Measures of the degree of progressivity and the redistributive effects of
taxation are usually based on cross-sectional data for a single year. However,
the income distribution in any single year consists of individuals from a large
number of separate cohorts, distinguished by date of birth. This would not
matter if all individuals obtained the same income in each year of their life
cycle, but it is well known that age profiles of mean earnings display a
“humped” shape and there is a considerable amount of mobility of individuals
within the distribution of contemporaries. At any moment in time, the low
income groups may contain large numbers of both relatively young and old
people, as well as many who, in terms of their lifetime incomes, may be in the
higher deciles. Much redistribution in a tax and transfer system may be
between stages of the life cycle rather than between the members of a cohort,
or between cohorts. Cross-sectional studies of taxation are therefore
potentially misleading as indicators of the progressivity in terms of lifetime
incomes. In addition, the existence of income variability from year to year can
produce a degree of re-ranking when using a longer period framework. For
example, of two individuals with the same present value of pretax income but
different time profiles, the person with the greater variability pays more tax in
a system having increasing marginal tax rates[1]. This is the basis of
arguments for averaging of income for tax purposes: see for example Creedy
(1979) and Moffitt and Rothschild (1987)