Semiparametric estimation of equivalence scales using subjective information

Household equivalence scales are not identified from consumer demand data alone. We estimate household equivalence scales using two types of subjective information. First, we use the answers to questions on the income required to attain a given utility level. This is the type of information often used in this type of research. We compare the results for the usual linear model with semiparametric estimates, in which the functional form of the relationship between required income and family size and actual income is left unspecified. Second, we use answers to the question: how satisfied are you with actual household income? We present parametric and semiparametric estimates for the ordered response model explaining this discrete variable, which has possible outcomes 1,2,...,10. We find that according to the second type of information, costs of children are much larger than according to the first.