Abstract
This paper studies distributional effects of environmental policies in Swedish coastal environments, in monetary and environmental quality terms, for different dimensions: income, gender, age, non-users vs. users, distance, familiarity, and origin (if people have a Swedish background or not). The study area is widely used for different recreational activities and has a mix of different visitors. The data come from a choice experiment study. The results indicate that latent class modelling can be used to identify how monetary preferences vary between different groups of respondents, and largely confirm the limited existing knowledge from the previous research on distributional effects of environmental policies. However, the previous literature on distributional effects related to background is very limited, making it hard to draw comparisons. The results in our paper also show that the distributional effects differ depending on the environmental amenity. These results are of policy relevance since coastal environments are important for people's well-being and associated with positive health effects.
Acknowledgements
The authors would like to thank the past and present colleagues: Sofia Ahlroth, Susanne Baden, Scott Cole, Frida Franzén, Gerda Kinell, Petter Lundin, Leif Pihl, Åsa Soutukorva, and Tore Söderqvist. The linking of scenarios to policy and ecology was facilitated by valuable input from Tina Elfwing, Ragnar Elmgren, Kerstin Harvenberg, Ulf Larsson, Ewa Lawett, Ingrid Nordemar, and Jacob Walve. The authors are grateful to Fredrik Carlsson who gave valuable advice in CE design.
Disclosure statement
No potential conflict of interest was reported by the authors.
Notes
1. See Östberg, Hasselström, and Håkansson (Citation2010) for details regarding background to how different attributes and levels were chosen.
2. In the fall of 2009 when the valuation study was carried out, 10 SEK corresponded to approximately 1 EUR.
3. We used ‘very low’, ‘low’, ‘moderate’, ‘good’, and ‘very good’ as descriptors of the water quality levels to the respondents. This corresponds to ‘bad’, ‘poor’, ‘moderate’, ‘good’, and ‘high’, as used in WFD policy documents.
4. The scenarios are presented in detail in Appendix 1.
5. Details about the pre-testing is presented in Östberg, Hasselström, and Håkansson (Citation2010).
6. This definition is somewhat broader than the definition used by Statistics Sweden (cf. Section 1).
7. Locals were sampled from the municipalities of Botkyrka, Nynäshamn, and Trosa, while Non-locals were sampled from the municipalities of Stockholm, Nacka, Nykvarn, Tyresö, Haninge, Salem, and Huddinge.
8. A simple t-test reveals that this difference is not significant within a 95% confidence interval.
9. The Akaike Information Criteria (AIC) is computed as −2LL + 2K, where LL is the log-likelihood of the model and K is the number of estimated parameters.
10. Tables showing descriptive statistics for the LCM classes are presented in Appendix 2.