The attitudes of students towards scholarships, graduate income, and freedom of study: An attempt at empirical construction of a social welfare function

Conclusion

Even though the regression-equations and other estimated parameters are uncertain owing to the instability of the individual answers, obvious simplifications in the 8 assumptions, etc., the numbers show so much systematism, that the answers of the IP's make sense. They are hardly forced by the formulation of the questions and express without doubt, what is the IP's wish, all things considered. But what this more explicity signifies is hazy, thus e.g. what a low fairness value signifies with regard to organizational policy for graduates, the accession to the universities, political activity, etc. Neither is it known, whether the IP's will be of the same opinion when they are themselves graduates and reach the age 50.

The choice of the welfare function is arbitrary, but that the interpersonal utility comparison is more meaningless than the other assumptions, is not self-evident, and the jump from the ordinal to the cardinal utility does not seem especially unacceptable. Then it seems more proper to revise the concept of rationality completely ¹.

Often attempts have been made to avoid the dilemma of choosing a welfare function, but without much success [cf. Riker, 1961]. As the choice seems to be inevitable, the relevant problem must be, how far the results are robust with regard to variations in the choice of the welfare function.

The reluctance to make this choice seems exaggerated, when simultaneously many assumptions, which are not only disputable, but also fundamentally erroneous (e.g. the assumption of individualistic preferences), are accepted without much ado ².

In the last resort the welfare-problem can only be solved, if everybody desires the same total state of things [Arrow, 1963: 69], and therefore a liberalistic welfare theory based upon individualistic preferences may be viewed as a self-contradiction. The method used here reflects these difficulties and has no general applicability. Apart from more special problems, such as the possibility that the IP's may answer deliberately deceitfully for tactical reasons [Arrow, 1963: 80], the level of precision aimed at is probably also fundamentally impossible. But some qualitative conclusions appear from the data, the most interesting of which seems to be, that beyond a certain point a smaller income is preferred to a larger: the utility for the individual depends on the situation of other persons, so that the thoughtless self-interest is insufficient to describe at least some of the IP's. Probably, this is no great surprise to anyone, but it is contrary to one usual and often tacit assumption in economic theory.