Food consumption of low income groups in Poland and Belgium

Full text

SGGW
Food consumption
of low income groups
in Poland and Belgium
Warsaw University of Life Sciences Press
Food consumption of low income groups in Poland and Belgium

Food consumption
of low income groups
in Poland and Belgium
Warsaw University of Life Sciences Press
Warsaw 2007
Editors
Renata Januszewska, Krystyna Rejman, Jacques Viaene

© Copyright by Warsaw University of Life Sciences Press, Warsaw 2007
Cover design: Daria Kieruczenko
Editorial staff: Zofi a Orowska, Ewa Ramus
ISBN 978-83-7244-909-2
Warsaw University of Life Sciences Press
166 Nowoursynowska St., 02-787 Warsaw, Poland
phone (0 22) 593 55 20, fax (0 22) 593 55 21
e-mail: wydawnictwo@sggw.pl
www.wydawnictwosggw.pl
Printed by: Agencja Reklamowo-Wydawnicza A. Grzegorczyk, www.grzeg.com.pl

Acknowledgements
The authors of this publication express their gratitude to the University
of Gent, Programme of Bilateral Scientifi c and Technological Cooperation
2003, for fi nancial support during three years of research.
Part of the analysis was prepared within the frame of the “social
cohesion” programme of the Belgian Federal Public Planning Service Science
Policy entitled “Explorative policy-oriented research with respect to social
stratifi cation in the purchase and consumption of food items”.
We would like to appreciate the contribution of the State Committee for
Scientifi c Research (KBN) of the Republic of Poland to the support on the
part of Warsaw University of Life Sciences (WULS – SGGW).
Finally, we wish to thank the low income people who contributed to the
survey and organisations involved in supporting the nutrition of the poor.
The authors,
Prof. Dr. ir. Jacques Viaene Prof. WULS Dr Barbara Kowrygo
Dr. ir. Renata Januszewska Dr in. Krystyna Rejman
Prof. Dr. ir. Xavier Gellynck Dr in. Ewa Halicka
Department of Organisation
Department Agricultural Economics and Economics of Consumption
Gent University Warsaw University of Life Sciences
(WULS – SGGW)
Coupure Links 653 ul. Nowoursynowska 166
9000 Gent 02-787 Warsaw
Belgium Poland
Tel: +32 (9) 264 59 45 Tel: +48 (22) 59 37 140 (141)
Fax: +32 (9) 264 62 46 Fax: +48 (22) 5937147
renata.januszewska@UGent.be barbara_kowrygo@sggw.pl
jacques.viaene@UGent.be krystyna_rejman@sggw.pl
xavier.gellynck@UGent.be ewa_halicka@sggw.pl

Contents
Foreword ............................................................................................ 5
Chapter I. Determinants of food consumption ................................ 7
Chapter II. Specifi cities of poverty phenomenon in Poland ............... 14
Chapter III. Methodology applied to analyse food consumption
of low income population groups .................................... 31
Chapter IV. Food consumption in Poland and Belgium on the base
of secondary data ........................................................... 42
Chapter V. Analysis of food consumption and nutritional behaviour
of low income consumers in Poland ................................ 62
Chapter VI. Analysis of food consumption and nutritional behaviour
of low income consumers in Belgium ............................. 88
Chapter VII. Comparison of food consumption and nutritional behaviour
of low income groups in Poland and Belgium ................. 104
Chapter VIII. Food policy programmes for low income groups ............. 115
Conclusions ........................................................................................ 123
References .......................................................................................... 125
Annexes .............................................................................................. 136

Foreword
Food consumption in the last decades has changed signifi cantly in the
European countries. The directions of changes were not always in compliance
with contemporary dietary recommendations and guidelines. The different
factors infl uencing dietary changes include: policies of food supply, pricing
and technology, production, marketing activities and public health messages.
However, it should be stressed that for low-income consumers economical
status of the household and cost of food are still decisive determinants
infl uencing food choice. The abundance of food assortment, including highly
processed and convenience products on the one hand and low purchasing
power of vulnerable population groups on the other hand, push the choice
towards products which are not valuable from the nutritional point of view.
Changing food consumption patterns, lifestyle and consumer behaviour
caused wide-spreading prevalence of obesity and other non-communicable
diseases, contributing to overall mortality and morbidity in Europe.
In the nineties of the XX century two major processes infl uenced the
general conditions of European societies living, namely the transformation
of Central and East European countries into market economies and as
a consequence of this the enlargement of the European Union. These two
phenomena lead to economical development and improvement of markets
of all goods and services, including food products and out of home food
consumption.
In spite of visible symptoms of an overall progress, economical and
social deterioration of part of population, and poverty and hunger are among
the most important social problems of today’s Europe. It is very diffi cult to
effectively counteract these issues and to prevent them. Having in mind
these circumstances it was valuable to focus on understanding the
structural changes in the two countries representing the “old” and “new”
Europe, namely Belgium and Poland. The book presents the outcomes of
research on socio-economic aspects of quality of life, directed towards the
improvement of food consumption and upgrading health status of low-
-income population groups in both countries.

Research teams from Gent University in Belgium and Warsaw
University of Life Sciences (WULS – SGGW) in Poland developed and carried
out the project entitled “Social stratifi cation in food consumption in Poland
and Belgium” that was accepted in the frame of bilateral scientifi c and
technological cooperation of UGent (BOF No 01150904). The project started
in February 2004 and ended in December 2006.
The fi rst and second chapter of the book give a background to the issues
investigated within the project. In chapter three the methodology applied
in the research to investigate food consumption of low-income groups is
described. Comparative analysis of food consumption in Poland and
Belgium on the base of secondary data is the topic of chapter four. The
next two chapters (number fi ve and six) present the results of fi eld research
regarding food consumption and nutritional behaviour of low income
consumers both in Poland and Belgium. Food consumption and nutritional
behaviour of low income groups in both countries are compared in chapter
seven. The last chapter, number eight, deals with food policy programmes for
low income groups on the background of different countries’ experiences.

Chapter I
DETERMINANTS OF FOOD CONSUMPTION
Renata Januszewska, Barbara Kowrygo, Jacques Viaene
1. Introduction
Food is the most important need of human being. The process of addressing
food is multidimensional and serves many purposes. The fi rst and foremost
one is to satisfy hunger and keep the person alive. Second, it allows to
enhancing peoples’ physical growth and keeps them in good health.
Subsequently, it makes them fi t for work, and helps stay intellectually and
physically active. And the last one is to enable the person to perform social
functions and feel good, which is related to hedonistic feelings.
Human behaviour is governed by economic, biological, demographical,
socio-psychological and cultural factors which form an integrated system.
These factors determine consumption changes as well as hierarchy of
nutritional needs, and may act as stimuli, constraints or circumstances.
The factors vary in importance from society to society and from individual
to individual.
2. General determinants of food consumption
Existing literature offers various detailed classifi cations and descriptions
of determinants which affect overall consumption (demand) structure, and
food generally (Meulenberg and Viaene, 2005). Many divisions or models are
similar to one another, although there are differences in the way hierarchy
is stressed or in the infl uence exerted by particular factors.
Figure 1 presents a general synthesis of food consumption determinants
– cause and effect relation. A general food consumption framework is
defi ned by external environment of household and household members. It is
shaped by country’s natural resources such as food production capacity,
overall market and economic situation, law and fi nance systems, marketing
actions undertaken by food producers and traders, organizations of social
life, and government policy concerning agricultural, food and nutrition.
Consuming subjects, in this case households and individual consumers, do
not have a direct infl uence on the above mentioned determinants, therefore

EXTERNAL INTERNAL
LEVEL AND STRUCTURE
FOOD CONSUMPTION
HOUSEHOLD ENVIRONMENT
Geo-biological
determinants
Climate
Soil and topography
Natural environment
A
gricultural
production
Economi
c
and political
determinants
Social and economic
policy of the State
National income
Income breakdown
A
gri-food policy
Foreign trade
Food
econom
y
related
determinants
Educational
determinants
Progress
determinants
A
gricultural technology
Competitiveness o
f
domestic
food industry
Food supply
Trade structure
Catering services
Goods prices and thei
r
relations combined with
marketing actions
done by companies
School education
Outside school education
Research and development
Tertiary education
Number of household
members
A
ge and se
x
Family development stage
Location
Physical development state
Physiological state
State of health
Housing resources
(including kitchen facilities
and equipment)
Income level and
income source
Effective food demand
Education
Job
Social group membership
Working activities
of women
Food and eating
knowledge
Eating habits and
preferences
Eating tradition
Fashion and its pursuance
Loyalty towards
producer/brand
Demographical
and environmental
determinants
Health and
physiological
determinants
Economi
c
determinants
Social and
work related
determinants
Psychological
and social
determinants
Method of nutrition
State of nutrition
State of health
FIGURE 1. Food consumption determinants
Source: Kowrygo, 2000.

Determinants of food consumption 9
their behaviour is marked by adjusting to the existing situation. As far as
internal environment of consumer is concerned, food consumption is mostly
determined by household income, price and availability of products and
services.
The internal conditions of consumer behaviour are linked with motives
which drive a particular consumer: his or her personality, learning and
memorizing skills, preferences, likings, and nutritional habits. In the second
half of the twentieth century it was observed that some of food consumption
determinants are at the same time tools which can be used for shaping food
consumption policy. The most signifi cant ones are income, price, and food
supply (Koodziejek, ed., 1980).
It ought to be stressed that the two groups of determinants related to
macro- and micro-environment of each consumer/household are integrated
into a dependency chain – method of nutrition – state of nutrition – state
of health – marked by consumption level and structure which is also
infl uenced by other determinants (e.g. genetic ones).
The last few decades have seen changes in European dietary patterns;
not always positive. Several factors shape dietary changes: policies on
food supply, pricing and technology, production, promotion activities and
public health messages. A combination of consumer demand and commercial
investment in mass production largely determines the direction of changes
(WHO, 2002).
Almost forty years ago Périssé et al. proved that there are relationships
between the development of a given country expressed as GDP per capita
and the structure of an average diet (Drewnowski and Popkin, 1997). The
authors declared that increased GDP is accompanied by bigger share of
sugar, animal and vegetable fat in the daily food intake as well as a sharp
drop in energy obtained from complex carbohydrates. The share of proteins
in daily diet energy remains almost unchanged, but in rich countries
as opposed to poor countries, animal proteins prevail. There are also
alterations within the energy obtained from animal fat. The richer the
country the bigger the share of visible fats as well as of the invisible fat
from animal products; and the lower the share of invisible vegetable fat.
The trends in changes in the energy structure of the diet have been proved
among others by FAO/WHO research done in the 1980s, as well as studies
by Posner et al. (1994) and Chevassus-Agnes (1994) at the beginning of
1990s.
The need of changes in contemporary human eating patterns call for
a profound knowledge of both general determinants of food consumption
and particular factors determining food choice and nutrition.

10 R. Januszewska, B. Kowrygo, J. Viaene
Figure 2 shows the relationship of a wide set of factors infl uencing the
choice of food. It also indicates the importance of public policy in many
sectors ranging from agriculture and food processing, manufacturing and
trade to retailing, catering and advertising all of which shape the availability
and accessibility of food.
FIGURE 2. Factors infl uencing food choice
Source: WHO, 2002.
The Pan-European Survey of Consumer Attitudes to Food, Nutrition and
Health found that the top fi ve infl uences on food choice in 15 European
Union countries are quality /freshness (74%), price (43%), taste (38%),
trying to eat healthy (32%) and what my family wants to eat (29%). In the USA
the following order of factors affecting food choice has been reported: taste,
cost, nutrition, convenience and weight concerns (Glanz et al., 1998).

Determinants of food consumption 11
European Food Information Centre, whose goal is to enhance the public
understanding of nutrition and food safety made a list of food choice factors,
which include (EUFIC Review, 2005):
• biological determinants such as hunger, appetite and taste,
• economic determinants such as cost, income, availability,
• physical determinants such as access, education, skills (e.g. cooking)
and time,
• social determinants such as culture, family, peers and meal patterns,
• psychological determinants such as mood, stress and guilt,
• attitudes, beliefs and knowledge on food.
The complexity of food choice is obvious from the list above, which is
in itself not exhaustive. Food choice factors also vary according to life stage
and the power of one factor will vary from one individual or group of people
to the next. Thus, one type of intervention to modify food choice behaviour
will not suit all population groups. Rather, interventions need to be geared
towards different groups of the population with consideration to the many
factors infl uencing their decisions on food choice (EUFIC Review, 2005).
3. Socio-economic status as a determinant of food choice
It must be underlined that household income and the cost of food is an
important factor infl uencing food choice, especially for low income consumers.
Low income may restrict the ability to buy food on the basis of health and
limited access to healthy foods (Dowler et al., 1997; James et al., 1997).
When socio-economic differences in consumption of food and nutrition are
observed, they are generally in line with health inequalities. However, the
poor in affl uent societies do not suffer from dramatic energy and nutrient
defi ciencies; in fact over-nutrition is more common than under-nutrition.
A special report on health inequalities published in the journal of
Social Science and Medicine in 1990 revealed striking differences in many
countries between social groups in regard to structural and lifestyle factors.
Studies show that lower socio-economic groups have greater incidences
of premature babies, low birth-weight babies, hearth disease, stroke, and
some cancers (Carr-Hill, 1990; Spruit, 1990; Wnuk-Lipiski, 1990; Aiach
and Curtis, 1990; Duch and Sokoowska, 1990; Rodriguez and Lemkow,
1990; Lagasse et al., 1990; James et al., 1997). Risk factors including lack
of breast feeding, smoking, physical inactivity, obesity, hypertension, and
poor diet are clustered in lower socio-economic classes. The diet of these
groups provides energy from foods such as meat products, full cream milk,

12 R. Januszewska, B. Kowrygo, J. Viaene
fats, sugars, canned products, potatoes and cereals. This diet is also low in
vegetables, fruit and whole-wheat bread (Krebs-Smith et al., 1995; Li et al.,
2000; Krebs-Smith and Kantor, 2001; Simila et al., 2003). As a consequence,
poor people have fewer intakes of essential nutrients such as calcium, iron,
magnesium and vitamin C than people in higher social classes.
Socio-economic status is represented by multiple indicators including
income, education, and occupation, all of which may operate independently
or interact in leading to inequalities that infl uence food choices. Socio-
-economic indicators are proposed in many studies (Stronks et al., 1996;
Cantillon et al., 1999; Duncan et al., 2002). A review of fi fteen European
countries showed that high educational level is associated with a more
healthy diet (Roos and Prättälä, 1999; Irala-Esteves et al., 2000). A higher
education level tends to be associated with a wider knowledge of a healthy
diet (Margetts et al., 1997). Differences in the intakes of energy yielding
nutrients are less evident but those with high education tend to have
a smaller intake of fat. Obesity also varies by socio-economic status. A study
based on twenty six (mostly European) populations shows that lower edu-
cation is associated with higher Body Mass Index (BMI) in approximately
half of male and in almost all of female populations (Molarius, 1999). Many
studies indicate the importance of socio-demographic indicators such as
employment status and age (Frank-Stromborg et al., 1990), urbanisation
and region (Cook, 1990), and social and cultural determinants (Lagasse
et al., 1990). Other studies show the positive relationship between gender
and food choices (Furst, 1991; Lupton, 1996; Grogan et al., 1997). Socio-
-economic indicators are also proposed in many studies (Cantillon et al.,
1990; Stronks et al., 1996; Duncan et al., 2002).
Additionally, prices and incomes have been reported as signifi cant
explanatory variables for food purchase in developed countries (Ritson
and Petrovici, 2001). Prices and consumer incomes could account for 97%
of variation in the US demand for food (Huang, 1985). This implies that
economic factors are key indicators of food choices, particularly in low
income groups (Dobson et al., 1994, Kaufman et al., 1997). The recent study
of Duncan and others (2002) shows that economic indicators are considerably
more sensitive than traditional ones, such as education and occupation.
However, the study also shows the difference in relative rates of morality
between two groups of respondents under and above 64 years of age. In this
regard, the earlier fi ndings of Lehmann and others (1990), which show that
people in lower socio-economic groups live shorter, are confi rmed.
Many analyses indicate that low income groups spend a very high share
of the total budget for food, confi rming Engel’s Law (Engel et al., 1986).

Determinants of food consumption 13
According to Maseide (1990), it is more important to focus upon vulnerable
groups, like low income groups, than on all socio-economic classes in
order to plan effective health policies that aim to improving health
conditions of the whole population. Importance of such research is
recognised and national studies have been launched inside (FSA, 2004) and
outside Europe (CBPP, 2004).
4. Final remarks
There are many infl uences on food choice and state of nutrition and health
of the consumer. Among them the education level, employment situation,
demographical and socio-cultural determinants are most often underlined.
However the economical factors, especially household’s income and the cost
of food are the decisive factors determining the access to food products and
possibility of purchase. The food choice of low income group of consumers
depends in the highest degree on these factors implying a greater tendency
to consume unbalanced diets resulting in under- (micronutrient defi ciency)
or over-nutrition (overweight and obesity).
For this reason it is important to pay attention and to analyze the
direction of changes in nutrition of low-income consumers, representing
mainly so called vulnerably groups: children, young, older, living alone,
homeless or unemployed people.
Increasing number of low income population in modern/affl uent
society shows that it is necessary to develop both governmental and non
governmental actions, which integrate local committees and aim at improving
living conditions including accessibility to shops and availability of food
affordable to people with low income.
Literature indicates that general changes in nutrition behaviour are
not always positive. Therefore, there is a need for multi-aspect monitoring
of food consumption and formulating as well as implementing a new and
well-focused national nutrition policy.

Chapter II
SPECIFICITIES OF POVERTY PHENOMENON IN POLAND
Magorzata Radziukiewicz
Institute of Home Market and Consumption (IRWiK), Warsaw
1. Introduction
Among priority goals of the social policy in many countries, there is
a struggle against poverty and social exclusion. The programmes devoted
to this goal require identifi cation of the social groups most endangered by
poverty and requiring help. There is the need to answer the questions:
Who is poor? What is the incidence and degree of poverty? Has the sphere
of poverty extended since the time of economic transformations? In this
chapter, an attempt to answer these questions is undertaken from both the
methodological and empirical points of view.
After a short presentation of methodological assumptions adopted
in the survey1 – the empirical results of the analysis of the sphere of poverty
in Poland are presented. An assessment of the scope and degree of poverty
is made in the separated socio-economic and demographical profi les of
households, and the results obtained are compared with the results of
research carried out earlier (based on household budgets in the 1990s).
Identifi cation of poverty and determination of its dimensions are not
simple tasks for the reason that poverty is a multidimensional phenomenon.
The analysis of the sphere of poverty is extended by factors other than
income2 describing the situation of separated households in terms of their
housing conditions, health and education. An alternative to use single
measures of poverty might be the synthetic measure3, based on information
on separated aspects of poverty, i.e. income, housing conditions, health and
education.
1 There are defi ned: poverty, data sources, ways of measuring its scope and depth.
2 The analyses based on income have the longest tradition among studies on poverty
worldwide.
3 A way that takes into account the multidimensional concept of poverty is aggregation
(combination) of various measures into one index.

Specifi cities of poverty phenomenon in Poland 15
2. Poverty concepts and measures
2.1. Defi nition of poverty
A few dozens years ago, an answer to the question “What is poverty?” was
somehow simple and precise. The majority of poverty defi nitions include
inability to obtain the goods essential for living, of which food is the most
obvious element. Poverty was measured with a degree of under nutrition
or with per capita consumption. An alternative approach defi ned the
person (or household) as indigent if he (it) did not have at his (its) disposal
a suffi cient income to obtain food to meet basic needs. Despite many
discussions on a precise defi nition of both ‘economic resources’ and ‘basic
needs’, this approach has been popular and used in surveys till now. In
1976, the International Labour Offi ce (ILO), presenting their programme for
fi ght against poverty and unemployment, established that the basic needs
of the population in each country should be taken into account. In result
of broadening the concept of basic needs by additional aspects (dwelling,
education, health care and culture), the defi nition of poverty has also been
broadened. The poor is that human being who has failed to reach the
minimum level of consumption of goods or services considered as basic.
The current concept of poverty is defi ned in a broader way – it includes
social and political aspects of life deterioration. The Council of Ministers of
the European Community was the fi rst to formulate it in 1984 (Eurostat,
1990): “Poverty relates to persons, families or groups of people whose
material, cultural and social measures are limited to such a degree that the
level of their life declines beyond an acceptable minimum in the country
of residence”.
The United Nations Development Programme in its Report in 1991
states that: “Most people are poor as they do not have land, capital, credit
or the possibility to obtain a fair job (…). They do not have access to welfare
services on an adequate level (…). Moreover, they admit that poverty relates
also to freedom and human rights” (World Bank, 2003).
Defi ning poverty in this broad context, as a complex social phenomenon,
they admit existence of the three dimensions of poverty: economic, human
and political. Economic poverty is a result of an unequal distribution of
resources: capital and land, and improper exploitation of those resources.
Human poverty is connected with an insuffi cient access to the essential
facilities indispensable for just existence. A low or none share in the
decision-making process is defi ned as the political dimension of poverty.

16 M. Radziukiewicz
With broadening the scope of basic needs, a shift is made from the
concept of understanding poverty as merely lack of resources for meeting, to
the concept of understanding poverty as lack of opportunities and limitation
of choices resulting from social and personal contingencies, necessary for
conducting a valuable life.
All the efforts aimed at broadening the traditional defi nition of poverty
explain the essence of the phenomenon but make measurement more
diffi cult (Kordos, 1996).
2.2. Sources of information on poverty
Poverty has many faces, depending on the place, time, and it may be
described in many ways. However, the most important in surveys on poverty
are statistical sources, published mainly by the Central Statistical Offi ces.
Information on consumption and income is obtained by representative
surveys where households are asked to reply to detailed questions on
expenditure and sources of income.
2.3. Measurement of poverty
To measure poverty, several highly important issues appear:
• defi nition of the degree of wealth,
• choice of the poverty line, i.e. the threshold below which a given
household or person is considered as poor,
• choice of the indicator of poverty that is only applied in relation to the
whole population or subgroups.
Selection of the degree of wealth, the poverty line, equivalence scale
and basic measures of poverty are very important as they infl uence the
results of measurement. Different methods sometimes lead to contradictory
conclusions.
2.3.1. Indicators of wealth – income or consumption
Income or consumption are indicators of wealth estimating poverty with the
use of money-based measures.
It is known that consumption better refl ects a current living standard
and the possibility to meet the basic needs in the household. Consumer
expenditure refl ects not only goods and services the household may have
at its disposal based on its current income, but also it is important to have
access to the credit market.

Specifi cities of poverty phenomenon in Poland 17
The main benefi t to use income as the measurement of poverty is the
possibility to compare income with the data from other sources, e.g., with
salaries, pensions, etc. Irrespective of whether income is used or consumption,
it is highly important that the data collected from the household or individuals
are fair and reliable.
2.3.2. Equivalent scales
Households of a different size and demographic composition have different
needs and this fact is not easy refl ected through the measures of poverty.
Variety of the needs is taken into account by the equivalence scales.
The use of the equivalence scales, budgets of households of various
types may be altered accordingly to their needs4 adjusted due to the social
and demographic features of the household as the scales equivalence are
rates (indicators) that allow determining volumes of income for different
types of the households based on the income fi xed for the representative
household. Overall, this leads to dividing the total income by the number
of persons adjusted with value of scale related to the number of persons5.
If the scale adopts value 1, and the reference family is one person, then value
of the adjusted income is equal to the per capita income amount.
In practice, the set of equivalence scales that are used is wide (Szulc,
2004). Thus, e.g., in international surveys on wealth, inequalities and
poverty related to households, carried out in the second half of the 1980s
within the framework of Luxemburg Income Study, the scales are classifi ed
in terms of the purposes they are aimed at (Buchman et al., 1988). Two
groups of the scales are distinguished namely, the scales of an expert nature
and the scales based on surveys and constructed with use of the multidi-
mensional analysis. In the fi rst group of scales, there are the scales that
pursue only statistical purposes, for example to estimate the number of
people who live below the poverty line (the STAT scale) as well as the
scales serving as the tool to determine benefi ts of social programmes (the
PROG scale). The second group of scales is estimated based on surveys on
consumption expenditure (the CONS scale) or based on assumptions of
suffi cient consumer’s income (the SUBJ scale). To build the latter, questions
concerning estimates of one’s income, minimum income, have to be
answered. Moreover, values of the scales differ as follows:
4 Since it is known that per capita minimum expenditure in the several-person household will
be lower due to certain fi xed costs, irrespective of the number of persons.
5 DE = D/LOSe, where: DE – adjusted (equivalent) income; D – total income in the household,
LOS – number of persons in the household; e – the scale in respect of the number of persons.

18 M. Radziukiewicz
• the STAT scale: 0.72,
• the CONS scale: 0.36,
• the PROG scale: 0.55,
• the SUBJ scale: 0.25.
Therefore, different values of income in relation to the number of
persons in the household depend on the construction of equivalent scale.
An equivalent scale takes into account the size of the household and
other demographic features of the household, e.g., age of the household’s
members. The OECD scale assigns the value 1 to the fi rst adult person,
the value 0.7 to each further adult person, and the value 0.5 to each child
aged below 14. In this case, the 3-person family consisting of two adults and
a child below 14 will have the value of the scale equal to 2.2. The equivalence
scale being equal to 2.7 for the household consisting of 3 adults and one
child means that this household must spend 2.7 times more than the
one-man household to achieve the same level of consumption. It is considered
that the OECD scale assigns a relatively too large importance to many-
-person families. Therefore, in the modifi ed OECD scale, the second and
every further person has the value of 0.5, and to each child below 14 the
value of 0.36. It is proper to add that the above-mentioned scales are standard
in the EU countries’ statistics.
2.3.3. Poverty lines
When income, consumption or non-pecuniary measures are defi ned for
an individual or household, the next step is to defi ne one or more poverty
lines. Poverty lines are the terminal points separating the poor from those in
better situation. The lines may be pecuniary (e.g., a defi nite level of income
or consumption) or non-pecuniary (e.g., a defi nite level of literacy). There are
two main ways for determining the poverty line: relative and absolute.
Relative poverty lines are defi ned in relation to an overall distribution
of income or consumption in the country, e.g., the poverty line may be fi xed
at 50% of the national average income or consumption.
Absolute poverty lines are set up within certain absolute standards
that households should agree to meet their basic needs. For the pecuniary
measures, these absolute poverty lines are often based on estimates of costs
of the basic nutrition needs, i.e. the food-related basket considered as the
minimum for health of the typical family, to which there is added provision
for non-food-related needs.
6 This change results from a diminishing share of expenses on foods in the households’
budgets. With an increase in the number of persons in the household, expenses on foods grow
signifi cantly faster than, e.g., expenses on housing.

Specifi cities of poverty phenomenon in Poland 19
Sometimes, alternative poverty lines are based on subjective measures
of poverty determined by respondents.
Setting up a proper poverty line is a diffi cult matter. Poverty lines differ
from one another over time and space, and each country uses the lines that
are relevant to its level of development, social standards and values7.
In the European Union countries, the poverty lines amount to 60%
of the median of average equivalent income or the arithmetic mean of
households’ expenditure.
2.3.4. Poverty indicators
It is obvious that income or expenditure may not adequately refl ect
a complex, multidimensional phenomenon as poverty. Nevertheless, when
it comes to a quantitative analysis of poverty, it is usually restricted to the
income aspect, and mainly monetary.
The income aspect of poverty (income poverty) comprises both the
synthetic and specifi c measures (Kordos, 1991). Having the information on
income and poverty line, the problem is to apply adequate total measure
summing up individual measures for assessing the phenomenon of poverty.
Such measures are aggregated measures of poverty (Panek et al., 1999).
There are many alternative measures of poverty and mostly three of them
are used (World Bank, 2004):
1. Incidence of poverty (headcount index) – is the share of population
whose income (or consumption) is below the poverty line. This means the share
of population that may not afford the basket of essential commodities.
2. Poverty gap index – is a measure giving information on how far poor
households are from the poverty line8. It is obtained by adding all “defi cits”
of the poor households and dividing this sum by the number of poor
households. The above measure is described as the cost of elimination
of poverty (in relation to the poverty line), as expressed in money9. This
measure is used for non-pecuniary indicators assuming that the distance
is signifi cant. e.g., the poverty gap index in education may be a number of
years in education necessary or required to reach the defi nite threshold of
education.
3. Poverty severity index – this index measures not only the distance
that separates the poor from the poverty line (poverty gap index) but also
7 The poverty line value fl uctuates from 30 cents per day in the poorest countries to a few
dollars or so in rich countries.
8 The concept of “income defi cit”.
9 It means that every poor will receive exactly such amount, which is necessary for them to get
rid of poverty.

20 M. Radziukiewicz
inequality among the poor. It means that a greater attention should be paid
to those households, which are placed further from the poverty line.
An alternative for use of the above-specifi ed measures of poverty may
be combining the information about various aspects of poverty, such as
creating the measure called “well-being composite index” that takes into
account income, employment, health and education.
2.4. Other measures of poverty
2.4.1. Human poverty index
Evaluation of poverty in all the dimensions is the best way of poverty
measurement. This goal is achieved by a system of human indices.
The United Nations Development Programme (UNDP) developed an index
measuring the social aspect of poverty that is called the human poverty
index (HPI) (World Bank, 2003). This aggregate index is constructed for
both the developing countries (HPI-1) and for the OECD countries of high
incomes (HPI-2). Both indices include information on the three dimensions
of quality of life: long and healthy life, knowledge, and decent standard
of living. The HPI-2 index additionally takes into account another aspect of
life – social exclusion.
In case of the advanced countries, the human poverty index HPI-2 is:
HPI-2 =
1
1234
1()
4PPPP
a
aaaa
⎡⎤
+++
⎢⎥
⎣⎦
where:
P1 – percentage of people whose life expectance does not exceed 60 years;
P2 – percentage of people being functional illiterates;
P3 – percentage of people whose income is below the poverty line, which is
below 50% of median of per capita disposable income;
P4 – rate of a long-term (lasting 12 months or longer) unemployment;
α – indices level of poverty.
The human poverty index is a concept providing an indication about
poverty in both rich and poor countries. It is a standardised measure, with
a value ranging from 0 to 100.
2.4.2. Social exclusion indices
In 1999 at the summit in Lisbon, the EU developed a programme of social
actions. Marginalisation, besides poverty and unemployment, is considered
as a barrier for socio-economic development. Having effective methods to

Specifi cities of poverty phenomenon in Poland 21
reduce these phenomena is one of the main objectives of the social policy
of EU member states.
In order to carry out an effective policy of counteracting social exclusion,
a system of information is needed that would allow to diagnose precisely and
to compare the situation in the EU Member States. Adoption by the Member
States of a set of jointly developed and agreed indicators10 covering the four
most important areas concerning social exclusion (income, employment,
health and education) serves this objective (MLSP, 2004).
The following indicators are connected with income (poverty gap):
− the poverty line fi xed at the level of 60% of median of an equivalent
income disposable in a given country;
− incidence of poverty prior to and after taking into consideration of social
transfers according to the gender;
− incidence of poverty according to:
• socio-economic groups,
• age,
• type of economic activities,
• gender,
• type of ownership of fl at;
− incidence of poverty with adoption of different poverty lines (the so-called
index of dispersion around the poverty line);
− degree of poverty;
− differentiation of equivalent income distribution (Gini index, relation
of quintile V to quintile I).
The indicators related to employment are:
• households without employed persons,
• rate of long-term unemployment,
• coeffi cient of variability of the regional employment rate.
The indicators related to health are:
• life expectancy,
• self-assessment of the state of health.
Education indicator is only represented by the number of young people not
continuing education in secondary and higher-level schools.
Research is under way on widening and supplementing the list of indices
of social exclusion by new ones, specifi c for each country. The list of indicators
in the area of diagnosis of material situation of the poor includes the indices
measuring threat of poverty, the indicators on effectiveness of various forms
10 At the summit in Laeken in December 2001, 18 statistical indices on income, poverty and
social exclusion are approved.

22 M. Radziukiewicz
of assistance. The measures on the issue of employment relate the “invisible
categories”, i.e. the people looking for a part-time job, temporarily employed
on contracts, or those employed for a defi nite time-period.
3. Social geography of poverty in Poland
3.1. Premises
This part of the study consists of poverty incidence in Poland and its spread
among the groups of households according to income level and other
selected socio-demographic characteristics.
An analysis of the relative poverty was carried out setting up the poverty
line at the level of 60% of the value of median of equivalent income for the
whole set of households that consists of 529.5 PLN. Below this poverty line,
there were 16.8% of households in Poland in 2003. The index measuring
the degree of poverty (determining an average “defi cit” of income in relation
to the poverty line for the poor population) reached the value of 27.1%. This
means that a household with the equivalent income below 60% of value of
median has an average defi cit in the monthly budget of about 143.3 PLN.
3.2. Poverty in Poland
3.2.1. Poverty by socio-economic group in 2003
The group least vulnerable to poverty are, besides retired people’s house-
holds (only 6.5% of the households are poor in this group), the households
of self-employed people and employees’ households (Fig. 1). The households
living on non-earnings sources dominate the group of households with
income below 60% of the value of median. Also for farmers’ households, the
group of poor is 2.5 times larger than in the case of total population. Thus,
earlier conclusions are confi rmed on dominance of farmers’ households and
the households living on non-earnings sources in the group of the poor.
The social groups with high percentages of the poor are characterised by
high degree of poverty. The degree of poverty – as for the European standards
– is quite signifi cant, and in the poorest households, it amounts to 38.6%.
“Defi cit” to equivalent income in the agricultural households exceeds 200 PLN,
whereas in the households living on non-earnings sources approaches
190 PLN. The households of retired people need an amount of 108.8 PLN
to leave the sphere of poverty.

Specifi cities of poverty phenomenon in Poland 23
3.2.2. Poverty by household composition
A high number of children in the family is a main determinant for
a disadvantageous position of the household; i.e. the most disabled in terms
of poverty are families with the highest number of children11. It turns out
FIGURE 2. Poverty incidence and poverty gap according to the type of family
11 With three or more children as opposed to not numerous families (with one child or two
children).
FIGURE 1. Poverty incidence and poverty gap by socio-economic status of the household head
0
10
20
30
40
50
percent (%)
employees
employees-farmers
farmers
self-employed
retirees
pensioners
non-earned sources
headcount poverty gap
incidence
0
5
10
15
20
25
30
35
40
45
50
percent (%)
witho ut childr en 1 child 2 child ren 3 & mor e
childr en
single parent
with c hildren
other
headcount poverty gap
incidence

24 M. Radziukiewicz
that the presence of the third child in the family increases the probability of
a signifi cant decline in the material living standard of the family and drives
such family in the poverty area.
In 2003, more than 46% of the families with 3 or more children were
in the sphere of poverty. In the case of families with four or more children,
almost half of them had at their disposal a monthly equivalent income lower
or at least equal to 529.5 PLN. In addition, the remaining indices – depth and
severity of poverty – point out to a diffi cult situation of numerous families.
The smallest poverty incidence is among the childless families (Fig. 2).
3.2.3. Poverty by level of education of the head of household
The greatest poverty incidence (above 24%) occurs among the households
with the lowest status of education – elementary, technical and primary
education. Just these groups are characterised by both a high frequency of
occurrence of poor households and a signifi cant depth of poverty.
Quite a large depth of poverty is attributed to the household, in which
its head has a higher education; however, there are not many poor households
in this group, only 2.4%. The fact of signifi cant differentiation of income takes
also place among poor households whose head has secondary education
(Fig. 3).
FIGURE 3. Poverty incidence and poverty gap according to the level of education of the head
of household
0
5
10
15
20
25
30
percent (%)
university secondary vocational primary
headcount poverty gap
incidence

Specifi cities of poverty phenomenon in Poland 25
3.2.4 Poverty by the province
An analysis of the data from Table 1 gives a signifi cant differentiation of the
incidence of poverty in provinces although not so signifi cant as in the case
of poverty in the socio-economic groups, or in the households with different
number of children.
TABLE 1. Incidence, depth, severity of poverty and HPI in households by province in 2003 (%)
No Province / Voivodship Poverty
incidence1
Poverty gap
index
(P1)
Poverty severity
index
(P2)
HPI
α = 3
1 Mazowieckie 12.58 25.25 10.21 22.13
2 lskie 14.76 25.76 10.13 21.83
3 Opolskie 14.95 30.20 14.99 24.70
4 Wielkopolskie 15.71 27.31 11.67 22.02
5 Lubuskie 15.92 23.84 9.25 23.49
6 Pomorskie 16.23 26.37 11.21 22.75
7 ódzkie 16.25 28.02 12.32 25.13
8 Maopolskie 16.43 25.87 10.44 22.68
9 Zachodniopomorskie 17.44 27.52 11.57 24.58
10 Dolnolskie 18.17 29.26 13.02 23.30
11 Kujawsko-Pomorskie 18.65 29.05 13.27 24.34
12 Lubelskie 19.34 28.45 12.57 26.09
13 Warmisko-Mazurskie 20.53 26.09 10.53 27.94
14 Podkarpackie 21.57 26.11 10.83 25.62
15 witokrzyskie 22.60 27.18 12.37 26.64
16 Podlaskie 23.44 28.00 12.70 28.12
1Percent of poor households.
Source: Radziukiewicz, 2003.
The Mazowieckie Province has the lowest percent of poor households,
whereas the agricultural provinces of eastern and southeastern Poland, i.e.
Podlaskie, Podkarpackie and witokrzyskie Provinces are characterised
by the highest extent of poverty. More than 1/5 of households in these
provinces are in the sphere of poverty. In addition, the agricultural
Warmisko-Mazurskie Province has a signifi cant incidence of poverty
exceeding 20% of the households. Perhaps, it relates to the fact that
unemployment and poverty is larger in the countryside than in towns.
In six provinces: Kujawsko-Pomorskie, Dolnolskie, Zachodnio-
pomorskie, Maopolskie, ódzkie and Pomorskie, 16% to 19% of households
are below the poverty line. The lowest percent of poor households (below
16%) is in the following provinces: Mazowieckie, lskie, Opolskie,
Wielkopolskie and Lubuskie. Difference in terms of the poverty incidence
between the Mazowieckie Province and the Podlaskie Province amounts to as

26 M. Radziukiewicz
many as 10.86%, and shows that regional approach is an equally important
factor in the policy to reduce poverty in the country and to alleviate its
effects.
3.3. Cumulating of poverty
Poverty in Poland is obviously not limited to a single element, which is low
income. Another factor of poverty relates to housing conditions (Beskid,
1995). Housing conditions are mainly determined by the fact of possession
of house or an apartment.
An indicator on “housing poverty” is the information about the house-
holds dwelling in premises without bathroom. Housing conditions are also
determined by the per capita living space. It is assumed that there is poverty
in those households having a per capita fl oor not bigger than 10 m2. Poor
families are usually renting municipality fl ats or sub-renting residential
premises and paying social rent (Table 2).
Discussing each of the above-mentioned elements of housing poverty
separately, gives the result of a highly dispersed situation. About 15.5%
of households have cheap residential premises, almost 11% of households
have overcrowded fl ats (having less than 10 m2 per capita), and 8.3% of
households have no bathroom.
The relationship between human poverty and housing poverty is not
obvious. Diffi cult housing conditions for one-person households correspond
with almost 21% of poor households, i.e. being below the income-related
poverty line. About 25% of poor households have no bathroom, and 22% are
renting the cheapest residential premises. Housing poverty combined with
poverty gap concerns only 3% of households. About 50% of households are
located below the income-related poverty line and not belonging to the
housing property group.
In order to reach poor households, access to knowledge is an important
criterion. Percentage of households that do not have computer and access to
Internet refl ects poverty in the sphere of education (“education poverty”).
Another sphere of poverty – besides income, bad housing conditions,
and low education – is health. Even without clear analyses, the situation is
that illness moves families towards poverty.
The symptoms of poverty result in different groups of poor households
as presented in Table 2. Obvious is the fact that the families may have low
income though they have computers or good housing conditions. It also
appears that the poverty gap is related to one of the forms of housing

TABLE 2. Cumulating of poverty in households in 2003
Specifi cation
Percent of poor households
total
with symptoms of housing poverty with
income
< than
expendi-
ture
without
computer
or Internet
with
a disabled
person
without
bath-
room
minimum
fl oor
cheap
premises
at least
with one of
symptoms
Total 16.82 24.26 20.89 21.65 47.50 58.63 85.24 17.64
Socio-economical type of household:
Employees 13.45 17.93 29.61 27.22 51.52 52.75 79.91 0.88
Employees – farmers 22.56 19.65 18.86 1.96 33.01 59.14 85.66 1.18
Farmers 40.94 24.47 11.28 0.01 29.36 77.45 87.45 5.53
Self-employed people 13.15 20.00 12.83 23.40 39.62 63.02 71.70 1.13
Retired people 6.48 36.22 8.86 17.32 49.02 60.04 92.72 10.04
Pensioners 21.98 25.79 16.58 21.98 47.35 57.85 90.79 81.69
Living on non-earnings
sources 49.69 30.72 23.64 33.21 57.51 60.57 87.66 9.86
Biological type of a family:
Married couple, childless 5.59 28.84 3.13 27.59 43.89 64.89 97.49 31.35
with 1 child 15.20 17.56 10.89 20.16 38.05 64.07 85.69 16.91
with 2 children 23.98 16.23 22.68 18.67 40.37 58.94 75.37 10.95
with 3 children 40.39 17.25 32.93 20.38 55.53 58.01 75.78 8.89
with 4 and more children 61.10 23.77 56.83 20.77 65.85 49.73 83.33 9.56
Single parents with children 32.88 27.20 21.24 38.60 56.22 61.14 87.05 13.99
Others 14.30 30.64 16.09 19.98 48.86 58.04 90.54 23.27
Number of persons in the household:
1 7.72 48.33 1.00 30.39 60.78 66.96 95.27 27.65
2 9.61 32.41 3.83 28.96 49.89 65.68 94.26 26.92
3 15.07 22.99 9.23 22.24 42.21 62.38 87.75 19.79
4 21.44 18.58 21.46 19.21 40.04 57.70 79.03 14.36
5 31.51 19.55 29.90 18.70 46.81 54.52 79.71 11.83
6 and more 36.21 21.64 50.73 17.24 59.29 51.95 85.58 13.45
Type and size of place of living:
Town ≥ 500 thousand
inhabitants 6.40 29.11 27.74 47.95 61.64 65.41 75.68 21.92
≥ 200–500 thousand 10.61 15.66 21.97 39.14 55.81 61.62 75.76 15.15
≥ 100–200 thousand 12.64 24.66 25.20 44.77 60.86 52.55 83.65 16.35
20–100 thousand 13.71 21.39 24.28 36.76 54.22 57.57 84.05 18.61
below 20 thousand 16.99 19.78 22.98 31.62 50.70 56.55 82.03 18.66
Countryside 26.51 26.97 17.86 6.06 40.14 59.97 88.99 17.14
Education:
Higher 2.44 5.56 13.89 8.33 22.22 67.59 50.93 9.26
Secondary 9.55 11.17 15.53 17.87 32.69 65.58 72.49 12.99
Elementary technical 24.86 19.67 23.04 21.10 45.49 59.12 85.10 14.78
Primary 24.54 38.80 21.24 25.25 59.77 54.88 94.35 24.59
Age of the head of household:
24 years and less 21.85 24.37 26.47 24.37 55.04 52.52 89.50 8.40
25–34 20.07 22.06 31.24 22.58 52.68 55.05 85.88 4.95
35–44 23.79 18.74 25.61 21.18 44.80 57.70 77.79 9.05
45–54 18.96 23.98 16.56 23.15 45.46 63.55 85.74 23.79
55–64 12.07 32.09 9.80 19.59 44.43 60.14 92.57 39.53
65–74 8.20 32.03 10.58 17.83 47.63 59.61 94.15 28.41
75 years and more 6.60 49.01 9.93 17.88 62.91 56.29 96.03 27.15
Position of the head of household:
Blue-collar worker 21.95 19.51 31.18 28.02 53.78 50.89 83.17 0.89
White-collar worker 4.20 8.98 20.70 22.66 38.67 63.28 61.33 0.78
Source: Radziukiewicz, 2006a–d.

28 M. Radziukiewicz
poverty in case of 40.7% of households. In the majority of households
(85.4%), poverty gap corresponds with lack of computer or with lack of
access to Internet and in some households (17.6%) it relates to both income
and disability. A simultaneous occurrence of all symptoms of poverty –
poverty gap (BD), housing (BM), education (BE) and health (BZ) – is recorded
in 1.5% of all households.
Finally, the question rises what are the factors to cumulate various
symptoms of poverty. Taking into account the multidimensional nature of
poverty, primarily a synthetic index of poverty (WSKU) is constructed in the
following form:
WSKU = BD + BM + BE + BZ
When a given syndrome of poverty has occurred in the household,
a relevant partial factor gets the value 1, and in the case of absence of
this syndrome – the value 0. For the households touched with each of the
syndromes of poverty, the poverty index has the value 4 and the value 0 for
the households without any of the “symptoms” of poverty. The above-specifi ed
measure is a sum of partial indices and informs which household is “poorer”,
i.e. when one household has a higher value of the poverty index WSKU than
the other household.
In this way, a specifi c value of the poverty index is assigned to each
household. In 62.4% of households, none or only one of the symptoms of
poverty occurs; in 26% of households 2 symptoms of poverty is a fact, and
10% of households have 3 symptoms. The value of the poverty index (WSKU)
equalling 2 is arbitrarily assumed as a threshold poverty, and this results
in 11.5% of poor households. This index gives an idea about the relation
between poverty gap and poverty factors.
In order to fi nd out the infl uence of each of the factors, a logit model
is developed (Radziukiewicz, 2006a–d). Analysis of the results of assessed
logit function mostly confi rmed the conclusions from the analyses of poverty
carried out so far.
4. Conclusion
Analysis of impact of social and economic factors on a differentiation of
poverty in Poland and comparison with the surveys carried out earlier allow
summarising and presenting the most important conclusions.
The analysis of relative poverty established the poverty line at the level of
60% of median value of an equivalent income of households in 2003. On the

Specifi cities of poverty phenomenon in Poland 29
one hand, there were 16.8% of households below the so formulated poverty
line in Poland. On the other hand, the index determining an average income
“defi cit” in relation to the poverty line for the population of poor reached the
value of 27.1% what indicates that poverty in Poland is moderately deep.
Households of pensioners, and the families composed of more than fi ve
persons (with more than three children) are often exposed to poverty. The
households, whose head has vocational or primary education, being at the
age of 25–34 are also exposed to greater poverty. Defi nitely less poverty
endangers households of the southern and northwestern regions that those
of the eastern region.
In Poland, the incidence of poverty due to unemployment of members of
the household is the largest. The incidence of poverty in case of the house-
holds whose head is unemployed is very high (50%). In diffi cult situation
are also the households whose head remain unemployed for a period logger
than one year (79.7%). Among them, a highly unfavourable situation is in
case of households not receiving unemployment benefi t (80.7%).
The Mazowieckie Province has the lowest percent of poor households,
whereas agricultural provinces of eastern and southeastern Poland, i.e.
Podlaskie, Podkarpackie and witokrzyskie are characterised by the greatest
scope of poverty. In addition, the agricultural Mazursko-Warmiskie Province
has a considerable incidence of poverty, exceeding 20% of household. In
six provinces: Kujawsko-Pomorskie, Dolnolskie, Zachodniopomorskie,
Maopolskie, ódzkie and Pomorskie, below the poverty line there are from
16 to 19% of households. The lowest percent of poor households (16%) is
attributed to the provinces: Mazowieckie, lskie, Opolskie, Wielkopolskie
and Lubuskie.
The index of human poverty calculated for 16 provinces shows that
human poverty in Poland fl uctuates from 24.9% in the Warmisko-Mazurskie
Province to 18.5% in the Wielkopolskie Province. In the two provinces:
Podlaskie and Warmisko-Mazurskie, this index is close to 25% what
indicates that almost 1/4 of the population of these provinces are touched by
human poverty. The largest human poverty is attributed to the witokrzyskie
and Podkarpackie Provinces. The Lubuskie, Opolskie, Maopolskie and
Pomorskie Provinces (value of HPI amounts to 20.1, 20.1, 19.6 and 19.3%,
respectively) constitute the population with the lowest human poverty.
The incidence of poverty was rapidly growing in the period of transition,
and currently, the growing trend is considerably slower. Certainly, optimistic
may be the fact, observed in 2003, of declining of the fraction of poor
households living on pensions and on non-earning-related sources, as

30 M. Radziukiewicz
well as large households. However, irrespectively of positive changes, the
dimension of poverty remains at a high level.
Poverty is divided equally among the households with income below
60% of the median value of equivalent income. Poor households living on
non-earnings-related sources had the lowest average equivalent income and
poor households with higher education had the highest income. The lowest
differentiation of income among poor households is assigned to the house-
holds of pensioners and those whose head is at the age of 65–74, and the
highest to households of farmers and married couples having two children.

Chapter III
METHODOLOGY APPLIED TO ANALYSE FOOD
CONSUMPTION OF LOW INCOME POPULATION GROUPS
Renata Januszewska, Krystyna Rejman, Jacques Viaene, Barbara Kowrygo,
Xavier Gellynck
1. Introduction
The research focuses on problem detecting and policies related to food
consumption patterns among the low social classes in Belgium and Poland.
Even though the background of the relation between nutrition and health is
extremely complicated and until today still largely unknown, the impact of
social inequality on the life and health expectancy is a fi eld of research that
has been growing rapidly during the past decades. Since the publication of
the so called “Black Report” on inequalities in health in the United Kingdom
(Townsend et al., 1979) the scientifi c interest in this topic within the
industrialised world increased.
Social groups have different priorities toward food consumption and
different ideas about what constitutes healthy eating (Lang, 1998). Low
income may restrict the ability to buy food on the base of health and limited
access to healthy foods (Dowler et al., 1997; James et al., 1997). When
socio-economic differences in food consumption are observed, they are
generally in line with health inequalities. However, the poor in affl uent
societies do not suffer from dramatic energy and nutrient defi ciencies; in
fact over-nutrition is more common than under-nutrition.
The term “social class” covers however a quite extensive number of
mutually interacting person-specifi c characteristics of demographic,
economical, psycho-social and cultural nature, the impact of which can
fl uctuate over time and space. In many researches the variables used as
indicators of social class are measured in a different way and indicate
various aspects of social class. An important current topic within this
context concerns the question to what extend differences in health behaviour
(nutritional behaviour) are determined by economical factors like income or
rather by cultural factors that are related to education (Power et al., 1991;
Glendinning et al., 1994). Prices and incomes have been reported as
signifi cant explanatory variables for food purchase in developed countries
(Ritson and Petrovici, 2001).

32 R. Januszewska et al.
Multiple indicators including income, education, and occupation, all of
which may operate independently or interact in leading to inequalities that
infl uence food choices, represent socio-economic status. A review of fi fteen
European countries showed that high educational level is associated with a
more healthy diet (Roos and Prättälä, 1999; Irala-Estevez et al., 2000). Those
with higher education, with the exception of Southern European countries,
tend to consume more vegetables, fruits and cheese and less fats and oils.
A higher educational level tends to be associated with a wider knowledge of
a healthy diet (Margetts et al., 1997). Differences in the intakes of energy
yielding nutrients are less evident but those with high education tend to
have a smaller intake of fat. Obesity also varies by socio-economic status.
The Body Mass Index (BMI) is important in the food consumption studies
since obesity status explains a large share of overall mortality and morbidity
in Europe (Stronks et al., 1996; Puska, 2000). A study based on twenty-six
(mostly European) populations shows that lower education is associated
with a higher BMI in approximately half of male and in almost all of female
populations (Molarius, 1999). Different energy needs as well as cultural
and social factors have been suggested as causes of nutritional inequalities
(Hulshof et al., 1991; Roos et al., 1996).
The diet of the low-income groups provides energy from foods such
as meat products, full cream milk, fats, sugars, preserves, potatoes and
cereals. This diet is also low in vegetables, fruit and whole-wheat bread
(Krebs-Smith et al., 1995; Krebs-Smith and Kantor, 2001; Simila et al.,
2003). As a consequence, poor people have lower intakes of essential
nutrients such as calcium, iron, magnesium and vitamin C than people in
higher social classes. A recent study shows that food consumption among
poor people in Belgium is mostly related to age of respondents and is not
infl uenced by education, profession or region of living (Januszewska and
Viaene, 2005). Results indicate that the total amount of food purchased per
person in a low-income group is not signifi cantly lower than in the medium
and high-income groups.
The objective of the current research is to focus on the low income
respondents in Belgium and Poland and to analyse their food consumption
patterns and nutritional status.
The project called “Social stratifi cation in food consumption in Poland
and Belgium” was developed by University of Gent in Belgium and Warsaw
University of Life Sciences in Poland with the aim to analyse food consump-
tion of low income groups based on primary and secondary data.

Methodology applied to analyse food consumption... 33
2. Structure of the project
2.1. Identifi cation of consumption determinants and existing food
consumption patterns included in chapter IV
There are fi ve specifi c tasks of desk research:
− to analyse and describe differences and similarities between social
classes in the purchase and consumption of food as well as nutrients;
− to identify groups of food items that have the highest proportional
contribution to the intake of nutrients, and to determine in what way
these foods contribute to this intake – high intake, frequent intake,
nutritional density and index of food quality (Sorenson and Hansen,
1975);
− to study differences between social classes on the intake of macro- and
micro- nutrients in foods groups;
− to study the evolution of the proportional budget share within the total
budget for all foods and for specifi c food groups;
− to compare the nutrient supply per head in households with the
recommended daily allowances (RDA) (Murphy et al., 2002) and to
establish the vulnerable groups.
The relevant databases that include nutrition and economic data between
1980 and 2000, in Poland and Belgium, are explored. Budget panel data
developed by the Polish Central Statistical Offi ce (GUS) and Belgian National
Institute of Statistics (NIS) as well GfK (market research organisation) are
examined. These analyses provide insights into three levels of data:
− national food balances (NIS, GUS) and FAO’s Food Balance Sheet database
show trends in food consumption patterns, i.e. food supply per head;
− the expenditure patterns for food and specifi c groups of food – market
purchase, self-provision, and supplementary alimentation;
− the household budget data on food consumption in Poland and Belgium
give some insights in a social gap in food consumption patterns.
Food consumption volume, income, food prices and household compo-
sition data are explored and incorporated into descriptive and econometric
analysis. The analyses are performed within and between the participating
countries. Standard statistical procedures as described in Malhotra (1996),
and econometric analyses (Greene, 1997) are implemented. The examples
of databases that are used for this project are:
• databases on purchase and consumption of food items on household
level,

34 R. Januszewska et al.
• databases on consumption determinants: demographic, economical and
other characteristics of social groups (income, economic growth, GDP,
unemployment, system of social subsidies, food price changes, etc.),
• databases on the expenditure share for food items within the household
budget.
2.2. Empirical research: nutritional behaviour and food consumption
patterns included in chapters V and VI
Empirical research is based on the fi ndings of the project “European Food
Consumption Survey Method” (EFCOSUM), which was undertaken within
the framework of the EU Programme on Health Monitoring in 2001. The
project emphasises the need for co-ordinating nutritional surveillance
activities within the European Union (Ballard-Barbash, 2001). Pan-European
food consumption surveillance on an individual level based on uniform
procedures and 24 hours recalls appears the fi rst choice to provide
comparable dietary intake data and dietary indicators appropriate for use in
the Health Information Exchange and Monitoring System (HIEMS).
A special importance of the current project is attributed to the fact that
Poland is a new Member State and tries to harmonise many regulations
including food and beverages sector. Thus, all efforts in implementation of
research on internationally comparable new data on population food intake
are supported. According to Brussaard et al. (2002) “It is recommended that
any country that carries out a (national) food consumption survey includes the
minimum amount of 24 h recalls to allow a calibration with other countries”.
In the frame of undertaken empirical research three aims were assigned:
− to characterise the low-income respondents in the Belgian and Polish
sample;
− to establish the relationships between BMI (Body Mass Index) respondent
groups on the one hand, and physical activity and health as well as
social functioning related variables on the other hand;
− to fi nd out the relationships between BMI groups, food behaviour, frequency
of food consumption and the nutritional value of diet consumed.
2.2.1. Research model
Taking into account the aims showed above the model developed by Mela
(1996) was adapted for analysing the research results. The model shows
a schematic representation of prominent aspects of eating behaviour, food
preferences and nutrients intake to the development of obesity (Fig. 1). The

Methodology applied to analyse food consumption... 35
model allows for interplay of physiological and cognitive forces which might
act to promote higher fat intakes in association with the susceptibility or
presence of obesity. It is argued that fat containing foods may have greater
reinforcing psycho-biological effects for certain individuals or under certain
conditions. In this way, these foods became more potential stimuli for the
acquisition and maintenance of conditioned preferences. Further on, it is
suggested that increased liking may be mediated through variations in the
stimulation or function of neural mechanisms responsible for hedonic
responses. This assumption leads to the theoretical scenario in which
predisposed individuals gain weight on such a diet.
Mela suggests that it is possible to test the links between characteristics
of metabolism and the existence or acquisition of preferences. Based on
FIGURE 1. Human nutrition and health survey research model
Source: Based on Mela, 1996.
Consumption of a
high fat, energy-
dense die
t
Overeating and
positive fat balance
Increased body fat:
BMI
Weight concern
Cognitive restraint
and/or dieting
Preference for
high fat foods
Exposure to high fat foods: social position
(income, place of living, education, employment)
Genetic/ socio-
demographic
p
redis
p
osition
Successful
weight
control
Low
physical
activity
Weak effect
on satiet
y
Blunted fat
oxidation
and/or other
biological
defects
Experience/
Dietary
management
p
ractices
Food availability
(budget limitation)
and social norms
Loss of dietary
control
Heightened
response to
palatability
Eating
environment
Emotional eating/
well-being
Restoration of
fat balance at
higher weight
Physical Activity and
Health Survey
Food Behaviour Survey
Social Functioning Survey

36 R. Januszewska et al.
such approach, the empirical research in the current project was developed.
The variables measuring the nutritional and psychological conditions for
excessive intakes and poor weight control are combined with physiological
effects of energy-dense, high-fat foods and a heightened responsiveness to
such foods. Mela’s model is linked to three parts of the research questionnaire.
These parts are associated with three aspects: health and physical activity,
social functioning and food related behaviour. Some elements of the original
model are however not investigated. To measure fat oxidation or restoration of
fat at higher weight requires another methodological approach. Understanding
of the relationships between foods and behavioural characteristics and their
links to overeating and obesity can potentially contribute to formulating and
predicting responses to prevention and treatment strategies.
2.2.2. Research questionnaire
The main questionnaire, Food Frequency Questionnaire and 24 hours
Recall Method are integrated into one research questionnaire based on
the approach developed by Ballard-Barbash (2001) and Food and Drug
Administration (FDA, 2003).
The main questionnaire consists of three parts (see the Annex 1):
1. Physical Activity and Health Survey (PAHS) includes the following
variables: internal and external health factors; health behaviour related to
physical activity; diet-related diseases and medical history.
2. Social Functioning Survey (SFS) focuses on three aspects: emotional
eating (well being), social involvement and effects of illness (consequences
of diseases).
3. Food Behaviour Survey (FBS) focuses on three areas:
• knowledge, attitude and preference: knowledge and importance of dietary
guidelines, awareness of diet-disease links; self-assessment of nutrition;
sources of diet and health information,
• food environment: out-of home free meals consumption and food
provision; budget limitation of food consumption,
• actual food behaviour: dietary management practices, food choice and
meal patterns; FFQ and 24 h RM.
In Food Frequency Questionnaire (FFQ) the respondent is asked
about usual frequency of consumption of foods chosen as specifi c dietary
indicators during the last year (Buzzard, 1998b). For the estimation of food
quantities consumed, questions regarding the portion size are included. The
development of the food list is crucial for a successful and reliable data
collection. Ideally, the food list is adapted to the studied population. The

Methodology applied to analyse food consumption... 37
applied food frequency questionnaire consists of 42 food and drink products
representing all food groups. Different food items are recognized as specifi c
dietary indicators e.g. fats and oils, fruit and vegetables, fi sh and fi sh
products. Consumption frequency is recorded on a six-point scale. The
results are presented in two ways: as percentage of occurrence of each
frequency category and as mean frequency for each product. Then, food and
drink products are classifi ed into six types of frequency (Table 1).
TABLE 1. The method of food consumption frequency evaluation
Frequency scale Mean frequency
1 – a few times per year or never 1.00–1.49
2 – one-three times a month 1.50–2.49
3 – once a week 2.50–3.49
4 – a few times per week 3.50–4.49
5 – once a day 4.50–5.49
6 – a few times per day 5.50–6.00
A 24 hours recall method (24 h RM) is conducted in parallel with
FFQ. The method is selected by EFCOSUM group as most suitable to get
internationally comparable new data on population means and distributions
of actual energy and nutrients’ intake. The method is originally attributed to
Wiehl (1942) and includes an interview. A few aspects are important when
applying this method:
1. Training of interviewers. The method is dependent on well-trained
interviewers skilled in the identifi cation of available foods and meals, in
preparation practices used generally in particular population groups. The
interviewers are familiar with the nutritional habits (eg. foods usually
eaten together) in order to be able to get detailed and complete answers
and to control the accuracy of data. For an evaluation, a coding system of
foodstuffs and meals, and also a computerised program are needed
(Beaton et al., 1979; Nelson and Bingham, 1997; Slimani and Valsta, 2002;
De Henauw et al., 2002). The investigator asks the respondent to enumerate
the foods and beverages consumed in the preceding full day, including their
quantity. Due to intra-individual variability, a single 24 h recall does not
represent the usual individual intake but it characterises the average intake
of a group or population.
2. Dietary indicators. Since 24 h recall method registers total daily
consumption, a few determinants of nutrition are selected for comparison
of consumption between population groups and countries. Critical indices
of nutrition are defi ned as the consumption of food products or nutrients
which are signifi cantly related to the current recommendations for proper

38 R. Januszewska et al.
nutrition and prevention of diet-related diseases. The following list of dietary
indicators has been selected as the most relevant, to start with:
• food groups – vegetables, excluding potatoes; fruit, excluding fruit juices;
whole grain bread; and fi sh and shellfi sh,
• nutrients – total fat, saturated (and trans) fatty acids, polyunsaturated
and monounsaturated fatty acids (PUFA and MUFA) as well as proteins
and carbohydrates intakes expressed as percentage of dietary energy
supply (%DES); and alcohol as g/day (WHO, 1990; FAO, 1994).
3. Album of food portions. To facilitate the 24 h recall, the adapted pic-
ture book of portion sizes including country-specifi c dishes, with additional
household and other relevant measures, was presented to the respondents.
The album was developed by the National Institute of Food and Nutrition in
Poland (Szponar et al., 2000).
4. The Nutrient Database and the software. There are many food clas-
sifi cation systems in Europe and therefore, it is diffi cult to fi nd out common
grounds for food classifi cation at the European level (Ireland et al., 2002).
Thus, the national food composition tables are used in the research. These
food composition tables are applied for calculation of the content of nutritional
components in the consumed food in both countries. The food composition
tables are published by the National Institute of Food and Nutrition in
Poland (Kunachowicz et al., 1998) and by NUBEL (2004) in Belgium.
2.3. Comparison of low income consumer groups in both countries
included in chapter VII
This part of the research includes the comparative analysis of the results
of primary data, which are collected by the consumer survey in Poland and
Belgium (the main questionnaire, FFQ and 24 h RM). A pooled data set is
constructed in order to realise cross-country analyses. Body Mass Index
is used to determine food consumption and behaviour of investigated
consumer (Bray, 1990; Berger, 1995; Hunt and Hillsdon, 1996). BMI groups
are calculated utilizing the formula:
BMI = weight (kg)/ height2 (m)
Respondents are grouped in the four standard BMI groups:
• underweight: BMI < 18.5,
• normal weight: BMI = 18.5–24.9,
• overweight: BMI = 25–29.9,
• obese: BMI  30.

Methodology applied to analyse food consumption... 39
The obtained insights in consumer attitude and behaviour form the
input for policy recommendations. In order to explore existing food and
nutrition policies and measures in both countries, four steps were taken:
− fi rst: the types of policies identifi ed;
− second: the measures and interventions taken in both countries
identifi ed;
− third: the effect and outcomes of these policies and interventions
evaluated;
− fourth: the comparative perspective to policies and intervention in each
country assessed.
The intention is to use the framework, developed by Diderichsen (1998)
and refi ned by Whitehead and colleagues (2000), for mapping the impact
of policies on the social pathways to health inequalities. In this framework
(Fig. 2), the pathways leading to illness or health are approached from the
perspective of the individual or the society.
SOCIETY INDIVIDUAL
Context
Policy
Social position
Specific exposure
Disease/ injury
Social consequences of disease
I
II
A
III
IV
B
C
D
FIGURE 2. Framework for researching policy impact on health inequalities
Source: Whitehead et al., 2000.
On the one hand, an individual’s social position (for example defi ned by
gender, occupational class or ethnic origin) may infl uence the exposure to
important health risks such as poverty, nutritional defi ciencies, dangerous
working conditions, health damaging behaviours etc. According to Figure 2,
four important mechanisms to create different probabilities of being exposed
to specifi c health hazards or risk conditions are:
(I) The impact of social position on health through differential exposure.
(II) A specifi c exposure may lead to ill-health or disease of an individual,
depending on whether other contributory risk factors or risk conditions
are present (differential vulnerability).
(III) The effects of the illness on the ability to stay employed, live independently
and participation in the community (differential consequences of disease).

40 R. Januszewska et al.
(IV) Consequences of disease might feed back into a causal pathway, like the
etiological process (investigation of disease causes) and infl uence social
stratifi cation and insurance policies.
On the other hand, the societal perspective focuses on how the prevailing
social context interacts with and infl uences the pathways from social position
to ill-health. Four distinct entry points, at which policy may infl uence the
pathways between social position and health consequences, are, in accordance
with Figure 2:
(A) Policy may infl uence the social position individuals occupy in society
(modifying effect of social context and policy on social stratifi cation).
(B) Policy may infl uence exposure to health hazards faced by people in
different social positions (differential exposure).
(C) Policy may infl uence the effect of being exposed to a hazardous factor
(differential vulnerability).
(D) Policy may infl uence the impact of being ill (differential social conse-
quences of disease).
Policies and interventions to reduce nutritional inequalities can be
universal or selective. Universal measures are directed to the population as
a whole with the idea that all citizens should have equal access to adequate
food and proper diet. The universal approach does not always explicitly aim
to diminish inequalities. The selective measures aim at improving the
conditions of the least advantaged. Food related inequity and the balance
between the universal and selective principles have been central themes in
European food and nutrition policies since the nineteenth century onwards.
Most recent interventions have followed the selective approach, i.e. targeted
at the less advantaged groups, such as low income mothers, ethnic groups
in deprived areas or homeless young people (Prättälä et al., 2002).
2.4. Food policy recommendations for low income groups included
in chapter VIII
The important aspect is to fi nd out the knowledge of the specifi c
characteristics of each country involved in this project in regard to social
circumstances of healthy eating. Cross-national differences and product
category effects (Frewer et al., 2000) are captured and presented. Analysing
such information at the national level permits consumers to make informed
choices about consumption of specifi c foods that are especially interesting
for health status.

Methodology applied to analyse food consumption... 41
The extensive review of food policy in different countries is done in
regard to the dietary goals, dietary guidelines, cooperation in the food chain,
and communication with consumers (FAO/WHO, 1992b, c). Taking these
fi ndings into consideration a proposal of food and nutrition policy is
elaborated for the vulnerable groups of population, based on the information
gained from primary and secondary research.

Chapter IV
FOOD CONSUMPTION IN POLAND AND BELGIUM
ON THE BASE OF SECONDARY DATA
Ewa Halicka, Renata Januszewska, Barbara Kowrygo, Jacques Viaene,
Xavier Gellynck
1. Introduction
Popkin (1994) described the phenomenon of nutritional transition i.e.
the changes in dietary patterns that follow economical development and
urbanization, as being complete in Europe, North America and Australasia.
This concept implies a transition of diet towards an “industrial-type” model
which is relatively low in starch staple foods and high in foods of animal
origin, sugar and processed products. However analysis of population-based
data shows that signifi cant changes are still continuing to take place. In
some cases consumption patterns refl ect the infl uence of non-economical
determinants on consumer behaviour and can be linked to health-oriented
choices of some population groups. Additionally in Poland and other Central
European countries consumption trends were in the nineties radically
modifi ed or even reversed by economical factors due to the development of
market economy.
Assessing, analysing and monitoring nutritional situations are the major
priorities of food policy measures (FAO/WHO, 1992a). The most comprehensive
and widely comparable database used to assess diets internationally comprises
the FAO’s Food Balance Sheets (FBS). The FBS are compiled from a highly
disaggregated set of supply-utilisation accounts (Schmidhuber and Trail,
2006). This source of data delivers internationally comparable data that
refl ects the mean per capita supply of food, available for consumption.
The FAO-based data analysis presented in this chapter focuses on the
period from 1989, i.e. from the introduction of market economy in Poland,
to 2003. For Belgium 1989–1999 FAO data show the combined average
per capita supply of food in Belgium and Luxembourg, from 2000 – data
concerns only the Belgian population.
In the second part of the chapter analysis of household budget data
is shown indicating signifi cant socio-economic diversifi cation of food
consumption in Poland and Belgium. Household Budget Surveys (HBS)

Food consumption in Poland and Belgium... 43
occupy a position between the food balance sheets of FAO and the specially
designed individual nutrition surveys (Elmadfa and Weichselbaum, 2005).
HBS collect data on food availability among nationally representative
samples of households and provide a more detailed and valid description
of the dietary choices of the population. The exploitation of HBS-derived
data for nutritional purposes was evaluated and implemented in the context
of the DAFNE (Data Food Networking) initiative in which national datasets
from 16 European countries were classifi ed into 56 sub-groups. However
in the presented analysis more current HBS data collected in Poland and
Belgium are presented.
2. Comparison of food quantities available for consumption
in Poland and Belgium
In order to compare the FAO data in both countries the most important
food groups were selected and analysed: plant-derived (vegetal) foodstuffs,
animal-derived foodstuffs, fats and oils as well as main nutrients and energy
in diet.
2.1. Trends in vegetal food products’ consumption
According to Food Balance Sheet data the average per capita supply
quantities of cereals and potatoes – the traditional staple foods in Poland are
signifi cantly higher than in Belgium. In comparison to other EU countries
Poland is ranked third (136 kg) in the case of cereals, following Italy and
Lithuania. In Belgium the supply level of cereals and potatoes is below
EU-25 average which was 123 kg in 2003. As shown in Figure 1 (Annex 2)
the mean cereal supply in Belgium constituted only 70% of the Polish level.
The supply of potatoes in Poland is also very high in comparison to all
EU countries and is bigger only in Latvia. In the second part of the nineties
the falling tendency in consumption stabilized at 130–135 kg/capita/year
what constitutes almost 140% of the Belgian level.
Sugar and sweeteners consumption in Belgium was on average 49.4 kg/
/capita/year vs. 44.3 kg in Poland. For Poland, radical price increase in 1990
determined a drop to below 40 kg/capita/year in 1991/1992. In the
following years the consumption leveled out at 42–45 kg/capita. Currently
sugar supply in Poland constitutes circa 82% of the Belgium level which
fl uctuates between 55 and 59 kg in the last decade showing a growing trend.

44 E. Halicka et al.
The consumption of vegetables and fruit is characterized by signifi cant
yearly fl uctuations due to weather conditions and changes in price and
supply in both countries (Fig. 2, Annex 3).
According to FAO data the per capita quantity of fruit available for
consumption in Poland has shown an increase from 32 and 57 kg per year
in the analysed period; however it ranks last among EU-25 countries, after
Slovakia. In comparison to Belgium it is about 20 kg/capita less. The growing
trend in fruit consumption in Poland comprises the increase of imported
southern-grown species. The demand for apples, which constitute 50%
of overall fruit consumption and berries, remains relatively unchanged
(Halicka, 2005).
FIGURE 1. Consumption of main vegetal products in Poland and Belgium, 1989–2003,
kg/capita/year
Source: FAO, 2006.
0
30
60
90
120
150
180
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
kg/capita/year
B-cereals PL-cereals B-potatoes
PL-potatoes B-sugar PL-sugar
FIGURE 2. Consumption of fruits and vegetables in Poland and Belgium, 1989–2003,
kg/capita/year
Source: FAO, 2006.
0
30
60
90
120
150
180
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
kg/capita/year
PL-vegs B-vegs PL-fruit B-fruit

Food consumption in Poland and Belgium... 45
FAO data show that the supply of vegetables in the Polish and Bel-Lux
markets was quite similar in the nineties. According to WHO recommenda-
tions the daily supply of fruits and vegetables on Balance Sheet level should
be at least 600 g, i.e. 219 kg per year. Neither of the studied countries meets
this recommendation. In 2003 the yearly supply was only 148 kg/capita in
Poland and 199 kg/capita in Belgium.
The remaining groups of vegetal products, i.e. pulses and tree nuts,
constitute a very small share of consumed plant-derived products. The
supply of tree nuts in Poland is since the introduction of market economy
and the development of imports 5 times bigger than in the previous decades
while the consumption of pulses remains stable.
2.2. Consumption of main animal products in Poland and Belgium
In Poland, after a sharp fall in animal products’ consumption due to
liberalization of prices in 1990, it has increased in the period 1989–2003,
with the exception of milk. As shown in Figure 3 the per capita level of milk
and milk products’ has decreased to 173 kg in 2003 (Annex 4).
In the years 1992–1995 the consumption of milk and its products in
Poland and Belgium was almost at the same level. Since 1996 consumption
of this group of foodstuffs in Belgium continued to grow, reaching 256 kg in
2003 that is almost 150% of the Polish supply level. Besides important eco-
nomical (price and income) factors the falling trend in Poland was determined
by the growing supply of juices, soft drinks and mineral water. The effect of
FIGURE 3. Dairy products consumption in Poland and Belgium, 1989–2003, kg/capita/year
Source: FAO, 2006.
90
120
150
180
210
240
270
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
kg/capita/year
B-milk PL-milk B-milk trend PL-milk trend

46 E. Halicka et al.
decreasing dairy products’ consumption is particularly negative in the
context of observed fall in animal protein consumption (Kowrygo, 2000).
Meat and meat products supply level was lower in Poland through the
whole analysed period (Fig. 4, Annex 5).
0
30
60
90
120
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
kg/capita
/year
B-meat PL-meat B-fish PL-fish
kg/capita/year
FIGURE 4. Meat and fi sh products’ consumption in Poland and Belgium, 1989–2003, kg/
/capita/year
Source: FAO, 2006. B-fi sh in 2001 – domestic data.
Due to sharp price increase in the beginning of the nineties, meat and
its products’ consumption in Poland decreased to a level below 70 kg/year
while in Belgium the average consumption level was 87 kg. It is interesting
to notice that in 2003 the yearly per capita supply of meat products in
Poland was only 4 kg lower than in Belgium.
Fish and seafood consumption in Poland shows major yearly fl uctuations,
with a low (9 kg/capita) in 1992 and slowly growing trend. In Belgium the
consumption is higher and fl uctuates around 20 kg per year. Since 2001
FAOSTAT does not show Belgian data but according to domestic data the
consumption remained stable, between 18.5 and 20.4 kg/capita/year.
2.3. Consumption of fats and oils in Poland and Belgium
Comparison between Poland and Belgium shows that the total per capita
supply of fats and oils in Poland in 2003 was two times below the Belgian
level (Fig. 5, Annex 6). Due to nutritional aspects and high energy value of
fats this situation has a major infl uence on the diet structure and health of
consumers. Belgian per capita supply of animal fats, which include butter
and cream, constitutes 175% of the Polish level.

Food consumption in Poland and Belgium... 47
The trend in butter consumption in both countries was downward since
1989, however in Poland it is currently lower than in Belgium and constitutes
about three quarters of the Belgian level. The same situation describes the
supply of cream, which is about 1.6 times higher in Belgium than in Poland.
In the years 1989–1995 the supply of vegetable oils in Poland increased
steadily (almost twofold) due to growth of production potential and marketing
activities of new investors while supply of animal fats gradually decreased.
In 1995 the consumption of animal fats constituted 60% of the 1990 level.
The main vegetable oils consumed in Poland are traditionally rapeseed and
mustard oils (also processed to margarine), although the consumption of
soybean and palm oils show a growing trend.
In Belgium a growing trend can be observed in the case of vegetable
oils but the tendency in quantity of animal fats available for consumption
is rather stable. The consumption of both types of fat exceeds 20 kg and is
high compared to other EU countries.
2.4. Trends in daily energy (DES) and nutrients’ supply
FAO Food Balance Sheets are also a source of information on the diet
structure of populations at country level. The collected data show that
per capita daily energy supply in Belgium is higher than in Poland (Fig. 6,
Annex 7).
FIGURE 5. Vegetable and animal fat consumption in Poland and Belgium, 1989–2003,
kg/capita/year
Source: FAO, 2006.
0
5
10
15
20
25
30
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
kg/capita
/year
B-anfat PL-anfat B-vegoil PL-vegoil
kg/capita/year

48 E. Halicka et al.
During the period 1989–2003, the total energy supply in Poland
fl uctuated between 3290 kcal and 3500 kcal, which means on average 3354
kcal and was highest in 1989. In the two decades proceeding economical
transformation the energy value of the Polish diet was estimated at even
higher level.
In Belgium during the period 1989–2003 the average daily energy value
of diet was about 3605 kcal and was 7% higher than in Poland. In Belgium
the energy in diet shows a growing trend, while in Poland it is rather
stabilized. In comparison to other EU countries in 2003 Belgium was ranked
6th according to amounts of calories in diet, Poland was 16th. The
majority of energy in both Poland and Belgium is derived from vegetal
products. In Poland since 1993 less than 30% of energy comes from animal
foods, in Belgium the share of animal products in energy supply has been
gradually decreasing from 35% in the late 1980s to 31% in late 1990s.
According to FAO data in Poland carbohydrates are the major source
of energy (57–58% of calories), followed by dietary fats (30%) and proteins
(12%). In Belgium the role of dietary fats in energy supply is signifi cantly
bigger (39–40%), while carbohydrates deliver less than 50% of calories and
proteins about 11–12%.
The level of consumption of fats and proteins in daily diet in Poland and
Belgium since 1989 does not show signifi cant changes (Fig. 7). The Belgian
fat intake constituted about 146% of the Polish level in 2003. Fat intake in
Belgium is also high compared to other EU countries (second place, after
France), Poland ranks 17th.
FIGURE 6. Diet energy supply in Poland and Belgium, 1989–2003, kcal/capita/day
Source: FAO, 2006.
3000
3200
3400
3600
3800
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
kcal/capita
B-energy PL-energy
B-energy trend PL-energy trend
kcal/capita/day

Food consumption in Poland and Belgium... 49
In the last 15 years and until 2000, total daily protein consumption
stabilized at 99–102 g/day in Poland and 102–106 g/day in Belgium. However,
since 2000 the intake of protein according to FAO is lower in Belgium than
in Poland. This situation is linked to the lack of data concerning Belgian fi sh
supply in the FAO database.
3. Food availability at household level in Poland
Secondary data on food consumption (availability) on household level in
Poland are collected yearly by the Central Statistical Offi ce GUS (Gówny
Urzd Statystyczny). The Household Budget Survey sources come from
about 33 thousand households, which conduct records of all food purchases,
contributions from the household’s own production and other people. The
surveys are based on a sampling method which allows a generalization
of the results to all Polish households. The data obtained have several
limitations. For example no records are collected on the type and quantity
of food items and beverages consumed outside the home. Data before 1993
can not be directly compared with later years due to signifi cant methodology
changes, related to the classifi cation of households and method of survey.
Since 1993 the data are collected with the use of the monthly rotation
method instead of yearly or quarterly ones and households are classifi ed into
six socio-economic groups, i.e. households of employees, employee-farmers,
farmers, self-employed, retirees & pensioners and persons with unearned
sources of maintenance.
Changes in food consumption in terms of food availability in an average
Polish household in the years 1993–2005 are presented in Table 1.
FIGURE 7. Consumption of dietary fat in Poland and Belgium, 1989–2003, g/capita/day
Source: FAO, 2006.
0
30
60
90
120
150
180
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
g/capita/day
B-fat PL-fat B-protein PL-protein

50 E. Halicka et al.
Overall in the analysed years the consumption of most groups of
products showed a falling trend. The biggest decreases were observed in the
case of animal fats, sugar, milk and potatoes.
Traditionally food consumption is highest in the households of pensioners
and retirees. These types of households typically include only adults, are
usually smaller sized (one or two-persons) and have a low level of out-of-
-home consumption (Table 2).
TABLE 2. Consumption of main food groups in different types of Polish households in 2005,
kg/capita/year
Food groups Employees Farmers Self-
-employed
Retirees and
pensioners
Cereals 91.3 121.6 84.1 119.2
Potatoes 68.2 101.8 63.1 102.1
Vegetables 60.1 79.2 61.4 86.9
Fruits 40.7 43.9 46.2 54.4
Meat 59.0 82.0 59.5 77.9
Fish 4.4 4.4 5.3 6.6
Edible fats:
animal fats
vegetable oils
butter
16.0
1.9
10.9
3.2
19.7
4.5
11.8
3.4
11.3
1.6
9.4
4.3
23.0
3.5
14.6
4.9
Milk (litres) 44.2 79.9 44.0 66.4
Eggs (pcs) 163 215 162 217
Sugar 15.2 26.2 13.9 23.9
Source: GUS, 2006.
TABLE 1. Consumption of main food groups in Polish households, 1993–2005, kg/capita/
/year
Food groups 1993 1995 1999 2002 2004 2005
Cereals 101 94 113 107 104 101
Potatoes 114 105 93 90 83 80
Vegetables174 71 67 66 64 69
Fruits 51 45 45 49 47 45
Meat 65 61 67 65 65 66
Fish 5.5 5.3 4.4 4.8 4.9 5.0
Fats and oils, incl.:
animal fats
vegetable oils
butter
20.9
4.9
10.6
5.4
19.1
3.7
12.0
3.4
19.0
3.0
11.8
4.2
18.9
2.6
12.1
4.1
18.8
2.6
12.2
4.0
18.3
2.5
12.0
3.7
Milk (litres)289 84 69 62 62 53
Eggs (pcs) 183 181 182 181 179 181
Sugar 31 28 28 25 24 18
1Including legumes.
2Excluding condensed and powdered milk.
Source: GUS, 1994–2006.

Food consumption in Poland and Belgium... 51
The differentiated income and food expenditure levels in Polish
households in 2005 (Table 3) shows that the highest income characterizes
households of the self-employed. On average food and non-alcoholic
beverages comprise 28% of the total expenditures in households – its share
being 36% in the farmers’ and almost 24% in self-employed households.
TABLE 3. Income and expenditures in different types of Polish households in 2005
Specifi cation Grand
total Employees Farmers Self-
-employed
Retirees and
pensioners
Average monthly per capita in PLN
Available income 761.5 770.0 606.2 977.1 800.2
Expenditures:
food and non alcoholic
beverages
alcoholic beverages and
tobacco
690.3
194.1
18.8
684.4
180.2
19.5
533.9
192.8
14.1
869.8
204.9
23.6
746.0
227.2
18.3
In percent1
Share in expenditures:
food and non alcoholic
beverages
alcoholic beverages and
tobacco
28.1
2.7
26.3
2.9
36.1
2.6
23.5
2.8
30.4
2.5
1Percentage of expenditure (total = 100%).
Source: GUS, 2006.
According to the “Social Diagnosis” Report (Czapiski and Panek, 2005)
the share of households which are not able to fulfi l nutritional needs due
to insuffi cient income increases with the number of children in the family
(Table 4).
TABLE 4. Share of households which are not able to fulfi l basic nutritional needs
Products Total
Married couples (%)
without
children
with one
child
with 2
children
with 3 or more
children
incomplete
families
Vegetables112.0 6.7 6.0 10.1 17.9 16.2
Fruits119.0 11.2 11.1 17.3 27.5 26.3
Meat222.6 13.9 14.7 20.5 32.3 31.2
Meat products 22.5 14.9 14.0 19.0 32.1 28.6
Fish134.3 23.8 26.4 31.4 46.8 42.5
Butter and
edible fats 8.8 4.9 5.2 7.0 14.4 12.1
Milk 6.4 3.4 4.0 4.9 10.2 9.8
Dairy products 15.0 8.7 11.4 12.6 22.1 21.5
Sugar 5.9 3.0 3.3 4.4 8.8 9.2
Sugar products 30.4 20.6 22.3 28.1 40.3 40.6
1Including preserves.
2Including poultry meat.
Source: Czapiski and Panek, 2005.

52 E. Halicka et al.
The largest share of households does not meet the needs in the case of fi sh
(34.3%) and sugar products (30.4%). More than 40% of incomplete families
can not fulfi l the nutritional needs regarding these groups of products.
Analysis of quintiles (groups of households, each consisting of 20% of
total group, according to income) show that the expenditure on food and
beverages in Polish households in 2004 varied from 20.7% in the highest
income households (V quintile) to 42 % in the lowest income households
(I quintile). Together with the increase of income the expenditure on most of
food products increases, with the exception of fl our, full-fat milk and mixed
(rye-wheat) bread (Gulbicka and Kwasek, 2006).
The most signifi cant differences in food consumption between house-
hold with highest and lowest income levels can be observed in the case of
relatively more expensive food products. These include fruit and vegetable
juices, mineral water, veal and beef, dried fruits, nuts, rye bread, fi sh and
processed potatoes as well as milk products. For example in the highest
income group the consumption of veal was 5 times bigger than in the lowest
one and in the case of beef 3.8 times bigger. The differences are much smaller
in the case of poultry and pig meat (33 and 43%). Smaller diversifi cation of
food consumption is also characteristic for relatively cheaper products, such
as fl our, milk, bread, potatoes, vegetal fats, cream and pasta.
In the lowest income group (I quintile) the calculated average dietary
energy value is 1750 kcal/day, what is below the minimum level of food
security of 2000 kcal per capita on household level as recommended by
FAO/WHO (1992a). The infl uence of household composition on food
consumption in Poland is presented in Table 5.
TABLE 5. Share of food and non alcoholic beverages in expenditures on consumer goods and
services according to number of persons in households in 2005
Number
of persons
Grand
total
Share of expenditure by type of households (%)
employees farmers self-
-employed
retirees and
pensioners
1 25.9 18.9 – – 29.8
2 28.4 23.3 37.1 21.4 32.2
3 28.1 26.4 36.6 23.6 33.3
4 29.2 28.4 35.8 25.0 33.8
5 33.4 32.0 38.8 28.8 37.2
6 and more 37.6 35.5 40.5 32.6 40.7
Source: GUS, 2006.
The share of food and alcoholic beverages in expenditures on consumer
goods increases with the number of persons in households. It exceeds 40%
in the case of large families, in which the head of household is a retiree,
pensioner or farmer and is lowest in households of self-employed.

Food consumption in Poland and Belgium... 53
4. Household budget data on food consumption in Belgium
4.1. Methodology of Belgian analysis
In Belgium two sources of information – NIS (Nationaal Instituut voor de
Statistiek) and GfK (Gesellschaft für Konsum und Absatzforshung) can be
used to collect food purchase and expenditure data.
NIS organises regular household budget surveys with the representative
sample of about 3000 families. GfK is a private company and collects
systematic purchase data on a monthly base for about 2000 families since
1995. In the current study the data from the year 2000 by NIS and data
from the year 2001 by GfK are analysed.
Due to the fact that the two databases were set up with different
objectives the following limitations of these databases can be noticed:
1. The two databases consist of data about purchase and expenditure
for food products. The household panel data of GfK does not take into
account the out of home consumption and no one of the two databases
takes into account the food self-produced. Therefore, the original fi gures do
not give the effective consumption of food products.
2. The database of GfK does not include certain food products such as
bread, pastry and sugar products, while in the NIS database these products
are included. It means that these databases are not comparable. Furthermore,
it is impossible to formulate hard conclusions about the nutritional content
of the diet per person because there is no record about the dietary intake.
3. The sampling methods of each of the two databases contain limited
information about socio-economic criteria. The used socio-economic crite-
ria are adapted to the norms of social stratifi cation, e.g. age or profession
groups.
4. The extended databases of 2000 or 3000 families on the one hand
and about 150 to 200 food products on the other hand, produce huge data
matrixes, which are not easy to handle. Therefore, the purchase data of food
products are aggregated into seven food groups (meat, fi sh, poultry, dairy,
edible fats, vegetables, fruit). The analysis is made with the aggregated data
per family.
5. The classifi cation of social classes into groups in each of the databases
occurs ex post and is adapted to the available data. It means that social
stratifi cation is not homogeneous for the two databases.
The major problem is that NIS and GfK collect data at the household
level. To obtain the data at the individual level, the adaptation is made by
using a quantity or Q-factor and income equivalent:

54 E. Halicka et al.
Quantity or Q-factor: the Q-factor indicates the weight of food purchased by
adults and children (Cantillon et al., 1999):
Q-factor = 1 + 0.5 (number of adults – 1) + 0.3 (number of children)
Income equivalent: income per household is divided by the Q-factor resulting
in income equivalent or income per adult person:
Y’ = Y/Q-factor
Quantity of food purchased per person (PP):
PP = Purchase per household / Q-factor
Expenditure on food per person (EP):
EP = Expenditure per household / Q-factor
For NIS, the income classes are established from frequencies in the
original database. The income classes are based on the income equivalent
(Y’), and four groups of households are considered in relation to the Relative
Poverty Level (RPL). RPL relates to half of the average expenditure of the total
population. Thus, the RPL draws the line between relative rich and relative
poor households. The lowest 25% of the total sample is called “Poor Income”,
the next 25% “Low Income”, the sample over RPL till next 25% is called
“Medium Income”, and the top 25% is “High Income” (Table 6).
TABLE 6. Mean monthly income per person in 2000, NIS
Income groups Income (€) N (%)
minimum maximum mean
Poor 280 988 736 125 (25%)
Low 1000 1352 1201 125 (25%)
Medium 1368 1785 1598 125 (25%)
High 1814 7916 2970 125 (25%)
Total 500 (100%)
In regard to GfK, the three distinctive groups of households are separated
by cluster analysis taking into account the income equivalent (Y’). Table 7
indicates the mean monthly income per person in the three income groups.
TABLE 7. Mean monthly income per person in 2001, GfK
Income groups Income (€) N (%)
minimum maximum mean
Low 222 806 627 208 (42 %)
Medium 841 1308 1032 215 (43%)
High 1363 2355 1627 76 (15%)
This later procedure corresponds to the Relative Poverty Level method,
however only three income groups are chosen, because the high income
class is underrepresented in GfK database.

Food consumption in Poland and Belgium... 55
4.2. Results and discussion of the Belgian data
The analysis procedure comprises three steps. First, the descriptive analyses
show differences between quantity of food purchased per person (PP) and
expenditure on food per person (EP) for the total sample of respondents and
within income groups.
Second, by regression analysis the relationship between PP or EP and
socio-economic variables is determined. Third, the possible relationship
between income group and the other variables is found out.
4.2.1. Average quantity of food purchased and expenditure on food
Quantity of food purchased per person (PP) and expenditure on food per
person (EP) are calculated for all food categories. Further the PP and EP for
the income groups are considered and the difference between these groups
is tested by the analysis of variance (ANOVA). Finally, the budget share of
EP is calculated and the ANOVA test is applied.
The collected information from NIS data relates only to expenditure for
food items. This expenditure is not signifi cantly different between income
classes. However, there is a difference in budget share in expenditure for all
food and drink categories between income classes. Thus, the law of Engel is
again confi rmed. The low income group spends a higher share of their budget
(23%) than high income class (8%). The additional analysis is computed
with the insight to the expenditure for 43 groups of food and drink items.
The signifi cant difference between income classes is tested. There are only
fi ve groups of products with a signifi cant difference between income classes.
Consumers with a higher income clearly spend more for cakes and pastry,
fresh meat, fi sh and crustaceans in can, as well as cheese. Expenditure for
milk and milk products is higher for the people in the class with a medium
income.
The results of GfK database are presented in Table 8 that shows the
average quantity of food purchased per person in Belgium. There is no
signifi cant difference in purchase of different food groups between the three
income groups, although it is noticed that a total consumption of food by the
low income group is lower than the one calculated for other income groups.
The only signifi cant difference is the greater quantity of fi sh purchased
by the high income group.
Table 9 presents the average expenditure on food per person in Belgium
in 2001, which is 1825 €. The most expensive is meat since the annual
expenditure on meat is 624 € while expenditure on fruit (394 €), dairy (262 €),
vegetables (253 €), fi sh (131 €), poultry (104 €) or fat (58 €) is lower.

56 E. Halicka et al.
Food expenditure between the three income groups is diversifi ed, i.e.
the low income group spends signifi cantly less money on fi sh, dairy and
vegetables. A total food expenditure per person is lower in the low income
group (1695 €) than in medium (1806 €) or high income ones (1974 €).
In Table 10 the budget share in expenditure for different kinds of foods
over the three income groups is presented. The analyses confi rm Engels’s
Law, i.e. the low income group spends a higher share (%) of the total budget
for food, than the medium and the high income groups. The low income
group has 2.3 times higher budget share (23.6%) in food expenditure than
the high income group (10.1%).
TABLE 10. Budget share (%) in food expenditure by income group, in 2001, GfK
Income groups Budget share Total
meat fi sh poultry dairy fat vegs fruit total
Low 8.6 1.4 1.3 3.2 0.7 3.1 4.9 23.6 100
Medium 4.8 1.0 0.8 2.2 0.4 2.0 3.1 14.6 100
High 3.4 0.8 0.5 1.4 0.3 1.4 2.2 10.1 100
ANOVA
Df = 2
Sig. 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 ×
F 86.4 11.1 31.6 63.7 28.1 53.4 44.8 148.4
4.2.2. Signifi cant socio-economic variables explaining PP and EP
In the regression analysis, the variance in PP and EP, and per income group
is explained by the fi ve socio-economic variables: income (IN), region (RE),
education (ED), profession (PR) and age (AG). In this way it is possible to
TABLE 8. Quantity (kg per capita) of food purchased by income group, in 2001, GfK
Income groups Quantity of food purchased N
(valid)
meat fi sh poultry dairy fat vegs fruit total
Low 78 13 17 117 20 196 238 679 208
Medium 74 15 19 126 21 206 260 721 215
High 78 20 17 122 22 207 277 743 76
ANOVA
Df = 2
Sig. 0.541 0.000 0.507 0.536 0.543 0.606 0.211 0.498 Total
499
F 0.6 8.1 0.6 0.6 0.6 0.5 1.5 0.6
TABLE 9. Expenditure (€ per capita) on food by income group, in 2001, GfK
Income groups Expenditure on food N
(valid)
meat fi sh poultry dairy fat vegs fruit total
Low 615 103 99 237 51 228 362 1695 208
Medium 597 126 108 272 60 252 391 1806 215
High 659 164 104 276 62 280 429 1974 76
ANOVA
Df = 2
Sig. 0.283 0.000 0.534 0.011 0.073 0.008 0.128 0.001 Total
499
F 1.2 11.7 0.6 4.5 2.6 4.8 2.0 6.7

Food consumption in Poland and Belgium... 57
indicate which variable is relevant for the total amount of food purchased,
or expenditure on food within each income group:
PP or EP = a + b1(IN) + b2(RE) + b3(ED) + b4(PR) + b5(AG)
Analysis of NIS data show that variation explained is lower than 10%,
with exception for the high income class. Income is the only signifi cant
variable having an infl uence on food expenditure.
For GfK, the regression analyses is calculated for the four variables
(region, education, profession and age) that have a potential infl uence
on quantity of food purchased and expenditure on food. Age is the only
one variable that has a positive signifi cant impact on the quantity of food
purchased. This means that the quantity of food purchased and the
corresponding expenditure on food increase with the age of people. Older
people in a high income groups purchase twice more food than older people
in the low income group. The wealthier elderly people spend also twice more
money on food in total than poorer ones. However, the explained variance
remains below 10% for the low and medium income group, and is about
20% for the high income group.
4.2.3. Relationship between income group and the other variables
The relationships between income groups and region, education, profession
and age are analyzed by chi-Square test.
Analysis of NIS data show the relation between income classes on the
one hand and region, education, profession and age on the other hand. Only
the link between low education and low income is signifi cant. Additionally,
not active people can be again found in the lowest income class.
In GfK database, the relationships between income groups and the four
social variables (region, education, profession and age) are calculated. It
is important to fi nd out how the low income group can be characterized in
terms of region of living, education, profession and age. The link between
income and region is not signifi cant. For the three regions in Belgium, the
distribution of respondents over the income groups is almost the same.
In Table 11 the relationship between education and income groups
is presented. In the case of Belgium, the primary education refers to the
education at the elementary grade while the secondary education indicates
all degrees obtained after additional fi ve years of studies but before entering
the college or the university.
Table 12 shows that low income group is mainly composed of workers
who are technical employees without higher or university education. The

58 E. Halicka et al.
offi ce employees and people with an independent or free profession have
mostly medium incomes.
In Table 13 the relationship between income groups and age is presented.
It is interesting that the low income group is composed of younger rather
than older people. The pensioners above the age of 65 years have rather
medium incomes when comparing to other age groups.
TABLE 13. Relationship between income group and age
Income groups Age (%) Total
(%)
 44 years 45–64 years  65 years
Low 42.3 44.2 33.3 42
Medium 37.1 39.9 58.6 43
High 20.6 15.8 8.1 15
N = 499, X2Test: Df = 6, p = 0.006 100
4.3. Conclusions from analysis of Belgian data
First, the food purchase per person in the low income class is not signifi -
cantly lower than for medium and high income classes. It is shown in the
sample of Belgian consumers, and contradictory to the presented literature,
that people having low income do not purchase signifi cantly less vegetables
and fruit than people with medium and high incomes. The exception is the
quantity of fi sh purchased that is much lower for the low income group than
other income groups. The data suggest that the difference from the amount
TABLE 11. Relationship between income group and education
Income groups Education (%) Total
(%)
primary secondary higher and university
Low 50.6 38.7 37.6 42
Medium 38.3 47.0 41.0 43
High 11.0 13.5 21.4 15
N = 490, X2Test: Df = 6, p = 0.026 100
TABLE 12. Relationship between income group and profession
Income groups
Profession (%) Total
(%)
workers offi ce-employees independent or
free profession
Low 53.9 41.7 26.1 42
Medium 36.7 45.6 49.3 43
High 9.4 12.8 24.6 15
N = 446, X2Test: Df = 6, p = 0.000 100

Food consumption in Poland and Belgium... 59
of food purchased between the three income groups should not be vital for
any health inequalities in Belgium. Since the methodological problem was
already indicating diffi culties in reaching the most vulnerable and really low
income group by secondary data, such a conclusion cannot be made in this
research.
Second, with regard to expenditure on food in Belgium, two aspects
are explained. First, it is shown that people with low income spend signifi -
cantly less money on total food in general, and specifi cally less on fi sh, dairy
products, and vegetables. The analysis also indicates that regardless of the
level of income Belgians spend approximately the same amount of money
on fruit. Second, the budget share on food expenditure for the low income
group is indeed 2.3 times higher than in the high income group. This fact
shows that poorer people may have other experiences with food availability
and food choice, i.e. its quality and variety, than richer people. Since the
budget share in food expenditure is higher for the poor and low income
classes the medium and high income classes may focus more on quality and
variation while purchasing food items.
Third, the socio-economic variables explaining the quantity of food
purchased and expenditure on food per person are related to people’s age
regardless the income group. The cross-tabulation analyses show signifi cant
relationships between education, profession and income groups. Primary
education and low professional status are characteristic for low income
groups in Belgium.
The comprehensive statistical data of NIS and GfK related to purchase
and expenditure for food items give indication and insight in the fact that
the low income classes spend less money on fi sh, dairy, vegetables and fruit.
These are food products, which are considered generally as “healthy” food.
In this perspective, it is clear that people from the low income classes have
less healthy food consumption behaviour.
As it was indicated before, really low income food consumers were not
interviewed in the NIS and GfK panels. Therefore the current research does
not generate relevant conclusions over the unhealthy food consumption
pattern of this social group. Even though Belgium is known as a relative
homogeneous country, where the range of income difference between rich
and poor is a factor 5 as in the current sample, the results of this study may
not be overestimated. Since both household panel data of NIS and GfK are
not quite relevant to come up with conclusions about a healthy diet, more
specifi c data have to be collected in Belgium in order to test these general
fi ndings.

60 E. Halicka et al.
5. General conclusions
Although the nutritional status of Polish and Belgian populations is not
refl ected in the “average” descriptions in statistical databases, the FAO
Balance Sheets are a valuable source of data that enable comparison
between countries. Research focusing on differences between socio-economic
groups brings more understanding of the consumer decision-making process
in regard to the quantity of food purchased and food expenditure.
The analysis of FAO Food Balance Sheet FBS data for Poland and
Belgium implies that the economical transition in Poland in 1989 had
a major infl uence on the consumption patterns of the population. The most
important changes in 1989–2003 concerned the average per capita supply
of fruits, edible fats – of both animal and plant origin – and milk. The levels
of consumption of these products are one of the most important diet-related
factors infl uencing health. In 2003 the average consumption of fruits and
vegetable oils in Poland was about 150% of the 1989 level and 60% in the
case of animal fats. These changes can be evaluated as positive. However,
the tendency in the case of full fat milk, of which supply in 2003 was only
70% of the 1989 level, is negative from the health perspective.
In Belgium, the observed changes in food consumption are less
signifi cant. However the majority of them may have important health
implications, contributing to increased obesity, heart disease and diabetes
prevalence. These negative trends include decrease in fruit and increase in
fat consumption.
The analysis based on FBS data shows that food consumption is
currently lower in Poland than in Belgium, both in the case of animal and
vegetable products. Two exceptions from these general characteristics are
identifi ed – consumption of cereals and potatoes is higher in Poland. Two
groups of products i.e. eggs and pulses have currently a similar consumption
level in both countries. There is also a growing trend for higher consumption
of fi sh and vegetable oils in both countries.
In Poland, the total energy intake was lower than in the previous decades,
reaching its minimum in 1994 and later stabilizing at slightly above
3350 kcal/capita/day. In Belgium the energy intake shows a growing trend.
The research shows that overall foods purchase per person in low
income households in Belgium is not signifi cantly lower than for medium
and high income class. In Poland differences are observed in the case of
fruit and vegetable juices, mineral water, veal and beef, dried fruits, nuts,
rye bread, fi sh, potato and milk products. In Belgium the socio-economic
variables explaining the quantity of food purchased and expenditure on food

Food consumption in Poland and Belgium... 61
per person are related to age regardless the income group. In this research,
the relationship between education and income groups is highly signifi cant
confi rming other studies showing that lower education is linked with
lower income. In Poland the share of food in expenditures increases with
the number of people in household and is the highest in the households in
which the head is a retiree, pensioner or farmer.
The observed trends indicate that food policy measures are needed to
redirect some of the negative (unhealthy) trends in food consumption.
According to the basic lessons of food security planning in preparing food
strategies clear, short-term goals should be set (Maxwell, 2001). These
should focus on the increase of fruit and dairy products consumption in
Poland and reduction of fats and sugar in Belgium. Food consumption and
expenditure data on household budget level collected in Poland and Belgium
due to methodological reasons were not used for comparison purposes but
enabled the identifi cation of consumption patterns needed to set policy goals
for different socio-economical and income groups of households.

Chapter V
ANALYSIS OF FOOD CONSUMPTION AND NUTRITIONAL
BEHAVIOUR OF LOW INCOME CONSUMERS IN POLAND
Krystyna Rejman
1. Sample, size and fi eldwork
The research based on interviews with 240 low income respondents was
carried out in a Central Poland medium-size city (38,000 inhabitants) during
the fi rst quarter of 2005. In order to select the target group of consumers,
the city was chosen taking into consideration high indices of unemployment
and decreasing standard of living due to economic transition. The selection
factor to sample was monthly income per person which varied depending on
household size. For a single person, the income ceiling was set at 1100 PLN1
per month. For the respondent living in households with two or more persons
the ceiling was 800 PLN per person, taking into consideration monthly
income of the Polish households amounted to 1020 PLN on average in 2004
(GUS, 2006).
The data were collected by face-to-face interviews. Each interview took
roughly one and a half hour. To assure the best practise of 24 h recall dietary
interviews, two publications of the National Institute of Food and Nutrition
in Poland (I) were used. The fi rst was a video fi lm “Instruction on
Conducting 24 Hours Dietary Recall Interview” (I, 2000), providing a very
detailed guidance for the fi eldwork. The second was “Album of Photographs
of Food Products and Dishes” (Szponar et al., 2000) used during the
interviews to facilitate respondent to give a precise description of the
volume/amount of daily food consumed.
The research has the same steps as in the Belgium study, namely: Physical
Activity and Health Survey (PAHS), Food Behaviour Survey (FBS), Social
Functioning Survey (SFS), Food Frequency Questionnaire (FFQ) and 24 hours
Recall Method (24 h RM).
1 1 euro = ±4 PLN in 2005.

Analysis of food consumption and nutritional behaviour... 63
2. Research design
Data were collected through the questionnaire consisting of four parts:
1. Socio-demographic characteristics of low income respondents and BMI
groups.
2. Physical Activity and Health Survey (PAHS) and Social Functioning
Survey (SFS):
− PAHS include the following variables: internal and external health
factors; health behaviour related to physical activity; diet-related
diseases and medical history;
− SFS focuses on: emotional eating (well being), social involvement
and effects of illness (consequences of diseases).
3. Food Behaviour Survey (FBS):
− knowledge, attitude and preference (knowledge and importance of
dietary guidelines, awareness of diet-disease links; self-assessment
of nutrition; sources of diet and health information; dietary
management practices, food choice and meal patterns);
− food environment (out-of home free meals consumption and food
provision; budget limitation of food consumption).
4. Food Consumption:
− frequency of food consumed – Food Frequency Questionnaire (FFQ);
− amount of food consumed – 24 h Recall Method (24 h RM).
3. Data analysis
The data obtained from the respondents were analysed in two steps. In the
fi rst step Body Mass Index was calculated for 215 respondents, as 25 persons
from the sample refused to indicate their weight. The assessment of the diet
by comparison with the recommended dietary allowances (RDA) was applied
to this smaller sample of the consumers (NRDA = 215).
In the second step the data collected in 24 h RM were entered to DIETA
2000 Program (Rybarczyk, 2000), which enables to calculate for each
respondent the following values:
1. Nutritional value of the diet including:
− dietary energy supply (DES) in kcal/day and its structure;
− daily nutrition intake of the following:
• energy nutrients in grams/day: total, animal and vegetable protein,
total fat, saturated fatty acids (SFA), monounsaturated fatty acids

64 K. Rejman
(MUFA), polyunsaturated fatty acids (PUFA) and total carbohy-
drates,
• minerals in mg/day: sodium, potassium, calcium, phosphor,