Comparable Indicators of Inequality Across Countries
be relevant in different contexts, especially when data are collected by using household surveys and when wages
and salary income and social security benefits are paid on a regular weekly or fortnightly basis.
In EU-SILC, the income reference period for most countries is the calendar year previous to the year in which
the survey is carried out, but there are two exceptions: in Ireland the income reference period is the twelve months
before the date of interview, while in the United Kingdom current income is annualised i.e. weekly figures for
employee income and transfers are multiplied by 52, monthly by 12.
This brings out first that even when an annual accounting period is deemed most suitable, it matters when this
starts and ends: in assessing the impact of the economic crisis, for example, 2008 EU-SILC data mostly refer to
incomes in the calendar year 2007, that is before its main effects will have been felt. A survey carried out in 2008
and focused on income in the 12 months leading up to the survey date, on the other hand, will capture much of
2008 incomes. Secondly, an annual accounting period may not always be used, and may not always be the most
appropriate either. Some surveys focus primarily on “current” income, i.e. income this week or month, and use a
longer period only for more irregular or fluctuating sources such as self-employment income. The most appropriate
reference period depends on the use to which the data are to be put, and current income derived in this fashion may
be more suitable when the focus is on poverty and social exclusion than an annual reference period, which in effect
assumes that households can even out their consumption over the year as income fluctuates.5
The Unit of Analysis and the Income Recipient Unit
As well as the period over which it is measured, one has to decide on the appropriate recipient unit in analysing
income: should one be focusing on individuals and the income accruing to them, or adding up the incomes of mem-
bers of a family or wider household and analysing the distribution across those larger units? (A household includes
individuals at the same address who are not part of the nuclear family, such as grandparents, adult children, or
unrelated lodgers.) There are in fact two different issues here: the core interest is in the situation of persons, which
speaks for the individual being the unit of analysis, but the income accruing to others in their family or household
may be very important to their living standards and command over resources, so the income recipient unit may be
broader - we may analyse individuals but in terms of their family or household income.
The question is then how best to capture the situation of the individual given the data generally available. In-
dividuals such as women working in the home and dependent children may have no income in their own name, but
the income of their household ensures some access to resources via sharing within the household. Standard prac-
5 Atkinson et al. (2002) from this perspective recommended use of ‘current modified income’, i.e. aggregating annualized current regular
components (wages, regular social benefits, pensions, multiplied by twelve if paid monthly) and, for irregular components or those best
collected on an annual basis, figures for the most recent and appropriate period (e.g. self-employment income, capital income, annual
bonuses).
Page • 15