Changes in tax/GDP ratio also have the drawback of mixing the effect of tax policy and changes in GDP
composition. Let us assume for example that gross compensation as a share of GDP increases while
income from property of households as a share of GDP decreases. As in many EU countries taxation of
labour is higher than the taxation of income from capital, the tax/GDP ratio will be pushed up by the
change in the structure of GDP, even if no change in tax policy.
The tax mix may give some insights into the quality of public finance. The usual categories (direct and
indirect taxes, social security contributions) are perhaps not the most relevant ones, since they essentially
rely on administrative concepts. In such an approach payroll taxes, social security contributions and
personal income tax should be split in the three categories while they are part on the tax wedge on wages.
We prefer to use the split provided by the European Commission that separates out taxes and social
security contributions in five categories (a) labour employed, (b) income from self-employment, (c)
capital, (d) consumption and (e) social transfers, with a sixth but non-additive category for the
environmental taxes.
Implicit tax rates (ITR) are more interesting indicators since they use the national accounts to assess
separately the taxation of labour, capital and consumption. They build on the categorisation of taxes that
we suggest to use for the description of the tax mix and relate each of these sub-categories of taxes to a
macro-economic measurement of the tax base.
It is worthwhile to remind that the political debate from which they originate is related to the quality of
public finance. The work on these indicators started in the mid-nineties, following the White Paper of the
European Commission “Growth, Employment and Competitiveness”. Chapter 9 of this publication
suggested changing the tax mix in a way that could be more pro-employment. During the discussion on
the follow-up of this policy recommendation, the point was made that we lacked from indicators to
assess the taxation of labour, capital and consumption. The work culminated in a regular publication of
The European Commission on the structure of taxation. The last edition, renamed “Taxation trends in the
EU” contains interesting improvements of this methodology: taxation of companies and taxation of
savings are estimated separately and an indicator for the taxation of energy is suggested^5.
The main advantage of the ITR methodology is its disaggregated approach. It illustrates how the change
in the tax mix, if any, translates into changes in the taxation of labour, consumption and capital or simply
reflects GDP composition effects. It obviously says more on the tax policy stance than the single
tax/GDP ratio does. The ITR methodology may help in the assessment of the quality of tax revenue
because it disentangles the tax revenue in components that have different effects on growth and
employment.
Some of ITR are however more meaningful than others. The ITR on wages is a straightforward indicator
and it is generally easy to make the link between the tax policy stance and the changes in the ITR over
time. This is not so straightforward for the ITR on capital. Even when using the disaggregated approach,
that separates out the taxation of companies and the taxation of household savings, it might be very
difficult to explain the changes in the ITR and to link them with the tax policy stance. The main reasons
for this is that the tax base might be very different from national accounts definition of profits of
corporations and of property income of household, and the business cycle also affects the ITR on
corporations.
The main drawback of the ITR methodology is that implicit tax rates are backward looking, while an
assessment of the effect of taxes on economic behaviour requires forward-looking indicators.
The OECD “Taxing wages” methodology 6 provides a more disaggregated indicators for the taxation of
labour. It illustrates how the tax wedge varies according to the wage level (low, average, high) and to
specific family situations (single or married couples, with or without children). This approach has several
(^5) Cf. EUROPEAN COMMISSION (2007) for the most recent results.
(^6) Cf. OECD (2007b).