Valuation of subjective wellbeing and the role of marital status: Linear versus ordinal estimators

Subjective wellbeing measures have been increasingly used in policy evaluation. Although subjective wellbeing is measured on an ordinal scale, it has typically been treated as cardinal. The choice of a linear estimator is justified by the claims that there is no difference between the results of linear and ordinal estimators. Using the UK's British Household Panel Survey between 1991 and 2010, this paper assesses the robustness of this claim by applying linear and ordinal estimators to a well-established relationship between marital status and subjective wellbeing. The results reveal that although the direction of the effects is consistent across the two estimators, the magnitudes of the effects are different. When evaluating policies as part of Social Cost Benefit Analysis, the UK Government now highlights the importance of policy impacts on people's wellbeing expressed in monetary terms. Our research shows that such monetary valuations are sensitive to the choice of the estimator.