DIRECT TAXATION, TRANSFERS AND RERANKING: SOME EMPIRICAL RESULTS FOR THE UK
| Published date | 01 August 1987 |
| Author | Brian Nolan |
| DOI | http://doi.org/10.1111/j.1468-0084.1987.mp49003002.x |
| Date | 01 August 1987 |
OXFORD BULLETIN OF ECONOMICS AND STATISTICS, 49,3 (1987)
0305-9049 $3.00
DIRECT TAXATION, IRANSFERS AND
RERANKING: SOME EMPIRICAL RESULTS FOR
THE UK
Brian Nolan*
I. INTRODUCTION
While the redistributive effects of taxes and public expenditure have
received considerable attention, most analyses have concentrated on the
allocation of taxes and expenditures over the ranges of the income
distribution. Similarly, in examining the distributional impact of alter-
native policies or reforms, the focus is generally on changes in the
income distribution in terms of decile shares or aggregate measures of
vertical inequality. Recently, however, there has been an increasing
concern with another dimension, namely the reranking of individuals
and households within the distribution brought about by the operation
of taxes and public spending. In going from a pre-tax to post-tax
distribution, for example, in addition to the changes in the shares of
total income going to the bottom/top decile etc., there may be signifi-
cant changes in the income ranking of recipients, and mobility being
reflected in the different composition of the various deciles by post-
rather than pre-tax income.
Much of this recent discussion in the literature has been in terms of
horizontal equity - the extent to which those who are equally situated
are equally treated by, for example, the income tax system. The issues
which arise in attempting to define this concept and apply it, as well as
its relationship with the more usually considered aspect of vertical
equity, have been analysed in Feldstein (1976), Rosen (1978), Atkinson
and Stiglitz (1980), Atkinson (1980, 1981), Plotnick (1981, 1982) and
King (1983). Relatively little is known, however, about the actual
quantitative importance or pattern of reranking within the distribution
as the tax and transfer systems operate. It is this empirical question
which the present paper addresses, using data for the UK.
In particular, the reranking of households brought about by personal
direct tax (income tax plus employees' National Insurance contribu-
tions) and social security transfers is quantified. In doing so, data on a
large sample of households gathered in the Family Expenditure Survey
* This paper draws on material contained in a doctoral thesis for the University of London.
I am grateful to my supervisor, A. B. Atkinson, for comments. An earlier version of the paper
was presented to the Congress of the European Economic Association in 1986.
273
274 BULLETIN
(FES) is used, but with one important difference compared with most
analysis based on this source. The income concept used in the FES is,
for the most part, current weekly income. In examining the reranking
brought about by the tax and transfer systems, though, income over a
longer period - usually annual income - would generally be considered
more appropriate. This is particularly true for taxation, where the
income tax system is based on annual income, and weekly income and
tax paid may differ substantially from the annual totals. The results
presented here therefore make use of estimates which may be made of
the annual income of FES respondents, based on a methodology
described in Nolan (1983).
The results on the extent of reranking within the distribution will be
presented largely in the form of transition matrices based on the deciles
of the distribution - that is, the position of households in the distri-
bution by decile of pre- and post-transfer and pre- and post-tax incomes
will be compared. As far as direct tax is concerned, the results lend
support to the suggestion by Atkinson (1980), made on the basis of
purely illustrative data for the US, that the diagonal elements of the
matrix would be dominant, with the off-diagonal elements adjacent to
the diagonal quite large but other off-diagonal elements close to zero.
In other words, direct tax shifts a significant number of households up
or down by one decile in ranking, but very few by more than that. For
social security transfers, the results show a somewhat greater degree of
reranking than for direct tax, though again only a small proportion are
shifted more than one decile.
One surprising result appears from the comparison of the extent of
reranking before and after taking account of differences in household
size and composition. Such differences are an important reason why
households might be treated differently by the tax and transfer systems,
and rerankings produced because of these differences would not neces-
sarily be considered inequitable. However, the results set out below
show that the degree of reranking is in fact greater, rather than less,
when incomes are adjusted for the size and composition of the house-
hold by using adult equivalence scales.
In the next section, the major features of the data-base used are
described. In Section III the results on reranking of households pro-
duced by direct tax are described, while Section IV sets out those for
social security transfers. Section V looks at the combined effects of
transfers and direct tax taken together, and finally the main findings are
brought together.
II. THE DATA
The Family Expenditure Survey has been widely used as the basis for
the analysis of distributional issues, in particular the redistributive
To continue reading
Request your trial