Poverty and Social Exclusion in Europe
European comparisons and the impact of enlargement
John Micklewright and Kitty Stewart
While the UK tackles poverty and social exclusion at home, these are also issues
prominent on the wider European agenda. At the Lisbon summit in March 2000,
building on a long-standing commitment to economic and social cohesion in the EU,
the European Council declared that the number of people living below the poverty
line and in social exclusion in the Union was unacceptable. The Council called for
decisive steps to eradicate poverty, beginning with the setting of targets for particular
indicators.
These indicators need first to be defined. Furthermore, no analysis of poverty and
social exclusion in the EU can be made without considering the impact of its
expansion to include some of the Central and Eastern European countries, whose
applications are currently on the table.
Defining indicators
Should poverty in any EU member be measured in relation to a national standard or a
European one? Poor households in the UK are conventionally defined as those with
income below some fraction of average household income, for example 50 per cent of
the median. Should a poor household in the European context be one with less than 50
per cent of the EU median?
In defining the poor as being those whose 'resources (material, cultural and social) are
so limited as to exclude them from a minimum acceptable way of life in the member
state in which they live’, the European Council came down clearly on the side of a
national approach to poverty. For the present, a national standard is more informative
– although at the same time the nature of the EU as a community makes it essential to
keep an eye on what a single poverty line implies.
Attempting to compare social exclusion across countries raises other difficulties.
Enormous questions remain about how to define and measure this slippery concept
and the problems are compounded in an international context. That which constitutes
exclusion in one country, it might be argued, may not do so in another.
The 1998 Eurostat Task Force on Social Exclusion and Poverty emphasised the need
for a variety of indicators focusing on the overlap between low income, lack of work
and disadvantages concerning a range of non-monetary aspects of life. For practical
reasons, the Task Force recommended use of the European Community Household
Panel (ECHP) to investigate exclusion, since it provides comparable data on many
dimensions of living standards for most EU members – allowing consideration of
overlaps between different aspects of deprivation. The drawback is that the analysis of
poverty and exclusion is restricted to what can be measured with these data.
Human poverty in the EU
Before drawing on Eurostat’s recent work with ECHP, we therefore consider other
work that has compared poverty and exclusion in rich industrialised countries. The
United Nations Development Programme's (UNDP's) Human Development Report
1
jumped into the debate in 1998 with the presentation of its Human Poverty Index-2
(HPI-2), which aimed ‘to capture the multiple dimensions of poverty in a composite
measure’. The index gives equal weight to four measures of deprivation, chosen to
represent four dimensions of life:
income – the percentage living below 50 per cent of national median income
length of life – the percentage of people not expected to live to age 60
education – the percentage of 16 to 65 year olds classified as functionally illiterate
social inclusion – the percentage of long-term unemployed in the labour force.
Side-stepping arguments over the technical details of the index, the HPI-2 is an
illuminating place to begin a cross-country comparison of deprivation. The income
measure is a conventional one (although Eurostat now classifies as poor all those
below 60 per cent of the national median). There are good arguments that long life
and literacy are examples of necessary social and cultural resources, though note that
in both cases the standard used is a common international line.
The illiteracy at issue here is not the narrowly defined inability to read, but a lack of
capacity to absorb appropriate information in a modern society: to read instructions on
a medicine bottle correctly, for example. Long-term unemployment must be, and
often is, considered a strong risk factor for social exclusion in many communities due
both to its impact on the skills of an individual, and hence future employment
prospects, and to separation from the social interaction that comes with employment.
Table 1 UNDP Human Poverty Index 2
HPI-2 Poverty Short life Illiteracy Long-term
(%) (%) (%) (%) unemployment
(%)
Sweden 7.6 8.7 8.5 7.5 2.7
Netherlands 8.2 6.2 9.2 10.5 1.9
Finland 8.6 3.9 11.1 10.4 3.1
Denmark 9.3 6.9 12.7 9.6 1.5
Germany 10.4 5.9 10.5 14.4 4.9
Luxembourg 10.5 4.1 10.4 n.a. 0.9
France 11.1 8.4 11.1 n.a. 5.2
Spain 11.6 9.1 9.9 n.a. 10.2
Italy 11.9 12.8 8.9. n.a. 8.1
Belgium 12.4 5.5 9.9 18.4 5.5
UK 14.6 10.6 9.6 21.8 2.1
Ireland 15.0 9.4 9.8 22.6 4.4
Notes HPI-2 values are a non-linear function of the constituent indicators – see source for further
details. In calculation of the index for France, Italy, Luxembourg and Spain, functional illiteracy is set
by UNDP at the average OECD value. Poverty is defined as income below half the national median
(year sometime in 1987-97). Short life is the percentage of persons not expected to survive until age 60
(1998 values). Illiteracy is the percentage of persons not reaching prose level 1 in the OECD
International Adult Literacy Survey. Unemployment is the number of long-term unemployed as a
percentage of the labour force in 1998. Darker shading indicates a worse performance.
Source UNDP (2000)
The performances of EU member countries on the index both as a whole and on its
component parts are shown in Table 1, with countries ranked by the index. Overall,
the UK ranks poorly on the HPI-2, coming eleventh out of twelve EU countries for
which sufficient data are available – only Ireland does worse. (It should be noted that
2
the data come from various years in the 1990s.) The Scandinavian countries and the
Netherlands top the table, with Germany, France and Spain falling in the middle. No
functional literacy data are available for France and Spain, however, and this
potentially affects the rankings.
The breakdown of the separate components shows mixed performances across the
board, with no country ranking consistently in either the top or bottom third of the
distribution. The UK’s performances are perhaps the most polarised, comparing
favourably on life expectancy and unemployment, but badly on poverty and illiteracy.
Finland, Denmark and Germany all do relatively well on illiteracy, poverty and
unemployment, but have shorter life expectancy than neighbouring member states.
Unlike Eurostat’s ECHP, each measure of deprivation within HPI-2 is based on data
from a different source. Hence the overlap at the household level between low income
and other dimensions of life cannot be drawn out.
Household deprivation
Table 2 draws on Eurostat’s recent major analysis of poverty and social exclusion
using ECHP (Eurostat, 2000). It shows the percentage of the poor (in terms of
income) in each country who are also classified as deprived on the basis of other
indicators. The countries are ranked by the income poverty rates in the first column,
the poverty threshold being defined as 60 per cent of the national median income.
Income poverty varies, as in Table 1, by a factor of about two to one. Note that both
the data source and the poverty line differ between the two tables.
Table 2 Income poverty and deprivation among the poor – 1996
Income per cent of poor who:
poverty have great can't afford a have bad or who meet friends or
rate difficulty mak- week's very bad relatives less than
(%) ing ends meet holiday away health once a month
Denmark 12 10 24 10 7
Netherlands 12 14 38 6 4
Luxembourg 12 14 44 16 14
Austria 13 15 49 14 12
Germany 16 8 29 10 8
France 16 18 67 11 13
Belgium 17 12 47 10 9
Ireland 18 29 68 4 1
Spain 18 36 80 14 3
UK 19 17 64 13 6
Italy 19 15 70 15 n.a.
Greece 21 43 85 15 3
Portugal 22 31 86 40 13
EU 17 18 59 15 8
Notes Finland and Sweden not included. Averages unweighted. Poverty defined as income below 60
per cent of national median. Health and social contact indicators refer only to those aged over 16.
Source (Eurostat, 2000a)
How do the poor cope in general terms with their low incomes? In no country does a
majority of the poor actually say they have difficulty making ends meet, although in
Ireland, Spain, Portugal and Greece around 30 to 40 per cent of those on low income
3
do. Much higher numbers feel excluded from the opportunity of a week’s holiday
away from home on account of their low income – over half the poor in 7 of the 13
countries and below 40 per cent in only Denmark, the Netherlands and Germany. The
higher figures in the countries with higher income poverty rates reflect in part that
these are countries with lower average incomes – more holidays away from home are
one of the benefits of rising national income. In Spain, Greece and Portugal, the three
countries with the lowest GDP per capita in the EU, over 40 per cent of the non-poor
say they cannot afford holidays away compared to only 10 per cent in the Netherlands
and Germany (not shown in the Table).
Self-reported bad health is more common among the poor than the non-poor in every
country but the table shows that only in Portugal does the figure for the poor get quite
high. The last indicator, social contact, seems a convincing measure of one aspect of
social exclusion. Although there are reasons why it could relate to income poverty, the
Eurostat report does not show large differences between the poor and non-poor for
this indicator within each country. Lack of contact is always higher for the poor but
the differences may not be significant.
As far as the differences across countries are concerned, Table 2 shows that social
networks appear in general to be stronger in countries with higher poverty rates. Only
three per cent or less of poor people in Ireland, Spain and Greece meet friends or
relatives less than once a month, compared to between 12 and 14 per cent in France,
Austria and Luxembourg. Portugal is the oddball here: one in eight poor Portuguese
claim to go for one month without meeting friends or relatives, separating them from
the norm for Southern Europe.
One possible extension of this sort of analysis is to define the poor as being those with
both low income and deprivation in other dimensions. This is exactly what is done in
the Irish National Anti-Poverty Strategy (Nolan, 1999).
A common EU poverty line?
So far, the discussion of income poverty has assumed a national poverty line, but
there is a case for considering the EU as a single community with a single poverty
line. This case will grow stronger as integration proceeds. How different would
national poverty rates look if life necessities were measured relative to a single
European norm, defined as 60 per cent of median household equivalised income?
The poverty rate in the 13 member states included in the ECHP would range from one
in every 20 persons in Denmark and Austria (and even fewer in Luxembourg) to
almost one in two in Portugal and one in three in Greece and Spain (Eurostat, 2000).
But while poverty rates are inevitably higher under an EU line in countries with
below-average national income, a sizeable share of the poor is still found in the richer
countries: 35 per cent in France, Germany and the UK (see also Atkinson, 1998).
The potential impact of enlargement
The question of whether poverty in Europe should be measured using a single EU
standard is particularly important when the prospective enlargement of the Union to
include new members from Central and Eastern Europe (CEE) is considered.
Disparities in national income per head, among current member states, are small
compared to those between members and applicants. GDP per capita in the ten CEE
4
countries being considered for membership ranged, in 1997, from 23 per cent
(Bulgaria) to 68 per cent (Slovenia) of the EU average. Slovenia falls just behind
Greece, the back-marker among the current members.
By extension, enlargement will mean dramatic shifts in where poverty is found if a
single EU poverty line is used, common to all members old and new. The majority of
people in most of the applicant countries will classify as poor. In a recent article,
David Piachaud argues that the five states likely to be admitted first will have an
average poverty rate of almost 60 per cent if poverty is defined as living on less than
50 per cent of median consumption in the expanded EU. In the second round states,
poverty rates are even higher, reaching 80 per cent in Latvia.
If a single EU-wide standard is used as the measure, enlargement would also result in
a drop in measured poverty in existing EU member states. Average incomes in the EU
as a whole would fall with the accession of the CEE countries and any poverty line
expressed as a percentage of the average (or median) would fall with it. Piachaud’s
figures show a poverty rate for the EU-15 of 17 per cent, falling to 13.5 per cent once
lower living standards in accession countries are taken into account.
Of course, there is an extent to which this is meaningful. If poverty depends on what
others have, our own needs become less urgent as we consider a wider world
community. At another level, it underlines why national poverty lines continue to be
the more informative: someone who has difficulty making ends meet in Ireland or
Portugal today will continue to have difficulty in an expanded EU.
How will enlargement affect the picture of poverty in the EU if national benchmarks
are used? The limited data easily to hand suggest that the situation in prospective
member states is far from uniform. Using a half-median income line, child poverty
rates in both the Czech Republic and Slovakia were at just 2 per cent in 1992, lower
than the best performance in the EU. This is an illustration of the low income
inequality inherited from the socialist period (Bradbury and Jäntti, 1999). But child
poverty was at 14 per cent in Poland in 1992 and 12 per cent in Hungary in 1994,
roughly on a level with rates in Germany and Spain, though considerably lower than
those in Italy and the UK. In the second half of the 1990s relative poverty rates will
have been affected by significant increases in income inequality. If more recent
estimates were available for all prospective entrants, the best would probably put
performance in many current members to shame, while the worst might set new low
standards.
Wider measures of poverty and social exclusion support the impression that the
accession states are a heterogeneous group. Figure 1 shows under-five mortality rates
(U5MR) for the EU-15 and the ten CEE applicants for 1997. Slovenia and the Czech
Republic are holding their own with current EU member states, while rates in
Bulgaria, Latvia and Romania are at least twice as high as in the worst EU performer.
It is worth noting, however, that child survival rates are impressive even for these
countries, when their levels of national income are taken into account. While Romania
looks to be out on a limb, its U5MR of 26 per 1,000 live births compares to 42 in
Brazil and Turkey (in 1998), with both countries having similar levels of GDP per
capita in purchasing power parity terms. Elsewhere, we show that the CEE applicants,
5
on average, had U5MRs in 1996 at nearly 60 per cent less than would be predicted on
the basis of per capita GDP and 25 per cent less than in EU member states
(Micklewright and Stewart, 2001). A strong history of investment in public services
means that straight comparisons of national income do not do these countries justice.
Slovenia may have per capita GDP at 68 per cent of the EU average, but it will bring
the challenge of positive example in areas like child mortality. This is something
completely overlooked by media and academic focus on the relative economic
positions of applicant states and current members.
Figure 1 Under 5 mortality in EU members and applicants – 1997
Luxembourg 5
Sweden 5
Finland 5
France 6
Austria 6
Germany 6
Netherlands 6
Spain 6
Denmark 7
Italy 7
UK 7
Ireland 7
Belgium 7
Greece 8
Portugal 9
Slovenia 6
Czech Republic 8
Slovakia 11
Hungary 12
Poland 12
Estonia 13
Lithuania 13
Latvia 19
Bulgaria 19
Romania 26
0 5 10 15 20 25 30
probability of dying before age 5, per 1000 live births
Notes 1995 data for Belgium. 1996 data for Ireland, Denmark, Finland and Sweden.
Source UNICEF TransMONEE and WHO Health for All databases
Widening the EU poverty focus
Hard outcome measures of living standards also get very little attention in the
European Commission’s annual reports on the progress of applicant countries towards
accession. The word ‘poverty’ appears a total of just six times in the ten 2000 country
reports, three of them in the report for one country (Bulgaria). Other parts of the
Commission make some effort, but more could undoubtedly be done. At a time when
the current Union members are placing increased emphasis on fighting poverty and
6
exclusion, as this article has shown, the situation in the applicant countries deserves
greater monitoring.
7
References
Atkinson AB (1998) Poverty in Europe Blackwell, Oxford
Bradbury B and Jäntti M (1999) ‘Child Poverty Across Industrialized Nations’
UNICEF Innocenti Research Centre, Florence (see www.unicef-icdc.org)
Eurostat (2000) European Social Statistics: Income, Poverty and Social Exclusion
Eurostat, Luxembourg
Eurostat (2000b) Analysis of wave 3 (1996) of the ECHP reported in European Social
Statistics: Income, Poverty and Social Exclusion
Micklewright J and Stewart K (2001) ‘Child Well-Being in the EU – and Enlargement
to the East’ in Child Well-Being, Child Poverty and Child Policy in Modern Nations:
What do We Know? Vleminckx K and Smeeding T (eds), The Policy Press, Bristol
Nolan B (1999) ‘Targeting Poverty: Lessons from Ireland on setting a national
poverty target’ New Economy, 6:1
Piachaud D (2001 forthcoming) Poverty and Enlargement of the European Union
Centre for the Analysis of Social Exclusion paper, LSE
UNDP (2000) Human Development Report 2000 UNDP, New York
8