A Statistical Matching approach to reproduce the heterogeneity of willingness to pay in benefit transfer

Researchers and policymakers seek a better understanding of the social demand for agri-environmental public goods (PGs), that, being nonmarket goods, are usually valuated by means of stated preference methods by eliciting people's willingness to pay (WTP). In actual policy design, benefit transfer (BT) is often preferred to novel surveys which are expensive and time demanding. Common BT approaches are value and function transfer that can provide good estimates of the mean WTP but disregard the heterogeneity of the individuals' preferences. The WTP distribution is thus flattened, leading to a misrepresentation of the PG demand. The objective of this paper is to improve BT in its ability to reproduce the actual WTP distribution at the policy site by means of the non-parametric micro Statistical Matching. We use this novel approach to transfer individual WTP values for soil erosion and carbon sequestration elicited by contingent valuation on people living in Emilia-Romagna, Italy. Comparing the results with the ones of value and function transfers, our approach outperforms the others, reflecting the actual WTP distribution and lowering the benefit transfer errors. In this way, BT can better support policymakers in designing new agri-environmental policy instruments, more targeted towards specific demand segments and hence with higher cost-effectiveness.