Property appraisal with a log-t predictive distribution

A multiplicative model with log-normal errors is used to appraise the market value of a property and a natural conjugate prior, describing the beliefs of an expert appraiser about the parameters, is elicited. Loss is an asymmetric function of the predicted price as a proportion of the actual price, and the optimal prediction is a multiplicative adjustment to the exponent of the predictive mean of the logarithm of price. The prediction of the price of a property is made using both the elicited prior and two reference priors. The robustness of the procedure is assessed by means of the evaluation of a number of elasticities of price prediction with regard to changes in a parameter value, either of the prior distribution or the loss function.