Non-market valuation of recreational services in Italy: a meta-analysis,

This paper reports on a meta-analysis study carried out in Italy to value recreational ecosystem services. The best meta-analysis model is identified in terms of functional form and robustness to non-independency of metadata. Estimates are obtained using Huber-White robust standard error models, fixed effect and random effects panel data models, and a multilevel (hierarchical) regression model. The hierarchical structure of the metadata was explored by considering as an upper cluster variable the lead author of the valuation study, then the primary study from which each benefit estimate was extracted, and as a lower cluster variable the typology of each recreational policy site.