Brian Nolan, Ive Marx and Wiemer Salverda
Group report states, “it is important to recognise at the outset that different measures of income may be the most
appropriate or the best available for different analytical purposes” (page 11).
Such practical and conceptual complexities may be highlighted by reference to several concrete issues that can
affect comparability across countries or over time. Perhaps the most important relates to income from “imputed
rent”. The Canberra Group recommends that such an income be attributed to people who own their own home and
thus do not have to pay rent, or are paying below-market rent, since they enjoy an advantage over those paying rent
which would otherwise go unrecognised. This can be estimated in various ways (with reference for example to the
market rent of an equivalent home), but in practice it is often difficult to do so consistently and reliably. Most coun-
tries do not include imputed rental income in their income distribution statistics, and that was also the case with the
ECHP and the earlier rounds of EU-SILC. Since 2007 the data produced for EU-SILC has to include an estimate of
imputed rent, with several alternative estimation methods put forward by Eurostat but not all countries following
them (see Eurostat, 2010). It appears that these amounts are not yet taken into account in the income inequality
and poverty indicators produced by Eurostat from EU-SILC, but users of the micro-data need to be aware that this
is included in the dataset.1 The amounts involved can be substantial (see e.g. Frick and Grabka, 2007), so knowing
whether an estimate is included or not is critically important from a comparability point of view. Where estimates
of imputed rent are available, there are conceptual issues involved in deciding whether one wants to include them.
Failure to do so can mislead as to the relative situation of for example older people (many of whom own their own
homes) versus young people who are renting or in the early stages of house purchase, and in measuring overall
inequality there is a strong case for its inclusion where possible, though where this cannot be done consistently
comparability considerations may be more important. However, imputed rent is not equivalent to cash income, in
that it cannot be used to meet other expenditure needs, and this may be very important when the focus is on low in-
comes and poverty. As Atkinson et al. (2002) emphasize, someone on a low income owning a large house benefits
in terms of housing costs, but the substantial rent imputed to them is not available to pay for their food, clothing
or heating; where poverty/low income is the focus, then, simply including imputed rent may not be satisfactory.
Other elements of income that may or may not be included in available measures include the value of home
production and employer benefits. Own production of agricultural produce is still important in some OECD coun-
tries, and in principle should be included in household income, but is ignored in the statistics gathered in many
countries where it is now much less common. Similarly, benefits from employers such as company cars, free or
subsidised meals, payment of housing-related expenses such as utilities or telephone bills, and free or subsidising
1 Eurostat’s instruction to national statistics offices in relation to EU-SILC 2008 is to collect this information but not to include it in total
household income (see Eurostat, 2010).
Page • 12