Joint models for noise annoyance and WTP for road noise reduction

Previous CV studies of the WTP for road noise reduction have used stated annoyance as an independent variable. We argue that this may be inappropriate due to potential endogeneity bias. Instead, an alternative model is proposed that treats both WTP and annoyance as endogenous variables in a simultaneous equation model as a combination of a linear regression with an ordered probit with correlated error terms and possibly common parameters. Thus, information on stated annoyance is utilised to estimate WTP without bias. Application of the model to a dataset from Copenhagen indicates a potential for improving the precision of the estimate of WTP for noise reduction with CV data.