Answers of sample survey respondents to questions like the following one: “In the circumstances of your household, what income would you regard as the minimum (or good/bad/sufficient, etc.) for your household?” can be used to estimate equivalence scales. It is sometimes claimed that this is a simple and cheap method to obtain “true” equivalence scales. Typically, these scales are much flatter than those derived from consumption behaviour data, or those incorporated in official poverty lines. A number of possible reasons have been advanced for this result. Yet, using subjective data does not necessarily result in a flat equivalence scale. Equivalence scales can also be estimated from answers to direct survey questions on how satisfied the respondent is with his or her household income, and these scales are often quite steep. In this paper, an application of both methods is reported. A number of possible reasons for the discrepancy are explored. In particular, the effects of a wide range of exogenous variables are examined. Data are from the 1985 and 1988 waves of the Belgian Socio-Economic Panel, a household income survey. The conclusion is that the discrepancy between the equivalence scales persists, even when controlling for a large number of possibly intervening factors.