The problem of the evaluation of 'real' national income has
recently received considerable attention. Hicks, ^† Kuznets, ^‡
Little ^§ and Samuelson ^∥ have made significant alterations to the
foundations laid by Pigou. ^¶ The purpose of this chapter is to
survey the ground fairly rapidly and to examine some of the
interpretations which have been suggested for index numbers of
the type the national income statisticians habitually construct.

In Chapter V we discussed in a 'comparative-statical' way the
effect of a given change on welfare. That is, we did not try to com-
pare welfare in one period with welfare in another period. We
asked instead what would happen to welfare in a given period if
some hypothetical change were made. We were concerned with
contemporaneous comparisons. We saw that the problem can be
tackled quite directly whenever we are given a well-defined W,
but that we have to content ourselves with occasional statements,
based on shifts in efficiency loci, about potential welfare (or
'potential desirability') whenever we know no more than that W
is a Paretian welfare function.

Very much the same applies when we turn to consider inter-
temporal comparisons of well-being--comparisons of welfare in
one period (or 'year') with welfare in another. But some additional
complications arise. These can be made to stand out most clearly

-156-