Ca (Zx11)
  24. Europe and the Media
  25. Logotoaster
  26. Predictive modeling in e-mental health
  27. Paradigm shift
  28. Gutes Leben vor Ort
  29. Demographic change in work organizations
  30. Generalising IRT to Discriminate Between Examinees
  31. The five-factor asset pricing model – A theoretical review and assessment
  32. Forecasting Government Bond Yields with Neural Networks Considering Cointegration