2.1.2. Broadness of the tax base
An alternative approach for the assessment of efficiency and neutrality might be test the broadness of the
tax base. Such an approach faces several difficulties. Statistical information on the tax base is scare
compared to data collection on tax revenue.
The OECD initiated a study on personal income tax base in the early nineties^12 but did not update this
very interesting approach. As far as we know, there has been no other tentative to gather comparable
information on the tax bases across countries and over time.
Comparison will obviously be made with figures from the national accounts but a departure from
national accounts does not necessarily mean a move away from neutrality. It might be the result of
conceptual differences between the tax system and national accounting.
This point is particularly relevant for the taxation of income from capital. Neutrality requires uniform
taxation of income and capital gains, subject to adjustment for inflation. Consequently, the tax base of a
neutral tax system should include capital gains that are not included in the national accounts.
The C-efficiency ratio of VAT^13 may be used to test the broadness of the VAT tax base. It divides an
implicit tax rate of VAT (VAT Revenue divided by national consumption) by the nominal tax rate. A
high registration threshold or an extensive use of exemptions will reduce the C-efficiency ratio, what is
consistent with the intention of the C-efficiency ratio to test for the broadness of the tax base. Reduced
rates will have a similar effect. The indicator might however be ambiguous: non-deductible VAT will
increase the C-efficiency ratio, what could be interpreted as a progress neutrality while this is more a
departure from neutrality.
2.2. Income distribution
KAKWANI (1977) provided a well-known methodology to assess the effect of taxes and transfer on
income distribution. Redistribution is computed as the difference between GINI Indexes of income
before tax (or transfer) and after tax (or transfer) and further separated out in two components: the effect
of the average tax rate and the effect of progressivity (defined as the difference between the distribution
of taxes and the distribution of pre-tax income). Refinements of the method have been proposed by
PFÄLHER (1990) to separate out the effect of the progressive tax schedule and of various tax credits.
2.3. Effects of tax incentives
Information on tax expenditures is available in most EU countries^14. The cost of the budget is usually
computed by using the “revenue forgone” approach. This method estimates the cost of tax expenditures,
one by one, without taking into account any behavioural effect.
As tax incentives result in tax expenditures, this could be a convenient way to assess the effect of tax
incentives. We however face two problems: (a) there is no common definition of the “benchmark
system” from which tax expenditures depart and this makes impossible any comparisons between
countries on the use of tax expenditures and on the magnitudes of tax incentives; (b) coverage varies
between countries and some important tax incentives may not be included in the tax expenditures reports.
(^12) OECD (1991)
(^13) See OECD (2006) for example. EUROPEAN COMMISSION (2007) also uses the C-efficiency ratio to discuss the
broadness of the VAT tax base and to discuss the figures from the ITR on consumption.
(^14) See POLACKOVA e.a (2003) for the theoretical aspects of tax expenditures reporting and recent experiences from
developed and transition economies.